NIST Special Publication 1018-2

Sixth Edition

Fire Dynamics Simulator

Technical Reference Guide
Volume 2: Verification
Kevin McGrattan
Simo Hostikka
Jason Floyd
Randall McDermott
Marcos Vanella
Eric Mueller

http://dx.doi.org/10.6028/NIST.SP.1018
 NIST Special Publication 1018-2
Sixth Edition

Fire Dynamics Simulator

Technical Reference Guide
Volume 2: Verification
Kevin McGrattan
Randall McDermott
Marcos Vanella
Eric Mueller
Fire Research Division, Engineering Laboratory, Gaithersburg, Maryland
Simo Hostikka
Aalto University, Espoo, Finland
Jason Floyd
Fire Safety Research Institute, UL Research Institutes, Columbia, Maryland

http://dx.doi.org/10.6028/NIST.SP.1018

April 18, 2023

Revision: FDS-6.8.0-0-g886e009

E N T OF C O M
TM M
AR

ER
D EP

CE
ICA
UN
IT

ER

D
E

ST AM
ATES OF

U.S. Department of Commerce

Gina M. Raimondo, Secretary

National Institute of Standards and Technology

Laurie E. Locascio, NIST Director and Undersecretary of Commerce for Standards and Technology
 Certain commercial entities, equipment, or materials may be identified in this
document in order to describe an experimental procedure or concept adequately.
Such identification is not intended to imply recommendation or endorsement by the
National Institute of Standards and Technology, nor is it intended to imply that the
entities, materials, or equipment are necessarily the best available for the purpose.

National Institute of Standards and Technology Special Publication 1018-2

Natl. Inst. Stand. Technol. Spec. Publ. 1018-2, 306 pages (October 2013)
CODEN: NSPUE2
FDS Developers

The Fire Dynamics Simulator and Smokeview are the products of an international collaborative effort led by
the National Institute of Standards and Technology (NIST) and Fire Safety Research Institute, UL Research
Institutes. Its developers and contributors are listed below.

Principal Developers of FDS

Kevin McGrattan, NIST, Gaithersburg, Maryland

Simo Hostikka, Aalto University, Espoo, Finland
Jason Floyd, Fire Safety Research Institute, UL Research Institutes, Columbia, Maryland
Randall McDermott, NIST, Gaithersburg, Maryland
Marcos Vanella, NIST, Gaithersburg, Maryland
Eric Mueller, NIST, Gaithersburg, Maryland

Principal Developer of Smokeview

Glenn Forney, NIST, Gaithersburg, Maryland

Collaborators and Contributors

Anthony Hamins, NIST, Gaithersburg, Maryland

Emanuele Gissi, Corpo Nazionale dei Vigili del Fuoco, Italy
Susan Kilian, hhpberlin, Germany
William Mell, U.S. Forest Service, Seattle, Washington
Julio Cesar Silva, Brazilian Navy, Brazil
Craig Weinschenk, Fire Safety Research Institute, UL Research Institutes, Columbia, Maryland

i
About the Developers

Kevin McGrattan is a mathematician in the Fire Research Division of NIST. He received a bachelor of
science degree from the School of Engineering and Applied Science of Columbia University in 1987
and a doctorate at the Courant Institute of New York University in 1991. He joined the NIST staff
in 1992 and has since worked on the development of fire models, most notably the Fire Dynamics
Simulator.

Simo Hostikka is an associate professor of fire safety engineering at Aalto University School of Engineer-
ing, since January 2014. Before joining Aalto, he worked as a Principal Scientist and Team Leader at
VTT Technical Research Centre of Finland. He received a master of science (technology) degree in 1997
and a doctorate in 2008 from the Department of Engineering Physics and Mathematics of the Helsinki
University of Technology. He is the principal developer of the radiation and solid phase sub-models
within FDS.

Jason Floyd is a Lead Research Engineer at the Underwriters Laboratories Fire Safety Research Institute
in Columbia, Maryland. He received a B.S. (1993), M.S (1995), and a Ph.D. (2000) from the Nuclear
Engineering Program of the University of Maryland. After graduating, he was awarded a National
Research Council Post-Doctoral Fellowship at the Building and Fire Research Laboratory of NIST.
He is a principal developer of the combustion, control logic, aerosol, droplet evaporation, and HVAC
sub-models within FDS.

Randall McDermott joined the Fire Research Division at NIST in 2008. He received a B.S. from the
University of Tulsa in Chemical Engineering in 1994 and a Ph.D. from the University of Utah in 2005.
His research interests include subgrid-scale models and numerical methods for large-eddy simulation,
turbulent combustion, immersed boundary methods, and Lagrangian particle methods.

Marcos Vanella joined the Fire Research Division at NIST in 2019. He received diplomas in Mechani-
cal and Aeronautical Engineering from the National University of Cordoba, Argentina, and M.S. and
Ph.D. degrees in Mechanical Engineering from the University of Maryland, College Park. His research
interests include computer simulation and scientific software development applied to engineering sys-
tems, mainly in the areas of fluid flow and multiphysics interaction problems.

Glenn Forney is a computer scientist in the Fire Research Division of NIST. He received a bachelor of
science degree in mathematics from Salisbury State College and a master of science and a doctorate in
mathematics from Clemson University. He joined NIST in 1986 (then the National Bureau of Standards)
and has since worked on developing tools that provide a better understanding of fire phenomena, most
notably Smokeview, an advanced scientific software tool for visualizing Fire Dynamics Simulation data.

Eric Mueller joined the Fire Research Division at NIST in 2021. He recieved a B.S. in Engineering Physics
from Tufts University (2010), an M.S. from Worcester Polytechnic Institute (2012), and a Ph.D. from
the University of Edinburgh (2017), both in fire safety engineering. His research interests include the

iii
 development of sub-models relevant to heat and mass transfer in wildland and wildland-urban interface
fires.

Emanuele Gissi is a professional fire chief of the Italian Fire and Rescue Service (CNVVF) since 2002.
He received a Ph.D. in Engineering Physics from the University of Ancona, Italy, in 2000. In addition
to managing fire brigades, his main interest is bridging the gap between fire research and fire safety
engineering practice. He develops open source tools for FDS.

Anthony Hamins joined the Fire Research Division at NIST in 1989. He received his B.S. from the Uni-
versity of California, Berkeley, in Physics and is Ph.D. from U.C. San Diego in Engineering Physics in
1985. His research interests include fire model validation, fire dynamics, heat and mass transfer pro-
cesses in fires of multiple scales, fire suppression, flame structure, wildland-urban interface fires, and
micro-gravity combustion.

Susan Kilian is a mathematician with numerics and scientific computing expertise. She received her
diploma from the University of Heidelberg and received her doctorate from the Technical University
of Dortmund in 2002. Since 2007 she has been a research scientist for hhpberlin, a fire safety engineer-
ing firm located in Berlin, Germany. Her research interests include high performance computing and
the development of efficient parallel solvers for the pressure Poisson equation.

William (Ruddy) Mell is an applied mathematician currently at the U.S. Forest Service in Seattle, Wash-
ington. He holds a B.S. degree from the University of Minnesota (1981) and doctorate from the Uni-
versity of Washington (1994). His research interests include the development of large-eddy simula-
tion methods and sub-models applicable to the physics of large fires in buildings, vegetation, and the
wildland-urban interface.

Julio Cesar Silva is a Lieutenant in the Naval Engineers Corps of the Brazilian Navy. He worked in the Fire
Research Division of NIST as a Guest Researcher from National Council for Scientific and Technologi-
cal Development, Brazil. He received a M.Sc. in 2010 and a doctorate in 2014 from Federal University
of Rio de Janeiro in Civil Engineering. His research interests include fire-structure interaction and he
develops coupling strategies between FDS and finite-element codes.

Craig Weinschenk is a Lead Research Engineer at the Underwriters Laboratories Fire Safety Research
Institute, in Columbia, Maryland. He worked in the Fire Research Division at NIST as a National
Research Council Postdoctoral Research Associate in 2011. He received a B.S. from Rowan University
in 2006 in Mechanical Engineering. He received an M.S. in 2007 and a doctorate in 2011 from The
University of Texas at Austin in Mechanical Engineering. His research interests include numerical
combustion, fire-structure interaction, and human factors research of fire-fighting tactics.

iv
Preface

This is Volume 2 of the FDS Technical Reference Guide [1]. Volume 1 describes the mathematical model
and numerical method. Volume 3 documents past and present experimental validation work. Instructions
for using FDS are contained in a separate User’s Guide [2].
The three volumes of the FDS Technical Reference Guide are based in part on the “Standard Guide
for Evaluating the Predictive Capability of Deterministic Fire Models,” ASTM E 1355 [3]. ASTM E 1355
defines model evaluation as “the process of quantifying the accuracy of chosen results from a model when
applied for a specific use.” The model evaluation process consists of two main components: verification
and validation. Verification is a process to check the correctness of the solution of the governing equations.
Verification does not imply that the governing equations are appropriate; only that the equations are being
solved correctly. Validation is a process to determine the appropriateness of the governing equations as a
mathematical model of the physical phenomena of interest. Typically, validation involves comparing model
results with experimental measurement. Differences that cannot be explained in terms of numerical errors
in the model or uncertainty in the measurements are attributed to the assumptions and simplifications of the
physical model.
Evaluation is critical to establishing both the acceptable uses and limitations of a model. Throughout its
development, FDS has undergone various forms of evaluation. This volume provides a survey of verification
work conducted to date to evaluate FDS.

v
Disclaimer

The US Department of Commerce makes no warranty, expressed or implied, to users of the Fire Dynamics
Simulator (FDS), and accepts no responsibility for its use. Users of FDS assume sole responsibility under
Federal law for determining the appropriateness of its use in any particular application; for any conclusions
drawn from the results of its use; and for any actions taken or not taken as a result of analysis performed
using these tools.
Users are warned that FDS is intended for use only by those competent in the fields of fluid dynamics,
thermodynamics, heat transfer, combustion, and fire science, and is intended only to supplement the in-
formed judgment of the qualified user. The software package is a computer model that may or may not have
predictive capability when applied to a specific set of factual circumstances. Lack of accurate predictions
by the model could lead to erroneous conclusions with regard to fire safety. All results should be evaluated
by an informed user.
Throughout this document, the mention of computer hardware or commercial software does not con-
stitute endorsement by NIST, nor does it indicate that the products are necessarily those best suited for the
intended purpose.

vii
Acknowledgments

The following individuals contributed to the verification process of FDS.

• Chris Lautenberger and Carlos Fernandez-Pello contributed the “two-reaction” test case.

• Matthias Münch of the Freie Universität Berlin provided useful test cases for the basic flow solver.

• Clara Cruz, a student at the University of Puerto Rico and Summer Undergraduate Research Fellow at
NIST, helped develop useful Matlab scripts to automate the process of compiling this guide.

• Bryan Klein, now at Thunderhead Engineering, developed the source code version control system that
is an essential part of the verification process.

• Joonas Ryynänen of VTT, Finland, designed and documented the FED/FIC test cases.

• Max Gould, Summer Undergraduate Research Fellow, helped develop and produce the “2D Vortex” test
case.

• Howard Baum, NIST Fellow Emeritus, contributed the analytical solution “Collapse of a Hot Layer in
a Micro-gravity Environment”.

• Tony Saad and James Sutherland of the University of Utah developed an analysis for temporal order of
accuracy for variable-density projection schemes.

• Chao Zhang of Tongji University provided the finite-element analysis for the HT3D I-beam test case.

• Salah Benkorichi of Ulster University, Belfast, developed the HT3D sphere test case.

• Kiyoung Moon and Jung-il Choi of Yonsei University, Seoul, South Korea, provided the Heat Channel
Flow case.

• Jesus Alberto Mejias Tuni of Politecnico di Torino provided the 1-D tunnel cases with heat release.

• Morteza Gholami Haghighi Fard of Aalto University, Finland, contributed to the development of 3-D
pyrolysis.

• Jonathan Hodges of Jensen Hughes developed the FED for a moving particle test case.

ix
Contents

FDS Developers i

About the Developers iii

Preface v

Disclaimer vii

Acknowledgments ix

Contents xi

List of Figures xix

List of Tables xxvii

1 What is Verification? 1

2 Survey of Past Verification Work 3

2.1 Analytical Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.2 Numerical Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.3 Sensitivity Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.3.1 Grid Sensitivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.3.2 Sensitivity of Large Eddy Simulation Parameters . . . . . . . . . . . . . . . . . 7
2.3.3 Sensitivity of Radiation Parameters . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.3.4 Sensitivity of Thermophysical Properties of Solid Fuels . . . . . . . . . . . . . . 8
2.4 Code Checking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

3 The Basic Flow Solver 11

3.1 Analytical Solution to the Navier-Stokes Equation (ns2d) . . . . . . . . . . . . . . . . . . 11
3.2 Analytical Solution to the Continuity Equation . . . . . . . . . . . . . . . . . . . . . . . . 14
3.2.1 Pulsating 1-D solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
3.2.2 Pulsating 2-D solution (pulsating) . . . . . . . . . . . . . . . . . . . . . . . . 15
3.2.3 Stationary compression wave in 1-D . . . . . . . . . . . . . . . . . . . . . . . . 16
3.2.4 Stationary compression wave in 2-D (compression_wave) . . . . . . . . . . . 16
3.3 Scalar Transport . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.3.1 Uniform Velocity Field (move_slug) . . . . . . . . . . . . . . . . . . . . . . . 18
3.3.2 Solid Body Rotation Velocity Field (soborot) . . . . . . . . . . . . . . . . . . 20
3.4 Temporal Error Analysis for a Variable-Density Projection (saad) . . . . . . . . . . . . . 23
3.5 Variable-Density Manufactured Solution (shunn3) . . . . . . . . . . . . . . . . . . . . . 24

xi
 3.6 Collapse of a Hot Layer in a Micro-gravity Environment (hot_layer_360) . . . . . . . . 28
3.7 2-D Vortex Simulation (vort2d) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
3.8 Boundedness Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.8.1 Species Bounds (bound_test_*) . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.8.2 Isothermal Helium Wave (helium_1d_isothermal) . . . . . . . . . . . . . . 34
3.8.3 Temperature Bounds for Isothermal Helium Jet (helium_air_jet_floor) . . 34
3.8.4 Temperature Bounds for Large Differences in Molecular Weight (mwtest_cfl) . 35
3.8.5 Temperature Lower Bound in a Fire Simulation (tmp_lower_limit) . . . . . . 36
3.9 Constant Specific Heat Ratio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
3.9.1 Helium Wave using Constant Specific Heat Ratio (helium_1d_const_gamma) . 37
3.9.2 Fire Plume using Constant Specific Heat Ratio (fire_const_gamma) . . . . . . 38
3.9.3 Evaporation with Constant Specific Heat Ratio (water_evap_1_const_gamma) 39
3.10 Tunnel Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
3.10.1 1-D Flow in a Tunnel (tunnel_const_gamma) . . . . . . . . . . . . . . . . . . 40
3.10.2 1-D Flow in a Tunnel (tunnel_linear_cp) . . . . . . . . . . . . . . . . . . . 42

4 Turbulence 45
4.1 Decaying Isotropic Turbulence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
4.1.1 Constant Smagorinsky (csmag) . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
4.1.2 Dynamic Smagorinsky (dsmag) . . . . . . . . . . . . . . . . . . . . . . . . . . 47
4.1.3 Deardorff (deardorff) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
4.1.4 Vreman (vreman) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
4.1.5 WALE (wale) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
4.2 Jet Centerline Velocity Decay (jet) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

5 Boundary Effects 55
5.1 The FDS Moody Chart (poiseuille, moody, z0) . . . . . . . . . . . . . . . . . . . . . 55
5.1.1 Laminar Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
5.1.2 Turbulent Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
5.2 Blasius boundary layer (blasius) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
5.3 Pohlhausen thermal boundary layer (Pohlhausen) . . . . . . . . . . . . . . . . . . . . . 60
5.4 Near-Wall Grid Resolution (yplus) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
5.5 Heated Channel Flow (heated_channel) . . . . . . . . . . . . . . . . . . . . . . . . . . 61
5.6 Forced Convection Flat Plate Boundary Layer (forced_conv_flat_plate) . . . . . . . 63
5.7 Natural Convection Horizontal Plates (natconh) . . . . . . . . . . . . . . . . . . . . . . 65
5.8 Natural Convection Vertical Enclosure (natconv) . . . . . . . . . . . . . . . . . . . . . . 66
5.9 Free Convection from a Sphere (free_conv_sphere) . . . . . . . . . . . . . . . . . . . 68
5.10 Ribbed Square Duct Flow (ribbed_channel) . . . . . . . . . . . . . . . . . . . . . . . 69
5.10.1 Cartesian Geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
5.10.2 Complex Geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

6 Atmospheric Flows 73
6.1 Monin-Obukhov Similarity Profiles (MO_velocity_profile) . . . . . . . . . . . . . . . 73
6.2 Ekman Layers (ekman_) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73

7 Mass and Energy Conservation 77

7.1 Mass Conservation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
7.1.1 Flow of a Gas through a Simple Duct (simple_duct) . . . . . . . . . . . . . . 77

xii
 7.1.2 Species Mass Flux (species_conservation) . . . . . . . . . . . . . . . . . . 77
7.1.3 Mass Flux through Domain Boundaries (mass_flux_wall) . . . . . . . . . . . 78
7.1.4 Mass Balance on a Gas Control Volume (mass_balance_gas_volume) . . . . 80
7.1.5 Mass Balance for Species in a Reacting Flow (mass_balance_reac) . . . . . . 80
7.1.6 Checking the Equation of State (particle_offgas) . . . . . . . . . . . . . . . 81
7.1.7 Realizability of Species Mass Fractions (realizable_mass_fractions) . . . 81
7.1.8 Integral Mass Test for Complex Geometry (geom_mass_file_test) . . . . . . 81
7.2 Energy Conservation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
7.2.1 The Heat from a Fire (energy_budget) . . . . . . . . . . . . . . . . . . . . . . 85
7.2.2 Gas Injection via an Isentropic Process (isentropic) . . . . . . . . . . . . . . 85
7.2.3 Gas Injection via a Non-Isentropic Process (isentropic2) . . . . . . . . . . . 86
7.2.4 Mixing Gases (energy_budget_dns) . . . . . . . . . . . . . . . . . . . . . . 88
7.2.5 Mixing with Variable Specific Heats (energy_budget_tmix) . . . . . . . . . . 89
7.2.6 Combustion Chamber (energy_budget_combustion) . . . . . . . . . . . . . 89
7.2.7 Solid phase energy conservation (energy_budget_solid) . . . . . . . . . . . 90
7.2.8 Energy Conservation at Walls (mass_heat_wall_device_test) . . . . . . . 91
7.2.9 Energy Conservation for Solid Particles (energy_budget_particles) . . . . 91

8 Checking for Coding Errors 93

8.1 Symmetry Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
8.1.1 Counter-Flowing Jets (symmetry_test) . . . . . . . . . . . . . . . . . . . . . 93
8.1.2 Fire in a Thin-Walled Pan (symmetry_test_2) . . . . . . . . . . . . . . . . . . 93
8.1.3 Axial symmetry (cyl_test) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
8.2 Velocity boundary condition implementation (velocity_bc_test) . . . . . . . . . . . . 96
8.3 Checking the divergence constraint . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
8.3.1 Isothermal flow (divergence_test_1) . . . . . . . . . . . . . . . . . . . . . 97
8.3.2 Sealed compartment (divergence_test_2) . . . . . . . . . . . . . . . . . . . 97
8.3.3 Creating/Removing Obstructions (divergence_test_3) . . . . . . . . . . . . 97
8.4 Multi-mesh Pressure and Temperature Lapse (lapse_rate) . . . . . . . . . . . . . . . . 98
8.5 Randomly Overlapping Meshes (random_meshes) . . . . . . . . . . . . . . . . . . . . . 99
8.6 Overlapping INIT Regions (init_overlap) . . . . . . . . . . . . . . . . . . . . . . . . 99
8.7 Spurious Obstruction Effects (hole) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
8.8 OpenMP Thread Checking (race_test) . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
8.9 Multi-Mesh Layer Height Calculation (layer) . . . . . . . . . . . . . . . . . . . . . . . . 105
8.10 The Restart Feature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106

9 Thermal Radiation 107

9.1 Radiation View Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
9.1.1 Hot Plate (plate_view_factor) . . . . . . . . . . . . . . . . . . . . . . . . . 107
9.1.2 Plates at a Right Angle (geom_rad) . . . . . . . . . . . . . . . . . . . . . . . . 107
9.1.3 Plates at a Right Angle, Part 2 (geom_rad_2) . . . . . . . . . . . . . . . . . . . 108
9.2 Radiation inside a Box (radiation_box) . . . . . . . . . . . . . . . . . . . . . . . . . . 109
9.3 Radiation from a Polygonal Plate (radiating_polygon) . . . . . . . . . . . . . . . . . 110
9.4 Radiation from a Layer of Hot Gas (radiation_plane_layer) . . . . . . . . . . . . . . 111
9.5 Wall Internal Radiation (wall_internal_radiation) . . . . . . . . . . . . . . . . . . 111
9.6 Radiation Emitted by Hot Spheres (hot_spheres) . . . . . . . . . . . . . . . . . . . . . 112
9.7 Radiation Absorbed by Liquid Droplets (droplet_absorption) . . . . . . . . . . . . . 113
9.8 Consistency of Particle Absorbed Energy and the Radiative Loss Term . . . . . . . . . . . 114

xiii
 9.9 Radiation Attenuation by Scattering Particles (part_attenuation) . . . . . . . . . . . 114
9.10 Radiation Attenuation by Non-Scattering Particles
(radiation_gas-veg_consistency) . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
9.11 Transient Thermocouple Response (TC_heating) . . . . . . . . . . . . . . . . . . . . . 116
9.12 Radiation and Thermocouples (TC_view_factor) . . . . . . . . . . . . . . . . . . . . . 117
9.13 Radiation Targets (target_test) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
9.14 Radiation Shield (radiation_shield) . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
9.15 Radiation from a gas-fired panel (radiation_gas_panel) . . . . . . . . . . . . . . . . 119

10 Species and Combustion 121

10.1 Gas Properties (species_props) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
10.2 Humidity (humidity) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
10.3 Product Yields (methane_flame) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
10.4 Reaction Rates (reactionrate) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
10.4.1 Mixing-Controlled Reactions (reactionrate_EDC) . . . . . . . . . . . . . . . 124
10.4.2 Finite Rate Reactions (reactionrate_arrhenius) . . . . . . . . . . . . . . . 129
10.4.3 Fast-Slow Reaction Scheme (reactionrate_fast_slow) . . . . . . . . . . . 133
10.5 Heat Release Rates from Reaction Mechanisms (hrrpuv_reac) . . . . . . . . . . . . . . 134
10.6 Mixture Fraction (burke_schumann) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
10.7 Ideal and Non-Ideal Heat of Combustion (HoC) . . . . . . . . . . . . . . . . . . . . . . . . 138
10.8 Single Step and Two Step Chemistry Heat of Combustion . . . . . . . . . . . . . . . . . . 139
10.9 Fractional Effective Dose (FED) and Fractional Irritant Concentration (FIC) ) . . . . . . . 139
10.9.1 FED and FIC of gas mixtures (FED_FIC) . . . . . . . . . . . . . . . . . . . . . 141
10.9.2 FED from post-combustion yield of CO and HCN (FED_CO_HCN) . . . . . . . . 141
10.9.3 FED for a moving particle (FED_moving) . . . . . . . . . . . . . . . . . . . . . 141
10.10 Smoke Detector Model (smoke_detector) . . . . . . . . . . . . . . . . . . . . . . . . . 145
10.11 Aerosol Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
10.11.1 Gravitational Settling and Deposition of Aerosols
(aerosol_gravitational_deposition) . . . . . . . . . . . . . . . . . . . 146
10.11.2 Thermophoretic Settling and Deposition of Aerosols
(aerosol_thermophoretic_deposition) . . . . . . . . . . . . . . . . . . . 147
10.11.3 Turbulent Deposition of Aerosols (aerosol_turbulent_deposition) . . . . 148
10.11.4 Mass Conservation of Depositing Aerosols (propane_flame_deposition) . . 148
10.11.5 Agglomeration of Aerosols (aerosol_agglomeration) . . . . . . . . . . . . 149
10.11.6 Scrubbing of Aerosols (aerosol_scrubbing) . . . . . . . . . . . . . . . . . . 151
10.11.7 Oxidation of Soot on Surfaces (soot_oxidation_wall) . . . . . . . . . . . . 151
10.12 Condensation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
10.12.1 Condensation and Evaporation in the Gas Phase
(condesnation_1 and condensation_2) . . . . . . . . . . . . . . . . . . . . 152
10.12.2 Condensation and Evaporation in the Solid Phase
(wall_cond) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
10.12.3 Radiation for Condensed Phase (condesnation_3) . . . . . . . . . . . . . . . 153

11 Heat Conduction 155

11.1 Heat Conduction Through a Solid Slab (heat_conduction) . . . . . . . . . . . . . . . . 155
11.2 Temperature-Dependent Thermal Properties (heat_conduction_kc) . . . . . . . . . . . 156
11.3 Convective Cooling (convective_cooling) . . . . . . . . . . . . . . . . . . . . . . . . 157
11.4 Simple Thermocouple Model (thermocouples) . . . . . . . . . . . . . . . . . . . . . . 158

xiv
 11.5 Heat Conduction through Insulated Steel (insulated_steel_x) . . . . . . . . . . . . . 159
11.6 Heat Conduction across Mesh Boundaries (back_wall_test) . . . . . . . . . . . . . . . 160
11.7 Heat Conduction through Immersed Boundaries (back_wall_test_2) . . . . . . . . . . 161
11.8 Specified Heat Flux Boundary Conditions (adiabatic_xxx_flux) . . . . . . . . . . . . 161
11.9 Solid Heat Transfer 3-D (Beta) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
11.9.1 Energy Conservation in a 3-D Solid (ht3d_energy_conservation) . . . . . . 163
11.9.2 Heat Transfer to a Steel Beam (ht3d_beam_heating) . . . . . . . . . . . . . . 164
11.9.3 Mass Conservation in a 3-D Solid (ht3d_mass_conservation) . . . . . . . . 165
11.9.4 Continuous Heat Flux Boundary (ht3d_slab) . . . . . . . . . . . . . . . . . . 165
11.9.5 3-D Heat Diffusion in a Steel I-Beam (ht3d_ibeam) . . . . . . . . . . . . . . . 166
11.9.6 3-D Heat Diffusion in Thin Steel Plates (ht3d_network) . . . . . . . . . . . . 168
11.9.7 3-D Heat Diffusion in a Sphere (ht3d_sphere) . . . . . . . . . . . . . . . . . . 168

12 Pyrolysis 171
12.1 Conservation of Pyrolyzed Mass (surf_mass_conservation) . . . . . . . . . . . . . . 171
12.1.1 Pyrolysis at a Solid Surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
12.1.2 Pyrolysis of Discrete Particles . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
12.1.3 Rate of Solid Particle Decomposition (part_baking_soda) . . . . . . . . . . . 177
12.2 Conservation of Pyrolyzed Energy (matl_e_cons) . . . . . . . . . . . . . . . . . . . . . 179
12.3 Burning Rate of a Small Object (cell_burn_away) . . . . . . . . . . . . . . . . . . . . 183
12.4 Evaporation of a Liquid Surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
12.4.1 Mass Conservation (surf_mass_vent_liquid) . . . . . . . . . . . . . . . . . 183
12.4.2 Evaporation Rate (water_pool) . . . . . . . . . . . . . . . . . . . . . . . . . . 184
12.5 Change in Surface Emissivity (emissivity) . . . . . . . . . . . . . . . . . . . . . . . . 185
12.6 Shrinking and swelling materials (shrink_swell) . . . . . . . . . . . . . . . . . . . . . 186
12.7 Enthalpy of Solid Materials (enthalpy) . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
12.8 A Simple Two-Step Pyrolysis Example (two_step_solid_reaction) . . . . . . . . . . 188
12.9 Interpreting Bench-Scale Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
12.9.1 General Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
12.9.2 Interpreting TGA Data (tga_sample) . . . . . . . . . . . . . . . . . . . . . . . 190
12.9.3 Effect of the TGA Heating Rate (birch_tga) . . . . . . . . . . . . . . . . . . . 192
12.9.4 TGA for a Charring Sample (Needle_TGA) . . . . . . . . . . . . . . . . . . . . 192
12.9.5 TGA of various Mediterranean vegetation (Morvan_TGA) . . . . . . . . . . . . . 198
12.9.6 Interpreting MCC Data (cable_XX_mcc) . . . . . . . . . . . . . . . . . . . . . 199
12.10 Cone Data Extrapolation Model (cone_demo_2) . . . . . . . . . . . . . . . . . . . . . . 202
12.11 Melting Ice Cube (ice_cube) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202

13 Lagrangian Particles 203

13.1 Particle and Droplet Drag . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
13.1.1 Dragless Particle (cannon_ball) . . . . . . . . . . . . . . . . . . . . . . . . . 203
13.1.2 Particles Swept Up in a Free Stream (fluid_part_mom) . . . . . . . . . . . . . 203
13.1.3 Particles Slowing Down a Gas Flow (particle_drag) . . . . . . . . . . . . . . 204
13.1.4 Stationary Spherical Particles in a Duct (sphere_drag) . . . . . . . . . . . . . 204
13.1.5 Drag Caused by Vegetation (vegetation_drag) . . . . . . . . . . . . . . . . . 207
13.1.6 Mesh Screen in a Duct (screen_drag) . . . . . . . . . . . . . . . . . . . . . . 209
13.1.7 Flow through Porous Media (porous_media) . . . . . . . . . . . . . . . . . . . 210
13.1.8 Particle Terminal Velocity (terminal_velocity) . . . . . . . . . . . . . . . . 210
13.1.9 Particle Trajectory (flat_fire) . . . . . . . . . . . . . . . . . . . . . . . . . . 211

xv
 13.2 Solid Particle Mass and Energy Transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
13.2.1 Radiation Absorption by Vegetation . . . . . . . . . . . . . . . . . . . . . . . . 213
13.2.2 Heating a Metal Sphere via Radiation and Convection . . . . . . . . . . . . . . . 213
13.3 Water Droplet Evaporation (water_evaporation) . . . . . . . . . . . . . . . . . . . . 215
13.3.1 Case 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
13.3.2 Case 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
13.3.3 Case 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
13.3.4 Case 4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
13.3.5 Case 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
13.3.6 Case 6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
13.3.7 Case 7 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
13.3.8 Case 8 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
13.4 Water Droplets on Solid Surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222
13.4.1 Basic Flow and Mass Conservation (geom_sprk_mass) . . . . . . . . . . . . . 222
13.4.2 Obstructions of Different Orientations (geom_particle_cascade_2) . . . . . 223
13.4.3 Adiabatic Surface Temperature (adiabatic_surf_temp_spray) . . . . . . . 223
13.5 Flow Rate and Pipe Networks (flow_rate_2) . . . . . . . . . . . . . . . . . . . . . . . 224
13.6 Turbulent Dispersion (random_walk) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224

14 Heating, Ventilating, and Air Conditioning (HVAC) 227

14.1 Duct Flow Losses (HVAC_flow_loss) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
14.2 Node Flow Losses (HVAC_tee_loss_1, HVAC_tee_loss_2) . . . . . . . . . . . . . . . 227
14.3 HVAC Mass Conservation (HVAC_mass_conservation) . . . . . . . . . . . . . . . . . 228
14.4 HVAC Energy Conservation and Pressure (HVAC_energy_pressure) . . . . . . . . . . . 229
14.5 HVAC Dampers (HVAC_damper) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
14.6 HVAC Filters (HVAC_filter) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
14.7 HVAC Fans (fan_test) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232
14.8 HVAC Operation Point Single Fan (qfan_test) . . . . . . . . . . . . . . . . . . . . . . . 233
14.9 HVAC Operation Point Multiple Fans (qfan_multi) . . . . . . . . . . . . . . . . . . . . 234
14.10 Leakage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
14.10.1 Single Zone Leakage (leak_test) . . . . . . . . . . . . . . . . . . . . . . . . 235
14.10.2 Multiple Zone Leakage (leak_test_2) . . . . . . . . . . . . . . . . . . . . . . 236
14.10.3 Leakage through Immersed Boundary Wall (leak_test_3) . . . . . . . . . . . 236
14.10.4 Leakage through Immersed Boundary Compartment (leak_test_4) . . . . . . 237
14.10.5 Localized Leakage (leak_enthalpy) . . . . . . . . . . . . . . . . . . . . . . . 237
14.10.6 Multiple Zone, Multiple Mesh Leakage (zone_shape_2) . . . . . . . . . . . . 238
14.10.7 Leak Pressure Exponent (HVAC_leak_exponent) . . . . . . . . . . . . . . . . 239
14.11 ASHRAE Sample Problem 7
(ashrae_7_fixed_flow, ashrae_7_quadratic, ashrae_7_table) . . . . . . . . . 240
14.12 HVAC transient mass transport . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241
14.12.1 Convergence study for HVAC mass transport (HVAC_convergence_study_x) . 241
14.12.2 Transport time (HVAC_mass_transport) . . . . . . . . . . . . . . . . . . . . . 242
14.12.3 Duct definition sensitivity analysis (HVAC_mass_transport_x) . . . . . . . . 243
14.12.4 Mass conservation 1 (HVAC_mass_transport_conservation_1) . . . . . . . 243
14.12.5 Mass conservation 2 (HVAC_mass_transport_conservation_2) . . . . . . . 244
14.12.6 Mass conservation 3 (HVAC_mass_transport_conservation_3) . . . . . . . 244
14.12.7 Transient transport branching ducts (HVAC_mass_transport_branch) . . . . 245
14.12.8 Transient transport combining ducts (HVAC_mass_transport_combine) . . . 246

xvi
 14.12.9 Energy conservation and pressure (HVAC_mass_transport_energy) . . . . . 247

15 Wildland Fire Spread 249

15.1 Level Set Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249
15.1.1 Elliptical Fire Spread . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249

16 Unstructured Geometry (Beta) 251

16.1 Verification of setup checks on broken geometries . . . . . . . . . . . . . . . . . . . . . . 251
16.2 Simple geometry and properties definitions . . . . . . . . . . . . . . . . . . . . . . . . . . 252
16.2.1 Simple Geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252
16.2.2 Cuboid Obstacle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252
16.2.3 Self-generated Sphere Geometry . . . . . . . . . . . . . . . . . . . . . . . . . . 253
16.2.4 Terrain from Digital Elevation Map . . . . . . . . . . . . . . . . . . . . . . . . . 256
16.2.5 Adding Texture Map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256
16.2.6 Adding Multiple Texture Maps . . . . . . . . . . . . . . . . . . . . . . . . . . . 257
16.2.7 Adding Texture Maps with Different Methods . . . . . . . . . . . . . . . . . . . 258
16.3 Computational Geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259
16.3.1 Cut-cell Definition Around Immersed Cone: Split Cells, Piercing and Alignment 259
16.4 Cut-cell Method for Scalar Transport - Stress Method . . . . . . . . . . . . . . . . . . . . 261
16.4.1 Temporal Error Analysis for Variable Density Projection . . . . . . . . . . . . . 261
16.4.2 Variable-Density Manufactured Solution . . . . . . . . . . . . . . . . . . . . . . 261
16.4.3 Stress Method on Manufactured Solution Around Rotated Cube . . . . . . . . . 262
16.4.4 Poiseuille Flow in Channel Created by Two GEOMs . . . . . . . . . . . . . . . 266
16.5 Geometry Tests Involving Heat and Mass Transfer . . . . . . . . . . . . . . . . . . . . . . 267
16.5.1 Radiating Sphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267
16.5.2 Leaking Sphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 268
16.6 Restart Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 268

17 Outputs 271
17.1 Statistical Quantities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271
17.1.1 RMS, Co-Variance, and Cross-Correlation (rms_cov_corr) . . . . . . . . . . . 271
17.1.2 RMS as a Running Average (rms_example) . . . . . . . . . . . . . . . . . . . 273
17.1.3 Favre Average (shunn3_FavreZ) . . . . . . . . . . . . . . . . . . . . . . . . . 273
17.2 Interpolation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274

Bibliography 275

A Summary of Verification Results 283

B Installation Testing 303

C Unstructured Geometry integral conservation tests 305

xvii
List of Figures

3.1 Initial and final states of Navier-Stokes solution . . . . . . . . . . . . . . . . . . . . . . . 12

3.2 Velocity time history, qualitative convergence . . . . . . . . . . . . . . . . . . . . . . . . 13
3.3 Navier-Stokes convergence study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.4 Pulsating wave convergence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.5 Stationary compression wave convergence . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.6 Multi-mesh square wave transport . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.7 Solid body rotation scalar transport . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.8 Solid body rotation square wave convergence . . . . . . . . . . . . . . . . . . . . . . . . 21
3.9 Solid body rotation cosine wave convergence . . . . . . . . . . . . . . . . . . . . . . . . 22
3.10 The saad temporal order test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
3.11 Evolution of the variable-density manufactured solution . . . . . . . . . . . . . . . . . . . 27
3.12 Convergence of the manufactured solution . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.13 The hot_layer_collapse test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
3.14 Vector field of a two-dimensional vortex . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
3.15 Velocity in the vort2d test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.16 Convergence study for the vort2d test case . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.17 Time history of velocity at a point showing qualitative convergence . . . . . . . . . . . . . 33
3.18 Results of the bound_test_1 case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.19 Results of the bound_test_2 case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.20 Results of the helium_1d_isothermal cases . . . . . . . . . . . . . . . . . . . . . . . 35
3.21 Results of the helium_air_jet_floor case . . . . . . . . . . . . . . . . . . . . . . . . 35
3.22 Results of the mwtest_cfl case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
3.23 Results of the tmp_lower_limit cases . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
3.24 Results of the helium_1d_const_gamma cases . . . . . . . . . . . . . . . . . . . . . . 38
3.25 Results of the fire_const_gamma case . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.26 Sample case water_evap_1_const_gamma . . . . . . . . . . . . . . . . . . . . . . . . 40
3.27 Sample case tunnel_const_gamma . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
3.28 Sample case tunnel_linear_cp . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

4.1 Initial and final states for isotropic turbulence . . . . . . . . . . . . . . . . . . . . . . . . 46

4.2 Integrated and spectral energy decay for CBC experiment using constant Smagorinsky . . 47
4.3 Dynamic Smagorinsky coefficient . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
4.4 Integrated and spectral energy decay for CBC experiment using dynamic Smagorinsky . . 49
4.5 Integrated and spectral energy decay for CBC experiment using Deardorff’s model . . . . 50
4.6 Integrated and spectral energy decay for CBC experiment using Vreman’s model . . . . . 51
4.7 Integrated and spectral energy decay for CBC experiment using the WALE model . . . . . 52
4.8 Jet centerline velocity decay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

5.1 Convergence for laminar (Poiseuille) flow in a 2-D channel . . . . . . . . . . . . . . . . . 56

xix
5.2 LES of square duct flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
5.3 FDS Moody chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
5.4 Blasius profile and convergence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
5.5 Pohlhausen similarity solution and vertical profile . . . . . . . . . . . . . . . . . . . . . . 60
5.6 Near-wall grid resolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
5.7 Mean velocity and temperature profiles for heated channel flow . . . . . . . . . . . . . . . 62
5.8 Forced convection flat plate boundary layer . . . . . . . . . . . . . . . . . . . . . . . . . 63
5.9 Nusselt correlation local turbulent flat plate . . . . . . . . . . . . . . . . . . . . . . . . . 65
5.10 Nusselt correlation Rayleigh-Bénard convection . . . . . . . . . . . . . . . . . . . . . . . 66
5.11 Nusselt correlation natural convection vertical enclosure . . . . . . . . . . . . . . . . . . . 67
5.12 Nusselt correlation free convection from a sphere . . . . . . . . . . . . . . . . . . . . . . 68
5.13 Bulk velocity ribbed_channel test case . . . . . . . . . . . . . . . . . . . . . . . . . . 69
5.14 The ribbed_channel test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
5.15 Mean and RMS velocity profiles for ribbed square duct flow . . . . . . . . . . . . . . . . 70
5.16 Bulk velocity ribbed_channel_geom test case . . . . . . . . . . . . . . . . . . . . . . 71
5.17 Mean and RMS velocity profiles for ribbed channel geom . . . . . . . . . . . . . . . . . . 72

6.1 Monin-Obukhov velocity profiles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73

6.2 Schoenberg Ekman Layers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75

7.1 The simple_duct test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77

7.2 The species_conservation test cases . . . . . . . . . . . . . . . . . . . . . . . . . . 79
7.3 The mass_flux_wall test cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
7.4 The mass_balance_gas_volume test case . . . . . . . . . . . . . . . . . . . . . . . . . 80
7.5 The mass_balance_reac test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
7.6 The particle_offgas test cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
7.7 The realizable_mass_fractions test case . . . . . . . . . . . . . . . . . . . . . . . 83
7.8 The geom_mass_file_test test case . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
7.9 The energy_budget_adiabatic_walls test case . . . . . . . . . . . . . . . . . . . . 86
7.10 The energy_budget_cold_walls test case . . . . . . . . . . . . . . . . . . . . . . . . 86
7.11 The isentropic test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
7.12 The isentropic2 test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
7.13 The energy_budget_dns test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
7.14 The energy_budget_tmix test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
7.15 The energy_budget_combustion test case . . . . . . . . . . . . . . . . . . . . . . . . 90
7.16 The energy_budget_solid test case . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
7.17 The mass_heat_wall_device_test results . . . . . . . . . . . . . . . . . . . . . . . 91
7.18 The energy_budget_particles test case . . . . . . . . . . . . . . . . . . . . . . . . . 92

8.1 The symmetry_test cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94

8.2 The symmetry_test_2 case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
8.3 The cyl_test test cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
8.4 Images of the velocity_bc_test test case . . . . . . . . . . . . . . . . . . . . . . . . 96
8.5 The velocity_bc_test test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
8.6 The divergence_test cases 1 and 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
8.7 The divergence_test case 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
8.8 The lapse_rate test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
8.9 The random_meshes test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99

xx
8.10 The overlapping init_overlap test case . . . . . . . . . . . . . . . . . . . . . . . . . . 101
8.11 The overlapping init_overlap test case . . . . . . . . . . . . . . . . . . . . . . . . . . 102
8.12 Results of the hole/no hole test case . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
8.13 Result of the race_test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
8.14 Result of the layer test cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
8.15 Result of the device_restart test case . . . . . . . . . . . . . . . . . . . . . . . . . . 106

9.1 The plate_view_factor test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108

9.2 Results of the geom_rad test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
9.3 Results of the geom_rad_2 test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
9.4 Radiation inside a box geometry. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
9.5 Incident heat flux. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
9.6 The radiating_polygon test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
9.7 Radiation from hot spheres . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
9.8 Radiation absorption by liquid droplets . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
9.9 The particle_isotropic_radi and particle_anisotropic_radi test cases . . . 114
9.10 Radiation attenuation by Lagrangian particles . . . . . . . . . . . . . . . . . . . . . . . . 115
9.11 The radiation_gas-veg_consistency test cases . . . . . . . . . . . . . . . . . . . . 116
9.12 Transient thermocouple heating . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
9.13 Radiation/thermocouple test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
9.14 The target_test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
9.15 The radiation_shield test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
9.16 The radiation_gas_panel case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119

10.1 Results of the species_props cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122

10.2 Test of 30 % relative humidity at 70 ◦ C . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
10.3 Yield of combustion products for alternative reactions . . . . . . . . . . . . . . . . . . . . 123
10.4 Species evolution in a 1-step methane EDC reaction . . . . . . . . . . . . . . . . . . . . . 124
10.5 Temperature and pressure evolution in a 1-step methane EDC reaction . . . . . . . . . . . 125
10.6 Species evolution in a 1-step propane EDC reaction . . . . . . . . . . . . . . . . . . . . . 125
10.7 Temperature and pressure evolution in a 1-step propane EDC reaction . . . . . . . . . . . 125
10.8 Species evolution in 1-Step methane EDC reaction . . . . . . . . . . . . . . . . . . . . . 126
10.9 Temperature and pressure evolution in a 1-step Methane EDC reaction . . . . . . . . . . . 126
10.10 Species evolution in 2-step methane EDC reaction . . . . . . . . . . . . . . . . . . . . . . 127
10.11 Results of the reactionrate_series_reaction case . . . . . . . . . . . . . . . . . . 127
10.12 Species evolution in two parallel EDC reactions . . . . . . . . . . . . . . . . . . . . . . . 128
10.13 Results of the reactionrate_lumped_two_species case . . . . . . . . . . . . . . . . 128
10.14 Species evolution in a 0-order 1-step finite rate reaction . . . . . . . . . . . . . . . . . . . 129
10.15 Species evolution in a 2-order 1-step finite rate reaction . . . . . . . . . . . . . . . . . . . 130
10.16 Species evolution in a 1.75-order 2-step finite rate reaction . . . . . . . . . . . . . . . . . 131
10.17 Species evolution in a 1.75-order 2-step reversible finite rate reaction . . . . . . . . . . . . 131
10.18 Species evolution in an equilibrium case . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
10.19 Temperature and pressure evolution for equilibrium case . . . . . . . . . . . . . . . . . . 133
10.20 Species evolution in a fast-slow reaction scheme . . . . . . . . . . . . . . . . . . . . . . . 134
10.21 Results of the hrrpuv_reac test cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
10.22 Mixture fraction state relations for species (burke_schumann) . . . . . . . . . . . . . . . 138
10.23 Mixture fraction state relations for temperature (burke_schumann) . . . . . . . . . . . . 138
10.24 HoC_Ideal and HoC_NonIdeal test cases . . . . . . . . . . . . . . . . . . . . . . . . . 139

xxi
10.25 Result of the 1_step_2_step_compare test case . . . . . . . . . . . . . . . . . . . . . . . . 139
10.26 The FED and FIC values in FED_FIC test cases . . . . . . . . . . . . . . . . . . . . . . . 142
10.27 The FED and FIC values in FED_FIC_SMIX test cases . . . . . . . . . . . . . . . . . . . 143
10.28 The FED value in FED_CO_HCN case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
10.29 The FED value in FED_CO_HCN case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
10.30 Smoke detector chamber obscuration for smoke_detector case . . . . . . . . . . . . . . 145
10.31 Gas phase soot mass fractions and wall surface densities for gravitational deposition . . . . 146
10.32 Gas phase soot densities and wall surface densities for thermophoretic deposition . . . . . 147
10.33 Wall surface densities for the aerosol_turbulent_deposition case . . . . . . . . . . 148
10.34 Soot mass conservation for propane_flame_deposition case . . . . . . . . . . . . . . 149
10.35 Gas phase soot mass fractions the aerosol_agglomeration cases . . . . . . . . . . . . 150
10.36 Gas phase soot mass fractions the aerosol_agglomeration_2 cases . . . . . . . . . . 150
10.37 Gas phase soot mass fractions the aerosol_scrubbing cases . . . . . . . . . . . . . . . 151
10.38 Results of the soot_oxidation_wall case . . . . . . . . . . . . . . . . . . . . . . . . 152
10.39 Results of the condensation_1 (top) and condensation_2 (bottom) cases . . . . . . . 153
10.40 Results of the wall_cond case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
10.41 Radiative heat flux for the condensation_3 case . . . . . . . . . . . . . . . . . . . . . 154

11.1 The heat_conduction test cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156

11.2 The heat_conduction_kc test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
11.3 The convective_cooling test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
11.4 Convergence study for the convective_cooling test case . . . . . . . . . . . . . . . . 158
11.5 The thermocouples test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
11.6 The insulated_steel test cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
11.7 The back_wall_test test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
11.8 The back_wall_test_2 test cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
11.9 The adiabatic_con_flux and adiabatic_net_flux test cases . . . . . . . . . . . . 162
11.10 The ht3d_energy_conservation test case . . . . . . . . . . . . . . . . . . . . . . . . 163
11.11 Additional ht3d_energy_conservation test cases . . . . . . . . . . . . . . . . . . . . 164
11.12 The ht3d_beam_heating test cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
11.13 The ht3d_mass_conservation test case . . . . . . . . . . . . . . . . . . . . . . . . . 165
11.14 The ht3d_slab test cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
11.15 The ibeam-ht3d test case: contours . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
11.16 The ibeam-ht3d test case: time history . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
11.17 The ht3d_network test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
11.18 The ht3d_sphere temperature profiles and convergence rate . . . . . . . . . . . . . . . 169

12.1 The surf_mass_vent_char test cases . . . . . . . . . . . . . . . . . . . . . . . . . . . 172

12.2 The surf_mass_vent_nonchar test cases . . . . . . . . . . . . . . . . . . . . . . . . . 173
12.3 The surf_mass_part_char test cases . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
12.4 The surf_mass_part_nonchar test cases . . . . . . . . . . . . . . . . . . . . . . . . . 175
12.5 The surf_mass_two_species test cases . . . . . . . . . . . . . . . . . . . . . . . . . 176
12.6 The part_baking_soda test cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
12.7 The matl_e_cons test cases 1-6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
12.8 The matl_e_cons test cases 7-9 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
12.9 The cell_burn_away test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
12.10 The surf_mass_vent_liquid_fuel test cases . . . . . . . . . . . . . . . . . . . . . . 184
12.11 The water_pool test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185

xxii
12.12 The emissivity test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
12.13 The shrink_swell test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
12.14 The enthalpy test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
12.15 The two_step_solid_reaction test case . . . . . . . . . . . . . . . . . . . . . . . . . 189
12.16 Results of a TGA analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
12.17 An example of TGA data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
12.18 Needle_TGA calibration curves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
12.19 Results of pine Needle_TGA calibration exercise, Cases A and B . . . . . . . . . . . . . . 194
12.20 Results of pine Needle_TGA calibration exercise, Cases C and D . . . . . . . . . . . . . . 195
12.21 Results of pine Needle_TGA calibration exercise, Cases E and F . . . . . . . . . . . . . . 196
12.22 Results of pine Needle_TGA calibration exercise, Cases G and H . . . . . . . . . . . . . . 197
12.23 Results of Morvan_TGA analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198
12.24 Results of a micro-calorimetry analysis of cable materials . . . . . . . . . . . . . . . . . . 200
12.25 Cone data extrapolation model (the cone_demo_2 case) . . . . . . . . . . . . . . . . . . 202
12.26 Check of the particle production by surface reaction feature (the ice_cube case) . . . . . 202

13.1 The cannon_ball case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204

13.2 Test of fluid-particle momentum transfer for nonstationary particles . . . . . . . . . . . . . 205
13.3 Sample cases particle_drag . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206
13.4 Sample case of particle_drag sum . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
13.5 The sphere_drag cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
13.6 The vegetation_drag test cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
13.7 Sample cases screen_drag . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
13.8 Results of the porous_media test case . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
13.9 Convergence analysis of the terminal_velocity case . . . . . . . . . . . . . . . . . . 211
13.10 Results of the flat_fire test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
13.11 Results of the vegetation_absorb test case . . . . . . . . . . . . . . . . . . . . . . . . 213
13.12 The particle_heating test cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214
13.13 Sample case water_evaporation_1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216
13.14 Sample case water_evaporation_2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217
13.15 Sample case water_evaporation_3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218
13.16 Sample case water_evaporation_4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
13.17 Sample case water_evaporation_5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
13.18 Sample case water_evaporation_6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
13.19 Sample case water_evaporation_6a . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
13.20 Sample case water_evaporation_7 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222
13.21 Sample case water_evaporation_8 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222
13.22 The geom_sprk_mass test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
13.23 The geom_particle_cascade_2 test case . . . . . . . . . . . . . . . . . . . . . . . . . 223
13.24 The adiabatic_surface_temperature_spray test case . . . . . . . . . . . . . . . . 224
13.25 The flow_rate_2 case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224
13.26 Turbulent dispersion of tracer particles . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225

14.1 The HVAC_flow_loss test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227

14.2 The HVAC_tee_loss test cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228
14.3 The HVAC_mass_conservation case . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
14.4 The HVAC_energy_pressure case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
14.5 Example case HVAC_damper . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231

xxiii
14.6 Results for the HVAC_filter sample case . . . . . . . . . . . . . . . . . . . . . . . . . . 232
14.7 Results of the fan_test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
14.8 Results of the qfan_test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234
14.9 Results of the qfan_multi case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
14.10 The leak_test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236
14.11 The leak_test_2 test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236
14.12 The leak_test_3 test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
14.13 The leak_test_4 test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238
14.14 The leak_enthalpy test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238
14.15 The zone_shape_2 test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239
14.16 The HVAC_leak_exponent cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240
14.17 Results of ashrae7 test cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241
14.18 HVAC_mass_transport_conv_x test case result . . . . . . . . . . . . . . . . . . . . . 242
14.19 HVAC_mass_transport_conv_x L1 and L2 error results . . . . . . . . . . . . . . . . . 242
14.20 HVAC_mass_transport test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243
14.21 HVAC_mass_transport_x test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243
14.22 HVAC_mass_transport_conservation_1 test case . . . . . . . . . . . . . . . . . . . 244
14.23 HVAC_mass_transport_conservation_1 test case . . . . . . . . . . . . . . . . . . . 244
14.24 HVAC_mass_transport_conservation_2 test case . . . . . . . . . . . . . . . . . . . 245
14.25 HVAC_mass_transport_conservtation_3 test case . . . . . . . . . . . . . . . . . . 245
14.26 HVAC_mass_transport_branch test case . . . . . . . . . . . . . . . . . . . . . . . . . 246
14.27 HVAC_mass_transport_branch test case . . . . . . . . . . . . . . . . . . . . . . . . . 246
14.28 HVAC_mass_transport_energy ∆ p and enthalpy results . . . . . . . . . . . . . . . . 247

15.1 Level set fire front ellipse shape . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250

15.2 Level set fire front ellipse error . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250

16.1 Simple triangulated geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252

16.2 Block geometry using XB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253
16.3 Recursive sphere discretization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254
16.4 Latitude/longitude sphere discretization . . . . . . . . . . . . . . . . . . . . . . . . . . . 255
16.5 Defining digital elevation maps using ZVALS . . . . . . . . . . . . . . . . . . . . . . . . . 256
16.6 Applying a texture map to a GEOM surface . . . . . . . . . . . . . . . . . . . . . . . . . . 257
16.7 Applying multiple texture maps to a GEOM surface . . . . . . . . . . . . . . . . . . . . . . 258
16.8 Applying texture maps to a GEOM surface, different methods . . . . . . . . . . . . . . . . . 259
16.9 Generating polyhedral cutcells around an immersed cone . . . . . . . . . . . . . . . . . . 260
16.10 The saad CC temporal order test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262
16.11 The Shunn CC accuracy order test case . . . . . . . . . . . . . . . . . . . . . . . . . . . 263
16.12 Reference frame for a rotated immersed body . . . . . . . . . . . . . . . . . . . . . . . . 264
16.13 The Rotated Cube CC accuracy order test case . . . . . . . . . . . . . . . . . . . . . . 265
16.14 The Rotated Cube CC accuracy order test case . . . . . . . . . . . . . . . . . . . . . . 265
16.15 The Rotated Cube CC accuracy order test case . . . . . . . . . . . . . . . . . . . . . . 266
16.16 Poiseuille flow between two unstructured geometries . . . . . . . . . . . . . . . . . . . . 267
16.17 Poiseuille flow convergence with unstructured geometry . . . . . . . . . . . . . . . . . . . 267
16.18 The sphere_radiate test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 268
16.19 The sphere_leak test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 268
16.20 The geom_restart test case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269

xxiv
17.1 Sample case rms_cov_corr . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272
17.2 Sample case rms_example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
17.3 Favre average test shunn3_FavreZ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274
17.4 Test of devc_interpolation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274

xxv
List of Tables

3.1 Parameters for manufactured solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

3.2 Shunn temporal error norms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.3 Helium constant specific heat ratio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

5.1 Case matrix for Moody chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

5.2 Forced convection boundary layer resolution . . . . . . . . . . . . . . . . . . . . . . . . . 63
5.3 Free convection correlation parameters for horizontal plates . . . . . . . . . . . . . . . . . 66
5.4 Free convection correlation parameters for a vertical enclosure . . . . . . . . . . . . . . . 67

6.1 Schoenberg Ekman Layer parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74

10.1 Arrhenius values for a single step CO reaction . . . . . . . . . . . . . . . . . . . . . . . . 129

10.2 Arrhenius values for a single step C3 H8 reaction . . . . . . . . . . . . . . . . . . . . . . . 129
10.3 Arrhenius values for a two-step C3 H8 reaction . . . . . . . . . . . . . . . . . . . . . . . . 130
10.4 Arrhenius values for a two-step reversible C3 H8 reaction . . . . . . . . . . . . . . . . . . 130
10.5 Arrhenius values for a two-step reversible C3 H8 reaction . . . . . . . . . . . . . . . . . . 132

12.1 Parameters for shrinking and swelling of materials . . . . . . . . . . . . . . . . . . . . . . 186

12.2 Parameters for pine needle TGA experiments . . . . . . . . . . . . . . . . . . . . . . . . 193
12.3 Parameters for kinetic constants for cable materials . . . . . . . . . . . . . . . . . . . . . 201

16.1 Conservation tests for polyhedra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261

A.1 Summary of verification cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284

B.1 FDS Installation Test Suite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 304

xxvii
Chapter 1

What is Verification?

The terms verification and validation are often used interchangeably to mean the process of checking the
accuracy of a numerical model. For many, this entails comparing model predictions with experimental
measurements. However, there is now a fairly broad-based consensus that comparing model and experiment
is largely what is considered validation. So what is verification? ASTM E 1355 [3], “Standard Guide for
Evaluating the Predictive Capability of Deterministic Fire Models,” defines verification as

The process of determining that the implementation of a calculation method accurately represents
the developer’s conceptual description of the calculation method and the solution to the calculation
method.

and it defines validation as

The process of determining the degree to which a calculation method is an accurate representation
of the real world from the perspective of the intended uses of the calculation method.

Simply put, verification is a check of the math; validation is a check of the physics. If the model predictions
closely match the results of experiments, using whatever metric is appropriate, it is assumed by most that
the model suitably describes, via its mathematical equations, what is happening. It is also assumed that the
solution of these equations must be correct. So why do we need to perform model verification? Why not
just skip to validation and be done with it? The reason is that rarely do model and measurement agree so
well in all applications that anyone would unquestionably just accept the results. Because there is inevitably
differences between model and experiment, we need to know if these differences are due to limitations or
errors in the numerical solution, or physical sub-models, or both.
Whereas model validation consists mainly of comparing predictions with measurements, as documented
for FDS in Volume 3 of the Technical Reference Guide, model verification consists of a much broader range
of activities, from checking the computer program itself to comparing calculations to analytical (exact)
solutions to considering the sensitivity of the dozens of numerical parameters. The next chapter discusses
these various activities, and the rest of the Guide is devoted mainly to comparisons of various sub-model
calculations with analytical solutions.

1
Chapter 2

Survey of Past Verification Work

This chapter documents work of the past few decades at NIST, VTT and elsewhere to verify the algorithms
within FDS.

2.1 Analytical Tests

Most realistic combustion processes, including fire, are turbulent and time-dependent. There are no closed-
form mathematical solutions for the fully-turbulent, time-dependent Navier-Stokes equations. CFD provides
an approximate solution for the non-linear partial differential equations that describe fluid flow by replacing
them with discretized algebraic equations that can be solved using a powerful computer. While there is
no general analytical solution for fully-turbulent flows, certain sub-models address phenomenoa that do
have analytical solutions, for example, one-dimensional heat conduction through a solid. These analytical
solutions can be used to test sub-models within a complex code such as FDS. The developers of FDS
routinely use such practices to verify the correctness of the coding of the model [4, 5]. Such verification
efforts are relatively simple and routine and the results may not always be published nor included in the
documentation. Examples of routine analytical testing include:

• The radiation solver has been verified for scenarios in which simple objects, like cubes or flat plates,
are positioned in simple, sealed compartments. All convective motion is turned off, the object is given
a fixed surface temperature and emissivity of one (making it a black body radiator). The heat flux to the
cold surrounding walls is recorded and compared to analytical solutions. These studies help determine
the appropriate number of solid angles to be set as the default.

• Solid objects are heated with a fixed heat flux, and the interior and surface temperatures as a function of
time are compared to analytical solutions of the one-dimensional heat transfer equation. These studies
help determine the number of nodes to use in the solid phase heat transfer model. Similar studies are
performed to check the pyrolysis models for thermoplastic and charring solids.

• Early in its development, the hydrodynamic solver that evolved to form the core of FDS was checked
against analytical solutions of simplified fluid flow phenomena. These studies were conducted at the
National Bureau of Standards (NBS)1 by Rehm, Baum and co-workers [6, 7, 8, 9]. The emphasis of
this early work was to test the stability and consistency of the basic hydrodynamic solver, especially the
velocity-pressure coupling that is vitally important in low Mach number applications. Many numerical
algorithms developed to that point in time were intended for use in high-speed flow applications, like
1 The National Institute of Standards and Technology (NIST) was formerly known as the National Bureau of Standards.

3
 aerospace. Many of the techniques adopted by FDS were originally developed for meteorological mod-
els, and as such needed to be tested to assess whether they would be appropriate to describe relatively
low-speed flow within enclosures.

• A fundamental decision made by Rehm and Baum early in the FDS development was to use a direct
(rather than iterative) solver for the pressure. In the low Mach number formulation of the Navier-Stokes
equations, an elliptic partial differential equation for the pressure emerges, often referred to as the Pois-
son equation. Most CFD methods use iterative techniques to solve the governing conservation equations
to avoid the necessity of directly solving the Poisson equation. The reason for this is that the equation
is time-consuming to solve numerically on anything but a rectilinear grid. Because FDS is designed
specifically for rectilinear grids, it can exploit fast, direct solvers of the Poisson equation, obtaining the
pressure field with one pass through the solver to machine accuracy. FDS employs double-precision
(8 byte) arithmetic, meaning that the relative difference between the computed and the exact solution
of the discretized Poisson equation is on the order of 10−12 . The fidelity of the numerical solution of
the entire system of equations is tied to the pressure/velocity coupling because often simulations can
involve hundreds of thousands of time steps, with each time step consisting of two solutions of the Pois-
son equation to preserve second-order accuracy. Without the use of the direct Poisson solver, build-up
of numerical error over the course of a simulation could produce spurious results. Indeed, an attempt
to use single-precision (4 byte) arithmetic to conserve machine memory led to spurious results simply
because the error per time step built up to an intolerable level.

2.2 Numerical Tests

Numerical techniques used to solve the governing equations within a model can be a source of error in
the predicted results. The hydrodynamic model within FDS is second-order accurate in space and time.
This means that the error terms associated with the approximation of the spatial partial derivatives by finite
differences is of the order of the square of the grid cell size, and likewise the error in the approximation of
the temporal derivatives is of the order of the square of the time step. As the numerical grid is refined, the
“discretization error” decreases, and a more faithful rendering of the flow field emerges. The issue of grid
sensitivity is extremely important to the proper use of the model and will be taken up in the next chapter.
A common technique of testing flow solvers is to systematically refine the numerical grid until the
computed solution does not change, at which point the calculation is referred to as a Direct Numerical
Solution (DNS) of the governing equations. For most practical fire scenarios, DNS is not possible on
conventional computers. However, FDS does have the option of running in DNS mode, where the Navier-
Stokes equations are solved without the use of sub-grid scale turbulence models of any kind. Because the
basic numerical method is the same for LES and DNS, DNS calculations are a very effective way to test the
basic solver, especially in cases where the solution is steady-state. Throughout its development, FDS has
been used in DNS mode for special applications. For example, FDS (or its core algorithms) have been used at
a grid resolution of roughly 1 mm to look at flames spreading over paper in a microgravity environment [10,
11, 12, 13, 14, 15], as well as “g-jitter” effects aboard spacecraft [16]. Simulations have been compared
to experiments performed aboard the US Space Shuttle. The flames are laminar and relatively simple in
structure, and the comparisons are a qualitative assessment of the model solution. Similar studies have been
performed comparing DNS simulations of a simple burner flame to laboratory experiments [17]. Another
study compared FDS simulations of a counterflow diffusion flames to experimental measurements and the
results of a one-dimensional multi-step kinetics model [18].
Early work with the hydrodynamic solver compared two-dimensional simulations of gravity currents
with salt-water experiments [19]. In these tests, the numerical grid was systematically refined until almost

4
perfect agreement with experiment was obtained. Such convergence would not be possible if there were a
fundamental flaw in the hydrodynamic solver.

2.3 Sensitivity Analysis

A sensitivity analysis considers the extent to which uncertainty in model inputs influences model output.
Model parameters can be the physical properties of solids and gases, boundary conditions, initial conditions,
etc. The parameters can also be purely numerical, like the size of the numerical grid. FDS typically requires
the user to provide several dozen different types of input parameters that describe the geometry, materials,
combustion phenomena, etc. By design, the user is not expected to provide numerical parameters besides the
grid size, although the optional numerical parameters are described in both the Technical Reference Guide
and the User’s Guide.
FDS does not limit the range of most of the input parameters because applications often push beyond
the range for which the model has been validated. FDS is used for research at NIST, VTT, and elsewhere,
and the developers do not presume to know in all cases what the acceptable range is of any parameter. Plus,
FDS solves the fundamental conservation equations and is much less susceptible to errors resulting from
input parameters that stray beyond the limits of simpler empirical models. However, the user is warned that
he/she is responsible for the prescription of all parameters. The FDS manuals can only provide guidance.
The grid size is the most important numerical parameter in the model, as it dictates the spatial and tem-
poral accuracy of the discretized partial differential equations. The heat release rate is the most important
physical parameter, as it is the source term in the energy equation. Property data, like the thermal conduc-
tivity, density, heat of vaporization, heat capacity, etc., ought to be assessed in terms of their influence on
the heat release rate. Validation studies [20] have shown that FDS predicts well the transport of heat and
smoke when the HRR is prescribed. In such cases, minor changes in the properties of bounding surfaces do
not have a significant impact on the results. However, when the HRR is not prescribed, but rather predicted
by the model using the thermophysical properties of the fuels, the model output is sensitive to even minor
changes in these properties.
The sensitivity analyses described in this chapter are all performed in basically the same way. For a given
scenario, best estimates of all the relevant physical and numerical parameters are made, and a “baseline”
simulation is performed. Then, one by one, parameters are varied by a given percentage, and the changes
in predicted results are recorded. This is the simplest form of sensitivity analysis. More sophisticated
techniques that involve the simultaneous variation of several parameters are impractical with a CFD model
because the computation time is too long and the number of parameters too large to perform the necessary
number of calculations to generate decent statistics.

2.3.1 Grid Sensitivity

The most important decision made by a model user is the size of the numerical grid. In general, the finer the
numerical grid, the better the numerical solution of the equations. FDS is second-order accurate in space and
time, meaning that halving the grid cell size will decrease the discretization error in the governing equations
by a factor of 4. Because of the non-linearity of the equations, the decrease in discretization error does not
necessarily translate into a comparable decrease in the error of a given FDS output quantity. To find out
what effect a finer grid has on the solution, model users usually perform some form of grid sensitivity study
in which the numerical grid is systematically refined until the output quantities do not change appreciably
with each refinement. Of course, with each halving of the grid cell size, the time required for the simulation
increases by a factor of 24 = 16 (a factor of two for each spatial coordinate, plus time). In the end, a
compromise is struck between model accuracy and computer capacity.

5
 Some grid sensitivity studies have been documented and published. Since FDS was first publicly re-
leased in 2000, significant changes in the combustion and radiation routines have been incorporated into the
model. However, the basic transport algorithm is the same, as is the critical importance of grid sensitivity. In
compiling sensitivity studies, only those that examined the sensitivity of routines no longer used have been
excluded.
As part of a project to evaluate the use of FDS version 1 for large scale mechanically ventilated en-
closures, Friday and Mowrer [21] performed a sensitivity analysis to find the approximate calculation time
based on varying grid sizes. A propylene fire with a nominal heat release rate was modeled in FDS. There
was no mechanical ventilation and the fire was assumed to grow as a function of the time from ignition
squared. The compartment was a 3 m by 3 m by 6.1 m space. Temperatures were sampled 12 cm below the
ceiling. Four grid sizes were chosen for the analysis: 30 cm, 15 cm, 10 cm, 7.5 cm. Temperature estimates
were not found to change dramatically with different grid dimensions.
Using FDS version 1, Bounagui et al. [22] studied the effect of grid size on simulation results to de-
termine the nominal grid size for future work. A propane burner 0.1 m by 0.1 m was modeled with a heat
release rate of 1,500 kW. A similar analysis was performed using Alpert’s ceiling jet correlation [23] that
also showed better predictions with smaller grid sizes. In a related study, Bounagui et al. [24] used FDS
to evaluate the emergency ventilation strategies in the Louis-Hippolyte-La Fontaine Tunnel in Montreal,
Canada.
Xin [25] used FDS to model a methane fueled square burner (1 m by 1 m) in the open. Engineering
correlations for plume centerline temperature and velocity profiles were compared with model predictions to
assess the influence of the numerical grid and the size of the computational domain. The results showed that
FDS is sensitive to grid size effects, especially in the region near the fuel surface, and domain size effects
when the domain width is less than twice the plume width. FDS uses a constant pressure assumption at open
boundaries. This assumption will affect the plume behavior if the boundary of the computational domain is
too close to the plume.
Ierardi and Barnett [26] used FDS version 3 to model a 0.3 m square methane diffusion burner with
heat release rate values in the range of 14.4 kW to 57.5 kW. The physical domain used was 0.6 m by 0.6 m
with uniform grid spacings of 15 cm, 10 cm, 7.5 cm, 5 cm, 3 cm, 1.5 cm for all three coordinate directions.
For both fire sizes, a grid spacing of 1.5 cm was found to provide the best agreement when compared to
McCaffrey’s centerline plume temperature and velocity correlations [27]. Two similar scenarios that form
the basis for Alpert’s ceiling jet correlation were also modeled with FDS. The first scenario was a 1 m by
1 m, 670 kW ethanol fire under a 7 m high unconfined ceiling. The planar dimensions of the computational
domain were 14 m by 14 m. Four uniform grid spacings of 50 cm, 33.3 cm, 25 cm, and 20 cm were used in
the modeling. The best agreement for maximum ceiling jet temperature was with the 33.3 cm grid spacing.
The best agreement for maximum ceiling jet velocity was for the 50 cm grid spacing. The second scenario
was a 0.6 m by 0.6 m 1,000 kW ethanol fire under a 7.2 m high unconfined ceiling. The planar dimensions
of the computational domain were 14.4 m by 14.4 m. Three uniform grid spacings of 60 cm, 30 cm, and
20 cm were used in the modeling. The results show that the 60 cm grid spacing exhibits the best agreement
with the correlations for both maximum ceiling jet temperature and velocity on a qualitative basis.
Petterson [28] also completed work assessing the optimal grid size for FDS version 2. The FDS model
predictions of varying grid sizes were compared to two separate fire experiments: The University of Canter-
bury McLeans Island Tests and the US Navy Hangar Tests in Hawaii. The first set of tests utilized a room
with approximate dimensions of 2.4 m by 3.6 m by 2.4 m and fire sizes of 55 kW and 110 kW. The Navy
Hangar tests were performed in a hangar measuring 98 m by 74 m by 15 m in height and had fires in the
range of 5.5 MW to 6.6 MW. The results of this study indicate that FDS simulations with grids of 0.15 m
had temperature predictions as accurate as models with grids as small as 0.10 m. Each of these grid sizes
produced results within 15 % of the University of Canterbury temperature measurements. The 0.30 m grid
produced less accurate results. For the comparison of the Navy Hangar tests, grid sizes ranging from 0.60 m

6
to 1.80 m yielded results of comparable accuracy.
Musser et al. [29] investigated the use of FDS for course grid modeling of non-fire and fire scenarios.
Determining the appropriate grid size was found to be especially important with respect to heat transfer at
heated surfaces. The convective heat transfer from the heated surfaces was most accurate when the near
surface grid cells were smaller than the depth of the thermal boundary layer. However, a finer grid size
produced better results at the expense of computational time. Accurate contaminant dispersal modeling re-
quired a significantly finer grid. The results of her study indicate that non-fire simulations can be completed
more quickly than fire simulations because the time step is not limited by the large flow speeds in a fire
plume.

2.3.2 Sensitivity of Large Eddy Simulation Parameters

FDS uses Deardorff’s isotropic eddy viscosity along with the gradient diffusion hypothesis to close the
subgrid stress in the LES equations. The turbulent viscosity has the form
p
µLES = ρ Cν ∆ ksgs (2.1)

where Cν is an empirical constant, ∆ is a length on the order of the size of a grid cell, and ksgs is the
unresolved (subgrid scale) kinetic energy. Related to the “turbulent viscosity” are comparable expressions
for the thermal conductivity and material diffusivity:
µLES c p µLES
kLES = ; (ρD)LES = (2.2)
Prt Sct
where Prt and Sct are the turbulent Prandtl and Schmidt numbers, respectively. Thus, Cν , Prt and Sct are
a set of empirical constants. Most FDS users simply use the default values of (0.1,0.5,0.5), but some have
explored their effect on the solution of the equations. The value of Cν = 0.1 is a theoretical value developed
from Lilly’s analysis. The value is confirmed for a canonical flow, decaying isotropic turbulence, within this
volume.
Previous versions of FDS used the constant coefficient Smagorinsky model for the eddy viscosity. In
an effort to validate FDS with some simple room temperature data, Zhang et al. [30] tried different combi-
nations of the Smagorinsky parameters, and suggested the current default values. Of the three parameters,
the viscosity coefficient is the most sensitive. Smagorinsky [31] originally proposed a value of 0.23, but re-
searchers over the past three decades have used values ranging from 0.1 to 0.23. There are also refinements
of the original Smagorinsky model [32, 33, 34] that do not require the user to prescribe the constants, but
rather generate them automatically as part of the numerical scheme.

2.3.3 Sensitivity of Radiation Parameters

Radiative heat transfer is included in FDS via the solution of the radiation transport equation for a non-
scattering gray gas, and in some limited cases using a wide band model. The equation is solved using a
technique similar to finite volume methods for convective transport, thus the name given to it is the Finite
Volume Method (FVM). There are several limitations of the model. First, the absorption coefficient for the
smoke-laden gas is a complex function of its composition, wavelength, and temperature. Because of the
simplified combustion model, the chemical composition of the smoky gases, especially the soot content, can
affect both the absorption and emission of thermal radiation. Second, the radiation transport is discretized
via approximately 100 solid angles. For targets far away from a localized source of radiation, like a growing
fire, the discretization can lead to a non-uniform distribution of the radiant energy. This can be seen in the
visualization of surface temperatures, where “hot spots” show the effect of the finite number of solid angles.
The problem can be lessened by the inclusion of more solid angles, but at a price of longer computing

7
times. In most cases, the radiative flux to far-field targets is not as important as those in the near-field, where
coverage by the default number of angles is much better.
Hostikka [35] examined the sensitivity of the radiation solver to changes in the assumed soot production,
number of spectral bands, number of control angles, and flame temperature. Some of the more interesting
findings were:

• Changing the soot yield from 1 % to 2 % increased the radiative flux from a simulated methane burner
about 15 %

• Lowering the soot yield to zero decreased the radiative flux about 20 %.

• Increasing the number of control angles by a factor of 3 was necessary to ensure the accuracy of the
model at the discrete measurement locations.

• Changing the number of spectral bands from 6 to 10 did not have a strong effect on the results.

• Errors of 100 % in heat flux were caused by errors of 20 % in absolute temperature.

The sensitivity to flame temperature and soot composition are consistent with combustion theory, which
states that the source term of the radiative transport equation is a function of the absorption coefficient mul-
tiplied by the absolute temperature raised to the fourth power. The number of control angles and spectral
bands are user-controlled numerical parameters whose sensitivities ought to be checked for each new sce-
nario. The default values in FDS are appropriate for most large scale fire scenarios, but may need to be
refined for more detailed simulations such as a low-sooting methane burner.

2.3.4 Sensitivity of Thermophysical Properties of Solid Fuels

An extensive amount of verification and validation work with FDS version 4 has been performed by Hi-
etaniemi et al. [36] at VTT, Finland. The case studies are comprised of fire experiments ranging in scale
from the cone calorimeter (ISO 5660-1) to full-scale fire tests such as the room corner test (ISO 9705).
Comparisons are also made between FDS results and data obtained in the SBI (Single Burning Item) Euro-
classification test apparatus (EN 13823) as well as data obtained in two ad hoc experimental configurations:
one is similar to the room corner test but has only partial linings and the other is a space to study fires in
building cavities.
All of the case studies involve real materials whose properties must be prescribed so as to conform to
the assumption in FDS that solids are of uniform composition backed by a material that is either cold or
totally insulating. Sensitivity of the various physical properties and the boundary conditions were tested.
Some of the findings were:

• The measured burning rates of various materials often fell between two FDS predictions in which cold
or insulated backings were assumed for the solid surfaces. FDS lacks a multi-layer solid model.

• The ignition time of upholstery is sensitive to the thermal properties of the fabric covering, but the steady
burning rate is sensitive to the properties of the underlying foam.

• Moisture content of wooden fuels is very important and difficult to measure.

• Flame spread over complicated objects, like cables laid out in trays, can be modeled if the surface area
of the simplified object is comparable to that of the real object. This suggests sensitivity not only to
physical properties, but also geometry. It is difficult to quantify the extent of the geometrical sensitivity.

8
There is little quantification of the observed sensitivities in the study. Fire growth curves can be linear to
exponential in form, and small changes in fuel properties can lead to order of magnitude changes in heat
release rate for unconfined fires. The subject is discussed in the FDS Validation Guide [20] where it is noted
in many of the studies that predicting fire growth is difficult.
Recently, Lautenberger et al. [37] developed a method to automate the process of estimating material
properties to input into FDS. The methodology involves simulating a bench-scale test with the model and
iterating via a “genetic” algorithm to obtain an optimal set of material properties for that particular item.
Such techniques are necessary because most bench-scale apparatus do not provide a complete set of thermal
properties.

2.4 Code Checking

An examination of the structure of the computer program can be used to detect potential errors in the nu-
merical solution of the governing equations. The coding can be verified by a third party either manually or
automatically with profiling programs to detect irregularities and inconsistencies [3].
At NIST, VTT, and elsewhere, FDS has been compiled and run on computers manufactured by IBM,
Hewlett-Packard, Sun Microsystems, Digital Equipment Corporation, Apple, Silicon Graphics, Dell, Com-
paq, and various other personal computer vendors. The operating systems on these platforms include Unix,
Linux, Microsoft Windows, and Mac OS X. Compilers used include Lahey Fortran, Digital Visual Fortran,
Intel Fortran, IBM XL Fortran, HPUX Fortran, Forte Fortran for SunOS, the Portland Group Fortran, and
several others. Each combination of hardware, operating system and compiler involves a slightly different
set of compiler and run-time options and a rigorous evaluation of the source code to test its compliance with
the Fortran 90 ISO/ANSI standard [38]. FDS is now compliant with the Fortran 2003 standard. Through this
process, out-dated and potentially harmful code is updated or eliminated, and often the code is streamlined
to improve its optimization on the various machines. However, simply because the FDS source code can be
compiled and run on a wide variety of platforms does not guarantee that the numerics are correct. It is only
the starting point in the process because it at least rules out the possibility that erratic or spurious results are
due to the platform on which the code is running.
Beyond hardware issues, there are several useful techniques for checking the FDS source code that have
been developed over the years. One of the best ways is to exploit symmetry. FDS is filled with thousands
of lines of code in which the partial derivatives in the conservation equations are approximated as finite
differences. It is very easy in this process to make a mistake. Consider, for example, the finite difference
approximation of the thermal diffusion term in the i jkth cell of the three-dimensional grid:
Ti+1, jk − Ti jk Ti jk − Ti−1, jk
 
1
(∇ · k∇T )i jk ≈ ki+ 1 , jk − ki− 1 , jk +
δx 2 δx 2 δx
Ti, j+1,k − Ti jk Ti jk − Ti, j−1,k
 
1
k 1 − ki, j− 1 ,k +
δ y i, j+ 2 ,k δy 2 δy
Ti j,k+1 − Ti jk Ti jk − Ti j,k−1
 
1
ki j,k+ 1 − ki j,k− 1
δz 2 δz 2 δz
which is written as follows in the Fortran source code:

DTDX = (TMP(I+1,J,K)-TMP(I,J,K))*RDXN(I)
KDTDX(I,J,K) = .5*(KP(I+1,J,K)+KP(I,J,K))*DTDX
DTDY = (TMP(I,J+1,K)-TMP(I,J,K))*RDYN(J)
KDTDY(I,J,K) = .5*(KP(I,J+1,K)+KP(I,J,K))*DTDY
DTDZ = (TMP(I,J,K+1)-TMP(I,J,K))*RDZN(K)
KDTDZ(I,J,K) = .5*(KP(I,J,K+1)+KP(I,J,K))*DTDZ

9
 DELKDELT = (KDTDX(I,J,K)-KDTDX(I-1,J,K))*RDX(I) +
(KDTDY(I,J,K)-KDTDY(I,J-1,K))*RDY(J) +
(KDTDZ(I,J,K)-KDTDZ(I,J,K-1))*RDZ(K)

This is one of the simpler constructs because the pattern that emerges within the lines of code make it fairly
easy to check. However, a mis-typing of an I or a J, a plus or a minus sign, or any of a hundred different
mistakes can cause the code to fail, or worse produce the wrong answer. A simple way to eliminate many of
these mistakes is to run simple scenarios that have perfectly symmetric initial and boundary conditions. For
example, put a hot cube in the exact center of a larger cold compartment, turn off gravity, and watch the heat
diffuse from the hot cube into the cold gas. Any simple error in the coding of the energy equation will show
up almost immediately. Then, turn on gravity, and in the absence of any coding error, a perfectly symmetric
plume will rise from the hot cube. This checks both the coding of the energy and the momentum equations.
Similar checks can be made for all of the three dimensional finite difference routines. So extensive are these
types of checks that the release version of FDS has a routine that generates a tiny amount of random noise
in the initial flow field so as to eliminate any false symmetries that might arise in the numerical solution.
The process of adding new routines to FDS is as follows: typically the routine is written by one person
who takes the latest version of the source code, adds the new routine, and writes a theoretical and numerical
description for the FDS Technical Reference Guide, plus a description of the input parameters for the FDS
User’s Guide. The new version of FDS is then tested with a number of benchmark scenarios that exercise
the range of the new parameters. Provisional acceptance of the new routine is based on several factors: (1) it
produces more accurate results when compared to experimental measurement, (2) the theoretical description
is sound, and (3) any empirical parameters are obtainable from the open literature or standard bench-scale
apparatus. If the new routine is accepted, it is added to the working version of the software and evaluated
by running the standard verification and validation test cases. Assuming that there are no intractable issues
that arise during the testing period, the new routine eventually becomes part of the release version of FDS.
Even with all the code checking, it is still possible for errors to go unnoticed. One remedy is the fact that
the source code for FDS is publicly released. Although it consists of on the order of 100,000 lines of Fortran
statements, various independent researchers have been able to work with it, add enhancements needed for
very specific applications or for research purposes, and report back to the developers bugs that have been
detected. The source code is organized into 27 separate files, each containing subroutines related to a partic-
ular feature of the model, like the mass, momentum, and energy conservation equations, sprinkler activation
and sprays, the pressure solver, etc. The lengthiest routines are devoted to input, output and initialization.
Most of those working with the source code do not concern themselves with these lengthy routines but
rather focus on the finite-difference algorithm contained in a few of the more important files. Most serious
errors are found in these files, for they contain the core of the algorithm. The external researchers provide
feedback on the organization of the code and its internal documentation, that is, comments within the source
code itself. Plus, they must compile the code on their own computers, adding to its portability.

10
Chapter 3

The Basic Flow Solver

In this chapter we present test cases aimed at exercising the advective, pressure, and viscous terms, as well
as the time integration for non-reacting flows.

3.1 Analytical Solution to the Navier-Stokes Equation (ns2d)

In this section we present an analytical solution that is useful for confirming the convergence rates of the
truncation errors in the discretization of the terms in the governing equations. Consider the 2-D incompress-
ible Navier-Stokes equations
∂u
+ u · ∇u = −∇p + ν∇2 u , (3.1)
∂t
where the velocity is given by u = [u, w]T , and the kinematic viscosity and pressure are denoted ν and p,
respectively. An analytical solution of these equations is given by [39]

u(x, z,t) = 1 − A cos(x − t) sin(z − t) e−2ν t , (3.2)

−2ν t
w(x, z,t) = 1 + A sin(x − t) cos(z − t) e , (3.3)

A2
p(x, z,t) = − [cos(2(x − t)) + cos(2(z − t))] e−4ν t . (3.4)
4
Here, A represents an arbitrary amplitude and is assumed to take a value of 2 in this example. Note that this
solution satisfies continuity for all time,
∇·u = 0, (3.5)
is spatially periodic on an interval 2π in each direction, and is temporally periodic on 2π if ν = 0; otherwise,
the solution decays exponentially. Below we present two series of tests which demonstrate the second-order
accuracy of the FDS numerical scheme and thus provide a strong form of code verification for the advective
and viscous terms which are exercised.
The physical domain of the problem is a square of side L = 2π. The grid spacing is uniform δ x = δ z =
L/N in each direction with N = {8, 16, 32, 64} for each test series. The staggered grid locations are denoted
xi = i δ x and zk = k δ z, and the cell centers are marked by an overbar, x̄i = xi − δ x/2 and z̄k = zk − δ z/2.
First, we present qualitative results for the case in which ν = 0. Thus, only the advective discretization
and the time integration are being tested. Figure 3.1 shows the initial and final (t = 2π) numerical solution
for the case N = 64. As mentioned, with ν = 0 the solution is periodic in time and this figure demonstrates
that, as should be the case, the FDS numerical solution is unaltered after one flow-through time.

11
 Figure 3.1: Initial and final states of the u-component of velocity.

Next, in Fig. 3.2, we show time histories of the u-component of velocity at the center of the domain
for the case in which ν = 0.1. It is clearly seen that the FDS solution (dashed line) converges to the
analytical solution (solid line). Note that the analytical solution is evaluated at the same location as the
FDS staggered grid location for the u-component of velocity, (xN/2 , z̄N/2 ), which is different in each case,
N = {8, 16, 32, 64}.
Figure 3.3 is the key quantitative result of this verification test. In this figure we plot the rms error, εrms ,
in the u-component of velocity against the grid spacing. The error is defined by
s
1 M h m i2
εrms ≡ ∑ Ui j − u(xi , z̄k ,tm ) , (3.6)
M m=1

where M is the number of time steps and k is the time step index. The spatial indices are (i = N/2, k = N/2)
and Uikj represents the FDS value for the u-component of velocity at the staggered storage location for cell
(i, j) at time step m; u(xi , z̄k ,tm ) is the analytical solution for the u-component at the corresponding location
in space and time. The figure confirms that the advective terms, the viscous terms, and the time integration
in the FDS code are convergent and second-order accurate.

12
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release

Analytical (u-vel) Analytical (u-vel)

Velocity (ns2d_8_nupt1) FDS (UVEL)
Velocity (ns2d_16_nupt1) FDS (UVEL)
2 2
Velocity (m/s)

Velocity (m/s)
1.5 1.5

1 1

0.5 0.5

0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7
Time (s) Time (s)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release

Analytical (u-vel) Analytical (u-vel)

Velocity (ns2d_32_nupt1) FDS (UVEL)
Velocity (ns2d_64_nupt1) FDS (UVEL)
2 2
Velocity (m/s)

Velocity (m/s)
1.5 1.5

1 1

0.5 0.5

0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7
Time (s) Time (s)

Figure 3.2: Time history of the u-component of velocity half a grid cell below the center of the domain for a range of
grid resolutions. The domain is a square of side L = 2π m. The N × N grid is uniform. Progressing from left to right
and top to bottom we have resolutions N = {8, 16, 32, 64} clearly showing convergence of the FDS numerical solution
(dashed line) to the analytical solution (solid line). The case is run with constant properties, ρ = 1 kg/m3 and µ = 0.1
kg/m/s, and a Courant-Friedrichs-Lewy (CFL) of 0.25.
FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
10 0 10 0
Convergence, Inviscid Case (ns2d) Convergence, Viscous Case (ns2d)

10 -1 10 -1
RMS Error (m/s)

RMS Error (m/s)

10 -2 10 -2
O(?x) O(?x)
O(?x²) O(?x²)
FDS (RMS Error) FDS (RMS Error)
10 -3 10 -3
-1 -1
10 10
Grid Spacing (m) Grid Spacing (m)

Figure 3.3: (Left) Convergence rate for the u-component of velocity with ν = 0 showing that the advective terms in the
FDS code are second-order accurate. The asterisks represent the RMS error in the u-component for grid spacings of
δ x = L/N where L = 2π m and N = {8, 16, 32, 64}. The solid line represents first-order accuracy and the dashed line
represents second-order accuracy. The simulation is run to a time of t = 2π s with a CFL of 0.25. The u-component
at the center of the domain is compared with the analytical solution at the same location. (Right) Same case, except
ν = 0.1, showing that the viscous terms in the FDS code are second-order accurate.

13
3.2 Analytical Solution to the Continuity Equation
Analytical solutions for primitive flow variables (density, velocity, pressure, etc.) are useful in the develop-
ment and testing of numerical schemes for CFD. For example, an analytical solution to the 2-D incompress-
ible Navier-Stokes (NS) equations is presented in Section 3.1 and is used to verify the spatial and temporal
accuracy of momentum equation.
The aim of the present work is to develop analytical solutions for the compressible NS equations that can
be used for verification of both compressible and variable-density low-Mach flow solvers. The main idea
is that, given a specified velocity field, the continuity equation can be rearranged into a linear hyperbolic
PDE for the logarithm of the density. Let ρ denote the density and let u = [u, v]T denote the velocity. The
continuity equation (conservation of mass) can then be written as

∂ ln ρ
+ u · ∇ ln ρ + ∇ · u = 0 . (3.7)
∂t
Further, for certain simple specifications of the velocity this PDE can be solved using the method of charac-
teristics.
In what follows we present 1-D and 2-D solutions to (3.7) for two basic irrotational flow fields. All the
solutions are periodic in space. The first configuration is a pulsating flow that cycles between compressing
the fluid toward the center and then the corners of the domain. In the second configuration, time periodicity
is achieved by using a constant and uniform advection velocity in combination with the compression waves.
This results in a solution with a qualitatively different character than the first.

3.2.1 Pulsating 1-D solution

We specify the velocity as
u(x,t) = B sin(x) cos(ωt) , (3.8)
where B is a constant amplitude and ω is the frequency of the compression cycle. The velocity divergence
in 1-D is then
∂u
= B cos(x) cos(ωt) . (3.9)
∂x
Let q ≡ ln ρ (this notation is used throughout this work). The 1-D continuity equation can then be
written as the following linear hyperbolic PDE:

∂q ∂q
+ B sin(x) cos(ωt) + B cos(x) cos(ωt) = 0 , (3.10)
∂t ∂x
which can be solved using the method of characteristics to obtain the solution
  

 1 + tan2 x0 [x,t]
2 exp 2B
ω sin[ωt]  B
q(x,t) = q(x0 [x,t], 0) + ln   − sin(ωt) , (3.11)
 1 + tan2 x0 [x,t]
2
 ω

where the initial position is given by

 h i  
x B
x0 (x,t) = 2 arctan tan exp − sin(ωt) . (3.12)
2 ω

Note that we have taken the initial time to be zero, as is done throughout this work.

14
3.2.2 Pulsating 2-D solution (pulsating)
There is a simple extension of the 1-D stationary wave solution to 2-D. In this section we consider the
velocity field u = [u, v]T with components and velocity divergence given by
u(x,t) = B sin(x) cos(ωt) , (3.13)
v(y,t) = B sin(y) cos(ωt) , (3.14)
∇ · u = B(cos[x] + cos[y]) cos(ωt) , (3.15)
where, again, B is a constant amplitude and ω is the compression frequency. The 2-D continuity equation
may then be written as
∂q ∂q ∂q
+ B sin(x) cos(ωt) + B sin(y) cos(ωt) + B(cos[x] + cos[y]) cos(ωt) = 0 , (3.16)
∂t ∂x ∂y
where, again, q ≡ ln ρ(x,t).
The solution to (3.16) can be obtained by adding the solutions of the following two PDEs:
∂ q1 ∂ q1
+ B sin(x) cos(ωt) + B cos(x) cos(ωt) = 0 , (3.17)
∂t ∂x
∂ q2 ∂ q2
+ B sin(y) cos(ωt) + B cos(y) cos(ωt) = 0 . (3.18)
∂t ∂y
Thus, utilizing (3.11) and (3.12), and replacing q1 (x0 [x,t], 0) + q2 (y0 [y,t], 0) with q0 (x, y,t), we find that the
solution to (3.16) is
q(x, y,t) = q0 (x, y,t)
(
)
1 + a(x0 [x,t]) exp 2B
ω sin[ωt]
+ ln
1 + a(x0 [x,t])
(
)
1 + a(y0 [y,t]) exp 2B
ω sin[ωt]
+ ln
1 + a(y0 [y,t])
2B
− sin(ωt) , (3.19)
ω
where z
a(z) ≡ tan2 , (3.20)
2
and the initial positions are given by
 h i  
x B
x0 (x,t) = 2 arctan tan exp − sin(ωt) , (3.21)
2 ω
 h i  
y B
y0 (y,t) = 2 arctan tan exp − sin(ωt) . (3.22)
2 ω
Note that the initial condition q0 (x, y,t) is restricted to cases where ∂∂qx0 is independent of y and ∂∂qy0 is
independent of x. That is, the function q0 must be additively separable.
An example of the solution to (3.19) is shown in Fig. 3.4. The initial condition for the density is specified
as ρ(x, 0) = 1 and the amplitude and frequency are set to unity, B = 1 and ω = 1. FDS is run with three
scalar transport schemes: central differencing, Superbee, and the CHARM flux limiter. The solution at
(x, y) = (3π/2, 3π/2) for successively finer grid resolutions is plotted as a time series on the left and may
be compared with the analytical solution (black line). On the right, we confirm second-order convergence
for the FDS implementation. Central differencing and the CHARM limiter outperform Superbee for this
problem because the solution is relatively smooth.

15
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
2
FDS N=16 FDS Central
FDS N=32 FDS Superbee
FDS N=64 -2 FDS CHARM
10
1.5 FDS N=128 O( x)

L2 Error (kg/m³)
Analytical Solution O( x2)
Density (kg/m³)

1
10 -4

0.5

0 10 -6
0 2 4 6 8 10 12 10 -2 10 -1 10 0
Time (s) Grid Spacing (m)

Figure 3.4: (Left) Time series of ρ at the position (x, y) = (3π/2, 3π/2) for several grid resolutions using the Superbee
limiter. (Right) Convergence plot for central differencing, Superbee, and CHARM. All schemes are second-order
accurate.

3.2.3 Stationary compression wave in 1-D

Another problem which can be solved analytically is that of a stationary compression wave. In this section
we consider a stationary compression wave combined with a constant and uniform advection velocity in
1-D. The velocity is specified to be
u(x) = c + sin(x) , (3.23)
where |c| > 1 is a constant. The 1-D continuity equation becomes

∂q ∂q
+ [c + sin(x)] + cos(x) = 0 . (3.24)
∂t ∂x
A solution to (3.24) is

q(x,t) = q(x0 [x,t], 0)

+ ln −c2 − cos bt + 2 arctan[γ(x,t)] + b sin bt + 2 arctan[γ(x,t)]


− ln −c2 − cos 2 arctan[γ(x,t)] + b sin 2 arctan[γ(x,t)] ,




(3.25)

where p
b≡ −1 + c2 > 0 , (3.26)
 
1 + c tan x0 [x,t]
2
γ(x,t) = , (3.27)
b
and      
b 1 + c tan[x/2] bt 1
x0 (x,t) = 2 arctan tan arctan − − . (3.28)
c b 2 c

3.2.4 Stationary compression wave in 2-D (compression_wave)

As with the pulsating flow, there is a simple extension of the 1-D stationary wave solution to 2-D. In this
section we consider the velocity field with components

u(x) = c1 + sin(x) , (3.29)

16
 v(y) = c2 + sin(y) . (3.30)

The continuity equation can be written as

∂q ∂q ∂q
+ [c1 + sin(x)] + [c2 + sin(y)] + cos(x) + cos(y) = 0 . (3.31)
∂t ∂x ∂y
A solution to (3.31) is

q(x, y,t) = q0 (x, y,t)

+ I1 (x,t) − I10 (x,t)
+ I2 (y,t) − I20 (y,t) , (3.32)

where

Ii (z,t) = ln −c2i − cos bit + 2 arctan[γi (z,t)] + bi sin bit + 2 arctan[γi (z,t)] ,



(3.33)
Ii0 (z,t) = ln −c2i − cos 2 arctan[γi (z,t)] + bi sin 2 arctan[γi (z,t)] ,



(3.34)

and q
bi ≡ −1 + c2i > 0 , (3.35)
 
1 + ci tan z0 [z,t]
2
γi (z,t) = , (3.36)
bi
     
bi 1 + ci tan[z/2] bit 1
z0 (z,t) = 2 arctan tan arctan − − . (3.37)
ci bi 2 ci
Note that z0 ⇒ x0 for i = 1 and z0 ⇒ y0 for i = 2. To be clear, no summation is implied over repeated indices.
Also, note that the same restrictions apply to the initial condition as did in Section 3.2.2. Namely, that ∂∂qx0
is independent of y and ∂∂qy0 is independent of x.
An example of the solution to (3.32) is shown in Fig. 3.5. In this case we set c1 = 2 and c2 = 3 to create
an asymmetry in the flow. The periodicity in time depends on the choices of c1 and c2 ; it is possible that no
periodicity exists. We have not found a case that generates a singularity. The analytical time series of the
density at the position (x, y) = (3π/2, 3π/2) is shown as the solid black line. FDS is run with the CHARM
flux limiter scheme (as this is a DNS flow field). The solution at successively finer grid resolutions is plotted
and compared with the analytical solution, demonstrating convergence of the scheme. On the right side of
the figure we demonstrate second-order convergence of the central, Superbee, and CHARM limiter schemes
in FDS.

17
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
8 10 -1
FDS N=16 Central
FDS N=32 Superbee
FDS N=64 CHARM
6 FDS N=128 MP5
-2

L2 Error (kg/m³)
Analytical Solution 10 O( x)
Density (kg/m³)

2
O( x )

10 -3
2

0 10 -4
0 2 4 6 8 10 12 10 -2 10 -1 10 0
Time (s) Grid Spacing (m)

Figure 3.5: (Left) Time series of ρ at the position (x, y) = (3π/2, 3π/2) for several grid resolutions using the CHARM
flux limiter. (Right) Convergence plot for three different scalar transport schemes in FDS: central, Superbee, and
CHARM. All schemes are second-order accurate. In addition, notice that Superbee gives the lowest error at the
coarsest resolution while CHARM gives the lowest error at higher resolution. This is one reason why Superbee is
recommended for VLES and CHARM is the default for LES and DNS.

3.3 Scalar Transport

3.3.1 Uniform Velocity Field (move_slug)
In this section we demonstrate the qualitative behavior of the Superbee flux limiter scheme for transport of a
square wave. The diffusivity is set to zero and the advecting velocity is constant and uniform u = [1.0 1.0].
The domain is the unit square with a passive scalar marker initialized to zero everywhere, except for two
‘slugs’ of mass. The first slug is set to unity over the region x × z = [0.125, 0.375] × [0.125, 0.375]. The
second slug is set to 1/2 over the region x × z = [0.500, 0.750] × [0.500, 0.750]. With the scalar bounds set
to [0,1], these slugs demonstrate both the boundedness and TVD (total variation diminishing) behavior of
the transport scheme.
This case also tests two different types of boundary conditions applied in FDS. First, the domain is
periodic and the simulation runs for one flow through time. The scalar slugs therefore ideally arrive back
to their original locations with as little diffusion as possible. Also, the domain is broken into four equally
sized meshes, each with 40 × 40 uniform cells. To increase temporal accuracy and focus on the potential
spacial error, we run the case with a CFL of 0.25. In FDS, we refer to the mesh interface as an ‘interpolated
boundary’.
The results of the test are shown visually in Fig. 3.6. The upper-left image shows the initial condition.
The black lines indicate the mesh interfaces. To the right of the initial condition we show the first slug
crossing the mesh interfaces without incurring spurious noise. The lower left image shows the final result
for the Superbee limiter. By comparison with a first-order scheme (lower right), this test case confirms that
relatively low levels of diffusion are incurred at both periodic and interpolated boundaries in FDS.

18
Figure 3.6: (Upper left) Initial condition. (Upper right) Superbee solution after 0.175 seconds showing the scalar slug
cleanly passing through the mesh interface. (Lower left) Final result for Superbee after one flow through time. (Lower
right) Result for first-order upwinding after one flow through time, illustrating the relatively low dissipation of the
Superbee scheme. Note that, though the first-order scheme is available as an option in FDS, it is presented here for
comparison purposes only. In practice, the higher order Superbee scheme is preferred for VLES (FDS default).

19
3.3.2 Solid Body Rotation Velocity Field (soborot)
In this series of tests, the velocity field is a simple solid-body rotation about the origin. The field is given by

u=z (3.38)
w = −x (3.39)

The domain is a unit square with the origin at the lower-right. Two wave forms are examined. The first
is a simple square wave with a passive scalar introduced at the z = 0 boundary plane with a mass fraction
value of 1 for 0.25 ≤ r ≤ 0.75, where r is the radial position from the origin. The second is a section of
a compressed cosine wave with the phase shifted so that the minimum aligns with the start and end of the
square wave. The initial condition is given by 12 (1 + cos(4πr)) for 0.25 ≤ r ≤ 0.75. The velocity field is
held fixed. The scalar fields are evolved for π/2 s (1/4 rotation). The scalar concentration is measured along
the upper-left diagonal as shown in Fig. 3.7. The constant velocity field is shown on the left and the solution
of the square wave for the CHARM scheme at δ x = 1/64 m is shown on the right.

Figure 3.7: (Left) Solid body rotation velocity field. (Right) Tracer concentration at time π/2 (1/4 rotation).

The steady state analytical solution to this problem is that the scalar field remains constant as a function
of radius from the origin. In Fig. 3.8 we plot the L2 Error for Godunov, CHARM, and Superbee flux limiter
schemes for the square wave initial condition. As shown in [40], the Godunov scheme shows O(δ x1/2 )
convergence for this discontinuous solution. Other schemes have smaller initial errors, but also ultimately
converge at the same rate for this problem.
For the consine wave intial condition the derivatives of the scalar field are continuous. Therefore, we
see O(δ x2 ) convergence of the CHARM and MP5 schemes, as shown in Fig. 3.9. Superbee shows smaller
error at coarse resolution, but the gradient steepening degenerates its accuracy at higher resolutions—hence
CHARM is selected for LES and DNS.

20
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
1.2 1.2
Godunov Exact CHARM Exact
n =16 n =16
1 n =32
1 n =32
n =64 n =64
Scalar Mass Fraction

Scalar Mass Fraction

0.8 n =128 0.8 n =128

0.6 0.6

0.4 0.4

0.2 0.2

0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
Radial Position (m) Radial Position (m)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
1.2
Superbee Exact
Square Wave Error
n =16
1 n =32
n =64 -1
10
Scalar Mass Fraction

0.8 n =128

L2 Error
0.6

1/2
0.4 O( x )
O( x )
Godunov
0.2 Superbee
10 -2 CHARM
0
0 0.2 0.4 0.6 0.8 1 -2 -1
10 10
Radial Position (m) Grid Spacing (m)

Figure 3.8: Solid body rotation square wave solution convergence. Scalar fields along upper-left diagonal for Godunov
(upper-left), CHARM (upper-right), and Superbee (lower-left). (Lower-right) L2 Error.

21
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
1.2 1.2
Superbee Exact CHARM Exact
n =16 n =16
1 n =32
1 n =32
n =64 n =64
Scalar Mass Fraction

Scalar Mass Fraction

0.8 n =128 0.8 n =128

0.6 0.6

0.4 0.4

0.2 0.2

0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
Radial Position (m) Radial Position (m)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
1.2
MP5 Exact Cosine Wave Error
n =16
1 n =32
n =64 -2
10
Scalar Mass Fraction

0.8 n =128

L2 Error
0.6
-3
10
0.4 O( x )
O( x 2 )
CHARM
0.2 Superbee
10 -4 MP5
0
0 0.2 0.4 0.6 0.8 1 -2 -1
10 10
Radial Position (m) Grid Spacing (m)

Figure 3.9: Solid body rotation cosine wave solution convergence. Scalar fields along upper-left diagonal for Superbee
(upper-left), CHARM (upper-right), and MP5 (lower-left). (Lower-right) L2 Error.

22
3.4 Temporal Error Analysis for a Variable-Density Projection (saad)
Tony Saad and James Sutherland, The University of Utah

This is a 1-D variable-density benchmark solution for an inviscid, non-reacting mixture of two fluids with
different densities (10:1 ratio) in a plug-flow system [41]. The diffusivity is set to zero. The governing
equations for the mixture fraction, Z, and the x component of momentum, u, are
∂Z ∂Z
= −u , (3.40)
∂t ∂x
∂u ∂ u 1 ∂ p̃
= −u − . (3.41)
∂t ∂x ρ ∂x
The equation of state gives the following relationship between the mixture fraction and the density:
1 (1 − Z) Z
= + . (3.42)
ρ ρ0 ρ1
The FDS input files may be found in
/Verification/Scalar_Analytical_Solution/saad_512_cfl*.fds.

The densities are set to ρ0 = 0.5 kg/m3 and ρ1 = 5 kg/m3 , respectively. The initial field for Z is specified as
a sine wave on a 1-D domain of length L = 2 m (see Fig. 3.10 below).
1
Z(x, 0) = (1 + sin(2πx/L)) (3.43)
2
All cases are run with Nx = 512 grid points. The effect of the spatial error is discussed below. The velocity
field is initialized to u(x, 0) = U0 = 1 m/s, which is the analytical solution for the velocity for all time,
u(x,t) = U0 . Note, however, that the velocity field is not “frozen” in the calculations, it is subject to variation
by the projection scheme (if we are doing things correctly, then the velocity will not change). The analytical
solution for the mixture fraction is given by a characteristic solution:

Z(x,t) = Z0 (x −U0t) . (3.44)

Computing Temporal Order without an Analytical Solution Despite having an analytical solution for
this case, it is difficult to use this solution to assess the temporal error in the time integration scheme alone.
This is because the spatial errors will typically dominate for an explicit scheme. To overcome this, we use
a method based on Richardson extrapolation (see, e.g., Moin [42]). The idea is to expand a quantity f in a
∆t f ine
Taylor series with different time increments with a constant ratio r = ∆tcoarse < 1.

f1 = f0 + a0 ∆t p + O(∆t p+1 ) (3.45)

f2 = f0 + a0 (r∆t) + O(r
p p+1
∆t p+1
) (3.46)
f3 = f0 + a0 (r2 ∆t) p + O(r2p+2 ∆t p+1 ) (3.47)

The temporal order of accuracy is given by p, which may be written as

 
− f2
ln ff23 − f1
p= (3.48)
ln r
This procedure effectively filters out the spatial error. It is, in fact, possible to run this case with a first-order
spatial scheme (e.g., Godunov) and still see second-order temporal convergence.

23
Temporal Order Results The saad*.fds cases are run with three different time steps, each refined by
a factor of 2, and integrated to a total time T = 0.390625 s. This time represents 100 time steps for a case
with CFL = 1. To assess the asymptotic temporal order from Eq. (3.48), we use the three finest time steps
corresponding to CFL = [0.25, 0.125, 0.0625] for the solutions [ f1 , f2 , f3 ], respectively.
The results are shown in Fig. 3.10. At the top of the figure you can see the initial and final fields for
the mixture fraction and density for the finest time step. Both fields have been advected to the right. In the
lower-left we plot the numerator and denominator in Eq. (3.48) for the density solution. On the bottom-right
we show the order p computed pointwise on the 1-D domain. The deviations from second-order behavior
correspond to degenerate in the formula for p (as shown in the lower-left plot) where either the denominator
is near zero (p spikes up) or the numerator and denominator cross (thus p = 0). The l2 norm of p computed
pointwise is 2.0053. Note that the l∞ norm (computed pointwise) is degenerate for the reasons just discussed.
A method to overcome this issue is discussed in Sec. 3.5.
FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
1 5
initial field initial field
final field final field
0.8 4
mixture fraction

density (kg/m³)
0.6 3

0.4 2

0.2 1

0 0
-1 -0.5 0 0.5 1 -1 -0.5 0 0.5 1
x (m) x (m)

10 -4 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
6 4
- 2
3
4 -
2 1 3
density difference (kg/m³)

2
p order density

0 2

-2
1
-4

-6 0
-1 -0.5 0 0.5 1 -1 -0.5 0 0.5 1
x (m) x (m)

Figure 3.10: Temporal order for a variable-density projection. (Upper-left) Initial and final field for mixture fraction.
(Upper-right) Initial and final field for density. (Lower-left) Plot of numerator and denominator in Eq. (3.48) for
density at final time. (Lower-right) Plot of p computed pointwise from Eq. (3.48). Fluctuations are due to degenerate
points in the formula for p. The temporal order is p = 2.0053.

3.5 Variable-Density Manufactured Solution (shunn3)

To demonstrate second-order accuracy of the FDS time-marching algorithm, here we present numerical
results for the manufactured solution proposed by Shunn et al. [43] (with a modified pressure solution).

24
The parameters for this problem are given in Table 3.1. In the 2-D sinusoidal solution given below, Z is
the mixture fraction, ρ is the density, u and v are velocity components, and H is the Bernoulli integral
(H ≡ p̃/ρ + 21 |u|2 where p̃ is the hydrodynamic pressure). The solution translates diagonally across the
domain with velocity [u f , v f ]. The translated positions are define by x̂ ≡ x − u f t and ŷ ≡ y − v f t. The
solution is spatially periodic on a square domain extending from −1 m to 1 m (L = 2 m) on each side and
has a time period of 1 s.

Table 3.1: Parameters for Shunn et al. [43] manufactured solution.

Parameter Value Units

ρ0 5 kg/m3
ρ1 1 kg/m3
k=ω 2 1/m, 1/s
uf = vf 0.5 m/s
ρD = µ 0.001 kg/(m · s)

1 + sin(πkx̂) sin(πkŷ) cos(πωt)

Z(x, y,t) = (3.49)
(1 + ρρ01 ) + (1 − ρρ01 ) sin(πkx̂) sin(πkŷ) cos(πωt)
Z(x, y,t) 1 − Z(x, y,t) −1
 
ρ(x, y,t) = + (3.50)
ρ1 ρ0
 
ρ1 − ρ0 −ω
u(x, y,t) = u f + cos(πkx̂) sin(πkŷ) sin(πωt) (3.51)
ρ(x, y,t) 4k
 
ρ1 − ρ0 −ω
v(x, y,t) = v f + sin(πkx̂) cos(πkŷ) sin(πωt) (3.52)
ρ(x, y,t) 4k
1
H(x, y,t) = (u(x, y,t) − u f )(v(x, y,t) − v f ) (3.53)
2
The source terms for the manufactured solution, Q̇ρ , etc., are defined as the residuals of the transport
equations:

∂ρ
+ ∇ · (ρu) = Q̇ρ , (3.54)
∂t
∂ (ρZ)
+ ∇ · (ρZu) − ∇ · (ρD∇Z) = Q̇Z , (3.55)
∂t
∂u 1
− u × (∇ × u) + ∇H − p̃∇(1/ρ) − ∇ · T = Q̇u , (3.56)
∂t ρ

where the components of the deviatoric stress tensor, T, are given by

   
1 1 ∂ ui ∂ u j
Ti j = 2µ Si j − Skk δi j ; Si j ≡ + . (3.57)
3 2 ∂ x j ∂ xi

The mass source terms (which are complex sinusoidal functions) may be obtained from [43], by looking at
the source code in the FDS repository [44] (see module MANUFACTURED_SOLUTIONS in turb.f90), or by
running the sympy (Symbolic Python) script provided under
/Utilities/Python/shunn3_stokes_mms_sym.py. A feature of this particular solution is that the

25
mass source is zero, Q̇ρ = 0, which provides a more realistic test case and simplifies the MMS implemen-
tation. Note that the pressure solution has been modified from Shunn et al. [43] so that the H solution is
compatible with periodic boundary conditions. A new momentum source term is derived Q̇u that differs
from Shunn et al.

Manufactured Solution Procedure

Two species transport equations are solved. The mixture fraction in the manufactured solution is set as
component 2 while the background species is set to component 1:
∂ (ρZ1 )
+ ∇ · (ρZ1 u) − ∇ · (ρD∇Z1 ) = −Q̇Z , (3.58)
∂t
∂ (ρZ2 )
+ ∇ · (ρZ2 u) − ∇ · (ρD∇Z2 ) = +Q̇Z . (3.59)
∂t
The density is obtained via ρ = (ρZ)1 + (ρZ)2 .
The RHS of the momentum equation is augmented as follows:
∂u
= −(F + ∇H) + Q̇u . (3.60)
∂t
The Poisson equation for H becomes
 
2

∂
∇ H = − ∇ · F − Q̇u + (∇ · u) . (3.61)
∂t

Simulations were performed for N = {32, 64, 128, 256, 512}, where N is the number of cells in each
direction, using an adaptive time step satisfying both convective and diffusive CFL numbers of 0.5. The
time step criterion and the simulation parameters in Table 3.1 were chosen to match Shunn et al. [43]. The
species mass density equations, (3.58) and (3.59), are solved using the CHARM flux limiter. Qualitative
results for the 2562 case are shown in Fig. 3.11. These images may be compared to the images presented in
[43]. In Fig. 3.12, we plot the L2 error at time t = 0.9 s1 as a function of grid spacing, ∆x = L/N, confirming
second-order accuracy of the solutions for density, mixture fraction, and velocity. As shown in Fig. 3.12,
schemes such as ours, where the pressure is set to zero at the beginning of each stage of the integration
(in other words, no pressure gradient term shows up in the force on the RHS of the Poisson equation), are
known to be first-order accurate for pressure [45]. This is not a severe limitation of the algorithm since the
hydrodynamic pressure does not factor into the equation of state for low-Mach flows.

Temporal Error Analysis

The convergence plot in Fig. 3.12 is based on a constant CFL with increasing spatial resolution. From
this plot, the temporal accuracy is hard to discern. Further, the temporal accuracy for the Saad problem in
Sec. 3.4 may be a special case since most of the terms in the divergence are zero for that problem. For the
Shunn case, the right-hand-side of the Poisson equation is non-trivial and so it is useful to re-examine the
temporal order for this case. As with the Saad case, here we run a series of tests at constant grid resolution
256 × 256 while decreasing the time step. The three finest time steps correspond to CFL = [0.25, 0.125,
0.0625] for the solutions [ f1 , f2 , f3 ] in Eq. (3.48). The simulations are run to a total time equivalent to 2
time steps for a CFL=1 calculation.
1 At t = 1 s the H manufactured solution is trivially zero everywhere and fortuitous error cancellation occurs.

26
 t =0s t = 0.125 s t = 0.250 s t = 0.375 s ρ, kg/m3
5.0

4.0

3.0

2.0

1.0
Z
1.0

0.8

0.6

0.4

0.2

0.0
u, m/s
0.8

0.6

0.4

0.2

Figure 3.11: Evolution of the variable-density manufactured solution. From top to bottom, the images show density,
ρ, mixture fraction, Z, and the u-velocity component from the 2562 simulation at the times shown at the top of the
columns. These results may be compared to [43] to confirm the validity of the numerical solution.

Several different norms are considered and the results are listed in Table 3.2. In the left column of the
table, the L1 and L2 norms are taken from the vector p computed pointwise on the domain. Note that, as
discussed in Sec. 3.4, the infinity norm for this approach is degenerate. The “Saad” norms are formed by
taking the norms of the differences in the numerator and denominator of Eq. (3.48),
 
ln kk ff23 −
− f2 k
f1 k
kpk Saad = (3.62)
ln r

27
 FDS-6.8.0-0-g886e009-release

FDS
-2 FDS Z
10 FDS u
FDS H
O( x)

L Error
O( x2)

2
10 -4

10 -6
10 -3 10 -2 10 -1
x (m)

Figure 3.12: Convergence for the manufactured solution. The L2 error at time t = 0.9 s is plotted as a function of
grid spacing for N = {32, 64, 128, 256, 512} points in each direction. The grid spacing is ∆x = L/N, where L = 2 m.
Calculations were performed with an adaptive time step satisfying both convective and diffusive CFL (Von Neumann)
limits of 0.5. These results confirm second-order accuracy of the flow solver for density, mixture fraction, and velocity.
As is known for projection schemes like ours, the pressure solution (represented by H) is first-order accurate [45].

Table 3.2: Shunn temporal error norms for the order p.

Norm ρ Z u H
L1 2.0266 2.0274 2.0309 1.0067
L2 2.0528 2.0518 2.0792 1.0274
L1 Saad 2.0229 2.038 2.0408 0.99928
L2 Saad 2.0608 2.1103 2.1235 0.99894
L∞ Saad 2.1745 2.2374 2.188 0.99125

3.6 Collapse of a Hot Layer in a Micro-gravity Environment (hot_layer_360)

Howard R. Baum, NIST Fellow Emeritus

This problem describes the collapse of a hot layer of depth d adjacent to a wall at plane y = 0 in zero gravity.
The wall is presumed to remain cool at the ambient temperature T0 while the hot layer is initially Th . The
initial velocity field is quiescent. Starting at time t = 0 the hot layer begins to diffuse into the upper cool
ambient atmosphere and to expand due to dilatation. There is also heat transfer to the wall. The problem
is 1-D in the spatial coordinate y and time dependent. We are interested in the evolution of the temperature
T (y,t) and the velocity v(y,t).
For the present problem the governing equations are
∂ρ ∂
+ (ρv) = 0 (3.63)
∂t ∂ y
   
∂T ∂T ∂ ∂T
ρCp +v = k (3.64)
∂t ∂y ∂y ∂y
ρT = ρ0 T0 (3.65)

Here, ρ(y,t) is the mass density, Cp is the heat capacity (assumed constant), and k is the thermal conductivity
of the gas, which we assume obeys ρk = ρ0 k0 . The initial conditions are

v(y, 0) = 0 (3.66)

28
 
Th for 0 < y ≤ d
T (y, 0) = (3.67)
T0 for y > d

The boundary conditions are

v(0,t) = 0 (3.68)
T (0,t) = T0 (3.69)
T (y,t) → T0 as y → ∞ (3.70)

Note that no boundary condition can be imposed on v far from the boundary. In FDS, an OPEN boundary is
applied.
Introducing the dimensionless coordinate λ and dimensionless time τ, the dimensionless physical coor-
dinate Y (λ , τ), velocity V (λ , τ), and temperature θ (λ , τ) are related to dimensional quantities as follows:

t = ρ0Cp d 2 /k0 τ (3.71)

y = d Y (λ , τ) (3.72)
v = k0 /(ρ0Cp d)V (λ , τ) (3.73)
T = T0 θ (λ , τ) (3.74)

The solution is then given by the following (for coding details see hot_layer_collapse.m):
p √
f (λ , τ, a) = −2 τ/π exp(−(a − λ )2 /(4τ)) + (a + λ ) erfc((a + λ )/(2 τ)) (3.75)
F(λ , τ, a) = f (λ , τ, a) − f (0, τ, a) (3.76)
Y (λ , τ, Th , T0 ) = λ + (Th − T0 )/T0 (−F(λ , τ, 0) + 12 (F(λ , τ, T0 /Th ) + F(λ , τ, −T0 /Th ))) (3.77)
√ √ √
U(λ , τ, Th , T0 ) = −erfc(λ /(2 τ)) + 12 (erfc((λ − T0 /Th )/(2 τ)) + erfc((λ + T0 /Th )/(2 τ))) (3.78)
θ (λ , τ, Th , T0 ) = 1 + ((Th − T0 )/T0 )U(λ , τ, Th , T0 ) (3.79)
√
G(λ , τ, a) = 1/ πτ (− exp(−(a + λ )2 /(4τ)) + exp(−a2 /(4τ))) (3.80)
1
V (λ , τ, Th , T0 ) = (Th − T0 )/T0 (−G(λ , τ, 0) + 2 (G(λ , τ, T0 /Th ) + G(λ , τ, −T0 /Th ))) (3.81)

Figure 3.13 shows the analytical solution (solid lines) plotted together with the FDS (DNS) numerical results
(open symbols). The plots on the left are of temperature profiles at several times (different colors demarcate
different nondimensional times τ). The plots on the right are velocity. As can be seen, the agreement is
excellent.

29
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
1.2 1.2

1 1

0.8 0.8
y/d

y/d
0.6 0.6
-4 -4
=2 10 =2 10
0.4 =4 10 -4 0.4 =4 10 -4
-4 -4
=6 10 =6 10
0.2 =8 10 -4 0.2 =8 10 -4
= 10 10 -4 = 10 10 -4
0 0
1 1.5 2 2.5 3 3.5 4 -200 -150 -100 -50 0
T/T 0 v/v 0
FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
1.2
=2 10 -3 1.5 =2 10 -3
-3 -3
1 =4 10 =4 10
-3 -3
=6 10 =6 10

0.8 =8 10 -3 =8 10 -3
= 10 10 -3 1 = 10 10 -3
y/d

y/d
0.6

0.4 0.5

0.2

0 0
1 1.5 2 2.5 3 3.5 4 -60 -50 -40 -30 -20 -10 0
T/T 0 v/v 0

Figure 3.13: Hot layer solution. Lines are the analytical solution. Symbols are the FDS results.

3.7 2-D Vortex Simulation (vort2d)

Max Gould, NIST SURF student
Ragini Acharya, United Technologies Research Center

In this section we present another case that demonstrates the second-order accuracy of the FDS transport
algorithm. We consider the analytically stable flow field consisting of a single vortex advected by a uniform
flow, a test case developed by CERFACS2 [46]. Maintaining the geometry of the vortex over time provides
a good measure of the order of accuracy of the transport scheme.
The vortex is initially defined as the gradient of the potential field,
 2 2
x +z
Ψ0 (x, z) = Γ exp − , (3.82)
2 R2c

where Γ determines its intensity and Rc its characteristic size. The velocity components are determined
by taking the gradient of the potential field superimposed on the constant flow field of velocity U0 in the
positive x direction,
 2 2
∂ Γz x +z
u(x, z) ≡ U0 + Ψ0 = U0 − 2 exp − , (3.83)
∂z Rc 2 R2c
2 Centre Européen de Recherche et de Formation Avancée en Calcul Scientifique

30
 
 
 0.04
- -


 
 0.02
 

- -

? 0
6

- -
PP
PP -0.02
PP
U0 = 35 m/s P PP
- P -P
P
P PP -0.04
PP
PP
L = 0.3112 m -0.04 -0.02 0 0.02 0.04

Figure 3.14: Vector field of a two-dimensional vortex in a constant flow field.

 2 2
∂ Γx x +z
w(x, z) ≡ − Ψ0 = 2 exp − , (3.84)
∂x Rc 2 R2c

where u and w refer to velocity in the x and z-directions, respectively. For our purposes we need only analyze
one component of the velocity field. We will focus our attention on the u-component of velocity.
We define the computational domain as a two-dimensional square region, L = 0.3112 m on a side, with
periodic boundary conditions. The domain is discretized for a range of square, two-dimensional meshes of
402 , 802 , 1602 , and 3202 grid cells. For the purposes of this test, we set the flow parameters as

U0 = 35 m/s
Rc = L/20 = 0.01556 m
√
Γ = 0.04 U0 Rc e = 0.0359157

The constant flow field and periodic boundary conditions cause the vortex to repeatedly pass through the
computational domain. The “pass-through” time, t f , is defined as the time period required for the stable
vortex to return to its original position,

t f = L/U0 ' 8.8914 × 10−3 s.

To ensure that the numerical solution converges to the analytical solution, we set the time step, δt, so that
the Courant-Friedrichs-Lewy (CFL) number is 0.5.
A plot of u-velocity values just along the z-axis provides a simple characterization of the vortex geome-
try. The extent to which this geometry changes over time provides a qualitative measure of the accuracy of
the transport algorithm. Figure 3.15 displays these plots for two different grid resolutions. Each line repre-
sents a plot taken for a different number of flow-through times such that the red lines represent the vortex
after it has undergone the most passes through the computational domain while the green lines represent the
vortex in the initial phase. The broken black line represents the analytical solution. As the vortex undergoes
more passes through the computational domain, its velocity profile diverges further and further from the
analytical profile. While divergence still occurs on the finer mesh, the extent to which it diverges after the
same number of flow-through times is significantly smaller.

31
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release

36.5 36.5
Analytical Analytical
36 FDS (t=0.000 s) 36 FDS (t=0.000 s)
FDS (t=0.009 s)
35.5 35.5
u-velocity (m/s)

u-velocity (m/s)
35 35

34.5 34.5

34 34

33.5 33.5
-0.02 -0.01 0 0.01 0.02 -0.03 -0.02 -0.01 0 0.01 0.02 0.03
z-coordinate (m) z-coordinate (m)

Figure 3.15: u-velocity along the z-axis at x = 0 plotted for each of the vortex’s first four loops through the computa-
tional domain. (Left) 802 grid cell model. (Right) 1602 grid cell model.

FDS-6.8.0-0-g886e009-release
10 0
Convergence (vort2d)

10 -1
RMS Error (m/s)

10 -2 O(?x)
O(?x²)
FDS (RMS error at t=0.0089 s)
FDS (RMS error at t=0.0178 s)
FDS (RMS error at t=0.0267 s)
10 -3
-3
10
Grid Spacing (m)

Figure 3.16: RMS error between simulated and analytical u-velocity values along the z-axis plotted for each grid
resolution at each of three subsequent passes of the vortex through the computational domain.

The rate with which the simulated profiles converge to the analytical one defines the order of accuracy
of the numerical scheme. In Fig. 3.16 we plot the rms error of the numerical solution as a function of
the grid resolution. The three colored curves represent the rms error at three different pass-through times.
The broken and solid black lines represent the plot gradient corresponding to first and second order error
respectively. While the error increases with each flow-through time, the gradients of the lines are roughly
parallel to the solid black line, indicating second-order accuracy of the numerical scheme.
To analyze the stability of the vortex at times other than discrete multiples of the pass-through time,
consider the time-dependent potential field and its corresponding u-velocity component:
2 U0 x t −U02 t 2
 
Ψ(x, z,t) = Ψ0 exp , (3.85)
2 R2c
2 U0 x t −U02 t 2
 
z
u(x, z,t) = U0 − Ψ0 2 exp . (3.86)
Rc 2 R2c
In Fig. 3.17, we show the u-velocity at a single point, on the lower left fringe of the vortex, for two different
mesh resolutions, 802 and 1602 . A mesh resolution of 3202 grid cells produces a plot (not shown) that is, to
the eye, a perfect match out to four pass-through times.

32
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
36 36

35.8 35.8

35.6 35.6
u-velocity (m/s)

u-velocity (m/s)
35.4 35.4

35.2 35.2

35 35

34.8 Analytical 34.8 Analytical

FDS FDS
34.6 34.6
0 0.01 0.02 0.03 0 0.01 0.02 0.03
Time (s) Time (s)

Figure 3.17: Simulated and analytical values of u-velocity at a point plotted as a function of time over a time period
equal to four times the flow-through time. (Left) 802 grid cell model. (Right) 1602 grid cell model.

3.8 Boundedness Tests

3.8.1 Species Bounds (bound_test_*)
When multiple infinitely fast reactions are present it is not trivial to correctly partition the limiting reactant
and maintain species bounds. FDS uses a special subcycling algorithm by Kahaner [47] to maintain bound-
edness for species when time integrating the ODEs for the chemical reaction step (the method is further
discussed in the FDS Tech Guide [48]). In this section, we examine two test cases designed to verify the
time integration method.
The first case, bound_test_1, consists of two independent reactions, one with air as the limiting
reactant, one with fuel as the limiting reactant. The stoichiometry of each reaction is the same: F + A → 2 P.
The initial volume fractions of determine which species are limiting. The results are shown in Fig. 3.18.
For the first reaction, XF1 is initially 0.3 and XA1 is initially 0.2. Therefore, all the A1 is converted to P1,
which has a final volume fraction of XP1 = 0.4. In the second reaction, the initial fuel volume fraction is
XF2 = 0.1 (limiting) and the initial air volume fraction is XA2 = 0.4. As can be see in Fig. 3.18, the resulting
final values give a product volume fraction of twice the limiting reactant, XP2 = 0.2.
FDS-6.8.0-0-g886e009-release

1
Boundedness Test (bound_test_1) F1
A1
0.8 P1
F2
Volume Fraction

A2
0.6 P2
Sum
FDS F1
0.4 FDS A1
FDS P1
FDS F2
0.2 FDS A2
FDS P2
FDS SUM
0
0 0.2 0.4 0.6 0.8 1
Time (s)

Figure 3.18: Results of the bound_test_1 case.

The second test case, bound_test_2, is more challenging—we add a third reaction that uses the same

33
fuel species as the first and the same air species as the second. The reaction scheme is as follows:

F1 + A1 → 2 P1 (3.87)
F2 + A2 → 2 P2 (3.88)
F1 + A2 → 2 P3 (3.89)

The initial conditions and final results are shown in Fig. 3.19. As discussed in the FDS Tech Guide [48],
the right-hand-side of the chemistry ODE is set up assuming a large Arrhenius constant and no activation
energy and (in this case) second-order concentration dependence for each reaction. It can be shown through
numerical experiments that this assumption is equivalent to applying Curl’s mixing model (Coalescence-
Dispersion) [49] in a batch reactor with infinitely fast chemistry.
FDS-6.8.0-0-g886e009-release

1
Boundedness Test (bound_test_2)
F1
F2
0.8 A1
A2
Volume Fraction

P1
0.6 P2
P3
Sum
0.4 FDS F1
FDS F2
FDS A1
0.2 FDS A2
FDS P1
FDS P2
0 FDS P3
0 0.2 0.4 0.6 0.8 1 FDS SUM
Time (s)

Figure 3.19: Results of the bound_test_2 case.

3.8.2 Isothermal Helium Wave (helium_1d_isothermal)

This is a test of isothermal advection and mixing of gases with different molecular weights using the default
algorithm. A companion case using a simplified algorithm for CONSTANT_SPECIFIC_HEAT_RATIO is
given below in Sec. 3.9.1. In the present case, a wave of helium propagates in one dimension down the length
of a square duct with frictionless walls. The initial and inlet temperatures are 25 ◦C and should remain so
throughout the course of the simulation. In Fig. 3.20 we plot the simulation results for temperature, sensible
enthalpy, and velocity. The temperature should remain constant at 25 ◦C. The sensible enthalpy, which is
referenced to 25 ◦C, should remain at 0 kJ/m3 . And the velocity should level off at 1 m/s and hold steady.
This output is checked to make sure that no boundary effects manage to pollute the simulation.

3.8.3 Temperature Bounds for Isothermal Helium Jet (helium_air_jet_floor)

In this test, a jet of helium is generated from a floor vent. The mass flux is ramped up from 0 m/s to 1 m/s in
1 s. The flow field should remain isothermal at 20 ◦C. The results are shown in Fig. 3.21.

34
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
50 1
Check Isothermal Flow (helium_1d_isothermal) Check Isothermal Flow (helium_1d_isothermal)
40
0.5

Sensible Enthalpy (kJ/m³)

Temperature (°C)

30
0
20

-0.5
10 Expected Temp Expected H_S
FDS (max T) FDS (max H_S)
FDS (min T) FDS (min H_S)
0 -1
0 5 10 15 0 5 10 15
Time (s) Time (s)

FDS-6.8.0-0-g886e009-release
2
Check Isothermal Flow (helium_1d_isothermal)

1.5
Velocity (m/s)

0.5
Expected W-VEL
FDS (max W)
FDS (min W)
0
0 5 10 15
Time (s)

Figure 3.20: Results of the helium_1d_isothermal cases.

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
50 50
Check Isothermal Flow (helium_air_jet_floor) Check Isothermal Flow (helium_air_jet_floor)
40 40
Temperature (°C)

Temperature (°C)

30 30
Isothermal Isothermal
FDS (min T) FDS (max T)
20 20

10 10

0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
Time (s) Time (s)

Figure 3.21: Results of the helium_air_jet_floor case.

3.8.4 Temperature Bounds for Large Differences in Molecular Weight (mwtest_cfl)

Hydrogen gas (2 g/mol) is injected into a small compartment filled with sulfur hexafluoride (146 g/mol).
There should be no change in temperature, as shown in Fig. 3.22.

35
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
50 50
Large MW Difference (mwtest_cfl) Large MW Difference (mwtest_cfl)
40 40
Temperature (°C)

Temperature (°C)
30 30
Isothermal Isothermal
FDS (min T) FDS (max T)
20 20

10 10

0 0
0 5 10 15 20 0 5 10 15 20
Time (s) Time (s)

Figure 3.22: Results of the mwtest_cfl case.

3.8.5 Temperature Lower Bound in a Fire Simulation (tmp_lower_limit)

A 2-D methane-air diffusion flame calculation is run three different ways to check that the temperature does
not fall below the ambient (20 ◦ C). The results are shown in Fig. 3.23. The upper left plot shows the global
minimum temperature for the case where the reaction stoichiometry is specified explicitly. The upper right
plot is for the case where the initial time step is lowered to 0.001 s. The lower plot is for the case where the
simple chemistry model is used.

36
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
40 40
Temperature Lower Limit (default) Temperature Lower Limit (?t=0.001 s)

30 30
Temperature (°C)

Temperature (°C)
20 20

10 10
Expected (min T) Expected (min T)
FDS (min T) FDS (min T)
0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
Time (s) Time (s)

FDS-6.8.0-0-g886e009-release
40
Temperature Lower Limit (simple)

30
Temperature (°C)

20

10
Expected (min T)
FDS (min T)
0
0 0.2 0.4 0.6 0.8 1
Time (s)

Figure 3.23: Results of the tmp_lower_limit cases.

3.9 Constant Specific Heat Ratio

The ratio γ = c p /cv is called the “ratio of specific heats”. For many simple gases, γ = constant is a good
approximation. The specific heat of species α in the gas phase is then given by
R γα
c p,α = (3.90)
Wα γα − 1
where R is the ideal gas constant. When this assumption is viable, many terms in the energy equation
drop out, greatly simplifying the divergence expression and thus speeding up the code. These simplifica-
tions require that the sensible enthalpy of the gas is referenced to 0 K (see the FDS Tech Guide [48]). In
this section, we examine three cases using the assumption of constant specific heat ratio: one tests mix-
ing of two gases with different molecular weights (helium_1d_const_gamma), one tests fire conditions
(fire_const_gamma), and one tests evaporation of water droplets (water_evap_1_const_gamma).
Each of these cases is a clone of cases we develop later in this guide for the default version of the code.

3.9.1 Helium Wave using Constant Specific Heat Ratio (helium_1d_const_gamma)

In this problem, a wave of helium propagates in one dimension down the length of a square duct with
frictionless walls. The initial and inlet temperatures are 25 ◦C and should remain so throughout the course
of the simulation. The ratio of specific heats is kept at the default value of γ = 1.4. The molar weights, heat

37
 Table 3.3: Helium constant specific heat ratio parameters.

Species Wα (kg/kmol) ρα (kg/m3 ) c p,α (kJ/kg/K) hs,α (kJ/m3 )

Air 28.73167 1.174 1.0128 355
Helium 4.00260 0.164 7.2705 355

capacities, and enthalpies for both air and helium are given in Table 3.3. Notice that the sensible enthalpies,
hs,α = ρα c p,α T , are the same for both species.
In Fig. 3.24 we plot the simulation results for temperature, sensible enthalpy, and velocity. The temper-
ature should remain constant at 25 ◦C. The sensible enthalpy should remain constant at 355 kJ/m3 . And the
velocity should level off at 1 m/s and hold steady. This output is checked to make sure that no boundary
effects manage to pollute the simulation.
FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
50 500
Check Isothermal Flow (helium_1d_const_gamma) Check Isothermal Flow (helium_1d_const_gamma)
40 400

Sensible Enthalpy (kJ/m³)

Temperature (°C)

30 300

20 200

10 Expected Temp 100 Expected H_S

FDS (max T) FDS (max H_S)
FDS (min T) FDS (min H_S)
0 0
0 5 10 15 0 5 10 15
Time (s) Time (s)

FDS-6.8.0-0-g886e009-release
2
Check Isothermal Flow (helium_1d_const_gamma)

1.5
Velocity (m/s)

0.5
Expected W-VEL
FDS (max W)
FDS (min W)
0
0 5 10 15
Time (s)

Figure 3.24: Results of the helium_1d_const_gamma cases.

3.9.2 Fire Plume using Constant Specific Heat Ratio (fire_const_gamma)

In this case a 128 m3 box is initially filled with air at 20 ◦C. A 1000 MW propane fire with a 5 % soot yield
is placed at the bottom of the box for 20 s. All the compartment surfaces are adiabatic. The net change
in enthalpy and pressure can be computed using the specific heat ratio based constant specific heat values
for the air and product species along with the pressure work required to inject the fuel. The FDS predicted

38
 10 4 FDS-6.8.0-0-g886e009-release 10 4 FDS-6.8.0-0-g886e009-release
4 8
Enthalpy Rise (fire_const_gamma) Pressure Rise (fire_const_gamma)

3 6
Enthalpy (kJ)

Pressure (Pa)
2 4

1 2
Exact Exact
FDS FDS
0 0
0 5 10 15 20 25 30 0 5 10 15 20 25 30
Time (s) Time (s)

Figure 3.25: Results of the fire_const_gamma case.

enthalpy and pressure change are shown in Fig 3.25 compared with the predicted values.

3.9.3 Evaporation with Constant Specific Heat Ratio (water_evap_1_const_gamma)

This test case is a replica of water_evaporation_1 in Sec. 13.3.1 using constant specific heat ratio.
A 1 m3 box is initially filled with dry air at 200 ◦C and monodisperse water droplets totaling 0.01 kg in
mass initially at 20 ◦C. In this case, all the water evaporates. The final gas temperature may be computed
from energy conservation. Pressure change may be computed from the ideal gas law. Results are shown
in Fig. 3.26. Details of the expected results may be found in water_evap_1_const_gamma.m in the
Utilities/Matlab/scripts/ directory in the FDS repository [44].

39
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
50 3
Enthalpy Change (water_evap_1_const_gamma) Relative Humidity (water_evap_1_const_gamma)
2.5

Humidity (%)
Enthalpy (kJ)

Expected (h_gas)
Exact (h_water)
0 FDS (h_gas)
1.5
FDS (h_water)
1

0.5 Exact (Rel. Hum)

FDS (humid)
-50 0
0 2 4 6 8 10 0 2 4 6 8 10
Time (s) Time (s)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
0.015 220
Density Change (water_evap_1_const_gamma) Temperature (water_evap_1_const_gamma)
200

0.01

Temperature (°C)
Density (kg/m³)

180

160
0.005

140
Exact (dens) Expected (temp)
FDS (dens) FDS (Temp)
0 120
0 2 4 6 8 10 0 2 4 6 8 10
Time (s) Time (s)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
0 0.015
Pressure Change (water_evap_1_const_gamma) Evaporated Mass (water_evap_1_const_gamma)
-2

0.01
Pressure (kPa)

-4
Mass (kg)

Expected (pres)
FDS (pres)
-6
0.005

-8
Exact (vapor)
FDS (WATER VAPOR)
-10 0
0 2 4 6 8 10 0 2 4 6 8 10
Time (s) Time (s)

Figure 3.26: Output of the water_evap_1_const_gamma test case.

3.10 Tunnel Flow

Jesus Alberto Mejias Tuni, Politecnico di Torino

3.10.1 1-D Flow in a Tunnel (tunnel_const_gamma)

Consider a long tube with rectangular cross section. Air with density ρ∞ = 1.200 kg/m3 , temperature T∞ =
293.15 K, and pressure p∞ = 101325 Pa, is blown at a constant rate, u∞ = 1 m/s, through the tube. A steady
volumetric heat source is located in the middle of the tube. Ignoring wall friction, gravity, radiation transport,

40
thermal conductivity, viscosity, and assuming that the heat source uniformly spans the cross section of the
tube, a set of 1-D equations can be derived for the steady-state velocity, u(x), density, ρ(x), perturbation
pressure, p̃(x), and temperature, T (x):

d u2
 
1 d p̃
ρu = ρ∞ u∞ ; + =0 (3.91)
dx 2 ρ dx
 
du 1 000 d dρ
p∞ = RρT /W ; = q̇ − u (ρhs ) − (c p T − hs )u (3.92)
dx ρc p T dx dx

Note that R = 8314.5 J/kmol/K and the sensible enthalpy, hs = TT∞ c p (T 0 ) dT 0 . The tube is 10 cm long, 1 mm
R

wide and 4 mm tall, and discretized with 1 mm grid cells. A vertical column of cells contains a volumetric
heat source that introduces heat at a rate of q̇000 = 2.5133 × 108 W/m3 within each 1 mm cell. The velocity,
density, pressure and temperature are constant starting at the inlet, and at the heat source, the velocity and
temperature jump up, and the density and pressure drop down over the span of a single cell and remain
steady until the outlet is reached, where the perturbation pressure is set to zero.
For the case where the ratio of specific heats are constant with γ = 1.4 and W = 28.85 kg/kmol for air,
c p = γ (R/W )/(γ − 1) = 1008.7 J/kg/K and ρc p T = γ p∞ /(γ − 1). In the second equation of (3.92), the
second two terms on the right hand side are identically zero; thus, the jump in velocity can be computed
γ − 1 000
∆u = q̇ ∆x ≈ 0.7087 m/s (3.93)
γ p∞
Using the first equations of (3.91) and (3.92), the density and temperature downstream of the heat source
are computed to be 0.7023 kg/m3 and 500.61 K, respectively. To determine the drop in pressure, Eq. (3.91)
must be written in discretized form. Assume that the subscript 0 refers to the cell containing the heat source,
1 to the first cell downstream, and -1 to the first cell upstream. The velocity in cell 0 is the average of the
velocity upstream and downstream, u0 = 1.3544 m/s.

u21 − u20
p̃1 − p̃0 = − ρ1 ≈ −0.3811 Pa (3.94)
2
u2 − u2−1
p̃0 − p̃−1 = − 0 ρ0 ≈ −0.3968 Pa (3.95)
2
Summing the two equations yields ∆ p̃ ≡ p̃1 − p̃−1 ≈ −0.7779 Pa. The jump in energy across the heat source
is
d
(uρc p T ) ≈ 2.5113 × 108 W/m3 (3.96)
dx
Plots of u(x), ρ(x), p̃(x), and T (x) are shown in Fig. 3.27.

41
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
1 2.5
Pressure (fire_const_gamma) Velocity (fire_const_gamma)
0.8
2

Velocity (m/s)
Pressure (Pa)

0.6
1.5
0.4

1
0.2
Exact (p) Exact (u)
FDS (p) FDS (u)
0 0.5
0 0.02 0.04 0.06 0.08 0.1 0 0.02 0.04 0.06 0.08 0.1
Distance (m) Distance (m)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
1.5 300
Density (fire_const_gamma) Temperature (fire_const_gamma)
250

1 200

Temperature (°C)
Density (kg/m³)

150

0.5 100

Exact (rho)
50 Exact (T)
FDS (rho) FDS (T)
0 0
0 0.02 0.04 0.06 0.08 0.1 0 0.02 0.04 0.06 0.08 0.1
Distance (m) Distance (m)

Figure 3.27: Output of the tunnel_const_gamma test case.

3.10.2 1-D Flow in a Tunnel (tunnel_linear_cp)

Considering the same layout and the same environmental conditions of the previous case. Air is blown at a
constant rate, u∞ = 1 m/s, through a 10 cm long, 1 mm wide and 4 mm tall tube. A steady heat source is
located in the middle of the tube, composed of a vertical column of cells containing a volumetric heat source
that introduces heat at a rate of q̇000 = 2.5133 × 108 W/m3 within each 1 mm cell . Ignoring wall friction,
gravity, radiation transport, thermal conductivity, viscosity, and assuming that the heat source uniformly
spans the cross section of the tube, a set of 1-D equations can be derived for the steady-state velocity, u(x),
density, ρ(x), perturbation pressure, p̃(x), and temperature, T (x). This equations are the same used in the
previous example in (3.91) and (3.92).
Remembering that R = 8314.5 J/kmol/K and the sensible enthalpy, hs = TT∞ c p (T 0 ) dT 0 . The velocity,
R

density, pressure and temperature are constant starting at the inlet, until the heat source, where the velocity
and temperature jump up, and the density and pressure drop down over the span of a single cell and remain
steady until the outlet is reached, where the perturbation pressure is set to zero.
For the case where the specific heat varies as a function of the temperature. In the second equation of
(3.92), using the first equation of (3.91) and (3.92), and some relationships among the variables is possible
to arrive to a simpler expression.
du d d dρ dρ
ρc p T = q̇000 − uρ(hs ) − uhs (ρ) − c p Tu + hs u (3.97)
dx dx dx dx dx
dT 000 d dT
c p ρ∞ u∞ = q̇ − uρ (hs ) + c p uρ (3.98)
dx dx dx

42
 d
q̇000 = uρ (hs ) (3.99)
dx
q̇000 d
= (hs ) (3.100)
ρ∞ u∞ dx

Through the steps, the continuity, in (3.91), and state equation, in (3.92), are used to make the temper-
ature the only variable and simplifying equivalent terms until the equation (3.100) is obtained. Equation
(3.100) corresponds to a 1-D form of the first law of thermodynamics, proving the divergence equation
as a consistent energy balance. Then, for the case where the c p is a linear function of the temperature,
c p = aT + b, we have

q̇000 a a
∆x = T 2 + bT − (T∞2 + bT∞ ) (3.101)
ρ∞ u∞ 2 2

Substituting in Eq. (3.101) for a = 0.1584 J/kg/K2 and b = 953.5650 J/kg/K (these coefficients give c p (T∞ ) =
1000 J/kg/K) the temperature downstream of the heat source is computed to be 499.2 K. Correspondingly,
the density and velocity are, ρ1 = 0.7046 kg/m3 and u1 = 1.7030 m/s. To determine the drop in pressure,
Eq. (3.91) is written in discretized form as in the previous verification. Then assuming that the subscript 0
refers to the cell containing the heat source, 1 to the first cell downstream, and -1 to the first cell upstream.
The velocity in cell 0 is the approximated as the average of the velocity upstream and downstream, u0 =
1.3515 m/s.
u21 − u20
p̃1 − p̃0 = − ρ1 ≈ −0.3783 Pa (3.102)
2
u2 − u2−1
p̃0 − p̃−1 = − 0 ρ0 ≈ −0.3936 Pa (3.103)
2
Summing both of the equations yields ∆ p̃ ≡ p̃1 − p̃−1 ≈ −0.7718 Pa. The jump in energy across the heat
source is
d
(uρc p T ) ≈ 2.6685 × 108 W/m3 (3.104)
dx
Plots of u(x), ρ(x), p̃(x), and T (x) are shown in Fig. 3.28.

43
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
1 2.5
Pressure (fire_linear_cp) Velocity (fire_linear_cp)
0.8
2

Velocity (m/s)
Pressure (Pa)

0.6
1.5
0.4

1
0.2
Exact (p) Exact (u)
FDS (p) FDS (u)
0 0.5
0 0.02 0.04 0.06 0.08 0.1 0 0.02 0.04 0.06 0.08 0.1
Distance (m) Distance (m)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
1.5 300
Density (fire_linear_cp) Temperature (fire_linear_cp)
250

1 200

Temperature (°C)
Density (kg/m³)

150

0.5 100

Exact (rho)
50 Exact (T)
FDS (rho) FDS (T)
0 0
0 0.02 0.04 0.06 0.08 0.1 0 0.02 0.04 0.06 0.08 0.1
Distance (m) Distance (m)

Figure 3.28: Output of the tunnel_linear_cp test case.

44
Chapter 4

Turbulence

4.1 Decaying Isotropic Turbulence

In this section we present a canonical flow for LES which tests whether the subgrid stress model has been
coded properly. In some cases the difference between verification and validation is not so clear. Once a
model is well-established and validated it may actually be used as a form of verification. Granted, such
a test is not as strong a verification as the convergence study shown in Section 3.1. Nevertheless, these
tests are often quite useful in discovering problems within the code. The case we examine in this section,
decaying isotropic turbulence, is highly sensitive to errors in the advective and diffusive terms because the
underlying physics is inherently three-dimensional and getting the problem right depends strongly on a
delicate balance between vorticity dynamics and dissipation. An even more subtle yet extremely powerful
verification test is also presented in this section when we set both the molecular and turbulent viscosities to
zero and confirm that the integrated kinetic energy within the domain remains constant. In the absence of
any form of viscosity, experience has shown that the slightest error in the advective terms or the pressure
projection will cause the code to go unstable. This verification is therefore stronger than one might initially
expect.
In this section we test the FDS model against the low Reynolds number (Re) data of Comte-Bellot
and Corrsin (CBC) [50]. Viscous effects are important in this data set for a well-resolved LES, testing the
model’s Re dependence. Following [51], we use a periodic box of side L = 9 × 2π cm (≈ 0.566 m) and
ν = 1.5 × 10−5 m2 /s for the kinematic viscosity. The non-dimensional times for this data set are: x/M = 42
(initial condition), 98, and 171, where M is the characteristic mesh spacing of the CBC wind tunnel and
x is the downstream location of the data station. Considering the mean velocity in the CBC wind tunnel
experiment, these correspond to dimensional times of t = 0.00 s, 0.28 s, and 0.66 s in our simulations.
The initial condition for the FDS simulation is generated by superimposing Fourier modes with random
phases such that the spectrum matches that of the initial CBC data. An iterative procedure is employed where
the field is allowed to decay for small time increments subject to Navier-Stokes physics, each wavenumber
is then injected with energy to again match the initial filtered CBC spectrum. The specific filter used here is
discussed in [52].
To provide the reader with a qualitative sense of the flow, Fig. 4.1 shows the initial and final states of the
velocity field in the 3-D periodic domain. The flow is unforced and so if viscosity is present the total energy
decays with time due to viscous dissipation.

4.1.1 Constant Smagorinsky (csmag)

Because the viscous scales are unresolved, a subgrid stress model is required. Here the stress is closed using
the gradient diffusion hypothesis and the eddy viscosity is modeled by the constant coefficient Smagorinsky

45
 Figure 4.1: Initial and final states of velocity magnitude for the isotropic turbulence field.

model with the coefficient taken to be Cs = 0.2 (see the Technical Reference Guide for further details).
The decay curves for two grid resolutions are shown on the left in Fig. 4.2. For an LES code such
as FDS which uses a physically-based subgrid model, an important verification test is to run this periodic
isotropic turbulence simulation in the absence of both molecular and turbulent viscosity. For so-called
“energy-conserving” explicit numerics the integrated energy will remain nearly constant in time. This is
demonstrated by the dash-dot line in the top-left plot in Fig. 4.2. The deviations from identical energy
conservation (to machine precision) are due solely to the time discretization (the spatial terms are conser-
vative as discussed in [53]) and converge to zero as the time step goes to the zero. Note that strict energy
conservation requires implicit time integration [54, 55] and, as shown by the dashed curve on the same plot
where only molecular viscosity is present in the simulation, this cost is unwarranted given that the molecular
dissipation rate clearly overshadows the relatively insignificant amount of numerical dissipation caused by
the explicit method. The FDS result using the Smagorinsky eddy viscosity (the black solid line) matches
the CBC data (red open circles) well for the 323 case (top-left). However, the FDS results are slightly too
dissipative in the 643 case (bottom-left). This is due to a well-known limitation of the constant coefficient
Smagorinsky model: namely, that the eddy viscosity does not converge to zero at the appropriate rate as the
filter width (here equivalent to the grid spacing) is decreased.
To the right of each decay curve plotted in Fig. 4.2 is the corresponding spectral data comparison. The
three black solid lines are the CBC spectral data for the points in time corresponding to dimensional times
of t = 0.00 s, 0.28 s, and 0.66 s in our simulations. As described above, the initial FDS velocity field
(represented by the black dots) is specified to match the CBC data up to the grid Nyquist limit. From there
the spectral energy decays rapidly as discussed in [52]. For each of the spectral plots on the right, the results
of interest are the values of the red and blue dots and how well these match up with the corresponding
CBC data. For the 323 case (top-right) the results are remarkably good. Interestingly, the results for the
more highly resolved 643 case are not as good. This is because the viscous scales are rather well-resolved
at the later times in the experiment and, as mentioned, the constant coefficient Smagorinsky model is too
dissipative under such conditions. (The choice of the model constant also affects these results—better
agreement with the 643 case could be achieved with a lower value of Cs , but the agreement in the 323 case
would then be worse.)
Overall, the agreement between the FDS simulations and the CBC data is satisfactory and any discrep-
ancies can be explained by limitations of the model. Therefore, as a verification the results here are positive
in that nothing points to coding errors.

46
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
0.05 10 -3
Time
0.045

0.04
10 -4
Kinetic Energy (m²/s²)

0.035

E(k) (m³/s²)
FDS zero visc
0.03 FDS mol visc
FDS eddy visc 10 -5
0.025
Filtered CBC data
0.02

0.015 10 -6
0.01

0.005 10 -7
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
10 1 10 2 10 3
Time (s)
k (1/m)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
0.07 10 -3
FDS Time
0.06 Filtered CBC data

-4
10
Kinetic Energy (m²/s²)

0.05

E(k) (m³/s²)
0.04
10 -5
0.03

0.02
10 -6
0.01

0 10 -7
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 1 2 3
10 10 10
Time (s)
k (1/m)

Figure 4.2: (Left) Time histories of integrated kinetic energy corresponding to the grid resolutions on the right side of
the figure. In the 323 case (top), the CBC data (open circles) are obtained by applying a filter to the CBC energy spectra
at the Nyquist limit for an N = 32 grid. Similarly, for the 643 case (bottom), the CBC data are obtained from filtered
spectra for an N = 64 grid. (Right) Energy spectra for the 323 case (top) and the 643 case (bottom). The solid black
lines are the spectral data of Comte-Bellot and Corrsin at three different points in time corresponding to downstream
positions in the turbulent wind tunnel. The initial condition for the velocity field (spectra shown as black dots) in the
FDS simulation is prescribed such that the energy spectrum matches the initial CBC data. The FDS energy spectra
corresponding to the subsequent CBC data are shown by the red and blue dots. The vertical dashed line represents the
wavenumber of the grid Nyquist limit.

4.1.2 Dynamic Smagorinsky (dsmag)

In the previous section, all calculations were performed with a constant and uniform Smagorinsky coeffi-
cient, Cs = 0.2. For the canonical case of homogeneous decaying isotropic turbulence – at sufficiently high
Reynolds number – this model is sufficient. However, we noticed that even for the isotropic turbulence prob-
lem when the grid Reynolds number is low (i.e., the flow is well-resolved) the constant coefficient model
tends to over predict the dissipation of kinetic energy (see Fig. 4.2). This is because the eddy viscosity
does not converge to zero at the proper rate; so long as strain is present in the flow (the magnitude of the
strain rate tensor is nonzero), the eddy viscosity will be nonzero. This violates a guiding principle in LES
development: that the method should converge to a DNS if the flow field is sufficiently resolved.
The dynamic procedure for calculating the model coefficient (set TURBULENCE_MODEL=‘DYNAMIC
SMAGORINSKY’ on MISC) alleviates this problem. The basis of the model is that the coefficient should be
the same for two different filter scales within the inertial subrange. Details of the procedure are explained in
the following references [56, 57, 34, 58, 48]. Here we present results for the implementation of the dynamic

47
 Figure 4.3: Smagorinsky coefficient for a 643 simulation of the CBC experiment.

model in FDS. In Fig. 4.3 we show contours of the Smagorinsky coefficient Cs (x,t) at a time midway
through a 643 simulation of the CBC experiment. Notice that the coefficient ranges from 0.00 to roughly
0.30 within the domain with the average value falling around 0.17.
Next, in Fig. 4.4, we show results for the dynamic model analogous to Fig. 4.2. For the 323 case the
result is not dramatically different than the constant coefficient model. In fact, one might argue that the 323
constant coefficient results are slightly better. But there are several reasons why we should not stop here and
conclude that the constant coefficient model is superior. First, as pointed out in Pope Exercise 13.34 [59],
383 is required to resolve 80 % of the total kinetic energy (for this flow) and thus put the cutoff wavenumber
within the inertial subrange of turbulent length scales. Pope recommends that simulations which are under-
resolved by this criterion should be termed “very large-eddy simulations”—weather forecasting is a typical
example. For a 323 LES, the test filter width in the dynamic model falls at a resolution of 163 , clearly
outside the inertial range. A tacit assumption underlying the original interpretation of the dynamic model
is that both the grid filter scale and the test filter scale should fall within the inertial range, since this is the
range in which the scales of turbulent motion (in theory) exhibit fractal-like, scale similarity (recently the
procedure has been derived from other arguments [60]). With this in mind, it is perhaps not surprising that
the dynamic model does not perform optimally for the low resolution case. In the higher resolution 643
case, however, the dynamic model does perform better than the constant coefficient model—and this is the
desired result: we want better performance at higher resolution. As can be seen from the energy spectra
(lower right), the energy near the grid Nyquist limit is more accurately retained by the dynamic model. This
equates to better flow structure with fewer grid cells. Thus, for practical calculations of engineering interest
the small computational overhead of computing the coefficient may be recuperated by a reduction in cell
count.

4.1.3 Deardorff (deardorff)

In this section we
p present the results for the FDS implantation of the Deardorff isotropic eddy viscosity
model, νt = Cν ksgs ∆, where Cν = 0.1 [59, 61]. Instead of transporting the subgrid kinetic energy, in FDS

48
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
0.05 10 -3
FDS Time
0.045 Filtered CBC data

0.04
10 -4
Kinetic Energy (m²/s²)

0.035

E(k) (m³/s²)
0.03
10 -5
0.025

0.02

0.015 10 -6
0.01

0.005 10 -7
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
10 1 10 2 10 3
Time (s)
k (1/m)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
0.07 10 -3
FDS Time
0.06 Filtered CBC data

-4
10
Kinetic Energy (m²/s²)

0.05

E(k) (m³/s²)
0.04
10 -5
0.03

0.02
10 -6
0.01

0 10 -7
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 1 2 3
10 10 10
Time (s)
k (1/m)

Figure 4.4: Dynamic Smagorinsky model results (analogous to Fig. 4.2) for integrated kinetic energy (left) and spectra
(right).

ksgs is estimated with an algebraic closure based on scale similarity [62]. As may be seen in Fig. 4.5, for this
case the Deardorff model is comparable to dynamic Smagorinsky. Deardorff is cheaper computationally,
however, and tends to perform better for low resolution fire plume dynamics than the dynamic model. For
these reasons, Deardorff is the default turbulence model in FDS.

4.1.4 Vreman (vreman)

The Vreman eddy viscosity model [63] is an efficient alternative to the dynamic Smagorinsky model (DSM).
In short, the spatial fields for the velocity components are expanded in Taylor series and test filtered analyt-
ically, avoiding the expensive test filtering operations required in DSM. The resulting model is superior to
the constant coefficient Smagorinsky model (CSM) because, unlike CSM, with Vreman’s model the subgrid
dissipation automatically vanishes if the flow is fully resolved by the grid.

4.1.5 WALE (wale)

The WALE (Wall Adapting Local Eddy-viscosity) model of Nicoud and Ducros [64] was developed as a
near-wall eddy viscosity model with the proper scaling behavior without the need for a damping function
like Van Driest damping (see [65]). Details of the model are provided in the FDS Technical Guide [48].
While WALE is designed for use a near-wall model, WALE can also perform as a bulk flow eddy viscosity

49
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
0.05 10 -3
FDS Time
0.045 Filtered CBC data

0.04
10 -4
Kinetic Energy (m²/s²)

0.035

E(k) (m³/s²)
0.03
10 -5
0.025

0.02

0.015 10 -6
0.01

0.005 10 -7
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
10 1 10 2 10 3
Time (s)
k (1/m)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
0.07 10 -3
FDS Time
0.06 Filtered CBC data

-4
10
Kinetic Energy (m²/s²)

0.05

E(k) (m³/s²)
0.04
10 -5
0.03

0.02
10 -6
0.01

0 10 -7
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 1 2 3
10 10 10
Time (s)
k (1/m)

Figure 4.5: Deardorff model results (analogous to Fig. 4.2) for integrated kinetic energy (left) and spectra (right).

model. Here we confirm the appropriate choice of WALE model constant, Cw = 0.60 in FDS, by performing
simulation of decaying isotropic turbulence, as was also performed in [64].

50
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
0.05 10 -3
FDS Time
0.045 Filtered CBC data

0.04
10 -4
Kinetic Energy (m²/s²)

0.035

E(k) (m³/s²)
0.03
10 -5
0.025

0.02

0.015 10 -6
0.01

0.005 10 -7
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
10 1 10 2 10 3
Time (s)
k (1/m)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
0.07 10 -3
FDS Time
0.06 Filtered CBC data

-4
10
Kinetic Energy (m²/s²)

0.05

E(k) (m³/s²)
0.04
10 -5
0.03

0.02
10 -6
0.01

0 10 -7
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 1 2 3
10 10 10
Time (s)
k (1/m)

Figure 4.6: Vreman model results (analogous to Fig. 4.2) for integrated kinetic energy (left) and spectra (right).

51
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
0.05 10 -3
FDS Time
0.045 Filtered CBC data

0.04
10 -4
Kinetic Energy (m²/s²)

0.035

E(k) (m³/s²)
0.03
10 -5
0.025

0.02

0.015 10 -6
0.01

0.005 10 -7
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
10 1 10 2 10 3
Time (s)
k (1/m)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
0.07 10 -3
FDS Time
0.06 Filtered CBC data

-4
10
Kinetic Energy (m²/s²)

0.05

E(k) (m³/s²)
0.04
10 -5
0.03

0.02
10 -6
0.01

0 10 -7
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 1 2 3
10 10 10
Time (s)
k (1/m)

Figure 4.7: WALE model results (analogous to Fig. 4.2) for integrated kinetic energy (left) and spectra (right).

52
4.2 Jet Centerline Velocity Decay (jet)
Gregor Jäger, BFT Cognos GmbH, Germany

The jet_ series in the Validation/Turbulent_Jet directory tests the ability of subgrid stress closure (turbu-
lence model) to correctly capture the transition from laminar to turbulent flow and subsequent momentum
decay for an incompressible jet.
The FDS solution is compared with experimental observations from Kümmel [66]. In the initial region
of the jet, x < x0 , the velocity u is constant near the axis and is equal to the velocity at the nozzle orifice (u0 ).
For a rectangular nozzle orifice of height h = 0.8 and width b = 0.8, the initial length is calculated by

h
x0 = (4.1)
m
where the mixed number m is between 0.12 and 0.20.
In the transition region, viscous mixing spreads over the entire jet flow and the flow velocity along the
axis decreases. The mean flow velocity at the jet centerline umax (x) is calculated by

1 √
r
umax (x) x0 b
= = bh (4.2)
u0 x h mx
Four turbulence models are tested: (1) Constant Smagorinsky (csmag), (2) Dynamic Smagorinsky (ds-
mag), (3) Deardorff (FDS default), and (4) Vreman. The results are plotted in Fig. 4.8. For each model, two
grid resolutions are run, corresponding to h/δ x = 8 (colored dashed lines) and h/δ x = 16 (colored solid
lines).
FDS6.7.9-1883-g2e2d47d-master
1.2
Jet Centerline Velocity Decay
1

0.8

0.6

0.4

0.2
0 5 10 15 20 25

Figure 4.8: Centerline velocity decay for a high Reynolds number jet (Reh = 1 × 105 ) with a square orifice of side h.
FDS results (colored lines) are shown for various turbulence models at two grid resolutions. For comparison, analytical
results (black lines) are shown for two values of the mixing number m.

53
Chapter 5

Boundary Effects

Wall flows have long been a challenge for large-eddy simulation (LES) [67, 68, 69, 59, 70]. In spite of their
promise and sophistication, practical LES codes are resigned to model the wall shear stress as opposed to
resolving the dynamically important length scales near the wall. FDS uses a log law velocity profile [59] for
modeling turbulent flow near a wall.

5.1 The FDS Moody Chart (poiseuille, moody, z0)

At a minimum, a wall model should accurately reproduce the mean wall stress for flow in a straight square
duct. We verify that this is true for FDS by reproducing the Moody chart, a plot of friction factor versus
Reynolds number for pipe flow [71].
Details of the FDS formulation are given in the Technical Guide [48]. Here we provide only the salient
components of the model necessary for treatment of constant density channel flow. The filtered continuity
and momentum equations are:
∂ ūi
= 0, (5.1)
∂ xi
1 dp ∂ p̃ ∂ τ̄i j ∂ τirj
 
∂ ūi ∂ ūi ū j
+ =− + + + , (5.2)
∂t ∂xj ρ dxi ∂ xi ∂ x j ∂xj
where τirj is the deviatoric part of the residual stress tensor. In this work we specify a constant pressure
drop dp/ dx in the streamwise direction to drive the flow. The hydrodynamic pressure p̃ is obtained from
a Poisson equation which enforces (5.1). Details of the model used for the viscous stress at the wall, e.g.,
τ̄xz |z=0 , are provided in the FDS Technical Reference Guide [1].

5.1.1 Laminar Results

As verification of the no-slip boundary condition and further verification of the momentum solver in FDS, we
perform a simple 2-D laminar (Poiseuille) flow calculation of flow through a straight channel. The height of
the channel is H = 1 m and the length of the channel is L = 8 m. The number of grid cells in the streamwise
direction x is Nx = 8. The number of cells in the wall-normal direction z is varied Nz = {8, 16, 32, 64}. The
fluid density is ρ = 1.2 kg m−3 and the viscosity is 0.025 kg m−1 s−1 . The mean pressure drop is prescribed
to be dp/ dx = −1 Pa m−1 resulting in ReH ≈ 160. The (Moody) friction factor f , which satisfies

L1 2
∆p = f ρ ū , (5.3)
H2

55
 FDS6.7.9-1300-gbeeb4d8-ht1d
10 -2

Friction Factor Error

10 -3

10 -4

10 -2 10 -1

Figure 5.1: FDS exhibits second-order convergence for laminar (Poiseuille) flow in a 2-D channel.

is determined from the steady-state mean velocity ū which is output by FDS for the specified pressure drop.
The exact friction factor for this flow is fexact = 24/ReH . The friction factor error | f − fexact | is plotted
for a range of grid spacings δ z = H/Nz in Fig. 5.1 demonstrating second-order convergence of the laminar
velocity field.

5.1.2 Turbulent Results

Smooth Walls To verify the wall model for turbulent flow we perform 3-D LES of a square duct with
periodic boundaries in the streamwise direction and a constant and uniform mean pressure gradient driving
the flow. The problem set up is nearly identical to the laminar cases of the previous section except here we
perform 3-D calculations and maintain cubic cells as we refine the grid: we hold the ratio 8:1:1 between
Nx : Ny : Nz for all cases. The cases shown below are identified by their grid resolution in the z direction.
The velocity field is initially at rest and develops in time to a mean steady state driven by the specified
mean pressure gradient. The presence of a steady state is the result of a balance between the streamwise
pressure drop and the integrated stress from the wall model. FDS outputs the planar average velocity in the
streamwise direction and once a steady state is reached this value is used to compute the Reynolds number
and the friction factor. Table 5.1 provides a case matrix: nine cases for three values of specified pressure
drop and three grid resolutions. The nominal Reynolds number (obtained post-run) is listed along with
the friction factor from the most refined FDS case and the friction factor computed (iteratively) from the
Colebrook equation,  
1 s/H 2.51
√ = −2.0 log10 + √ , (5.4)
f 3.7 ReH f
which is a fit to the turbulent range of the Moody chart (for example, see Ref. [72]). The parameter s/H
is the relative roughness where H is the hydraulic diameter of the pipe or channel and ReH is the Reynolds
number based on H. To provide a qualitative picture of the flow field, Fig. 5.2 shows contours of streamwise
velocity for the case dp/ dx = −1 Pa m−1 and Nz = 32.

56
Figure 5.2: LES of square duct flow with smooth walls and periodic streamwise boundaries. For this image Nz = 32
and the mean pressure drop is dp/ dx = −1 Pa m−1 resulting in ReH = 7.5 × 105 and a friction factor of f = 0.0128.

Table 5.1: Case matrix and friction factor results for turbulent channel flow with smooth walls. The height
of the first grid cell δ z is given in viscous units z+ for each case. Additionally, the table gives the nominal
Reynolds number ReH and the FDS friction factor results compared to the Colebrook equation (5.4).

dp/ dx z+ ReH f FDS f Colebrook Rel. error

(Pa/m) Nz = 8 Nz = 16 Nz = 32 (Nz = 32) Eq. (5.4) %
-0.01 190 95 47 5.9 × 104 0.0212 0.0202 4.8
-1. 1.9 × 103 950 470 7.5 × 105 0.0128 0.0122 4.6
-100. 1.9 × 104 9.5 × 103 4.7 × 103 9.8 × 106 0.0077 0.0081 6.0

57
Rough Walls With the same grid set up as described above the smooth walls, a series of cases labeled as
z01 in the repository [44] were run at various roughness heights, grid resolutions, and Reynolds numbers.
The results are presented together with the smooth wall cases in Fig. 5.3. The laminar cases provide accurate
results for two different Reynolds numbers. Both the smooth wall and rough wall treatments behave well
over the range tested.

FDS6.7.9-1300-gbeeb4d8-ht1d

10 -1 0.1

0.01

0.001

0.0001
10 -2

10 2 10 4 10 6 10 8

Figure 5.3: The FDS Moody chart: friction factor, f , versus Reynolds number, ReH , based on duct height. The solid
line for Re < 2000 is the analytical result for 2-D Poiseuille flow, f = 24/Re. The solid lines for Re > 2000 (from the
Colebrook equation (5.4)) are Relative roughness s/H shown on the right axis.

5.2 Blasius boundary layer (blasius)

Hyun-Wook Park and Jung-il Choi, Yonsei University, Seoul, South Korea

Consider a steady flow over a semi-infinite flat plate aligned with the flow with free stream velocity, u0 . The
Prandtl boundary layer equations [73] are:
∂u ∂u ∂ 2u
u +w =ν 2 , (5.5)
∂x ∂z ∂z
∂u ∂w
+ = 0. (5.6)
∂x ∂z
Paul Richard Heinrich Blasius found a solution to Eqs. (5.5) and (5.6) by introducing a similarity variable:
z
η≡p . (5.7)
νx/u0
Substituting η and the stream function (u = ∂ ψ/∂ z, w = −∂ ψ/∂ x) into Eq. (5.5) yields the following
ODEs:
1 00 000
ff +f = 0 (5.8)
2
1 The atmospheric community typically uses z0 to denote the roughness height. However, we also use s interchangeably with z0
to conform to the notation used in Pope [59].

58
 0
u = u0 f (5.9)
r
1 ν u0 0
w = (η f − f ) (5.10)
2 x
The solution of Eq. (5.8) is the Blasius solution [73].
The physical domain of the present verification case is a rectangle with length L = 0.15 m in the stream-
wise direction and H = 0.3 m in the wall-normal direction. The grid spacing is uniform (δ x = L/Nx ,
δ z = H/Nz ) in each direction, and Nx = 1.5 Nz with Nz = {16, 32, 64} for each test series. The Boundary
conditions for the inlet and outlet boundaries are OPEN with the dynamic pressure set to zero at the inlet
boundary. On the top boundary, the gradient of w in the z direction is zero combined with a free-slip con-
dition that combines to yield an irrotational field. The boundary condition along the bottom of the domain
is free-slip upstream of the plate (x < 0.5 m) and no-slip on the plate. The initial velocity is 1 m/s and u0 is
taken as the value of u at the top boundary since boundary conditions at z = ∞ cannot be directly applied.
The case is run with constant properties: ρ = 1.2 kg/m3 , µ = 0.001 kg/m/s.
Figure 5.4 (left) shows a comparison of the streamwise velocity from FDS with the Blasius solution.
The FDS solution (lines) converges to Blasius (open circles) as the mesh resolution is increased. Figure 5.4
(right) shows the root mean square (RMS) error, ε, in u (streamwise component) at x = 0.1 m (0.05 m from
the leading edge of the plate) for the dimensional Blasius solution. The error is defined by
v
u 1 Nz
u
ε ≡ t ∑ [uk − uB,k ]2 (5.11)
Nz k=1

where k is the z index. The value uk represents the FDS result for u at the staggered grid storage location
for cell k, uB,k is the dimensional Blasius solution at the corresponding location. The slight departure from
second-order accuracy is likely due to numerical issues in representing the singularity in the Blasius solution
at the leading edge of the plate as well as errors in the boundary conditions due to truncation of the domain.
FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
0.15 10 0
Blasius FDS
N z =16 O( z )
N z =32 O( z 2 )
0.1 10 -1
RMS Error (m/s)

N z =64
z (m)

0.05 10 -2

0 10 -3
0 0.2 0.4 0.6 0.8 1 10 -3 10 -2 10 -1
u (m/s) Grid Spacing, z (m)

Figure 5.4: (Left) Comparison of theoretical and numerical results for the Blasius velocity profile (u component).
(Right) Error between FDS and Blasius solution for u(z) plotted for each grid resolution.

59
5.3 Pohlhausen thermal boundary layer (Pohlhausen)
This write up follows Appendix B of [74]. The nondimensional temperature as a function of the similarity
variable η is taken to be
T (η) − Tw
θ (η) ≡ (5.12)
T∞ − Tw
The energy equation may then be written as

d2 θ 1 dθ
2
+ Pr f =0 (5.13)
dη 2 dη

Given the boundary conditions θ = 0 at y = 0, η = 0 and θ = 1 at y = ∞, η = ∞, and the function f (η)

from the Blasius solution in Sec. 5.2, the Pohlhausen solution for the temperature profile is
 
Pr η
Z η Z
exp − f dη dη
0 2 0
θ (η) = Z ∞   (5.14)
Pr η
Z
exp − f dη dη
0 2 0

This solution is plotted in Fig. 5.5 (left) for different Prandtl numbers (Pr) and may be compared to the plot
in Fig. B-2 of [74] for verification.
In this test series, the 2-D FDS domain is set 10 m in length and 1 m in height. The simulation is run as
a DNS with the viscosity, conductivity, and specific heat set to provide Prandtl numbers of [0.5, 1, 2]. The
grid resolution (after a convergence study) is set to δ x = δ z = 1.25 cm. The inlet velocity is set to 1 m/s
with an ambient air temperature of T∞ = 20 ◦C. The wall boundary is set to a fixed temperature of Tw = 21
◦C. The outflow is set to OPEN. The top boundary is homogeneous Neumann for velocity and Dirichlet for

temperature at T∞ = 20 ◦C. The simulation is run to steady state. The resulting temperature profiles T (z) at
x = 5 m are shown in Fig. 5.5 (right).
FDS6.7.9-1798-gfdcf6c9-master FDS6.7.9-1798-gfdcf6c9-master
1 1
Pr=0.5 Pohlhausen Pr=0.5
Pr=1 FDS Pr=0.5
0.8 Pr=2 0.8 Pohlhausen Pr=1
FDS Pr=1
Pohlhausen Pr=2
0.6 0.6 FDS Pr=2

0.4 0.4

0.2 0.2

0 0
0 2 4 6 8 20 20.2 20.4 20.6 20.8 21

Figure 5.5: (Left) Pohlhausen similarity solution for nondimensional temperature. (Right) FDS results for vertical
profile of temperature at x = 5 m for different Prandtl numbers compared to the Pohlhausen solution.

60
5.4 Near-Wall Grid Resolution (yplus)
Abbas Jaber, NIST SURF student
The distance from the wall measured in “viscous wall units” is called y+ . In order to verify that y+ is
calculated correctly within FDS, we analyze an example of Poiseuille flow in a square channel with periodic
boundaries. The pressure gradient in the x-direction, the dynamic viscosity, and the dimensions of the
channel are known. This allows us to easily determine the expected values of wall shear stress,
dp
τw = −δ , (5.15)
dx
where p is pressure and δ is half the height of the channel. This, in turn, allows for determining the friction
velocity, r
τw
uτ ≡ . (5.16)
ρ
From here we calculate the exact value of y+ for the given flow,
y uτ y
y+ ≡ = . (5.17)
δν ν
In FDS, τw is obtained only from the wall model. The resolution of the wall-normal streamwise velocity
gradient increases with mesh resolution; the value of y+ decreases accordingly. In this verification case, a
wall device is used to output y+ for different mesh resolutions in the y-direction. y+ is output for a given
δ y, height of a cell, δ y = [1/8, 1/16, 1/32]. Then, the device values obtained are compared with the exact
value of y+ for a given mesh resolution (Fig. 5.6). An FDS wall device outputs the mid-cell value of the
first cell next to the wall, hence, the y value (y in Eq. (5.17)) used to determine the exact y+ varies with δ y
(y = δ y/2).
FDS-6.8.0-0-g886e009-release
0.05

0.04
y + predicted

0.03

0.02
exact
FDS
0.01
0.01 0.02 0.03 0.04 0.05
+
y specified

Figure 5.6: Verification of FDS y+ output for δ y = [1/8, 1/16, 1/32].

5.5 Heated Channel Flow (heated_channel)

The simulation of turbulent channel flow (Reτ = 180) with a heat source was performed to verify the log law
near-wall model for heat transfer [1]. The height of the channel is H = 2 m, the width is W = 6 m, and the
length is L = 12 m. There are 96 cells in the streamwise direction (δ x = 0.125 m) and 48 cells in the spanwise

61
and wall-normal directions (δ y = δ z = 0.125 m). The midpoint of the first cell in viscous units is z+ ≈ 18,
putting the first velocity unknown within the buffer layer. For this reason, the dynamic Smagorinsky SGS
model is used for the residual stress. Additionally, these cases were run with (SIMULATION_MODE=’LES’
on MISC), which accounts for both molecular the turbulent transport coefficients, a must for distinguishing
between fluids with different molecular Prandtl or Schmidt numbers (here the Reynolds number is so low
that molecular effects may dominate locally). The flow is maintained by a mean pressure gradient (9.0 ×
10−6 Pa/m) in the streamwise direction (follows from Reτ = 180). A constant volumetric heat source (0.5
W/m3 ) is imposed on the fluid. The boundary conditions for momentum equations are no-slip at top and
bottom walls and periodic for the streamwise and spanwise directions. The boundary condition for the
energy equation is constant temperature (Tw = 20 °C) at the wall. To evaluate the effects of Prandtl number
(Pr) and compare the DNS study [75] with FDS, the specific heat (c p = 1 kJ/(kg · K)) and viscosity (µ =
1.8216 × 10−5 kg/(m · s)) are fixed and the conductivity, k, of the fluid (air) is adjusted to:
Pr k (W/(m · K))
0.10 1.82 × 10−1
0.71 2.59 × 10−2
1.00 1.82 × 10−2
2.00 9.11 × 10−3

In Figure 5.7, FDS results are compared to DNS data for mean streamwise velocity, u+ = hui/uτ , and
mean temperature, T + = (hT i − Tw )/Tτ . The temperature scale is given by
00
hq̇w i
Tτ ≡ , (5.18)
ρc p uτ
00
where hq̇w i, ρ, and c p are the mean heat flux at the wall, density, and specific heat, respectively. Fully
developed, statistically stationary flow is achieved after approximately 20 flow through times. Statistics are
gathered between 40 and 60 flow through times.
As seen in the left side of Fig. 5.7, the velocity profile shows good agreement with the DNS result. The
near-wall temperature profile for each Pr is shown on the right side of Fig. 5.7. The mean wall heat flux is
correct, as indicated by the accurate result of the first off-wall grid value (in the viscous region).
FDS6.7.9-2220-gb05582a-master FDS6.7.9-2220-gb05582a-master
30 30
FDS FDS
DNS Re =180 DNS Re =180 Pr=2.0
25 25

20 20
u+

T+

15 15 Pr=0.71

10 10

5 5 Pr=0.10

0 0
0 1 2 3 0 1 2 3
10 10 10 10 10 10 10 10
+ +
z z

Figure 5.7: Comparison of FDS (lines with squares) with DNS (circles) for heated channel flow. (Left) Wall-normal
velocity profile compared to DNS of Moser et al. [76]. (Right) Wall-normal mean profiles of temperature for molecular
Prandtl numbers (Pr) = 0.10, 0.71, and 2.0, compared to DNS of Kim and Moin [75].

62
5.6 Forced Convection Flat Plate Boundary Layer (forced_conv_flat_plate)
In this section, we set up a series of cases to test the forced convection model in FDS. The scenario is a
constant temperature turbulent flat plate boundary layer. Gravity is set to zero in order to explicitly exercise
the forced convection model. The inlet flow velocity is constant and uniform at the leading edge of the plate.
Three velocities are considered, [0.1, 1, 10] m/s. The free stream fluid is nitrogen (ρ = 1.165 kg/m3 ) with
constant properties: dynamic viscosity, µ = 1 × 10−5 kg/m/s, thermal conductivity, k = 0.01 W/m/K, and
specific heat, c p = 1 kJ/kg/K. The free stream temperature is 20 ◦C and the wall temperature is a constant 30
◦C. The plate is 16 m long with constant and uniform inflow velocity and an OPEN outflow boundary. The

domain is 2 m in width and 2 m in height. The spanwise boundaries are PERIODIC and the top boundary is
MIRROR.

Figure 5.8: Smokeview rendering of a temperature slice through the center of the bounary layer as well as the heat
transfer coefficient on the constant temperature wall boundary for the 10 m/s case at 10 cm resolution.

The boundary layer thicknesses for each of the three velocity inlet conditions are [0.5227, 0.3298,
−1/5
0.2081] m, respectively, evaluated at x = 16 m from the correlation δx = 0.37 x Rex (see [77] Table
7.7). Three grid resolutions are considered, [25, 10, 2.5] cm. The resolution of the boundary layer at the
end of domain (δ16 /δ x) is provided below in Tab. 5.2. Note that the boundary layer is generally under-
resolved (δx /δ x < 10). Thus, accuracy of the wall heat flux relies heavily on the model for the heat transfer
coefficient. The simulation is set to run for 20 flow through times with averaging of results over the last half
of the simulation. An image of the spatially developing thermal boundary layer with medium resolution and
10 m/s inflow is shown in Fig. 5.8. Also shown is the local heat transfer coefficient along the wall.

Table 5.2: Forced convection boundary layer resolution, δx=16 m /δ x.

free stream velocity

δx 0.1 m/s 1 m/s 10 m/s
25 cm 2.1 1.3 0.83
10 cm 5.2 3.3 2.1
2.5 cm 21.0 13.0 8.3

63
 The results are plotted in Fig. 5.9. We are comparing FDS output to the local turbulent forced convection
correlation,
4/5
Nux = 0.0296 Rex Pr1/3 (5.19)
In these cases, the Prandtl number, Pr, is taken to be 0.7. The correlation is plotted as the line labeled “forced
local” in the legend of the plots. The plots are comparisons of “local” Nusselt number as a function of
“local” Reynolds number, as indicated by the subscript x. In these formulae, the length scale is given by the
distance from the leading edge of the plate. Of course, FDS does not know this information when it is trying
to determine the heat transfer coefficient on a given wall cell. Hence, it uses a constant “convective length
scale”, in this case set to 1 m (the default value). Also included in each plot is a dashed line representing
2Lx/δ x. If this curve is larger than the correlation this is an indication that the “DNS” value was used for
the heat transfer coefficient, that is h = 2k/δ x. This only happens for the 0.1 m/s case at the highest grid
resolution.
In order to make a comparison with the local correlation we need to scale the FDS heat transfer coeffi-
cient by the ratio of the local temperature difference to the free stream temperature difference assumed in the
correlation. We add device outputs of ’HTC’ and ’CONVECTIVE HEAT FLUX’ along the length of the plate.
From these two outputs we can discern the temperature difference at the wall, ∆TFDS . The correlation tem-
perature difference is Tw − T∞ . For a given heat flux the following holds, q̇00w = hFDS ∆TFDS = hcor (Tw − T∞ ).
For a given location, x, along the plate the FDS results for Nux are taken to be,
 
hcor,x x hFDS,x ∆TFDS x
Nux = = (5.20)
k Tw − T∞ k

64
 FDS6.7.9-2196-g979fd84-master FDS6.7.9-2196-g979fd84-master
1400 5000
forced local forced local
Velocity = 0.1 m/s Velocity = 1 m/s
1200 FDS x =25cm FDS x =25cm
FDS x =10cm 4000 FDS x =10cm
FDS x =2.5cm FDS x =2.5cm
1000
2*1/0.25* x 2*1/0.25* x
2*1/0.10* x 3000 2*1/0.10* x
800
Nu x

Nu x
2*1/0.025* x 2*1/0.025* x

600 2000
400
1000
200

0 0
0 0.5 1 1.5 2 0 0.5 1 1.5 2
Re x 5 Re x 6
10 10
4 FDS6.7.9-2196-g979fd84-master
10
3
forced local
Velocity = 10 m/s
FDS x =25cm
2.5 FDS x =10cm
FDS x =2.5cm
2 2*1/0.25* x
2*1/0.10* x
Nu x

2*1/0.025* x
1.5

0.5

0
0 0.5 1 1.5 2
Re x 10 7

Figure 5.9: FDS results for local Nusselt number in turbulent boundary layer flow over a flat plate. Clockwise from
top-left the flow velocity increases from 0.1 m/s to 10 m/s. Dashed lines indicate local Nusselt number based on
2L/(δ n), where L = 1 m is the convective length scale and δ n is the wall-normal grid resolution.

5.7 Natural Convection Horizontal Plates (natconh)

This section examines the problem of thermal convection between horiozontal plates of fixed temperature
with the hot plate on bottom. The following correlation for the Nusselt number is taken from [74]:
Nu = C Ran (5.21)
where the Rayleigh number (product of Grashof and Prandtl numbers), Ra, is given by
gβ (T1 − T2 )δ 2
Ra = (5.22)
να
Here g is the magnitude of the gravitational acceleration, T1 and T2 are the respective hot and cold plate tem-
peratures, δ is the distance between the plates, ν is the kinematic viscosity, and α is the thermal diffusivity
of the fluid.
The values of C and n depend on the Rayleigh number range, as shown below in Tab. 5.3.

Case setup
The fluid is set to ’LJ AIR’ with constant viscosity, specific heat, and thermal conductivity: µ = 1.8216 ×
10−5 kg/m/s, c p = 1 kJ/kg/K, and k = 0.018216 W/m/K, which provides a Pr = 1 fluid. (Note that the

65
 Table 5.3: Free convection correlation parameters for horizontal plates [74].

Ra C n
< 1700 1 0
1700-17000 0.059 2/5
7000-3.2 × 105 0.212 1/4
> 3.2 × 105 0.061 1/3

correlations are valid for Pr in the range 0.5-2.)

There are 19 cases ranging in Ra from roughly 1 - 1014 . The Rayleigh number is varied by changing
both the distance between the plates, δ ranges from 2 mm to 20 m, and the temperature difference. The top
plate is held fixed at 20 ◦C and the bottom plate temperature varies from 30 ◦C to 230 ◦C.
Three different grid resolutions are tested for the Cartesian cases: δ /δ z = [8,16,32]. The resolution is
uniform with δ x = δ y = δ z. The width and depth of the plates are each 8δ . The boundary conditions in the
x and y directions are periodic. The average heat flux on each plate is monitored and the simulation is run
until a clear stationary state is reached. The results are presented below in Fig. 5.10 (left).
For the complex geometry cases, the grid and gravity vector are rotated by 18 degrees. A hollow cavity
is formed using the GEOM namelist format with a hot surface below and a cold surface on top and insulated
side walls. Two grid resolutions are tested: δ /δ z = [8,16]. The results are shown in Fig. 5.10 (right).
FDS6.7.9-1883-g2e2d47d-master FDS6.7.9-2047-g0fe35fe-master
10 4 10 4
n n
Correlation, Nu = C Ra Correlation, Nu = C Ra
FDS H/ x =8 GEOM H/ x =8
FDS H/ x =16 GEOM H/ x =16
FDS H/ x =32
Nusselt Number

Nusselt Number

10 2 10 2

10 0 10 0
0 5 10 15 0 5 10 15
10 10 10 10 10 10 10 10
Rayleigh Number Rayleigh Number

Figure 5.10: FDS results for Nusselt number in Rayleigh-Bénard convection (natural convection between horizontal
plates)—(left) Cartesian geometry; (right) cutcell geometry.

5.8 Natural Convection Vertical Enclosure (natconv)

This section examines the problem of thermal convection in an enclosure with vertical plates of fixed tem-
perature and insulated top and bottom boundaries. The Nusselt number correlation requires an additional
factor to account for the ratio of the length L to the depth δ [74]:
 m
n L
Nu = C Ra (5.23)
δ

The values of C, n, and m are given below in Tab. 5.4.

66
Table 5.4: Free convection correlation parameters for a vertical enclosure, valid for the ranges Pr=[0.5-2]
and L/δ =[11-42] [74].

Ra C n m
< 2000 1 0 0
6000-2 × 105 0.197 1/4 -1/9
2 × 105 -1.1 × 107 0.073 1/3 -1/9

Case setup
The setup for this series is very similar to the horizontal configuration discussed above (see Sec. 5.7). The
fluid is set to ’LJ AIR’ with constant viscosity, specific heat, and thermal conductivity: µ = 1.8216 × 10−5
kg/m/s, c p = 1 kJ/kg/K, and k = 0.018216 W/m/K, which provides a Pr = 1 fluid.
There are 17 cases ranging in Ra from roughly 1 - 1014 . The Rayleigh number is varied by changing
both the distance between the plates, δ ranges from 2 mm to 20 m, and the temperature difference. The right
(cold) plate is held fixed at 20 ◦C and the left plate temperature varies from 30 ◦C to 230 ◦C.
Three different grid resolutions are tested for the Cartesian cases: δ /δ x = [8,16,32]. The resolution is
uniform with δ x = δ y = δ z. The length (vertical dimension) and depth (spanwise dimension) of the plates
are 16δ and 4δ , respectively. The boundary conditions in the y and z directions are homogeneous Neumann
(MIRROR or “insulated”). The average heat flux on each plate is monitored and the simulation is run until a
clear stationary state is reached. The results are presented below in Fig. 5.11 (left).
For the complex geometry cases, the grid and gravity vector are rotated by 18 degrees. A hollow cavity
is formed using the GEOM namelist format with a hot surface on the left and a cold surface on the right and
insulated side walls. Two grid resolutions are tested: δ /δ z = [8,16]. The results are shown in Fig. 5.11
(right).
FDS6.7.9-1883-g2e2d47d-master FDS6.7.9-1823-gf57ced9-master
10 4 10 4
n m n
Correlation, Nu = C Ra (L/H) Correlation, Nu = C Ra (L/H) m
FDS H/ x =8 GEOM H/ x =8
FDS H/ x =16 GEOM H/ x =16
FDS H/ x =32
Nusselt Number

Nusselt Number

10 2 10 2

10 0 10 0

10 0 10 5 10 10 10 15 10 0 10 5 10 10 10 15
Rayleigh Number Rayleigh Number

Figure 5.11: FDS results for Nusselt number in natural convection within a vertical enclosure—(left) Cartesian
geometry; (right) cutcell geometry.

67
5.9 Free Convection from a Sphere (free_conv_sphere)
In this section we present results for a set of four cases ranging in Rayleigh number from approximately
1 to 109 . Yuge [78] and Amato and Tien [79] propose the following Nusselt number correlations for free
convection from a sphere [74].
1/4
Nu f = 2 + 0.43 Ra f for 1 < Ra f < 105 (5.24)
1/4
Nu f = 2 + 0.50 Ra f for 3 × 105 < Ra f < 8 × 108 (5.25)

Case setup
The sphere is generated using the GEOM namelist with three levels. The surface temperature is uniform and
constant at 30 ◦C. The Rayleigh number is changed by adjusting the diameter of the sphere from 1 mm to 1
m. Fluid properties are from ’LJ AIR’ and held constant: µ = 1.8216 × 10−5 kg/m/s, c p = 1 kJ/kg/K, and
k = 0.018216 W/m/K, which provides a Pr = 1 fluid. The convective heat flux is integrated over the surface
of the sphere. All external boundaries are OPEN and the grid resolution is uniform. Results are presented in
Fig. 5.12 for two grid resolutions corresponding to D/δ x = [8,16].

FDS6.7.9-1883-g2e2d47d-master

1/4
Yuge (1960), Nu = 2 + 0.43 Ra
1/4
Amato & Tien (1972), Nu = 2 + 0.5 Ra
2 FDS D/ x =8
10 FDS D/ x =16
Nusselt Number

1
10

0
10
0 5 10
10 10 10
Rayleigh Number

Figure 5.12: FDS results for free convection from a sphere.

68
5.10 Ribbed Square Duct Flow (ribbed_channel)
5.10.1 Cartesian Geometry
In this test case, we compare FDS results with particle-image velocimetry (PIV) data from Casara and Arts
[80, 81] for flow in a square duct with a ribbed obstruction. This configuration was also studied numerically
by [82]. The rib height is h = 0.03 m and the duct height is D = 0.1 m. The distance between ribs in the
periodic duct (the ,“pitch”) is 10 h. The length of our domain is therefore 0.3 m. The bulk velocity Ub is
6.2 m/s with ambient air at 20 ◦C (density is 1.2 kg/m3 and the dynamic viscosity is 1.86 × 10−5 kg/m/s).
The Reynolds number based on D is nominally 40,000. The streamwise mean pressure drop is adjusted to
achieve the desired mean bulk flowrate (see Fig. 5.13). The FDS mesh is uniform in each direction. We test
grid resolutions of h/δ x = {3, 6, 12, 24}.
A side view image of mean velocity vectors in the 3-D flow is shown in Fig. 5.14 for the h/δ x = 12
case. The goals of this calculation are to predict the flow profile above the ribs and the recirculation patterns
on the windward and leeward side of the ribs. To measure the FDS results, a line of devices (represented
by the green dots in the figure) is placed near the floor of the duct. A change in sign of the mean u velocity
component indicates a reattachment location. Results for the mean streamwise velocity near the surface
are shown in the upper-left plot of Fig. 5.15. The independent axis on the plots ranges from 0 m to 0.3 m,
representing the center to center distance between ribs. Note that the image is offset and ranges from -0.125
m to 0.175 m. As can be seen from the plot, the h/δ x = 12 resolution (green sideways triangles) provides
sufficient accuracy of the reattachment length (note that the correspondence with the vectors in the image).
Despite the fact that the coarser cases do not perfectly capture the reattachment region, it is remarkable that
with h/δ x = 3 we achieve any recirculation at all. The mean streamwise velocity profile at the rib centerline
is shown in the upper-right plot of Fig. 5.15. The finest resolution, h/δ x = 24, is sufficient to accurately
capture the root-mean-square velocity fluctuations on top of the rib, as can be seen in the bottom-right plot
of the figure.
Note that the discrepancy between the model and experimental results for the mean streamwise profile
above the rib (Fig. 5.15, upper-right) can be explained by an asymmetry in the PIV data, as can be seen in
Figures 11 and 12 of [81]. These results should not be interpreted as spanwise means, they are based on PIV
data along the centerline plane.
FDS6.7.9-1300-gbeeb4d8-ht1d

8
Bulk Velocity (m/s)

4
Exp Bulk Velocity
FDS h/ x =3
2 FDS h/ x =6
FDS h/ x =12
FDS h/ x =24
0
0 2 4 6 8 10
Time (s)

Figure 5.13: Bulk velocity for the ribbed channel cases measured in the duct section in front of the rib compared with
the experimental value of 6.2 m/s (thick black line).

69
 Figure 5.14: Mean velocity vectors colored by velocity magnitude for the ribbed square duct case with h/δ x = 12.

FDS6.7.9-1300-gbeeb4d8-ht1d FDS6.7.9-1300-gbeeb4d8-ht1d
0.6
PIV data PIV data
FDS h/ x =3 3 FDS h/ x =3
0.3 FDS h/ x =6 FDS h/ x =6
FDS h/ x =12 FDS h/ x =12
FDS h/ x =24 2.5 FDS h/ x =24
U/U b

z/h

0
2

-0.3
1.5

-0.6 1
0 2 4 6 8 10 0 0.5 1 1.5 2
x/h U/U b

FDS6.7.9-1300-gbeeb4d8-ht1d FDS6.7.9-1300-gbeeb4d8-ht1d
0.6
PIV data PIV data
0.5 FDS h/ x =3 3 FDS h/ x =3
FDS h/ x =6 FDS h/ x =6
0.4 FDS h/ x =12 FDS h/ x =12
FDS h/ x =24 2.5 FDS h/ x =24
U rms /U b

z/h

0.3
2
0.2
1.5
0.1

0 1
0 2 4 6 8 10 0 0.2 0.4 0.6 0.8 1
x/h U rms /U b

Figure 5.15: Mean and RMS velocity profiles for ribbed square duct flow. Comparison with PIV data from Casara
and Arts [80, 81].

70
5.10.2 Complex Geometry
Below we plot the results for the ribbed_channel_geom test series. These cases are identical to the
Cartesian geometry discussed above, except they use a GEOM line with a grid-aligned XB region to model the
square rib. The same mean force vectors are used as in the analogous Cartesian cases.
FDS6.7.9-1300-gbeeb4d8-ht1d

Bulk Velocity (m/s) 6

4
Exp Bulk Velocity
FDS h/ x =3
2 FDS h/ x =6
FDS h/ x =12
FDS h/ x =24
0
0 2 4 6 8 10
Time (s)

Figure 5.16: Bulk velocity for ribbed channel geom cases measured in the duct section in front of the rib compared
with the experimental value of 6.2 m/s.

71
 FDS6.7.9-1300-gbeeb4d8-ht1d FDS6.7.9-1300-gbeeb4d8-ht1d
0.6
PIV data PIV data
FDS h/ x =3 3 FDS h/ x =3
0.3 FDS h/ x =6 FDS h/ x =6
FDS h/ x =12 FDS h/ x =12
FDS h/ x =24 2.5 FDS h/ x =24
U/U b

z/h
0
2

-0.3
1.5

-0.6 1
0 2 4 6 8 10 0 0.5 1 1.5 2
x/h U/U b

FDS6.7.9-1300-gbeeb4d8-ht1d FDS6.7.9-1300-gbeeb4d8-ht1d
0.6
PIV data PIV data
0.5 FDS h/ x =3 3 FDS h/ x =3
FDS h/ x =6 FDS h/ x =6
0.4 FDS h/ x =12 FDS h/ x =12
FDS h/ x =24 2.5 FDS h/ x =24
U rms /U b

z/h
0.3
2
0.2
1.5
0.1

0 1
0 2 4 6 8 10 0 0.2 0.4 0.6 0.8 1
x/h U rms /U b

Figure 5.17: Mean and RMS velocity profiles for ribbed square duct flow with complex geometry. Comparison with
PIV data from Casara and Arts [80, 81].

72
Chapter 6

Atmospheric Flows

6.1 Monin-Obukhov Similarity Profiles (MO_velocity_profile)

Atmospheric turbulence is affected by the stability of the boundary layer. A stable layer (cool, heavy air at
ground level) will suppress turbulence, while an unstable layer (warm, light air at ground level) will enhance
turbulent mixing as buoyant plumes rise. The theory governing these flows, Monin-Obukhov similarity
theory, is discussed at some length in the FDS User Guide [2]. Here we examine velocity profiles from
a stable boundary layer and an unstable boundary layer. When mean forcing is used for driving the wind
field the Monin-Obukhov parameters determine the shape of the mean streamwise velocity profile. These
cases use a very tight DT_MEAN_FORCING_2 of 0.1 s in order to drive the flow field directly to the specified
profile, therefore comfirming the target profile is being computed correctly in FDS.
FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
10 15
Monin-Obukhov profile stable Monin-Obukhov profile unstable
8

10
6
u (m/s)

u (m/s)

4
5

2
MO profile MO profile
FDS profile FDS profile
0 0
0 5 10 15 20 25 30 0 5 10 15 20 25 30
z (m) z (m)

Figure 6.1: Stable (left) and unstable (right) mean streamwise velocity profiles for specified Monin-Obukhov param-
eters on the WIND line.

6.2 Ekman Layers (ekman_)

In a simple turbulent channel flow, the resulting boundary layer profile results from the balance between the
streamwise pressure gradient and the turbulent stresses. In an Ekman layer, the Coriolis force is also part
of the balance. With Ekman layers the Coriolis force has the effect of twisting the atmospheric boundary
layer—near the surface there is a component of the wind aligned with the pressure gradient, but aloft the

73
winds tend to be geostrophic, meaning they run parallel to isobars (perpendicular to the pressure gradient).
In FDS, input of a geostrophic wind implies a mean pressure gradient in the orthogonal direction (see FDS
User Guide [2]).
In atmospheric flows, it is difficult to separate the forces for verification purposes. In this section, we
will undertake an admittedly weak form of verification, namely, code-to-code comparison. We compare
FDS with the results from Schoenberg [83] where OpenFOAM is modified for atmospheric flows. We
use the same domain, grid resolution, and—as closely as possible—initial conditions as Schoenberg. The
latitude is specified as 45◦ north (for specifying the Coriolis force). We use the Deardorff turbulence model
with constant turbulent Prandtl number, which differs from Schoenberg, who uses a dynamic turbulent
Prantdl number. The main difference in the models is that Schoenberg solves a transport equation for
ksgs whereas FDS uses a simple algebraic closure (see FDS Tech Guide [48]). Further, it was difficult
to discern the exact initialization procedure from the Schoenberg report, so there are likely some minor
differences that can account for the small discrepancies we see in the FDS results. It should also be noted
that the Schoenberg simulations apply a Boussinesq assumption and thus do not account for the pressure
stratification term in the energy equation. Hence, to match the Schoenberg results as closely as possible, we
have set STRATIFICATION=F for FDS, which is counterintuitive for an atmospheric flow. In practice, this
should not be done.
Three cases are run: neutral, stable, and unstable boundary layers. The domain size, grid resolution,
surface flux condition, and geostrophic wind driving force are presented in Table 6.1 (adapted from [83]).
Note that the density and specific heat are specified in the FDS input file such that the surface heat flux
in either K · m/s or W/m2 are roughly the same. The boundary conditions are periodic in x and y and
no flux at the top of the domain. Note that this makes achieving a steady state impossible for the sta-
ble and unstable cases. This is another reason why the code comparison is difficult—essentially we are
comparing the resulting profiles after approximately 40 hours of real time evolution. Schoenberg uses a
spin up process of running the neutral case for up to 30 hours before transitioning to stratification. The
surface aerodynamic roughness used by Schoenberg is z0 = 0.1 m for the unstable and neutral cases and
z0 = 0.13 m for the stable case. This translates to a sand grain roughness of 3.0 m and 3.9 m, respec-
tively, in FDS. Note that the input files for the ekman_* series are stored with the validation suite (see
fds/Validation/Schoenberg_Ekman_Layers/), since the cases use 16 processors and take a few
hours to complete.
The differences seen in the stable case may be attributed to (i) the use of a dynamic turbulent Prandtl
number by Schoenberg and (ii) FDS not making the Boussinesq approximation. It may be noted that the
winds aloft in the FDS run match the specified geostrophic wind Ug = 7 m/s for the stable case.

Table 6.1: Schoenberg Ekman Layer parameters.

Case Domain Size Nx × Ny × Nz Surface Heat Flux Ug

x × y × z (m) (W/m2 ) (m/s)
Neutral 4000 × 2000 × 1500 40 × 40 × 40 0 10
Stable 2000 × 1000 × 750 40 × 40 × 40 -0.00502 7
Unstable 4000 × 2000 × 1500 40 × 40 × 40 0.02008 15

74
 FDS6.7.9-2168-g16effce-master
1500
Neutral Case

1000

z (m)
500
Schoenberg U
Schoenberg V
FDS U
FDS V
0
0 5 10 15
Velocity (m/s)

FDS6.7.9-2168-g16effce-master FDS6.7.9-2168-g16effce-master
1500
700
Stable Case Unstable Case
600

500 1000
z (m)

z (m)
400

300
500
200 Schoenberg U Schoenberg U
Schoenberg V Schoenberg V
100 FDS U FDS U
FDS V FDS V
0 0
0 5 10 0 5 10 15 20
Velocity (m/s) Velocity (m/s)

Figure 6.2: Comparison Ekman layer profiles for neutral (top), stable (bottom-left), and unstable (bottom-right)
atmospheric boundary layers with the simulation results of Schoenberg [83].

75
Chapter 7

Mass and Energy Conservation

7.1 Mass Conservation

This section describes tests of global mass conservation.

7.1.1 Flow of a Gas through a Simple Duct (simple_duct)

In this example, a closed chamber with a simple recirculating duct is used to demonstrate that a gas such
as carbon dioxide can be transported through multiple meshes and its mass is conserved. An obstruction is
created that leaks 0.5 kg of CO2 over a time span of 10 s.
FDS-6.8.0-0-g886e009-release

0.7 CO 2 Mass (simple_duct)

0.6

0.5
CO 2 Mass (kg)

0.4

0.3

0.2

0.1 Ideal (Mass)

FDS (CARBON DIOXIDE)
0
0 5 10 15 20 25 30
Time (s)

Figure 7.1: The mass of CO2 contained in a simple room connected with a recirculating duct.

7.1.2 Species Mass Flux (species_conservation)

A specified species mass flux is a commonly used boundary condition, in particular when specifying a fire.
The specified heat release rate per unit area is divided by the heat of combustion, resulting in a specified fuel
mass flux, ṁ00α , at the surface:
∂Yα
ṁ00α = ρYα un − ρDα (7.1)
∂n

77
where un is the face normal mass-average velocity and ∂ /∂ n is the gradient normal to the face. The velocity
is determined from the total mass flux and the density at the face,

∑α ṁ00α
un = (7.2)
ρ(Y, T )

The density is computed using old values of the face mass fractions and the face temperature which may
be specified or computed depending on choices made by the user. The new face mass fraction is then set
to satisfy (7.1). The test cases below are designed to exercise contributions from both the advective and
diffusive terms in this boundary condition.

Case 1: Low Flux, Hot Gas Injection

Propane is injected into a 4 m cube with 1 m2 vents on all six faces. The mass flux is 0.0001 kg/(m2 ·s) and the
surface temperature is 500 °C. The box is initially filled with air at standard conditions. The accumulation
of propane in the box is shown in Fig. 7.2.

Case 2: Air Injection into a Pressurized Compartment

Air is injected into one of two cubic chambers separated by a solid partition. The mass flux of air is
1 kg/(m2 ·s) through a vent that is 0.04 m2 in area. After 25 s, the mass of air introduced should be 1 kg. The
accumulation of air in the chamber is shown in Fig. 7.2.

Case 3: Gases Released by Solid Phase Reactions within a Solid Obstruction

A 0.2 m solid cube undergoes a reaction that releases four different gases into a sealed unit cube. The solid
has a density of 100 kg/m3 . The inert gases have different yields, all summing to one, meaning that the
solid cube ought to be vaporized creating a total mass of 0.8 kg. The accumulation of all the gases in the
chamber is shown in Fig. 7.2. Note that in this case, when the solid cube vaporizes, its volume is filled with
the background species, nitrogen. The total mass of the newly created nitrogen is 0.0093 kg.

Case 4: Gases Released by Solid Phase Reactions at the Boundary of the Domain
This case is similar to Case 3. A 0.2 m by 0.2 m solid surface undergoes a reaction that releases four different
gases into a sealed unit cube. The solid has a density of 100 kg/m3 and thickness of 0.1 m. The inert gases
have different yields, all summing to one, meaning that the solid surface ought to be vaporized creating a
total mass of 0.4 kg. The accumulation of all the gases in the chamber is shown in Fig. 7.2.

7.1.3 Mass Flux through Domain Boundaries (mass_flux_wall)

This set of tests considers a specified mass injection of water vapor from a VENT on the floor of a compart-
ment. The entire top of the compartment is OPEN to the ambient. The lighter water vapors are buoyant,
so they rise creating a turbulent flow into and out of the domain. Therefore, the input is steady, while the
outflow of water vapor is quite chaotic. To judge whether a proper mass balance is achieved by FDS we
must monitor not only the inflow and outflow, but also the total mass of water vapor in the domain at a given
time. Therefore, we add DEVC of quantity MASS_FLUX_WALL and STATITSTICS=‘SURFACE INTEGRAL‘
for inflow and outflow, and we also add MASS_FILE=.TRUE. to the DUMP line to monitor the total water
vapor at any point in time in the domain. With ρ and YH2O denoting the local mass density and mass fraction

78
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
0.04
Mass (species_conservation_1) 1 Mass (species_conservation_2)

0.03 0.8

0.6 Exact Zone 1 Mass

Mass (kg)

Mass (kg)
Exact Zone 2 Mass
0.02 FDS Zone 1 Mass
0.4 FDS Zone 2 Mass

0.01 0.2
Ideal
FDS (PROPANE) 0
0
0 10 20 30 40 50 60 0 5 10 15 20 25
Time (s) Time (s)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
1 0.5
Mass (species_conservation_3) Mass (species_conservation_4)
0.4
0.8

0.3
0.6 Excess N2 Excess Ar
Mass (kg)

Mass (kg)
Total Evaporated Mass Total Evaporated Mass
FDS (Mass N2)
0.2 FDS (Mass Ar)
0.4 FDS (Total) FDS (Total)
0.1

0.2
0

0 -0.1
0 20 40 60 80 100 0 20 40 60 80 100
Time (s) Time (s)

Figure 7.2: Results of the species_conservation test cases. Case 1: Hot gas is injected at a very low flow rate
into a 4 m cube. Case 2: Air is injected into one side of a partitioned compartment. Case 3: A solid cube vaporizes
yielding four gases whose total mass is 0.8 kg. A small amount of nitrogen is created to fill the void left by the
vaporized solid cube. Case 4: A patch of solid surface reacts to form four gases whose total mass is 0.4 kg.

of water vapor in a cell, and with ṁ00H2O denoting the flux of water vapor at a cell face on the domain boundary,
the mass balance may be written as,
Z  Z
d
Z
00
ρYH2O dV + ṁH2O dS − ṁ00H2O dS = 0 (7.3)
dt V Sout Sin

Figure 7.3 shows the results of two cases that test the mass balance of water vapor in the domain just
described. The plot on the left tests lumped species balance. In this case, the background humidity is set to
zero and so the only water vapor in the domain is that of the tracked species injected at the bottom boundary.
In the second test case on the right we are showing the balance for the primitive species water vapor. In
this test, the background humidity is the default (40 %) and so the output device must sum the water vapor
contribution from the background air and the tracked species injected at the bottom. In each plot, the blue
and red lines show the inflow (blue, steady) and outflow (red, turbulent) as well as the balance from Eq. 7.3
formed from the mass and devices files. The balance (yellow line) fluctuates slightly around zero because
of a first-order time error in reporting the boundary fluxes. That is, within the actual simulation the mass
transport equation is updated using a second-order Runge-Kutta scheme, whereas at the time the output
routine is called the code does not have the machinery to perfectly reconstruct the second-order flux terms.

79
 10 -3 FDS-6.8.0-0-g886e009-release 10 -3 FDS-6.8.0-0-g886e009-release
15 15
Inlet H2O Inlet H2O
Outlet H2O Outlet H2O
dm/dt+out-in dm/dt+out-in
10 10
mass flow (kg/s)

mass flow (kg/s)

5 5

0 0

-5 -5
0 500 1000 1500 2000 0 500 1000 1500 2000
time (s) time (s)

Figure 7.3: Results of the mass_flux_wall test cases. (Left) Lumped species balance, (Right) primitive species
balance.

7.1.4 Mass Balance on a Gas Control Volume (mass_balance_gas_volume)

Using the same computational setup as the mass_flux_wall series above, the mass_balance_gas_volume
test case defines a control volume (CV) internal to the computational domain, 6 m on a side (from 2 m to
8 m in each direction). The mass of water in the CV is output using a SPATIAL_STATISTIC=‘VOLUME
INTEGRAL‘ with QUANTITY=‘DENSITY‘ of the lumped species for water and the flows into and out of
the CV are obtained using QUANTITY=‘TOTAL MASS FLUX X‘, etc., with SPATIAL_STATISTIC=‘AREA
INTEGRAL‘. Figure 7.4 shows the resulting balance of accumulation (dm/dt) and the inflow and outflow of
the control volume.
-3 FDS-6.8.0-0-g886e009-release
10
4
dm/dt
out-in
dm/dt+out-in
2
mass flow (kg/s)

-2

-4
0 500 1000 1500 2000
time (s)

Figure 7.4: Results of the mass_balance_gas_volume test case showing the balance between mass accumulation
(dm/dt) and the flow in and out of the control volume.

7.1.5 Mass Balance for Species in a Reacting Flow (mass_balance_reac)

A 0.25 m2 propane burner ramps up to 500 kW linearly in 10 s and holds this heat release rate for another
10 s. Using a mass file, appropriate wall flux devices with surface integrals, and reaction source term devices
with volume integrals, we perform a detailed mass balance on all the primitive species in the reaction. The
results are presented in Fig. 7.5. The mass balance can be understood through the mnemonic, “Accumula-

80
tion equals In minus Out plus Generation”. For each species, we plot the “accumulation” term (first-order
numerical differentiation of the columns in the MASS_FILE, which are volume integrated over the domain),
the “in-out” term (surface intgral of the mass fluxes over the domain boundaries), the “generation” term
(volume integral of the reaction source term over the domain), and the sum of these as the “balance” (the
balance should be zero).

7.1.6 Checking the Equation of State (particle_offgas)

The “particle offgas” cases include 100 randomly distributed Lagrangian particles that inject different gas
species into an open domain initially filled with air. The aim of these cases is to test whether the gas tem-
perature is maintained at ambient (20 ◦ C) in the gas injection cases, and whether the minimum temperature
remains at ambient in the cases where fuel is injected and burned.

Case 1: Lagrangian Particles Blowing Air

In this case, each particle injects air into the domain, and air is the background species. The temperature is
not expected to rise above or drop below ambient, as shown in the upper left of Fig. 7.6.

Case 2: Lagrangian Particles Blowing Carbon Dioxide

In this case, each particle injects CO2 into the domain, and air is the background species. The temperature
is not expected to rise above or drop below ambient, as shown in the upper right of Fig. 7.6.

Case 3: Lagrangian Particles Blowing Propane that Combusts

In this case, each particle injects propane into the domain, and air is the background species. Combustion
is expected, but the gas temperature is not expected to drop below ambient, as shown in the lower left of
Fig. 7.6.

Case 4: A Small Solid Obstruction Blowing Propane that Combusts

In this case, the particles are replaced by a small, solid block blowing propane. Combustion is expected, but
the gas temperature is not expected to drop below ambient, as shown in the lower right of Fig. 7.6.

7.1.7 Realizability of Species Mass Fractions (realizable_mass_fractions)

In order for species mass fractions to be physically realizable they must all be positive (semi-definite) and
sum to unity, Yα ≥ 0 and ∑α Yα = 1. In this test case, a simple methane burner is used to test the realizability
of mass fractions. A device is placed near the base of the center of the flame to measure the local mass
fractions. In Fig. 7.7 we plot the time history of the mass fractions for Fuel, Air, and Products along with
the sum (dashed line). This sum should remain perfectly one at all times.

7.1.8 Integral Mass Test for Complex Geometry (geom_mass_file_test)

A 1 m3 box contains AIR at ambient conditions and has a sphere of radius 0.5 m leaving a volume of 0.4764
m3 . The FDS MASS_FILE output is tested to ensure the correct cut cell volumes are being computed and the
primitive species for AIR are correctly output. The results are shown below in Fig. 7.8.

81
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
0.1 1
Propane Mass Balance Oxygen Mass Balance
0.5
0.05
Mass Flow (kg/s)

Mass Flow (kg/s)

0
0
-0.5
-0.05
-1
accumulation accumulation
-0.1 in - out in - out
generation
-1.5 generation
balance balance
-0.15 -2
0 5 10 15 20 0 5 10 15 20
Time (s) Time (s)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
2 0.2
Nitrogen Mass Balance Water Vapor Mass Balance
1
0.1
0
Mass Flow (kg/s)

Mass Flow (kg/s)

-1
0
-2

-3 accumulation accumulation
-0.1
in - out in - out
-4 generation generation
balance balance
-5 -0.2
0 5 10 15 20 0 5 10 15 20
Time (s) Time (s)

FDS-6.8.0-0-g886e009-release -3 FDS-6.8.0-0-g886e009-release
10
0.4 2
Carbon Dioxide Mass Balance Soot Mass Balance

0.2 1
Mass Flow (kg/s)

Mass Flow (kg/s)

0 0

-0.2 accumulation -1 accumulation

in - out in - out
generation generation
balance balance
-0.4 -2
0 5 10 15 20 0 5 10 15 20
Time (s) Time (s)

Figure 7.5: Results of the mass_balance_reac test case.

82
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
25 25
Min/Max Temperature (particle_offgas_1) Min/Max Temperature (particle_offgas_2)
Temperature (°C)

Temperature (°C)
20 20

Expected (T_min) Expected (T_min)

Expected (T_max) Expected (T_max)
FDS (T_min) FDS (T_min)
FDS (T_max) FDS (T_max)
15 15
0 2 4 6 8 10 0 2 4 6 8 10
Time (s) Time (s)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
25 25
Min. Temperature (particle_offgas_3) Min. Temperature (particle_offgas_4)
Temperature (°C)

Temperature (°C)
20 20

Expected (T_min) Expected (T_min)

FDS (T_min) FDS (T_min)
15 15
0 2 4 6 8 10 0 2 4 6 8 10
Time (s) Time (s)

Figure 7.6: Results of the particle_offgas test cases. Upper left, Case 1: minimum and maximum tempera-
ture when air only is injected. Upper right, Case 2: min/max temperature when carbon dioxide is injected into air.
Lower left, Case 3: minimum temperature when propane gas is injected and burned. Lower right, Case 4: minimum
temperature when propane gas is injected from a solid block and burned.

FDS-6.8.0-0-g886e009-release
1.2
Realizable Mass Fractions
1

0.8
Mass Fraction

Expected
Z Fuel
0.6 Z Air
Z Prod
Sum Z
0.4

0.2

0
0 2 4 6 8 10
Time (s)

Figure 7.7: Results of the realizable_mass_fractions test case. The sum of the mass fractions (dashed line)
should remain perfectly one at all times.

83
 FDS-6.8.0-0-g886e009-release
1
Integrated Mass (geom_mass_file_test)
0.8

Integrated Mass (kg)

Exact Total
0.6 Exact N2
Exact O2
Exact CO2
0.4 Exact H2O
FDS Total
FDS N2
0.2 FDS O2
FDS CO2
FDS H2O
0
0 0.02 0.04 0.06 0.08 0.1
Time (s)

Figure 7.8: Results of the geom_mass_file_test test case.

84
7.2 Energy Conservation
FDS does not explicitly solve the energy conservation equation. Instead, mass transport equations are solved
for the the n independent gas species. The density is found from summing the individual gas species densi-
ties, ρ = ∑α ρα . Source terms from the energy equation are incorporated in the velocity divergence, ∇ · u,
which is factored out of the sensible enthalpy transport equation, i.e.,
 
1 D 000 00 000
∇·u = ( p̄ − ρhs ) + q̇ + ∇ · (k∇T ) − ∇ · ∑ hs,α Jα − ∇ · q̇r + q̇b . (7.4)
ρhs Dt α

Here ρ is the mass density, hs is the sensible enthalpy, p̄ is the thermodynamic pressure, q̇000 is the chemical
heat release, k is thermal conductivity, Jα is the diffusion flux vector, q̇00r is the radiant heat flux vector, and
q̇000
b is the bulk heat source due to convective heat transfer to Lagrangian particles and the heat associated
with mass sources. The divergence is a fairly complicated expression, and it is important that it is computed
properly to ensure global energy conservation. The examples in this section check that mass and energy are
conserved for relatively simple configurations.

7.2.1 The Heat from a Fire (energy_budget)

For the purpose of verifying that the basic FDS algorithm is energy conserving, it is useful to think of a single
compartment as a control volume into which energy is generated by a fire and out of which energy either
flows via openings or is conducted through the walls. If the fire’s heat release rate (HRR) is steady, eventually
the system will reach a quasi-steady state (in an LES calculation, there is never a true “steady state”). Two
simple cases, called energy_budget_adiabatic_walls and energy_budget_cold_walls, illustrate
that in the quasi-steady state, the energy generated by the fire is conserved. For the case with adiabatic walls,
a 1200 kW fire is simulated within a compartment that is 10 m by 10 m by 5 m tall. There is a single door
and a single horizontal vent in the ceiling. The walls are assumed to be adiabatic; that is, there is no net heat
flux through them. Another way to look at it is that the walls are perfect insulators. To simplify the case
even further, the radiation transport algorithm is turned off. It is expected that in this case 1200 kW ought
to flow out of the compartment either via the door or ceiling vent. The plot in Fig. 7.9 shows both the HRR
and enthalpy flow out of the compartment converging to 1200 kW. During the warm up phase, the enthalpy
flow is less than the HRR because energy is consumed heating up the air within the room.
Next, the same compartment with the same fire is now assumed to have cold (20 °C) walls, and the
radiation is switched back on. After a few minutes of simulation, the net enthalpy outflow is approximately
470 kW and the heat losses to the wall (both radiative and convective) are approximately 730 kW, summing
to 1200 kW, the HRR of the fire. Fig. 7.10 displays the HRR, the sum of the losses (radiation, convection,
conduction, species diffusion), and the net enthalpy change in the compartment. Note that energy losses are
expressed as negative quantities. In steady state, the enthalpy added to the compartment by the fire is offset
by losses to the walls and opening.

7.2.2 Gas Injection via an Isentropic Process (isentropic)

This example checks that if nitrogen is added to a sealed compartment with adiabatic (i.e., no heat loss)
walls, that the density, pressure and temperature ought to rise according to the ideal gas law for an isentropic
process:
 γ   γ−1 γ
p2 ρ2 T2
= = (7.5)
p1 ρ1 T1

85
 FDS-6.8.0-0-g886e009-release
2000
Energy Budget (adiabatic walls)
1500

1000

Energy Rate (kW)

Expected (HRR)
Expected (Net Enthalpy)
500
Expected (Q CONV)
FDS (HRR)
0 FDS (Q_TOTAL)
FDS (Q_CONV)
-500

-1000

-1500
0 50 100 150 200 250 300
Time (s)

Figure 7.9: The energy budget for a simple compartment fire with adiabatic walls.

FDS-6.8.0-0-g886e009-release
2000
Energy Budget (cold walls)
1500

1000
Energy Rate (kW)

Expected (HRR)
Expected (Net Enthalpy)
500
Expected (Losses)
FDS (HRR)
0 FDS (Q_TOTAL)
FDS (Sum of Losses)
-500

-1000

-1500
0 50 100 150 200 250 300
Time (s)

Figure 7.10: The energy budget for a simple compartment fire with cold walls.

The subscripts 2 and 1 refer to the final and initial state, respectively. Nitrogen is a diatomic gas for which
γ = 1.4. As an additional check, the nitrogen is injected at two different rates such that in Case A the
injection occurs in 10 s and in Case B the injection occurs in 50 s. The nitrogen is introduced into the
domain via small spheres that do not generate or absorb heat. They do not occupy volume either. They just
inject the nitrogen at a specified rate into the particular grid cell that each occupies. Results are plotted in
Fig. 7.11.

7.2.3 Gas Injection via a Non-Isentropic Process (isentropic2)

This example checks that if a given mass of nitrogen is added via an external vent to a sealed compartment
with adiabatic (i.e., no heat loss) walls, that the density, pressure and temperature ought to rise to the same
values regardless of the duration of the injection time. The results are plotted in Fig. 7.12. In the two cases,
0.2 kg of N2 are added to a compartment that is 1 m3 in volume. In the first case, the injection occurs in
10 s, in the second case in 50 s. The temperature of the gas in both cases is 500 °C. It is expected that the
pressure, p, should rise at the rate:
dp γ p u A
= (7.6)
dt V

86
 FDS-6.8.0-0-g886e009-release 10 4 FDS-6.8.0-0-g886e009-release
1.4 2.5
Density (isentropic) Pressure (isentropic)
1.35
2

1.3
Density (kg/m³)

Pressure (Pa)
1.5
1.25
1
1.2
Exact (Density_1) Exact (Pressure_1)
Exact (Density_2) 0.5 Exact (Pressure_2)
1.15 FDS (density_1) FDS (pressure_1)
FDS (density_2) FDS (pressure_2)
1.1 0
0 10 20 30 40 50 60 0 10 20 30 40 50 60
Time (s) Time (s)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
40 450
Temperature (isentropic) Enthalpy (isentropic)
35
Temperature (°C)

Enthalpy (kJ)
30
400
25

Exact (Temperature_1) Exact (Enthalpy_1)

20 Exact (Temperature_2) Exact (Enthalpy_2)
FDS (temperature_1) FDS (enthalpy_1)
FDS (temperature_2) FDS (enthalpy_2)
15 350
0 10 20 30 40 50 60 0 10 20 30 40 50 60
Time (s) Time (s)

Figure 7.11: Density, pressure, temperature, and enthalpy rise due to the injection of nitrogen into a sealed compart-
ment.

where u is the injection velocity and A is the area of the vent, 0.04 m2 . The density of the incoming N2 is
found from the equation of state:
Wp
ρ= (7.7)
RT
The injection velocity is the mass flux divided by the density, u = ṁ00 /ρ, in which case, the pressure rise can
be written
dp γ ṁ00 RT A
= (7.8)
dt WV
Note that the rate of pressure rise is constant. In both the fast and slow injection cases, the pressure is
expected to rise 64,253 Pa above ambient. The density and temperature rise are also linear. The density
increases from 1.165 kg/m3 to 1.365 kg/m3 . The temperature increases from 20 °C to 135.7 °C. The change
in the internal energy of the system, based on the mass and temperature of the added N2 , is:

∆E = (0.2 kg) × (1.039 kJ/(kg · K)) × (773.15 K) = 160.6 kJ (7.9)

Adding in the work due to the pressure yields the change in total enthalpy:

∆H = ∆E +V ∆P = 160.6 kJ + (1 m3 ) × ( 64.3 kJ/m3 ) = 224.9 kJ (7.10)

87
 FDS-6.8.0-0-g886e009-release 10 4 FDS-6.8.0-0-g886e009-release
1.5 8
Density (isentropic2) Pressure (isentropic2)

1.4 6
Density (kg/m³)

Pressure (Pa)
1.3 4

1.2 Exact (Density_1) 2 Exact (Pressure_1)

Exact (Density_2) Exact (Pressure_2)
FDS (density_1) FDS (pressure_1)
FDS (density_2) FDS (pressure_2)
1.1 0
0 10 20 30 40 50 60 0 10 20 30 40 50 60
Time (s) Time (s)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
160 650
Temperature (isentropic2) Enthalpy (isentropic2)
140 600
120
550
Temperature (°C)

Enthalpy (kJ)
100
500
80

60 450
Exact (Temperature_1) Exact (Enthalpy_1)
Exact (Temperature_2) Exact (Enthalpy_2)
40 400
FDS (temperature_1) FDS (enthalpy_1)
FDS (temperature_2) FDS (enthalpy_2)
20
350
0 10 20 30 40 50 60 0 10 20 30 40 50 60
Time (s) Time (s)

Figure 7.12: Density, pressure, temperature, and enthalpy rise due to the injection of nitrogen into a sealed compart-
ment.

7.2.4 Mixing Gases (energy_budget_dns)

In this example, two fictitious gases are forced through tiny ducts (1 mm by 1 mm) that meet at a T-joint.
The cross sectional areas of all the ducts are the same. Gas 1 has an initial temperature of 606.3 °C and
a constant specific heat of 1 kJ/(kg · K). Gas 2 has an initial temperature of 20 °C and a constant specific
heat of 10 kJ/(kg · K). Both gases have a molecular weight of 29 g/mol and a mass flow of 1 × 10−8 kg/s.
Typically gases of the same molecular weight have comparable, but not necessarily the same, specific heats.
This is an extreme case. In steady state, it is expected that the convection and diffusion of enthalpy at the
boundaries will be equal: Z Z
hs ρu · dS = ∑ hs,α ρDα ∇Zα · dS (7.11)
α

where Zα is the mass fraction of species α. Because of the large difference in specific heats, the average
velocity of the gas mixture flowing out, U3 , is not simply the sum of the velocities at the two inlets, U1 +U2 .
In fact, in order to have the proper enthalpy flow at the outlet, the volume integral of the divergence should
be:
(U1 −U2 ) (c p,2 − c p,1 )
Z
∇ · u dV ≡ −A (U1 +U2 −U3 ) = −A = −1.357 × 10−8 m3 /s (7.12)
c p,1 + c p,2

Figure 7.13 shows the resulting integrated divergence or net volume flow.

88
 10 -8 FDS-6.8.0-0-g886e009-release
0
Energy Budget (dns)

Net Volume Flow (m³/s)

-0.5

Expected (div)
FDS (div)

-1

-1.5
0 0.5 1 1.5 2
Time (s)

Figure 7.13: The integrated divergence or net volume flow.

7.2.5 Mixing with Variable Specific Heats (energy_budget_tmix)

This case tests the ability of FDS to properly account for temperature dependent specific heats. A coarse 2-D
channel mixing case is presented where hot gas at 900 °C with a constant specific heat of c p = 1 kJ/(kg · K)
enters from the bottom left vent with a flux of 1 kg/(m2 · s) and cold gas at 20 °C with a constant specific
heat of c p = 10 kJ/(kg · K) enters from a top left vent with the same mass flux. The gases mix in a center
channel and exit the domain flowing to the right.
The enthalpy of the mixture at the outlet is 1 × 1 × (900 + 273.15) + 1 × 10 × (20 + 273.15) = 4104.65
kJ/(m2 · s). The mixture heat capacity at the outlet is 5.5 kJ/(kg · K) and the flux is 2 kg/(m2 · s). Therefore
the outlet temperature is 4104.65/2/5.5 − 273.15 = 100 °C.
FDS-6.8.0-0-g886e009-release
500
Energy Budget (tmix)
400
Outlet Temperature (°C)

300
Expected (temp)
FDS (Temp)
200

100

0
0 5 10 15 20
Time (s)

Figure 7.14: Outlet temperature for energy_budget_tmix case. Black circles represent the correct result of
100 °C.

7.2.6 Combustion Chamber (energy_budget_combustion)

Methanol and air are mixed and burned within an adiabatic combustion chamber 1 m on each side. Because
the internal energy of the system does not change, we expect that

∆h = V ∆p (7.13)

89
where ∆h is the increase in the enthalpy of the system (kJ), V is the volume of the chamber (1 m3 ), and ∆p is
the increase in pressure (kPa). The enthalpy includes the heats of formation of the various gases. Figure 7.15
displays the rise in pressure and enthalpy.
FDS-6.8.0-0-g886e009-release
500
Energy Budget (combustion)
400

Energy (kJ)
300

200

100
Pressure Rise (kPa)
Enthalpy Rise (kJ)
0
0 0.2 0.4 0.6 0.8 1
Time (s) 10 -3

Figure 7.15: The rise in pressure and enthalpy in a sealed, adiabatic combustion chamber.

7.2.7 Solid phase energy conservation (energy_budget_solid)

This case tests the ability of FDS to properly conserve the total of the solid phase and the gas phase energy.
A 1 m3 box is filled with 1 kg of a gas with a specific heat of 1 kJ/(kg · K), a molecular weight of 28 g/mol,
and an initial temperature of 20 °C. All the walls, except the floor, are adiabatic. The floor is defined as
a 1 mm thick layer of a material with a conductivity of 100 W/(m · K), a specific heat of 1 kJ/(kg · K), an
initial temperature of 150 °C, and a density of 1,000 kg/m3 . The initial enthalpy is 1 × 1 × (20 + 273.15) +
1 × 1 × (150 + 273.15) = 716.3 kJ. At the end state, the gas will have risen in temperature, the solid will
have decreased in temperature, and the pressure will have risen. Since energy is conserved this occurs
at a temperature of 96.3 °C. The final enthalpy is 1 × 1 × (96.3 + 273.15) + 1 × 1 × (96.3 + 273.15) +
22670/1000 = 716.3 kJ.
FDS-6.8.0-0-g886e009-release 10 4 FDS-6.8.0-0-g886e009-release
100 2.4
Energy Budget (solid) Energy Budget (solid)
98
2.3
Temperature (°C)

Pressure (Pa)

96
2.2
94
Exact T g
Exact T w
2.1
92 FDS T g
P Exact
FDS T w P FDS
90 2
0 500 1000 1500 2000 0 500 1000 1500 2000
Time (s) Time (s)

Figure 7.16: Outputs of temperature (left) and pressure (right) for the energy_budget_solid case. Circles
represent the correct result.

90
7.2.8 Energy Conservation at Walls (mass_heat_wall_device_test)
To obtain tight mass and energy balances it is important to output the actual values used for boundary
conditions in the transport equations in FDS. Shown in Fig. 7.17 are the results for two cases. These cases
are the same except that the vents have opposite IOR (orientation) values. Each case has a prescribed outflow
mass flux of 1 kg/(m2 · s) and exhaust vent area of 4 m2 . The inflow balances the outflow with both equal to
4 kg/s. The specific heat of the air is set to 1 kJ/(kg · K). The ambient temperature is 100 ◦C. The reference
temperature is 25 ◦C. The heat flow at the vent is, therefore, (4 kg/s)(1 kJ/(kg · K))(75 K) = 300 kW.
FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
6 500
Mass Flow (mass_heat_wall_device_test) Heat Flow (mass_heat_wall_device_test)
4

2
Mass Flow (kg/s)

Heat Flow (kW)

Ideal Mass Flow Out Ideal Heat Flow Out
Ideal Mass Flow In Ideal Heat Flow In
0 FDS Mass Flow Out
0 FDS Heat Flow Out
FDS Mass Flow In FDS Heat Flow In
-2

-4

-6 -500
0 10 20 30 40 50 60 0 10 20 30 40 50 60
Time (s) Time (s)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
6 500
Mass Flow (mass_heat_wall_device_test_2) Heat Flow (mass_heat_wall_device_test_2)
4

2
Mass Flow (kg/s)

Heat Flow (kW)

Ideal Mass Flow Out Ideal Heat Flow Out

Ideal Mass Flow In Ideal Heat Flow In
0 FDS Mass Flow Out
0 FDS Heat Flow Out
FDS Mass Flow In FDS Heat Flow In
-2

-4

-6 -500
0 10 20 30 40 50 60 0 10 20 30 40 50 60
Time (s) Time (s)

Figure 7.17: The mass_heat_wall_device_test flow results.

7.2.9 Energy Conservation for Solid Particles (energy_budget_particles)

A sealed, adiabatic box, 1 m on a side, is filled with 10 kg of spherically-shaped solid metallic particles, 1 cm
in diameter, with a specific heat of 1 kJ/(kg·K), density of 5000 kg/m3 , and conductivity of 50 W/(m·K).
The initial air temperature is 100 ◦ C and the initial particle temperature is 20 ◦ C. The simulation is run until
thermal equilibrium is achieved. Figure 7.18 displays on the left the contributions to the sensible enthalpy
of the air, most of which are negative because the air loses energy to the particles. Q_RADI is the net
contribution of thermal radiation, Q_COND is the contribution of convection at the exterior wall of the box,
Q_PRES is the contribution of pressure work on the exterior wall, and Q_PART is the contribution by the
particles. The negative value of Q_PART means that the particles absorb radiative and convective energy.
The plot on the right of Fig. 7.18 compares the sum of these four contributions, Q_TOTAL, with the computed

91
change in the sensible enthalpy of the air, (Q_ENTH). These two quantities should be equal, as shown by the
overlaying curves.
FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
1 1
Energy Budget (particles) Energy Budget (particles)
0.5 0.5
Energy Rate (kW)

Energy Rate (kW)

0 0

-0.5 -0.5

-1 -1
Q_RADI
Q_COND
-1.5 Q_PRES
-1.5 Q_ENTH
Q_PART Q_TOTAL
-2 -2
0 50 100 150 200 250 300 0 50 100 150 200 250 300
Time (s) Time (s)

Figure 7.18: (Left) Contributions to change in sensible enthalpy. (Right) Sum of contributions versus total change in
enthalpy.

92
Chapter 8

Checking for Coding Errors

This chapter presents a number of cases that are designed to check whether or not various algorithms have
been programmed properly. A good way to do this is to simulate simple flows, particularly ones with
symmetric features. Errors in coding are easily identified in such cases.

8.1 Symmetry Tests

These examples are very simple test cases that are used to determine if there are coding errors in the basic
flow solver. In the absence of a small amount of “noise” that FDS uses to initialize the velocity field, a
symmetric geometry ought to produce symmetric patterns of velocity, temperature, etc.

8.1.1 Counter-Flowing Jets (symmetry_test)

A closed box, 1 m on a side, has 6 injection vents, one centered on each face. Air is pumped into the box at
a steady rate. Anything that might lead to an asymmetry in the flow field is turned off, for example, gravity,
atmospheric stratification, natural convection, and random noise. One case is performed on a single mesh;
the other on multiple meshes. The cases are not identical. The resulting flow fields are shown at the left in
Fig. 8.1. On the right are plots of the three components of velocity at equidistant corners of the enclosure.
Ideally, there should be two equal and opposite time histories of the three components. Even the slightest
error in coding will throw this off almost immediately.

8.1.2 Fire in a Thin-Walled Pan (symmetry_test_2)

In this case, a steel pan is placed inside a small compartment with a small door at one end. The pan has thin
walls (zero cells thick) which is particularly challenging for a symmetry test because boundary conditions
must be applied on opposite sides of the thin wall. The case also involves a fire and multiple species.
Radiation is not tested here. Figure 8.2 displays the temperature field in a slice plane across the pan, and at
two points equidistant from the centerline.

8.1.3 Axial symmetry (cyl_test)

FDS has an option to perform an axially-symmetric calculation, which can be useful for simulating a laminar
diffusion flame, like a bunsen burner or candle. When invoking this option, the user still specifies three
coordinates, x, y, and z, but it is understood that x is the radial coordinate, r, and z is the vertical coordinate.

93
 FDS-6.8.0-0-g886e009-release
0.01
Component Velocity (symmetry_test)

0.005

Velocity (m/s)
0

-0.005

-0.01
0 10 20 30 40 50
Time (s)

10 -3 FDS-6.8.0-0-g886e009-release
1
Component Velocity (symmetry_test_mpi)

0.5
Velocity (m/s)

-0.5

-1
0 10 20 30 40 50
Time (s)

Figure 8.1: Flow field (left) and velocity component plots (right) for the symmetry_test cases.

FDS-6.8.0-0-g886e009-release
500
Temperature (symmetry_test_2)
400
Temperature (°C)

300
Left
Right
200

100

0
0 1 2 3 4
Time (s)

Figure 8.2: Flow field (left) and temperature plot (right) for the symmetry_test_2 case.

The coordinate, y, is used only by Smokeview to draw the domain, and it should not influence the calculation
at all. However, since FDS is primarily a three-dimensional, cartesian code, there is a possibility that a
coding error might cause the specification of the y dimension to have an effect on the result. Figure 8.3
displays results of two sets of axially-symmetric simulations. On the left is an LES simulation in which the
lateral size of the cells, δ y, is 2 mm and 2 cm. The two curves ought to lie exactly upon each other, which
is indicated by a single red curve. On the right is a DNS simulation of a laminar diffusion flame, with δ y

94
equal to 2 mm and 4 mm.
FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
100 2000
Cylindrical Test, LES (cyl_test_1-2) Cylindrical Test, DNS (cyl_test_3-4)
80
1500
Temperature (°C)

Temperature (°C)
60
1000
40

500
20
?y=2 mm ?y=2 mm
?y=2 cm ?y=4 cm
0 0
0 5 10 15 20 0 0.5 1 1.5 2 2.5 3
Time (s) Time (s)

Figure 8.3: Temperature traces from axially-symmetric calculations with different values of δ y.

95
8.2 Velocity boundary condition implementation (velocity_bc_test)
In this example, a simple jet is created by pushing air through a square duct. The obstructions forming the
duct are built three different ways. The first way is to build the duct with thin (zero cell thick) plates. The
second is to carve the duct out of a solid block. The third way is to build the duct out of one cell thick
plates. The three approaches are illustrated in Fig. 8.4. Ideally, the three jets ought to be exactly the same,
but subtle differences in the pressure solution and the random generation of noise lead to three unique flow
patterns that are only the same in a time-averaged sense. To test this, the centerline pressure and velocity are
recorded for a 2 m/s jet emanating from a 0.8 m square duct that is 5 m long. The grid resolution is fairly
crude (0.2 m) because the purpose of this exercise is to show that the steady-state velocity and pressure
profiles are essentially the same. Figure 8.5 displays the profiles. The curves ought to overlap each other.
Errors in the implementation of the velocity boundary conditions are likely to throw the three cases out of
alignment.

Figure 8.4: Time-averaged contours of velocity for the velocity_bc_test case.

96
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
1 3
Centerline Pressure (velocity_bc_test) Centerline Velocity (velocity_bc_test)
2.5

0.5 2

Velocity (m/s)
Pressure (Pa)

1.5

0 1
Thin Plate (Pres1) Thin Plate (Vel1)
Solid Block (Pres2)
0.5 Solid Block (Vel2)
Thick Plate (Pres3) Thick Plate (Vel3)
-0.5 0
-5 0 5 10 15 20 -5 0 5 10 15 20
Distance (m) Distance (m)

Figure 8.5: Centerline time-averaged pressure (left) and velocity (right) profiles for the velocity_bc_test case.

8.3 Checking the divergence constraint

The test cases in this section involve the divergence of the flow, ∇ · u, an important quantity in the low Mach
number formulation of the Navier-Stokes equations.

8.3.1 Isothermal flow (divergence_test_1)

In the first example, air is pushed through a 1 m cube at a speed of 1 m/s. The air is extracted at the same
rate, and there should be no pressure rise and the flow should remain divergence free. There are also a few
obstructions that are created and removed during the simulation to check that this does not cause any errors.
The left hand plot of Fig. 8.6 displays the minimum and maximum divergence as a function of time. These
ought to remain below approximately 1 × 10−10 s−1 in absolute value.

8.3.2 Sealed compartment (divergence_test_2)

In the second example, air is introduced into a sealed compartment at a steady rate via a single solid La-
grangian particle. In other words, the air is not introduced at the boundary, but rather within the center of
the compartment. The pressure, density, and temperature increase, but the volume integral of the divergence
remains zero according to the Divergence Theorem:
Z Z
∇ · u dV = u · dS = 0 (8.1)
Ω ∂Ω

The right hand plot of Fig. 8.6 displays the volume integral of the divergence as a function of time. This
ought to remain within machine accuracy, certainly less than 1 × 10−10 m3 /s in absolute value.

8.3.3 Creating/Removing Obstructions (divergence_test_3)

In the third example, air is pushed through a two-dimensional channel where obstructions are created and
removed at random times. The divergence of the flow should remain nearly zero throughout the entire
domain. Figure. 8.7 displays the maximum and minimum divergence as a function of time. These ought to
remain less than approximately 10−12 s−1 in absolute value.

97
 10 -10 FDS-6.8.0-0-g886e009-release 10 -15 FDS-6.8.0-0-g886e009-release
1 1
Divergence (divergence_test_1) Divergence (divergence_test_2)

0.5

Volume Flow (m³/s)

Divergence (1/s)

0 0

Expected (div_min) -0.5

Expected (div_max)
FDS (div_min)
FDS (div_max)
-1 -1
0 2 4 6 8 10 0 2 4 6 8 10
Time (s) Time (s)

Figure 8.6: Minimum and maximum divergence for the divergence_test cases.

-12 FDS-6.8.0-0-g886e009-release
10
1
Divergence (divergence_test_3)

0.5
Divergence (1/s)

-0.5 Expected (div_min)

Expected (div_max)
FDS (div_min)
FDS (div_max)
-1
0 0.5 1 1.5 2
Time (s)

Figure 8.7: Minimum and maximum divergence for the divergence_test_3 case.

8.4 Multi-mesh Pressure and Temperature Lapse (lapse_rate)

A 10 m x 10 m x 40 m volume is initialized with LAPSE_RATE=-0.01. The volume is divided vertically
into four equal meshes with each mesh having a 10 cm grid resolution. There should be a smooth drop in
both temperature and pressure that matches the analytic values, as shown in Fig. 8.8.

98
 FDS-6.8.0-0-g886e009-release 10 5 FDS-6.8.0-0-g886e009-release
22 1.015
Exact Exact
Temperature FDS
Pressure FDS

21
Temperature (°C)

Pressure (Pa)
20 1.01

19

18 1.005
0 10 20 30 40 0 10 20 30 40
Elevation (m) Elevation (m)

Figure 8.8: Centerline temperature (left) and pressure (right) profiles for the lapse_rate case.

8.5 Randomly Overlapping Meshes (random_meshes)

In most FDS simulations, multiple meshes are positioned adjacent to each other to allow two-way commu-
nication. However, in some instances, meshes need to overlap, and this case verifies the functionality of
meshes that overlap in a variety of ways. It is a 2-D case that runs fast and is easy to visualize. There are
five overlapping meshes shown at left in Fig. 8.9. Air is blown in at the bottom and exhausted via an open
vent at the top. The volume flow rate is fixed at 0.012 m3 /s, and the plot at right in Fig. 8.9 verifies that
the simulation matches to within a few percent. The case uses the default normal velocity tolerance at solid
and interpolated boundaries; thus, there is some error expected in the volume flow due to the solution of the
Poisson equation. As a further check that the mesh allocations are being properly handled, there is a fine
mesh embedded within the coarse, and the case is run with four MPI processes rather than five.
FDS-6.8.0-0-g886e009-release
0.02
Volume Flow (random_meshes)

0.015
Volume Flow (m³/s)

0.01

0.005
Expected
FDS (Vdot)
0
0 1 2 3 4 5
Time (min)

Figure 8.9: Five overlapping meshes (left) and the net volume flow through the domain (right).

8.6 Overlapping INIT Regions (init_overlap)

Typically, an FDS simulation begins at time t = 0 with ambient conditions. However, ambient conditions
can be changed for rectangular regions of the domain using the namelist keyword INIT. If two rectangular

99
regions defined by INIT overlap, it is the first of the overlapping regions that takes precedence (i.e., first
come first serve) for all commonly defined inputs. That is, it is not possible to overwrite the initial conditions
explicitly specified by the first INIT line with the initial conditions explicitly defined by the second INIT
line. This verification case tests the overlap of INIT regions. Fig. 8.10 and Fig. 8.11 verify the predictions
by FDS match the expected conditions.

100
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
2 2
Density=0.42 (init_overlap) Density=0.50 (init_overlap)

1.5 1.5
Density (kg/m³)

Density (kg/m³)
1 1

0.5 0.5
Expected Expected
FDS (D0p42) FDS (D0p50)
0 0
0 0.02 0.04 0.06 0.08 0.1 0 0.02 0.04 0.06 0.08 0.1
Time (sec) Time (sec)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
2 2
Density=0.938 (init_overlap) Density=1.19 (init_overlap)

1.5 1.5
Density (kg/m³)

Density (kg/m³)
1 1

0.5 0.5
Expected Expected
FDS (D0p938) FDS (D1p19)
0 0
0 0.02 0.04 0.06 0.08 0.1 0 0.02 0.04 0.06 0.08 0.1
Time (sec) Time (sec)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
2 10
Density=1.51 (init_overlap) Integrated Intensity=6.0 (init_overlap)
8
Integrated Intensity (kW/m²)

1.5
Density (kg/m³)

6
1
4

0.5
2
Expected Expected
FDS (D1p51) FDS (I6p0)
0 0
0 0.02 0.04 0.06 0.08 0.1 0 0.02 0.04 0.06 0.08 0.1
Time (sec) Time (sec)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
15 20
Integrated Intensity=9.9 (init_overlap) Integrated Intensity=15.6 (init_overlap)
Integrated Intensity (kW/m²)

Integrated Intensity (kW/m²)

15
10

10

5
5
Expected Expected
FDS (I9p9) FDS (I15p6)
0 0
0 0.02 0.04 0.06 0.08 0.1 0 0.02 0.04 0.06 0.08 0.1
Time (sec) Time (sec)

Figure 8.10: Comparison of initial conditions from FDS compared with expected initial conditions.

101
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
100 1
Integrated Intensity=81.6 (init_overlap) Propane=0.00 (init_overlap)
80 0.8

Volume Fraction (mol/mol)

Integrated Intensity (kW/m²)

60 0.6

40 0.4

20 0.2
Expected Expected
FDS (I81p6) FDS (P0p0)
0 0
0 0.02 0.04 0.06 0.08 0.1 0 0.02 0.04 0.06 0.08 0.1
Time (sec) Time (sec)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
1 1
Propane=0.30 (init_overlap) Propane=0.50 (init_overlap)
0.8 0.8
Volume Fraction (mol/mol)

Volume Fraction (mol/mol)

0.6 0.6

0.4 0.4

0.2 0.2
Expected Expected
FDS (P0p3) FDS (P0p5)
0 0
0 0.02 0.04 0.06 0.08 0.1 0 0.02 0.04 0.06 0.08 0.1
Time (sec) Time (sec)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
40 400
TMPA (init_overlap) T=200C (init_overlap)

30 300
Temperature (C)

Temperature (C)

20 200

10 100
Expected Expected
FDS (T20) FDS (T200)
0 0
0 0.02 0.04 0.06 0.08 0.1 0 0.02 0.04 0.06 0.08 0.1
Time (sec) Time (sec)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
600
T=427C (init_overlap) 800 T=700C (init_overlap)
500

400 600
Temperature (C)

Temperature (C)

300
400
200
200
100 Expected Expected
FDS (T427) FDS (T700)
0 0
0 0.02 0.04 0.06 0.08 0.1 0 0.02 0.04 0.06 0.08 0.1
Time (sec) Time (sec)

Figure 8.11: Comparison of initial conditions from FDS compared with expected initial conditions.

102
8.7 Spurious Obstruction Effects (hole)
It is not uncommon for subtle differences between two nearly identical input files to lead to noticeable results
in the simulation output. In this particular case, a HOLE in an obstruction that was set to never open still
led to slightly different results when compared to a case with no HOLE. The reason for the difference was
attributed to slightly different boundary conditions at the edge of the HOLE. The problem has been fixed, as
confirmed by Fig. 8.12.
FDS-6.8.0-0-g886e009-release
200
Check Identical Solutions (hole, no_hole)

150
HRR (kW)

100

50
Expected (no_hole)
FDS (hole)
0
0 2 4 6 8 10
Time (s)

Figure 8.12: Heat release rates of two nearly identical calculations.

103
8.8 OpenMP Thread Checking (race_test)
OpenMP is used in FDS to sub-divide the work of the do-loops among multiple cores belonging to the
same processor. Figure 8.13 displays temperature and velocity traces from two calculations, one using one
OpenMP thread and one using four. The results of the two calculations ought to be identical, in which case
the respective pairs of temperature and velocity traces ought to completely overlap.
FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
1200 2
Temperature (race_test) Velocity (race_test)
1000
1.5
800
Temperature (°C)

Velocity (m/s)
600 1

400
0.5
200 1 Thread (TMP) 1 Thread (VEL)
4 Thread (TMP) 4 Thread (VEL)
0 0
0 1 2 3 4 5 0 1 2 3 4 5
Time (s) Time (s)

Figure 8.13: Temperature and velocity time histories of two calculations that use different numbers of OpenMP
threads. The traces ought to be identical.

104
8.9 Multi-Mesh Layer Height Calculation (layer)
The test cases called Miscellaneous/layer_1mesh.fds and layer_4mesh.fds simulate a fire in a
small compartment where the average smoke layer height and temperature are calculated. Figure 8.13
displays the results for both cases. Note that the 4 mesh case is run on 3 MPI processes to ensure that the
layer calculation works for an arbitrary number of meshes and MPI processes.
FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
2 100
1 mesh
Layer Height 4 mesh
Layer Temperatures
80
1.5

Temperature (°C)
60
Height (m)

T_low, 1 mesh
T_up, 1 mesh
1 T_low, 4 mesh
40 T_up, 4 mesh

0.5
20

0 0
0 10 20 30 40 50 60 0 10 20 30 40 50 60
Time (s) Time (s)

Figure 8.14: Average layer height (left) and temperatures (right) for simulations on 1 mesh and 4 meshes.

105
8.10 The Restart Feature
FDS has a feature with which users can stop and restart simulations. When a simulation is stopped, all
necessary variables are packed into arrays and written out to a file. This feature can be fragile when
the input file includes devices and controls that initiate actions at certain times. In the simulation named
Restart/device_restart_a.fds, a multi-room fire simulation is stopped and then restarted at 40 s.
Before the restart, at 30 s, a door is opened and high temperature gases flow from the fire room into a corridor.
After the restart, the case proceeds until 60 s. Another simulation called device_restart_base_case
runs uninterrupted. The temperature just outside of the fire room door (Fig. 8.15) should be similar in both
simulations. The temperatures are not exactly the same because restarting a simulation can lead to minor
changes in the algorithm at the time of the restart.
FDS-6.8.0-0-g886e009-release
100
Temperature (device_restart)
80
Temperature (°C)

60

40

20
Uninterrupted Case
Interrupted Case
0
0 10 20 30 40 50 60
Time (s)

Figure 8.15: Temperature outside of a door to a fire room that is opened at 30 s.

106
Chapter 9

Thermal Radiation

There are numerous examples in the heat transfer literature of exact solutions, for simple configurations of
hot and cold objects, of the radiation transport equation. This chapter presents a number of these examples,
plus others aimed at assessing the influence of absorptive media.

9.1 Radiation View Factors

The examples in this section test the basic radiation transport solver against exact solutions derived for
simple geometric shapes.

9.1.1 Hot Plate (plate_view_factor)

These verification cases test the computation of radiative heat flux from a hot plate to a target surface 1 m
away in different coordinate systems. The radiating surface is at 1000 °C and has an emissivity of 1.0. The
exact values are calculated using the analytical expressions for the view factors.
Co-ordinates Radiation source Heat flux (kW/m2 )
2D Cartesian Infinite plate of width 2 m 105.3
3D Cartesian Square plate of width 2 m 81.8
2D cylindrical Circular disk of diameter 2 m 74.1
A comparison of exact values and FDS predictions at three different angular resolutions is shown in Fig. 9.1.

9.1.2 Plates at a Right Angle (geom_rad)

In this case, two flat rectangular obstructions (0.98 m by 0.98 m, A = 0.9604 m2 ) abut at a right angle.
The view factor F12 = 0.20004. The vertical plate has a surface temperature of Thot = 1273.15 K and the
horizontal plate has a temperature of Tcold = 293.15 K. The emissivity of each surface is 1, and the convective
heat transfer coefficient is assumed to be 0. Two cases are considered. The first is where the plates are input
as conventional rectangular OBSTs, and the second where the plates are immersed boundary GEOMs. There
should be no difference in the total energy radiated from hot to cold plates, Fig. 9.2 (left), or the incident
heat flux in the center of the cold plate, Fig. 9.2 (right). The exact solution for the integrated incident (i.e.
4 A = 28.62 kW. The purpose of this exercise is
incoming) heat flux to the surface of the cold plate is F12 σ Thot
to ensure that both types of obstructions behave the same way.

107
 FDS-6.8.0-0-g886e009-release
120
Radiative Heat Flux (plate_view_factor)

100

Heat Flux (kW/m²)

Exact 2D
FDS 2D
80 Exact 3D
FDS 3D
Exact cyl.
60 FDS cyl.

40

20 40 60 80 100
Number of Radiation Angles

Figure 9.1: Convergence study showing increased accuracy of radiative transport equation with increasing number of
angles.

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
40 40
Integrated Heat Flux (geom_rad) Heat Flux (geom_rad)

30 30
Integrated Heat Flux (kW)

Heat Flux (kW/m²)

20 20

10 10
FDS (OBST) FDS (OBST)
FDS (GEOM) FDS (GEOM)
0 0
0 0.5 1 1.5 2 0 0.5 1 1.5 2
Time (s) Time (s)

Figure 9.2: A hot wall radiates onto a cold floor. At left is the total rate of energy transfer, and at right is the heat flux
to the center of the floor.

9.1.3 Plates at a Right Angle, Part 2 (geom_rad_2)

The case in the previous section is repeated, only now the two plates are not aligned with the Cartesian grid.
Figure 9.3 displays the comparison of exact and computed solutions for the integrated incident heat flux to
the cold plate.

108
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
40 40
Integrated Heat Flux (geom_rad_2) Heat Flux (geom_rad_2)

30 30
Integrated Heat Flux (kW)

Integrated Heat Flux (kW)

20 20

10 10
Exact Exact
FDS FDS
0 0
0 0.5 1 1.5 2 0 0.5 1 1.5 2
Time (s) Time (s)

Figure 9.3: The integrated (left) and point (right) heat flux from a hot plate to a cold plate at a right angle.

9.2 Radiation inside a Box (radiation_box)

This verification case tests the computation of three-dimensional configuration factor Φ inside a cube box
with one hot wall and five cold (0 K) walls. An overview of the test geometry is shown in Fig. 9.4. The

H4

H3 c a

(y,z)

H2 dA

H1

Figure 9.4: Radiation inside a box geometry.

configuration factors are calculated at the diagonal of the cold wall opposite to the hot wall. The exact values
of the configuration factor from plane element dA to parallel rectangle H are calculated using the analytical
solution [84]
(y, z) ΦH dA (y, z) ΦH dA
0.025 0.1457 0.275 0.2135
0.075 0.1603 0.325 0.2233
0.125 0.1748 0.375 0.2311
0.175 0.1888 0.425 0.2364
0.225 0.2018 0.475 0.2391

109
Different variations of the case include the mesh resolution (203 and 1003 cells) and the number of radiation
angles (50, 100, 300, 1000, 2000). The exact and FDS results are shown in Fig. 9.5.
FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
0.35 0.35
Incident Heat Flux (radiation_box) Incident Heat Flux (radiation_box)
0.3 0.3

0.25 0.25
Heat Flux (kW/m²)

Heat Flux (kW/m²)

0.2 0.2

0.15 0.15
Analytical (Phi_HdA) Analytical (Phi_HdA)
FDS (Flux_20_50) FDS (Flux_100_50)
0.1 FDS (Flux_20_100) 0.1 FDS (Flux_100_100)
FDS (Flux_20_300) FDS (Flux_100_300)
0.05 FDS (Flux_20_1000) 0.05 FDS (Flux_100_1000)
FDS (Flux_20_2000) FDS (Flux_100_2000)
0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
Position (m) Position (m)

Figure 9.5: Incident heat flux.

The results provide an illustration of numerical errors due to both angular resolution and spatial resolu-
tion. With 203 cells (5-cm spatial resolution), the discrepancies between the FDS and analytical solutions
decrease when increasing the number of radiation angles used in the Radiation Transfer Equation (RTE)
solver: this decrease may be explained by the increasingly accurate description of the angular region occu-
pied by the hot wall when viewed from the selected opposite wall locations. The results also show that at
5-cm spatial resolution (203 cells), the discrepancies between the FDS and analytical solutions do not totally
vanish and feature a residual error of approximately 10% even when using a large number of radiation an-
gles: this residual error may be explained by inaccuracies in the numerical description of the configuration
factors (configuration factors are defined as spatial integrals over the hot wall surface of an expression that
features two cosine angles and a separation distance; the implicit evaluation of configuration factors in the
RTE solver is subject to spatial integration errors). Consistent with this explanation, the results show that at
1-cm spatial resolution (1003 cells), the discrepancies between the FDS and analytical solutions decrease to
very small levels and feature a residual error of less than 1%.

9.3 Radiation from a Polygonal Plate (radiating_polygon)

This example is similar to the one presented in Sec. 9.2, except now the radiating plate is a polygon with
sides of even length. The incident radiation heat flux from an n-sided polygon with temperature, T = 1000 ◦ ,
is given by the formula [84]:
s
n H R2 − H 2 h r
q̇00r = σ T 4 √ tan−1 2
; H= ; R= (9.1)
π 1+H 2 1+H l l

where h is the distance from the center of the polygon to the center point of any of its sides, r is the distance
from the center to any vertex, and l is the distance from the center to the target surface along a perpendicular
line. Figure 9.6 displays calculations performed with a range of grid and angular resolutions. Note that the
number of angles used to solve the RTE is quadrupled when the number of grid cells in each coordinate
direction is doubled.

110
 FDS-6.8.0-0-g886e009-release

140 Radiative Flux from a Hot Square Plate

)
120

2
Heat Flux (kW/m
100

80
analytical
60 5 cm, 100 angles
2.5 cm, 400 angles
40 1.25 cm, 1600 angles

0 0.2 0.4 0.6 0.8 1

Distance from Plate (m)

Figure 9.6: Incident heat flux from a hot square plate as a function of distance to the center of the plate.

9.4 Radiation from a Layer of Hot Gas (radiation_plane_layer)

This case tests the computation of three-dimensional radiation from a homogeneous, infinitely wide layer
of hot gases. The temperature of the layer is 1273.15 K and the absorption coefficient, κ, is varied. The
thickness of the layer is fixed at 1 m, and the optical depth is τ = (1 κ). Wall temperatures are set to 0 K.
The results are compared against the exact solution S(τ) presented in [85]
S(τ) = Sb [1 − 2E3 (τ)] (9.2)
where Sb = σ T 4 is the black-body heat flux from the radiating plane and E3 (τ) is the exponential integral
function (order 3) of the optical depth τ.
The FDS results are computed at two mesh resolutions in the x-direction (I = 20 and I = 150). For
I = 20, both one-band and six-band versions are included to test the correct integration of heat fluxes over
multiple bands. For I = 20, 2-D versions are also computed (J = 1). The limiting case, τ = ∞, using a solid
wall of temperature 1273.15 K, is computed to test the wall heat flux computation. The exact values and
FDS predictions of the wall heat fluxes are given in the table below.
τ S(τ) FDS (I=20,J=20) FDS (I=20,J=1) FDS (I=150)
(kW/m2 ) 1 band 6 bands 1 band 6 bands 1 band
0.01 2.8970 2.9182 2.9072 2.8351 2.8244 2.9260
0.1 24.9403 25.5518 25.4556 25.1039 25.0094 25.7062
0.5 82.9457 83.1364 82.8234 84.3764 84.0587 84.0320
1.0 116.2891 115.4127 114.9782 117.8088 117.3652 116.7828
10 148.9698 148.9717 148.4108 148.9775 148.4167 148.9793
∞ 148.9709 147.9145 147.3577 148.0064 147.4492 148.1923

9.5 Wall Internal Radiation (wall_internal_radiation)

In-depth absorption of thermal radiation in a solid is computed using a two-flux model. In this example, the
accuracy of the two-flux model is tested in the computation of the emissive flux from a homogeneous layer
of material (thickness L = 0.1 m) at 1273.15 K temperature, surrounded by an ambient temperature of 1 K.
The absorption coefficient κ is varied to cover a range [0.01, 10] of optical depth τ = κL.
The exact solutions for radiative flux are the analytical solutions of plane layer emission [85]
S(τ) = Sb [1 − 2E3 (τ)] (9.3)

111
where Sb = σ T 4 is the black-body heat flux from the radiating plane and E3 (τ) is the exponential integral
function (order 3) of optical depth, τ. The exact solutions and FDS results are shown in the table below.

τ S(τ) FDS
(kW/m2 ) (kW/m2 )
0.01 2.897 2.950
0.1 24.94 26.98
0.5 82.95 93.90
1.0 116.3 128.4
10. 149.0 149.0

9.6 Radiation Emitted by Hot Spheres (hot_spheres)

This case tests the calculation of the radiation heat flux from a collection of hot objects. Within two com-
pletely open volumes that are 1 m on a side, hot spheres with a diameter of 1 cm and temperature of 500 °C
are situated within a smaller volume that is 0.5 m on a side. One of the open volumes contains 10 spheres,
the other contains 50,000. The simulation lasts for 10 s. In the first volume, the heat loss is just the sum of
the areas of the spheres multiplied by σ T 4 (emissivity is specified as unity in this case). See the upper left
plot in Fig. 9.7. In the second volume, the region is densely packed and it is expected that the collection of
spheres will radiate like a solid cube that is 0.5 m on a side and whose temperature is 500 °C. This is just
under 30 kW, as shown in the upper right of Fig. 9.7. The plot in the lower left of Fig. 9.7 shows the incident
heat flux to a gauge positioned at the center of the densely packed volume. It is expected that this gauge
would register a heat flux of σ T 4 = 20.26 kW/m2 . Finally, the plot in the lower right of Fig. 9.7 indicates
the integrated intensity, 4σ T 4 = 81.04 kW/m2 .

112
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
0 0
Loose Pack Radiation Loss (hot_spheres) Tight Pack Radiation Loss (hot_spheres)
-0.02
-20
Energy Loss (kW)

Energy Loss (kW)

-0.04
-40
-0.06

-60
-0.08
Exact (Rad Loss 1) Exact (Rad Loss 2)
FDS (rad loss 1) FDS (rad loss 2)
-0.1 -80
0 2 4 6 8 10 0 2 4 6 8 10
Time (s) Time (s)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
25 100
Incident Heat Flux (hot_spheres) Integrated Intensity (hot_spheres)
20 80
Heat Flux (kW/m²)

Intensity (kW/m²)
15 60

10 40

5 20
Exact (Rad Flux) Exact (Intensity)
FDS (rad flux) FDS (UII)
0 0
0 2 4 6 8 10 0 2 4 6 8 10
Time (s) Time (s)

Figure 9.7: The total heat losses, radiative heat flux, and integrated intensity of a collection of loosely and densely
packed radiating spheres.

9.7 Radiation Absorbed by Liquid Droplets (droplet_absorption)

These cases check that energy is conserved when radiation is absorbed by liquid droplets. We want to make
sure that the absorbed thermal radiation is accurately converted into internal energy of the droplets. The rate
of increase of the average droplet temperature, T p , is related to the rate of radiation absorption of all the
droplets, Q̇r , according to the following equation:

dT p
m cp = Q̇r (9.4)
dt
where m and c p are the total mass and specific heat of the droplets, respectively. The convective heat transfer
between the gas and the droplets is small because the gas phase Prandtl number has been set to an artificially
high value. If the radiative intensity remains constant during the time step, the average droplet temperature
over a time period, ∆t, should be:
∆t Q̇r
T p = T0 + (9.5)
m cp
By setting T0 = 0 °C, ∆t = 0.01 s, m = 0.01 kg, and c p = 1.0 kJ/(kg·K), the final value of T p should equal
Q̇r . Figure 9.8 shows comparisons of the predicted average droplet temperature and the expected value in
Cartesian and cylindrical coordinate systems.

113
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
10 10
Droplet Temperature (droplet_absorption_cart) Droplet Temperature (droplet_absorption_cyl)
8 8
Temperature (°C)

Temperature (°C)
6 6

4 4

2 2
Expected (Absorption) Expected (Absorption)
FDS (ADT) FDS (ADT)
0 0
0 0.002 0.004 0.006 0.008 0.01 0 0.002 0.004 0.006 0.008 0.01
Time (s) Time (s)

Figure 9.8: Transfer of absorbed thermal radiation energy into increased average droplet temperature.

9.8 Consistency of Particle Absorbed Energy and the Radiative Loss Term
Small spherical, cylindrical and planar particles with black (emissivity = 1) surfaces are placed in radiation
field. Radiation is coming from black surfaces, and the gas is pure Nitrogen, i.e. non-participating. The
amount of energy removed from the intensity field is measured by a device calculating spatial and tempo-
ral integrals of radiation loss term (Q_w). The reference solution (E_w) is the wall enthalpy increase, as
observed by the one-dimensional heat conduction solver. In particle_isotropic_radi, the radiation
field isotropic, i.e. all surfaces are black and constant temperature. In particle_anisotropic_radi,
an anistropic intensity field is geated by specifying one hot wall, one cold wall, and four other fields being
mirrors.
FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
12 12
Enthalpy (particle_isotropic_radiation) Enthalpy (particle_anisotropic_radi)
10 10

8 8
Enthalpy (kJ)

Enthalpy (kJ)

6 6
E_w ball E_w ball
4 E_w cyl 4 E_w cyl
E_w plate E_w plate
Q_r ball Q_r ball
2 Q_r cyl
2 Q_r cyl
Q_r plate Q_r plate
0 0
0 10 20 30 40 50 0 10 20 30 40 50
Time (s) Time (s)

Figure 9.9: Absorbed energy of spherical, cylindrical and cartesian particles in isotropic (left) and anisotropic (right)
radiation field.

9.9 Radiation Attenuation by Scattering Particles (part_attenuation)

This case is a qualitative assessment of the attenuation of thermal radiation by four different types of par-
ticles. In each case, the radiation is computed inside a one meter cube with one hot and five cold walls.

114
The radiative heat flux is monitored on the wall that is opposite the hot wall. In each case, 10 g of non-
evaporating, monodisperse, 100 µm particles with a density of 1,000 kg/m3 are placed inside the cube.
Five different cases are computed. The reference case contains no particles. The four other cases contain
transparent particles, water droplets, fuel droplets, and opaque particles. Different methods to specify the
refractive index are tested. For water and fuel droplets, the FDS internal tables are used. For transparent
particles, the refractive index is m = 1.0 + 0 i. For opaque particles, the refractive index is m = 1.5 + 0.5 i.
The resulting heat fluxes are plotted in Fig. 9.10. The transparent particles do not attenuate the radiation, and
the relatively opaque particles give the highest attenuation. The heat fluxes through water and fuel droplets
fall between the two extremes.
This case also demonstrates the finite convergence rate of the radiation calculations with scattering
media. As can be seen, the radiation field is not converged to its final value instantaneously because there
are no internal iterations in the radiation solver. As a result, a slight increase of the heat flux is observed
after 0.5 s, which is the third time step, i.e. the second call of the radiation solver. Further, smaller increase
can be seen three steps later. In real applications with small time steps, the time lag of the radiation field is
small.
FDS-6.8.0-0-g886e009-release
6
Radiation attenuation (part_attenuation)
Radiative heat flux (kW/m²)

5.5

Reference
4.5 Transparent
Water
Fuel
Opaque
4
0 0.5 1 1.5 2
Time (s)

Figure 9.10: Attenuation of thermal radiation by four different classes of Lagrangian particles.

9.10 Radiation Attenuation by Non-Scattering Particles

(radiation_gas-veg_consistency)
Relatively large solid particles, like idealized vegetation, act as opaque, non-scattering absorbers of thermal
radiation. For a cloud of particles contained within a region Ω with volume V , the absorption coefficient is
taken to be
As,m
κp = ∑ (9.6)
m∈Ω 4V
where As,m is the surface area of particle m. For this test case, the space between two infinitely wide parallel
plates, 1 m apart, is filled with either a gas with constant absorption coefficient, κ = 5.04 m−1 , or a uniform
distribution of mondisperse solid spheres of radius 0.7614 mm and density 514 kg/m3 such that the total
mass of spheres per unit volume is m000 = 2.63 kg/m3 . In the simulation, only one particle per grid cell is
used to represent all the particles in that cell. The effective absorption coefficient can be calculated:
m000 V As,m
Np = ; κp = Np = 5.04 m−1 (9.7)
mp 4V

115
Here, Np is the actual number of spheres represented by the single particle within a grid cell that is 10 cm
on each side.
Figure 9.11 displays the radiative heat flux at various distances from the hot plate, demonstrating that
the cloud of particles act the same way as a gas with constant absorption coefficient.
FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
80 80
Heat Flux (radiation_gas-veg_consistency_gas) Heat Flux(radiation_gas-veg_consistency_veg)

60 60
Heat Flux (kW/m²)

Heat Flux(kW/m²)
Expected (radx0) Expected (radx0)
Expected (radx0.5) Expected (radx0.5)
Expected (radx1) Expected (radx1)
40 FDS (radx0) 40 FDS (radx0)
FDS (radx0.5) FDS (radx0.5)
FDS (radx1) FDS (radx1)

20 20

0 0
0 10 20 30 40 0 10 20 30 40
Time (s) Time (s)

Figure 9.11: Heat flux at various distances from an infinitely wide hot plate where the radiation is attenuated by either
a gas with constant absorption coefficient (left) or a cloud of mondisperse, non-scattering spheres (right).

9.11 Transient Thermocouple Response (TC_heating)

A thermocouple is placed in the center of a 1 m cube with walls set to 600 °C and gas temperature set to
20 °C. The heat transfer coefficient is set to zero to eliminate convective heating. The gas temperature
remains at 20 ◦ C because there is no convective heat transfer from the walls. The thermocouple heats up
according to the equation:
dTTC
= εTC σ ATC Ts4 − TTC4


ρTC cTC VTC (9.8)
dt
whose analytical solution is:
TTC (t)
ln(x + 1) − ln(1 − x) + 2 tan−1 (x) εTC σ ATC TTC
t= ; β= ; x= (9.9)
4 β Ts3 TTC (0) ρTC cTC VTC Ts

Figure 9.12 displays the thermocouple response compared to the analytical solution.

116
 FDS-6.8.0-0-g886e009-release
700
Temperature (TC_heating)
600

500

Temperature (°C)
400

300

200

100 Analytical (Temp)

FDS (TC)
0
0 20 40 60 80
Time (s)

Figure 9.12: Increasing temperature of a thermocouple surrounded by walls at 600 °C.

9.12 Radiation and Thermocouples (TC_view_factor)

This case tests the radiation solver using the simple thermocouple model in FDS. Suppose that a small,
thermally thin metal bead is at the center of a 1 m cube. The left half of the cube boundary surface is at
20 °C, the right half at 500 °C. Assume that there is no convective heat transfer to the bead; only radiation.
Assume everything has an emissivity of 1. The bead temperature ought to rise to a steady value of
 14  14
T14 + T24 293.154 + 773.154
 
◦
TTC = − 273.15 = − 273.15 = 380.3 C (9.10)
2 2

The thermocouple model in FDS uses the integrated intensity divided by 4, U/4, as the average heat flux to
the bead surface. In the absence of convection, the steady value of the bead temperature ought to be
  14
U
TTC = − 273.15 (9.11)
4σ

This is essentially what is plotted in Fig. 9.13, compared to the ideal value as computed in Eq. (9.10).
FDS-6.8.0-0-g886e009-release
500
Temperature (TC_view_factor)
400
Temperature (°C)

300

200

100
Ideal (TC)
FDS (TC_1)
0
0 50 100 150
Time (s)

Figure 9.13: Temperature of a thermocouple equidistant from walls at temperatures of 20 °C and 500 °C.

117
9.13 Radiation Targets (target_test)
This case tests the functionality of particle “targets;” that is, massless Lagrangian particles that serve to
record the heat flux at a given point with a given orientation. Essentially, the particle is an infinitesimally
small heat flux gauge. Two unit cubes are set up, each with a top surface set to 1000 ◦ C. In one cube, a
small solid obstruction is created with a device embedded in its top facing upwards. In the other cube, a
plate-shaped particle is placed at the same location and given an upward facing orientation. The radiative
heat flux to both the obstruction and the particle should be the same, as shown in Fig. 9.14.
FDS-6.8.0-0-g886e009-release
100
Heat Flux (target_test)
80
Heat Flux (kW/m²)

60

40

20
Particle Target
Obstruction Target
0
0 1 2 3 4 5
Time (s)

Figure 9.14: Heat flux from a hot plate to two different types of targets.

9.14 Radiation Shield (radiation_shield)

A lead radiation shield with a thickness of 0.1 mm is placed between two infinitely wide parallel plates at
500 °C and 20 °C. The computational domain has dimensions of 10 m by 0.1 m by 1 m. The sides of the
domain are assumed to be symmetry planes. The emissivity, ε3 , of the lead shield is 0.48 on both sides. At
400 s, the shield is removed. Assuming that there is no convective heat transfer and that the emissivity of
the hot and cold plates is 1, the steady-state heat flux to the cold plate should be [77]:

σ (T14 − T24 )
q̇0012 =   = 4.76 kW/m2 (9.12)
1 1 1−ε3
ε1 + ε2 + 2 ε3

where T1 and T2 are the hot and cold plate temperatures, respectively, and ε1 , ε2 , and ε3 are the emissivity
values for the hot plate, cold plate, and lead shield, respectively. When the shield is removed at 400 s, the
heat flux should increase to 19.8 kW/m2 . The predicted heat flux is given in Fig. 9.15. Because the heat
flux from the hot plate to the shield is equal to the heat flux from the hot plate to the cold plate, the shield
temperature can be calculated using [77]:

σ (T14 − T34 )
q̇0012 = q̇0013 = 1 1
; T3 = 380.4 ◦ C (9.13)
ε1 + ε3 − 1

The predicted shield temperature is plotted in Fig. 9.15.

118
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
25 500
Heat Flux (radiation_shield) Temperature (radiation_shield)
20 400
Heat Flux (kW/m²)

Temperature (°C)
15 300
Ideal (Flux)
FDS (Flux)
10 200

5 100
Ideal (Temp)
FDS (Temp)
0 0
0 100 200 300 400 500 0 100 200 300 400 500
Time (s) Time (s)

Figure 9.15: (Left) Heat flux to a cold plate separated from a hot plate by a heat shield. At 400 s, the shield is removed.
(Right) Temperature of the heat shield.

9.15 Radiation from a gas-fired panel (radiation_gas_panel)

A gas-fired panel is often used as a source of thermal radiation in laboratory experiments. Here, the heat flux
from an h = 30.48 cm square panel is computed at distances of 10 cm, 15 cm, 25 cm, 38 cm, 46 cm, 61 cm,
and 76 cm along its perpendicular center axis, and along a perpendicular line that is displaced laterally 14 cm
and vertically 14 cm from the center axis, as illustrated in Fig. 9.16. We compare the results computed by
FDS to those calculated using configuration factors. The temperature and emissivity of the panel are set to
1250 K and 0.7, respectively, and the ambient temperature is 273 K.
The configuration factor, dF2→1 , describing the energy exchange between the radiant panel and a target
with differential area dA1 located on a perpendicular line extending from the panel at the point (x, y) is
computed from the following:

R2
Z h/2 Z h/2
dA1
dF2→1 (R; x, y) = dx dy (9.14)
π A2 −h/2 −h/2 [(x − x)2 + (y − y)2 + R2 ]2

The comparison of FDS and the exact solution is shown in Fig. 9.16.
FDS-6.8.0-0-g886e009-release
100

panel
Incident Heat Flux (radiation_gas_panel)
radiant 80
dA2
Heat Flux (kW/m²)

60 Exact (HF_cen)
y Exact (HF_off)
x θ FDS (HF_cen)
h R 40 FDS (HF_off)
r
20
R
h
θ 0
0 0.2 0.4 0.6 0.8 1
dA1 Distance (m)

Figure 9.16: Left: A radiant panel and target dA1 located a distance R from the panel on its center axis. Right:
Comparison of FDS and the exact result.

119
Chapter 10

Species and Combustion

This chapter contains examples that test the computations related to species concentrations, gas properties
and combustion.

10.1 Gas Properties (species_props)

This is a test of computing the gas properties of species mixtures. Three species are defined. The first species
has a molecular weight of 10 g/m, a viscosity of 0.00001 kg/(m · s), a conductivity of 0.01 W/(m · K), and a
specific heat of 1 kJ/(kg · K). The second species multiplies the first species’ inputs by two (e.g. molecular
weight of 20 g/m), and the third species multiplies the first species’ inputs by three. Five mixtures are tested:
pure species 1; pure species 3; equal volume fractions of species 1 and species 2; equal volume fractions of
all species; and volume fractions of 0.6, 0.3, and 0.1 respectively for the three species.

121
 10 -5 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
5 0.05
Viscosity Thermal Conductivity
4 0.04

Expected 1 Expected 1

k (W/m/K)
3 0.03
µ (kg/m/s)

Expected 2 Expected 2
Expected 3 Expected 3
Expected 4 Expected 4
2 Expected 5 0.02 Expected 5
FDS 1 FDS 1
FDS 2 FDS 2
1 FDS 3 0.01 FDS 3
FDS 4 FDS 4
FDS 5 FDS 5
0 0
0 0.005 0.01 0.015 0.02 0 0.005 0.01 0.015 0.02
Time (s) Time (s)

FDS-6.8.0-0-g886e009-release
5
Specific Heat
4

Expected 1
c p (kJ/kg/K)

3 Expected 2
Expected 3
Expected 4
2 Expected 5
FDS 1
FDS 2
1 FDS 3
FDS 4
FDS 5
0
0 0.005 0.01 0.015 0.02
Time (s)

Figure 10.1: Results of the species_props cases.

10.2 Humidity (humidity)

This is a test of the relative humidity computation. A simple case is run with the ambient temperature set
to 70 ◦ C and the ambient humidity set to 30 %. The figure compares the FDS initial mass fraction of water
vapor against a hand calculation.

122
 FDS-6.8.0-0-g886e009-release
0.1
Expected
Humidity Test (humidity) FDS

Water Vapor Mass Fraction (kg/kg)

0.08

0.06

0.04

0.02

0
0 0.2 0.4 0.6 0.8 1
Time (s)

Figure 10.2: Test of 30 % relative humidity at 70 ◦ C.

10.3 Product Yields (methane_flame)

A simple methane flame with 10 % CO yield is burned inside a closed cavity for 5 s. The total mass of
CO2 , H2 O and CO is recorded. The simulation is performed using the three alternative ways to define the
combustion reaction: simple chemistry, primitive species, and lumped species. The same amount of each
product species should be produced in each case. The following figure compares the mass of the three
combustion products. The figure compares the simple chemistry versus the primitive species versus the
lumped species.
FDS-6.8.0-0-g886e009-release
0.05
Species mass (simple v. primitive v. lumped)
0.04
CO2 simple
H2O simple
CO simple * 30
0.03
Mass (kg)

CO2 primitive
H2O primitive
CO primitive * 30
0.02 CO2 lumped
H2O lumped
CO lumped * 30
0.01

0
0 2 4 6 8 10
Time (s)

Figure 10.3: The combustion product masses using three alternative definitions of the reaction: Simple chemistry vs.
primitive species vs. lumped species.

10.4 Reaction Rates (reactionrate)

This section describes a series of test cases designed to check the reaction rate mechanism using the eddy
dissipation concept model for infinitely fast reactions (EDC) and finite-rate (Arrhenius) reactions. In the
most basic sense, if a reaction occurs the rate of the reaction is either infinite or finite-rate. For an infinite-
rate reaction, the rate is limited by the minimum reactant concentration. For i finite-rate reactions, the rate of
species α is dependent on factors such as local species concentration (Yα ), temperature (T ), and Arrhenius

123
parameters (E, A, a):
 
dYα 1 να,i Wα
=− ∑ A0i T ni e−Ei /RT ∏ Yα a α,i
where A0i = Ai ρ ∑ aα ∏ Wα −aα,i (10.1)
dt ρ i νF,i WF
Detailed discussion on the combustion routines can be found in the FDS Technical Reference Guide [1].

10.4.1 Mixing-Controlled Reactions (reactionrate_EDC)

The cases presented in the section are designed to examine species evolutions for reactions of varying
complexity using the EDC model with infinitely fast chemistry. The first test case is a one-step methane
reaction that occurs in a constant volume, adiabatic “reactor” with a fixed mixing time of τmix = 0.125 s.
Each computational cell contains a stoichiometric mixture of fuel and air. In this problem “air” is defined to
be 21 % oxygen and 79 % nitrogen by volume.
CH4 + 2 O2 → CO2 + 2 H2 O (10.2)

FDS-6.8.0-0-g886e009-release
0.3
Species Mass Fraction (EDC 1 Step)
0.25
Expected O2
0.2 Expected CH4
Mass Fraction

Expected CO2
Expected H2O
0.15 FDS O2
FDS CH4
0.1 FDS CO2
FDS H2O

0.05

0
0 0.5 1 1.5 2
Time (s)

Figure 10.4: Time evolution of species mass fraction for a one-step methane EDC reaction with a fixed mixing time
of τmix = 0.125 s.

In addition to species, we want to compare FDS predictions for temperature and pressure. For this
reacting system, the internal energy remains constant, thus the temperature can be found from:
uP,α − uR,α = 0 (10.3)
where P represents the products, R represents the reactants, and uα = hα − RT . The total enthalpy, hα , is
◦
defined as the sum of the enthalpy of formation (h f ) and sensible enthalpy (∆h) for each of the α species.
Assuming the products behave like an ideal gas, the final pressure can be found using the equation of state
for an ideal gas:
PV = nRT (10.4)

Under the same conditions as the methane case, species, temperature, and pressure are compared to
expected results for a one-step propane reaction:
C3 H8 + 5 O2 → 3 CO2 + 4 H2 O (10.5)

124
 FDS-6.8.0-0-g886e009-release 10 5 FDS-6.8.0-0-g886e009-release
3500 12
Temperature (EDC 1 Step CH4) Pressure (EDC 1 Step CH4)
3000 10
2500
8
Temperature (°C)

Pressure (Pa)
2000
6
1500
4
1000

500 Expected Temperature

2 Expected Pressure
FDS Temperature FDS Pressure
0 0
0 0.5 1 1.5 2 0 0.5 1 1.5 2
Time (s) Time (s)

Figure 10.5: Time evolution of temperature and pressure for a one-step methane EDC reaction with a fixed mixing
time of τmix = 0.125 s.

FDS-6.8.0-0-g886e009-release
0.3
Species Mass Fraction (EDC 1 Step)
0.25
Expected O2
0.2 Expected C3H8
Mass Fraction

Expected CO2
Expected H2O
0.15 FDS O2
FDS C3H8
0.1 FDS CO2
FDS H2O

0.05

0
0 0.5 1 1.5 2
Time (s)

Figure 10.6: Time evolution of species mass fraction for a one-step propane EDC reaction with a fixed mixing time of
τmix = 0.125 s.

FDS-6.8.0-0-g886e009-release 5 FDS-6.8.0-0-g886e009-release
10
3500 12
Temperature (EDC 1 Step C3H8) Pressure (EDC 1 Step C3H8)
3000 10
2500
8
Temperature (°C)

Pressure (Pa)

2000
6
1500
4
1000

500 Expected Temperature

2 Expected Pressure
FDS Temperature FDS Pressure
0 0
0 0.5 1 1.5 2 0 0.5 1 1.5 2
Time (s) Time (s)

Figure 10.7: Time evolution of temperature and pressure for a one-step propane EDC reaction with a fixed mixing
time of τmix = 0.125 s.

125
The results for the propane case are shown in Figs. 10.6 (species) and 10.7 (temperature and pressure).
In the previous two examples, each computational cell contained a stoichiometric mixture of fuel and
air. For this third example, the cells are initialized to be all fuel in one part of the reactor and all air in
the remaining volume. Globally, the reactor is fuel rich and follows the same one-step methane reaction as
described by 10.2. Figures 10.8 and 10.9 show the species concentrations, temperature, and pressure from
FDS and the expected values. In this case, the final state is not reached until approximately 35 s compared
to 1 s for the cases where each cell contained a flammable composition.
FDS-6.8.0-0-g886e009-release
0.3
Species Mass Fraction (EDC 1 Step)
0.25
Expected O2
0.2 Expected CH4
Mass Fraction

Expected CO2
Expected H2O
0.15 FDS O2
FDS CH4
0.1 FDS CO2
FDS H2O

0.05

0
0 10 20 30 40 50 60
Time (s)

Figure 10.8: Time evolution of species mass fraction for a one-step methane EDC reaction with a fixed mixing time
of τmix = 0.125 s.

FDS-6.8.0-0-g886e009-release 5 FDS-6.8.0-0-g886e009-release
10
3500 12
Temperature (EDC 1 Step CH4) Pressure (EDC 1 Step CH4)
3000 10
2500
8
Temperature (°C)

Pressure (Pa)

2000
6
1500
4
1000

500 Expected Temperature

2 Expected Pressure
FDS Temperature FDS Pressure
0 0
0 10 20 30 40 50 60 0 10 20 30 40 50 60
Time (s) Time (s)

Figure 10.9: Temperature (left) and pressure (right) evolution for a one-step methane EDC reaction with a fixed mixing
time of τmix = 0.125 s.

The EDC reaction implementation is also tested to check species evolution for chemical reactions of
increased complexity. The first is a consecutive two-step, fuel-limited methane reaction:
CH4 + 1.5 O2 → CO + 2 H2 O (10.6)
CO + 0.5 O2 → CO2
Each cell in the model reaction is initialized with an oxygen rich mixture and the mixing time, τmix , is fixed
at 0.05 s (Fig. 10.10).

126
 FDS-6.8.0-0-g886e009-release

0.8 Species Mass Fraction (EDC 2 Step)

Expected O2
Expected CH4
0.6 Expected CO

Mass Fraction
Expected CO2
Expected H2O
FDS O2
0.4
FDS CH4
FDS CO
FDS CO2
0.2 FDS H2O

0
0 0.2 0.4 0.6 0.8 1
Time (s)

Figure 10.10: Time evolution of species mass fraction for a series, fuel-limited EDC reaction with a fixed mixing time
of τmix = 0.05 s.

The second reaction test case is a consecutive three-step, fuel limited propane reaction:
C3 H8 + 1.5 O2 → 3 CO + 4 H2 (10.7)
CO + 0.5 O2 → CO2
H2 + 0.5 O2 → H2 O
The species evolutions are compared to expected values in Fig. 10.11.
FDS-6.8.0-0-g886e009-release
0.3
Species Mass Fraction (Series Reaction)
0.25
Expected O2
0.2 Expected C3H8
Mass Fraction

Expected CO2
Expected H2O
0.15 FDS O2
FDS C3H8
0.1 FDS CO2
FDS H2O

0.05

0
0 0.2 0.4 0.6 0.8 1
Time (s)

Figure 10.11: Results of the reactionrate_series_reaction case.

The third test case has two competing, oxygen-limited reactions; a one-step methane and one-step
propane reaction.
CH4 + 2 O2 → CO2 + 2 H2 O (10.8)
C3 H8 + 5 O2 → 3 CO2 + 4 H2 O
The species evolutions are compared to expected values in Fig. 10.12.
The fourth test case has propane (fuel), competing with two identical but separately lumped air species.
Each cell is initialized with stoichiometric amount of fuel and air; each air is initialized in equal quantities.
C3 H8 + 5 AIR1 → PRODUCTS (10.9)

127
 FDS-6.8.0-0-g886e009-release
0.5
Species Mass Fraction (EDC 2 Fuel)
0.4 Lumped CH4
Lumped C3H8
Lumped O2

Mass Fraction
0.3 Lumped CO2
Lumped H2O
Primitive CH4
0.2 Primitive C3H8
Primitive O2
Primitive CO2
0.1 Primitive H2O

0
0 0.2 0.4 0.6 0.8 1
Time (s)

Figure 10.12: Time evolution of species mass fraction for oxygen-limited EDC two fuel reaction set with a fixed
mixing time of τmix = 0.1 s.

C3 H8 + 5 AIR2 → PRODUCTS

The lumped species used in these reactions are defined by:

&SPEC ID = 'PROPANE' /
&SPEC ID = 'OXYGEN', LUMPED_COMPONENT_ONLY = .TRUE. /
&SPEC ID = 'NITROGEN', LUMPED_COMPONENT_ONLY = .TRUE. /
&SPEC ID = 'WATER VAPOR', LUMPED_COMPONENT_ONLY = .TRUE. /
&SPEC ID = 'CARBON DIOXIDE', LUMPED_COMPONENT_ONLY = .TRUE. /

&SPEC ID = 'AIR1', SPEC_ID = 'OXYGEN','NITROGEN', VOLUME_FRACTION =1,3.76,

BACKGROUND=.TRUE. /
&SPEC ID = 'AIR2', SPEC_ID = 'OXYGEN','NITROGEN', VOLUME_FRACTION =1,3.76 /
&SPEC ID = 'PRODUCTS', SPEC_ID = 'CARBON DIOXIDE', 'WATER VAPOR',
'NITROGEN',VOLUME_FRACTION = 3,4,18.8 /

The species evolutions are compared to expected values in Fig. 10.13.

FDS-6.8.0-0-g886e009-release
1
Species Mass Fraction (Two Lumped)
0.8
Mass Fraction

0.6
Expected C3H8
Expected AIR1
0.4 Expected AIR2
Expected Products
FDS C3H8
0.2 FDS AIR1
FDS AIR2
FDS Products
0
0 0.2 0.4 0.6 0.8 1
Time (s)

Figure 10.13: Results of the reactionrate_lumped_two_species case.

128
10.4.2 Finite Rate Reactions (reactionrate_arrhenius)
The main difference between Arrhenius rate reactions and the mixing-controlled reactions is that the reaction
rate in finite rate reactions is found from local parameters such as species concentration and temperature as
well as rate constants. The values of aα in Eq. (10.1) are specific to each reaction and for non-zero values,
dYα / dt becomes a non-linear differential equation. We consider both linear and non-linear cases with one
and two-step reaction mechanisms. First we examine two one-step cases; a zero-order and second-order
reaction. Equation (10.10) shows a one-step, zero-order test case of carbon monoxide oxidation:

CO + 0.5 O2 → CO2 (10.10)

The reaction rate input parameters are shown in the table below. E, aα , and nα are set to zero, and A is set
such that ki = 1. The resulting reaction rates simplify to a function of molecular weight and stoichiometric
coefficients. Species evolutions results from the CO reaction in Eq. (10.10) are shown in Fig. 10.14. The

Table 10.1: Arrhenius values for a single step CO reaction; α = [CO O2 CO2 ].

Reaction να E (J/mol) A (cm3 /mole/s) aα nα

10.10 [-1 -0.5 1] 0 4.244e-6 [0 0 0] [0 0 0]

FDS-6.8.0-0-g886e009-release
0.6
Species Mass Fraction (Arrhenius 0 Order)
0.5

0.4
Mass Fraction

0.3
Expected O2
0.2 Expected CO
Expected CO2
FDS O2
0.1 FDS CO
FDS CO2
0
0 2 4 6 8 10
Time (s)

Figure 10.14: Time evolution of species mass fraction for a one-step zero-order Arrhenius finite rate reaction.

second finite rate test case is a one-step, second-order propane reaction, Eq. (10.5). The table below shows
the reaction rate input parameters. In this case, aα = [1, 1, 0, 0] for propane, oxygen, carbon monoxide, and
water vapor respectively. This makes the reaction second-order as O = ∑ aα . Species evolutions for the
one-step second-order reaction are shown in Fig. 10.15. Multi-step Arrhenius finite rate reactions are also

Table 10.2: Arrhenius values for a single step C3 H8 reaction; α = [C3 H8 O2 CO2 H2 O].

Reaction να E (J/mol) A (cm3 /mole/s) aα nα

10.5 [-1 -5 3 4] 0 4e5 [1 1 0 0] [0 0 0 0]

129
 FDS-6.8.0-0-g886e009-release
0.3
Species Mass Fraction (Arrhenius 2 Order)
0.25
Expected O2
0.2 Expected C3H8

Mass Fraction
Expected CO2
Expected H2O
0.15 FDS O2
FDS C3H8
0.1 FDS CO2
FDS H2O

0.05

0
0 1 2 3 4 5
Time (s)

Figure 10.15: Time evolution of species mass fraction for a one-step second-order Arrhenius finite rate reaction.

examined. First, we consider a two-step forward propane reaction:

C3 H8 + 3.5 O2 → 3 CO + 4 H2 O (10.11a)
CO + 0.5 O2 → CO2 (10.11b)

with the following reaction input parameters:

Table 10.3: Arrhenius values for a two-step C3 H8 reaction; α = [C3 H8 O2 CO H2 O CO2 ].

Reaction να E (J/mol) A (cm3 /mole/s) aα nα

10.11a [-1 -3.5 3 4 0] 0 5e2 [0.1 1.65 0 0 0] [0 0 0 0 0]
10.11b [0 -0.5 -1 0 1] 0 8e3 [0 0.25 1 0.5 0] [0 0 0 0 0]

The species evolutions for the two-step 1.75 order reaction mechanism are shown in Fig. 10.16. We also
consider a two-step reversible propane reaction:

C3 H8 + 3.5 O2 → 3 CO + 4 H2 O (10.12a)
CO + 0.5 O2 → CO2 (10.12b)
CO2 → CO + 0.5 O2 (10.12c)

with the following reaction input parameters: Equations (10.12a)-(10.12c) includes the decomposition of

Table 10.4: Arrhenius values for a two-step reversible C3 H8 reaction; α = [C3 H8 O2 CO H2 O CO2 ].

Reaction να E (J/mol) A (cm3 /mole/s) aα nα

10.12a [-1 -3.5 3 4 0] 0 5e2 [0.1 1.65 0 0 0] [0 0 0 0 0]
10.12b [0 -0.5 -1 0 1] 0 8e3 [0 0.25 1 0.5 0] [0 0 0 0 0]
10.12c [0 0.5 1 0 -1] 0 4e-4 [0 0 0 0 0 1] [0 0 0 0 0]

CO2 into CO and O2 whereas the forward-only two-step mechanism, Eqs. (10.11a)-(10.11b), does not. This

130
reverse reaction mechanism is represented by three separate reactions on the REAC line in an FDS input file
as indicated by rows a, b and c in the preceding table. Species evolutions for the 2 step reverse reaction
are shown in Fig. 10.17. In the forward CO step, (row b in the two prior tables), we see that water vapor
contributes to the reaction rate based on its local concentration despite not being a participant in the reaction.
Note: the values of E and A used in these examples are not necessarily representative of the chemical
reactions presented. When using finite rate chemistry consult experimental data or literature for appropriate
reaction parameters.
FDS-6.8.0-0-g886e009-release
0.3
Species Mass Fraction (Arrhenius 1.75 Order)
0.25
Expected O2
Expected C3H8
0.2 Expected CO
Mass Fraction

Expected CO2
Expected H2O
0.15 FDS O2
FDS C3H8
0.1 FDS CO
FDS CO2
FDS H2O
0.05

0
0 2 4 6 8 10
Time (s)

Figure 10.16: Time evolution of species mass fraction for a two-step Arrhenius finite rate propane reactions.

FDS-6.8.0-0-g886e009-release
0.3
Species Mass Fraction (Arrhenius 1.75 Order)
0.25
Expected O2
Expected C3H8
0.2 Expected CO
Mass Fraction

Expected CO2
Expected H2O
0.15 FDS O2
FDS C3H8
0.1 FDS CO
FDS CO2
FDS H2O
0.05

0
0 2 4 6 8 10
Time (s)

Figure 10.17: Time evolution of species mass fraction for a two-step reversible Arrhenius finite rate propane reactions.

To couple verification of species mass fraction, compartment temperature, and compartment pressure,
an equilibrium example problem was setup. In this case, we examine a sealed box filled with a stoichiometric
mixture of fuel, oxygen, and nitrogen. The two-step reversible reaction mechanism, Eqs. (10.12a)-(10.12c),
is used to drive the FDS simulation. The initial temperature is set to 350 °C to ensure the reaction occurs
in a timely manner. The reaction rate parameters in the table below are based off values in literature from
Westbrook and Dryer [86]. Equilibrium values were determined using the NASA code Chemical Equilib-
rium with Applications (CEA) [87]. A constant volume, constant internal energy equilibrium calculation
was performed where the equilibrium values were found by minimizing Helmholtz Energy. The equilibrium

131
Table 10.5: Arrhenius values for a two-step reversible C3 H8 reaction for equilibrium; α = [C3 H8 O2 CO
H2 O CO2 ].

Reaction να E (J/mol) A (cm3 /mole/s) aα nα

10.12a [-1 -3.5 3 4 0] 125520 1.5e12 [0.1 1.65 0 0 0] [0 0 0 0 0]
10.12b [0 -0.5 -1 0 1] 167360 3.4e13 [0 0.25 1 0.5 0] [0 0 0 0 0]
10.12c [0 0.5 1 0 -1] 167360 8e8 [0 0 0 0 0 1] [0 0 0 0 0]

species were limited to C3 H8 , O2 , CO, H2 O, and CO2 ; the same species tracked in the FDS simulation. The
results are shown in Figs. 10.18 and 10.19.
FDS-6.8.0-0-g886e009-release
0.25
Species Mass Fraction (Equilibrium)
0.2 Expected O2
Expected C3H8
Expected CO
Mass Fraction

0.15 Expected CO2

Expected H2O
FDS O2
0.1 FDS C3H8
FDS CO
FDS CO2
0.05 FDS H2O

0
0 1 2 3 4
Time (s)

Figure 10.18: Time evolution of species mass fraction for a two-step Arrhenius finite rate reaction compared to equi-
librium values.

132
 FDS-6.8.0-0-g886e009-release 10 5 FDS-6.8.0-0-g886e009-release
3500 5
Temperature (Equilibrium) Pressure (Equilibrium)
3000
4
2500
Temperature (°C)

Pressure (Pa)
2000 3

1500 2
1000
1
500 Expected Temperature Expected Pressure
FDS Temperature FDS Pressure
0 0
0 1 2 3 4 0 1 2 3 4
Time (s) Time (s)

Figure 10.19: Time evolution of temperature (left) and pressure (right) for two-step Arrhenius finite rate propane
reactions compared to equilibrium values.

10.4.3 Fast-Slow Reaction Scheme (reactionrate_fast_slow)

A simple two-step reaction scheme for a clean-burning fuel like methanol can be written as follows:

CH3 OH + O2 → CO + 2 H2 O (10.13a)
CO + 0.5 O2 → CO2 (10.13b)

where the first reaction is infinitely fast and the second is not. Suppose the second reaction proceeds accord-
ing to the evolution equation:
d[CO]
= −1 × 1010 [CO] [O2 ] (10.14)
dt
where [CO] is the concentration of CO in units of mol/cm3 . To test the scheme, a sealed, adiabatic cube
1 m on a side is filled with nitrogen, fuel gas, and oxygen with initial volume fractions of 0.89, 0.05, and
0.06. Because the rates are independent of temperature, the reaction proceeds immediately assuming that
the gases are thoroughly mixed via the parameter INITIAL_UNMIXED_FRACTION=0. Figure 10.20 shows
the fuel volume fraction very quickly dropping to zero, leaving some oxygen which then reacts relatively
slowly with CO to form CO2 . The effective combined reaction is:

5 CH3 OH + 6 O2 → 2 CO2 + 3 CO + 10 H2 O (10.15)

The ’EXPLICIT EULER’ integration scheme is used with an initial gas phase time step of 1 × 10−6 s. The
final volume fractions of CO2 and CO are expected to be 2/104 ≈ 0.0192 and 3/104 ≈ 0.0288, respectively.

133
 FDS-6.8.0-0-g886e009-release
0.1
Exact (CH3OH)
reactionrate_fast_slow Exact (O2)
0.08 Exact (CO2)
Exact (CO)
FDS (CH3OH)

Volume Fraction
0.06 FDS (O2)
FDS (CO2)
FDS (CO)
0.04

0.02

0
0 0.2 0.4 0.6 0.8 1
Time (s) -3
10

Figure 10.20: Species evolution in a fast-slow reaction scheme.

10.5 Heat Release Rates from Reaction Mechanisms (hrrpuv_reac)

One of the most important quantities computed in FDS is the heat release rate per unit volume (HRRPUV),
q̇000 . HRRPUV shows up in the energy equation and hence the expression for the velocity divergence (see
the FDS Tech Guide [48]). The value of q̇000 within a computation cell can be computed two ways, which
should be equivalent. The first is to use individual species heats of formation times the respective chemical
production rates:
q̇000 = − ∑ ṁ000 ◦
α ∆hf,α (10.16)
α

Note that Eq. (10.16) is what is ultimately used in the FDS divergence expression.
The second is to sum the individual reaction rates times their respective heats of combustion, which are
precomputed in the read routines in FDS:

q̇000 = ∑ rF,i
0
∆Hc,i (10.17)
i

Here F denotes the fuel for reaction i.

In this section, we show results comparing Eqs. (10.16) and (10.17) for a series of cases with differ-
ent reaction mechanisms. This test series is meant to touch various pathways through the chemistry ODE
solver involving mixing, fast and slow chemistry, extinction, and series and parallel reactions. The re-
sult are presented in Fig. 10.21. The measure of correctness is that the black dots (hrrpuv), taken from
Eq. (10.16), exactly match with the sum of the individual reaction HRR curves denoted by the solid black
line, Eq. (10.17). Below we briefly describe the key features of each case.

Single Reaction (hrrpuv_reac_single)

This is a single reaction with specified lumped species without extinction.

Series Reactions (hrrpuv_reac_series)

This is a set of series reactions with primitive species.

Simple Reaction (hrrpuv_reac_simple)

This is FDS Simple Chemistry integrated using two Forward Euler substeps.

134
Reaction with Extinction (hrrpuv_reac_extinction)
This is a single reaction with specified lumped species with extinction.

Parallel Reactions (hrrpuv_reac_parallel)

This is a set of parallel reactions with two fuels going to one set of products.

Parallel Reactions (hrrpuv_reac_parallel_2)

This is a set of parallel reactions with one fuel going to two products.

Arrhenius Reaction (hrrpuv_reac_arrhenius)

This is simple Arrhenius rate law using the default ODE integrator for finite-rate chemistry, which is RK4.

Soot Production (hrrpuv_reac_soot)

This example combines a fast reaction for the primary reaction with finite-rate chemistry for the soot oxida-
tion reaction.

135
 4 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
10
2 10000
Reaction Test (hrrpuv_reac_single) Reaction Test (hrrpuv_reac_series)
8000
Heat Release Rate (kW/m³)

Heat Release Rate (kW/m³)

1.5
hrrpuv
6000 q1
hrrpuv
1 q1
q2
q3
4000 q1+q2+q3

0.5
2000

0 0
0 0.2 0.4 0.6 0.8 1 0 0.5 1 1.5 2
Time (s) Time (s)

10 4 FDS-6.8.0-0-g886e009-release 10 4 FDS-6.8.0-0-g886e009-release
2 2
Reaction Test (hrrpuv_reac_simple) Reaction Test (hrrpuv_reac_extinction)
Heat Release Rate (kW/m³)

Heat Release Rate (kW/m³)

1.5 1.5

hrrpuv hrrpuv
1 q1
1 q1

0.5 0.5

0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
Time (s) Time (s)

10 4 FDS-6.8.0-0-g886e009-release 10 4 FDS-6.8.0-0-g886e009-release
2 2
Reaction Test (hrrpuv_reac_parallel) Reaction Test (hrrpuv_reac_parallel_2)
Heat Release Rate (kW/m³)

Heat Release Rate (kW/m³)

1.5 1.5

hrrpuv hrrpuv
q1 q1
1 q2
1 q2
q1+q2 q1+q2

0.5 0.5

0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
Time (s) Time (s)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
2000 2000
Reaction Test (hrrpuv_reac_arrhenius) Reaction Test (hrrpuv_reac_soot)
Heat Release Rate (kW/m³)

Heat Release Rate (kW/m³)

1500 1500

hrrpuv
hrrpuv q1
1000 q1
1000 q2
q1+q2

500 500

0 0
0 1 2 3 4 0 1 2 3 4 5
Time (s) Time (s)

Figure 10.21: Heat release rate per unit volume for the various hrrpuv_reac test cases.

136
10.6 Mixture Fraction (burke_schumann)
For simple one-step reactions of the form F + O → P, there is an optional output quantity called the mixture
fraction. To test that this output is correctly implemented we consider a set of methane-air mixtures in a test
chamber that span the entire range of equivalence ratio. The complete combustion of methane is given by:

CH4 + 2 (O2 + 3.76 N2 ) → CO2 + 2 H2 O + 7.52 N2 (10.18)

|{z} | {z } | {z }
Fuel Oxidizer Products

The simulation results are compared to state relationships that express temperature and species as a function
of the mixture fraction, f [88]. For fuel rich mixtures ( fstoic < f ≤ 1), the expected species mass fractions
are:
f − fstoich
YF = (10.19)
1 − fstoich
YO = 0 (10.20)
1− f
YP = (10.21)
1 − fstoich
For fuel lean mixtures (0 ≤ f < fstoich ):

YF = 0 (10.22)
f
YO = 1 − (10.23)
fstoich
f
YP = (10.24)
fstoich
Figure 10.22 demonstrates that the predicted final fuel, oxidizer, and product mass fractions obey the ideal
state relations. The state relation for temperature is given by:

∆hc
T = T0 + ( f −YF ( f )) (10.25)
cv
The specific heat, c p , for each species is set to 1 kJ/(kg · K) as in the analysis by Turns [88]. The enthalpy
of formation of each of the species except for the fuel is set to zero, such that the heat of combustion of
the reaction equals the heat of formation of the fuel. To obtain true adiabatic conditions in FDS, these
cases were conducted inside a constant-volume reactor, so Eq. (10.25) is based on cv rather than c p . We
use the following relationship cv = c p − R/W to determine cv . The molecular weight of the species were
not assumed to be uniform; thus, the average molecular weight and cv vary with the mixture fraction. As a
result, the traditional linear temperature profile is not expected. The FDS result is shown in Fig. 10.23.

137
 FDS-6.8.0-0-g886e009-release
1
Burke Schumann (Species)
0.8

Expected Fuel

Mass Fraction
0.6 Expected Oxidizer
Expected Products
FDS Fuel
0.4 FDS Oxidizer
FDS Products

0.2

0
0 0.2 0.4 0.6 0.8 1
Mixture Fraction

Figure 10.22: (Left) Lumped species mass fractions as a function of mixture fraction.

FDS-6.8.0-0-g886e009-release
1
Burke Schumann (Temperature)
0.8
Non-Dimensional Temperature

0.6

0.4

0.2
Expected Temperature
FDS Temperature
0
0 0.2 0.4 0.6 0.8 1
Mixture Fraction

Figure 10.23: Non-dimensional temperature, (T − T0 )/(Tstoich − T0 ), as a function of the mixture fraction.

10.7 Ideal and Non-Ideal Heat of Combustion (HoC)

When using a non-predefined fuel, one may have heat of combustion data representing complete combustion
but species yield data representing typical large scale conditions. When using simple chemistry, the input
keyword IDEAL=.TRUE. on the REAC line will have FDS reduce the heat of combustion to account for
any minor products of combustion (e.g., CO and H2 O). This is done by removing energy from the heat of
combustion to account for carbon that goes to CO and soot rather than CO2 and for hydrogen that goes to soot
rather than H2 O. A custom fuel (methane) is defined with CO and soot yields of 0.1, a heat of combustion
of 50000 kJ/kg, and a 40 kW fire. Two input files are created, one with the HEAT_OF_COMBUSTION defined
with IDEAL=.TRUE. and the second with IDEAL=.FALSE.. In the IDEAL=.FALSE. case where the heat of
combustion is not adjusted we expect a fuel mass flow rate of 0.0008 kg/(m2 · s). With the heat of combustion
adjusted we expect a higher fuel mass flow rate of 0.000877 kg/(m2 · s), as shown in Fig. 10.24.

138
 10 -3 FDS-6.8.0-0-g886e009-release 10 -3 FDS-6.8.0-0-g886e009-release
1.2 1.2
Ideal Heat of Combustion Test Non-Ideal Heat of Combustion Test
1 1
Fuel Mass Flow (kg/m²)

Fuel Mass Flow (kg/m²)

0.8 0.8

0.6 0.6

0.4 0.4

0.2 Exact Ideal

0.2 Exact Non-Ideal
FDS Ideal FDS Non-Ideal
0 0
0 1 2 3 4 0 1 2 3 4
Time (s) Time (s)

Figure 10.24: Demonstration of correct adjust to heat of combustion for minor products of combustion.

10.8 Single Step and Two Step Chemistry Heat of Combustion

This verification case checks that single-step and two-step simple chemistry results in the same total heat of
combustion. The case consists of two 1 m3 domains with a 20 kW fire in each (defined using MLRPUA) that
burns for 8.5 s. Each burner is assigned the fuel C3 H8 (as a user defined species). In one mesh, single-step
chemistry is used, and in the other mesh, two-step chemistry is used. The total heat released in both meshes
should be the same as shown in Fig. 10.25.
FDS-6.8.0-0-g886e009-release
200
Q Check (1_step_2_step_compare)
Integrated Heat Release (kJ)

150

1-Step Q
100 2-Step Q

50

0
0 2 4 6 8 10
Time (s)

Figure 10.25: Comparison of the heat release rate for single-step an two-step simple chemistry with the
1_step_2_step_compare test case.

10.9 Fractional Effective Dose (FED) and Fractional Irritant Concentration

(FIC) )
The Fractional Effective Dose index (FED), developed by Purser [89], is a commonly used measure of
human incapacitation due to the combustion gases. The FED value is calculated as

FEDtot = (FEDCO + FEDCN + FEDNOx + FLDirr ) × HVCO2 + FEDO2 (10.26)

139
The fraction of an incapacitating dose of CO is calculated as
Z t
FEDCO = 2.764 × 10−5 (CCO (t))1.036 dt (10.27)
0

where t is time (min), and CCO is the CO concentration (ppm). The fraction of an incapacitating dose of CN
is calculated as    
Z t exp CCN (t)
43
FEDCN =  − 0.0045 dt (10.28)
0 220

where t is time (min), and CCN is the concentration (ppm) of HCN corrected for the protective effect of NO2 .
CCN is calculated as
CCN = CHCN −CNO2 (10.29)
The fraction of an incapacitating dose of NOx is calculated as
Z t
CNOx (t)
FEDNOx = dt (10.30)
0 1500
where t is time (min), and CNOx is the sum of NO and NO2 concentrations (ppm). The Fractional Lethal
Dose (FLD) of irritants is calculated as
Z t 
CHCl (t) CHBr (t) CHF (t) CSO2 (t) CNO2 (t) CC3 H4 O (t) CCH2 O (t)
FLDirr = + + + + + + dt
0 FFLD,HCl FFLD,HBr FFLD,HF FFLD,SO2 FFLD,NO2 FFLD,C3 H4 O FFLD,CH2 O
(10.31)
where t is time (min), the numerators are the instantaneous concentrations (ppm) of each irritant, and the
denominators are the exposure doses of respective irritants predicted to be lethal to half the population. The
lethal exposure doses [89] are given in the table below.
HCl HBr HF SO2 NO2 C3 H4 O CH2 O
FFLD (ppm × min) 114000 114000 87000 12000 1900 4500 22500
FFIC (ppm) 900 900 900 120 350 20 30
The fraction of an incapacitating dose of low O2 hypoxia is calculated as
Z t
dt
FEDO2 = (10.32)
0 exp [8.13 − 0.54 (20.9 −CO2 (t))]
where t is time (min), and CO2 is the O2 concentration (volume percent). The hyperventilation factor induced
by carbon dioxide is calculated as
exp(0.1903CCO2 (t) + 2.0004)
HVCO2 = (10.33)
7.1
where t is time (min), and CCO2 is the CO2 concentration (volume percent).
The Fractional Irritant Concentration (FIC), also developed by Purser [89], represents the toxic effect
which depends upon the immediate concentrations of irritants. The overall irritant concentration FIC is
calculated as
CHCl (t) CHBr (t) CHF (t) CSO2 (t) CNO2 (t) CC3 H4 O (t) CCH2 O (t)
FICirr = + + + + + + (10.34)
FFIC,HCl FFIC,HBr FFIC,HF FFIC,SO2 FFIC,NO2 FFIC,C3 H4 O FFIC,CH2 O
where the numerators are the instantaneous concentrations of each irritant, and the denominators are the con-
centrations of respective irritants expected to cause incapacitation in half the population. The incapacitating
concentrations [89] are given in the table above.

140
10.9.1 FED and FIC of gas mixtures (FED_FIC)
The FED and FIC values were computed from specified constant gas concentrations for four cases using an
external script. The gas concentrations are listed in the table below. The concentrations are in ppm.

Case ID O2, CO2, CO Asphyxiants Irritants All Gases

O2 97727 90218 209000 103055
CO2 343066 19189 0 7463
CO 3242 2456 0 1660
NO 0 134.9 0 89.34
NO2 0 0 1.139 0.569
HCN 0 265.3 0 204.0
HCl 0 0 68.33 34.17
HBr 0 0 68.33 34.17
HF 0 0 52.15 26.07
SO2 0 0 7.193 3.596
C3 H4 O 0 0 2.697 1.349
CH2 O 0 0 13.49 6.743

Two versions of the simulations are performed. In the first simulation, the species were defined as
individual tracked species. The FDS predictions and the expected FED and FIC values are compared in the
figures below. In the second simulation, each species was defined as a lumped species, being part of the
mixture using SMIX. Four different mixtures were defined. The results of this test are shown in Figs. 10.26
and 10.27.

10.9.2 FED from post-combustion yield of CO and HCN (FED_CO_HCN)

The calculation of FED resulting from post-combustion yields of CO2 , CO and HCN, and the corresponding
oxygen reduction is tested by specifying a combustion reaction of Polyisocyanurate (PIR) foam with 1 %
CO yield and 0.5 % yield of HCN. The computational volume is intialized with 15 vol-% of combustion
products. The species concentrations were measured by volume fraction devices, and the expected value of
FED index was calculated using an external Matlab script. The results are shown in 10.28.

10.9.3 FED for a moving particle (FED_moving)

The calculation of FED for a moving particle representing an occupant is tested by specifying species con-
centrations at different levels throughout the domain and moving the particle through these areas. The
species concentrations are initialized to known levels and the expected value of FED index was calculated
using an external Matlab script. The results are shown in 10.29.

141
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
0.7 1
O2, CO2 and CO (FED_FIC) Asphyxiants (FED_FIC)
0.6
0.8
0.5

0.4 0.6
FED

FED
0.3 0.4
0.2
0.2
0.1 Expected Expected
FDS (FED) FDS (FED)
0 0
0 20 40 60 80 100 0 20 40 60 80 100
Time (s) Time (s)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
0.01 0.7
Irritants (FED_FIC) All Gases (FED_FIC)
0.6
0.008
0.5
0.006 0.4
FED

FED
0.004 0.3

0.2
0.002
Expected 0.1 Expected
FDS (FED) FDS (FED)
0 0
0 20 40 60 80 100 0 20 40 60 80 100
Time (s) Time (s)

-3 FDS-6.8.0-0-g886e009-release -3 FDS-6.8.0-0-g886e009-release
10 10
1 1
O2, CO2 and CO (FED_FIC) Asphyxiants (FED_FIC)

0.5 0.5
FIC

FIC

0 0

-0.5 -0.5
Expected Expected
FDS (FIC) FDS (FIC)
-1 -1
0 20 40 60 80 100 0 20 40 60 80 100
Time (s) Time (s)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
1 0.5
Irritants (FED_FIC) All Gases (FED_FIC)
0.8 0.4

0.6 0.3
FIC

FIC

0.4 0.2

0.2 0.1
Expected Expected
FDS (FIC) FDS (FIC)
0 0
0 20 40 60 80 100 0 20 40 60 80 100
Time (s) Time (s)

Figure 10.26: Comparison of FED and FIC predictions with expected values.

142
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
0.7 1
O2, CO2 and CO (FED_FIC_SMIX) Asphyxiants (FED_FIC_SMIX)
0.6
0.8
0.5

0.4 0.6
FED

FED
0.3 0.4
0.2
0.2
0.1 Expected Expected
FDS (FED) FDS (FED)
0 0
0 20 40 60 80 100 0 20 40 60 80 100
Time (s) Time (s)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
0.01 0.7
Irritants (FED_FIC_SMIX) All Gases (FED_FIC_SMIX)
0.6
0.008
0.5
0.006 0.4
FED

FED
0.004 0.3

0.2
0.002
Expected 0.1 Expected
FDS (FED) FDS (FED)
0 0
0 20 40 60 80 100 0 20 40 60 80 100
Time (s) Time (s)

-3 FDS-6.8.0-0-g886e009-release -3 FDS-6.8.0-0-g886e009-release
10 10
1 1
O2, CO2 and CO (FED_FIC_SMIX) Asphyxiants (FED_FIC_SMIX)

0.5 0.5
FIC

FIC

0 0

-0.5 -0.5
Expected Expected
FDS (FIC) FDS (FIC)
-1 -1
0 20 40 60 80 100 0 20 40 60 80 100
Time (s) Time (s)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
1 0.5
Irritants (FED_FIC_SMIX) All Gases (FED_FIC_SMIX)
0.8 0.4

0.6 0.3
FIC

FIC

0.4 0.2

0.2 0.1
Expected Expected
FDS (FIC) FDS (FIC)
0 0
0 20 40 60 80 100 0 20 40 60 80 100
Time (s) Time (s)

Figure 10.27: Comparison of FED and FIC predictions with expected values using SMIX.

143
 FDS-6.8.0-0-g886e009-release
0.4
CO and HCN (FED_CO_HCN)

0.3

FED
0.2

0.1
Expected
FDS (FED)
0
0 20 40 60 80 100
Time (s)

Figure 10.28: Comparison of predicted and expected FED resulting from the post-combustion yields of CO and HCN.

FDS-6.8.0-0-g886e009-release
3
Moving Occupant FED (FED_moving)
2.5

2
FED

1.5

0.5 Expected
FDS (FED)
0
0 0.1 0.2 0.3 0.4
Time (s)

Figure 10.29: Comparison of predicted and expected FED for a moving particle.

144
10.10 Smoke Detector Model (smoke_detector)
The calculations described in this section check the smoke detector algorithm that computes the smoke
obscuration in the sensing chamber of a detector. For this example, the initial soot mass fraction in a 1 m3
box is set as 100 mg/kg via the INIT line. A vent on the left side injects soot at a velocity of 0.25 m/s and a
mass fraction of 100 mg/kg, and the soot exhausts on the right side of the box via an open vent. By default,
if a species is defined as an aerosol by setting AEROSOL=.TRUE. on the SPEC line, then all of the aerosol
deposition mechanisms (gravitational, thermophoretic, and turbulent) will act upon that species. For this
example, all deposition mechanisms are disabled by using the follows inputs on the MISC line:

&MISC GRAVITATIONAL_DEPOSITION = .FALSE.

THERMOPHORETIC_DEPOSITION = .FALSE.
TURBULENT_DEPOSITION = .FALSE. /

The transient mass fraction of smoke in the sensing chamber of the detector Yc is given as [1]

dYc Ye (t) −Yc (t)

= (10.35)
dt L/u

where Ye is the mass fraction of smoke in the free stream (kg/kg), L is the characteristic length of the
detector geometry (m), and u is the free stream velocity (m/s). For a constant mass fraction of smoke in the
free stream Y0 , Eq. 10.35 has an analytical solution of the form
h  tu i
Yc (t) = Y0 1 − exp − (10.36)
L
The resulting smoke mass fraction in the sensing chamber is converted into an obscuration by
 
O[%/m] = 1 − e−κρYc l × 100 (10.37)

where κ is the specific extinction coefficient, ρ is the density of the external gases in the ceiling jet, and l is
the preferred unit of length (1 m in this case). Fig. 10.30 shows a comparison of the analytical and simulated
results for the time evolution of the smoke detector chamber obscuration.
FDS-6.8.0-0-g886e009-release
80
Chamber Obscuration (smoke_detector)

60
Obscuration (%/m)

40

20
Analytical
FDS Chamber Obscuration
0
0 20 40 60 80 100
Time (s)

Figure 10.30: Time evolution of smoke detector chamber obscuration for smoke_detector case

145
10.11 Aerosol Behavior
10.11.1 Gravitational Settling and Deposition of Aerosols
(aerosol_gravitational_deposition)
This verification test consists of two test cases. The second case, aerosol_gravitational_deposition_2,
reverses the z-component of gravity. The case consists of a box 10 cm on side with adiabatic, free-slip side
walls. The box is filled with two gas species each having a molecular weight of 28.8 g/mol, a viscosity
of 0.00002 kg/(m · s, a thermal conductivity of 0.025 W/(m · K, and specific heat of 1 kJ/(kg · K, and zero
diffusivity. One of the gas species is defined as an aerosol with a diameter of 10 µm, a solid phase den-
sity of 2000 kg/m3 , and a solid phase conductivity of 1 W/(m · K. The initial mass fraction of the aerosol
is 0.00001. STRATIFICATION, NOISE, and all aerosol behaviors except for GRAVITATIONAL_SETTLING
and GRAVITATIONAL_DEPOSITION are turned off. Since the box has a constant density over its height, a
uniform settling rate over time is expected.
-5 FDS-6.8.0-0-g886e009-release -8 FDS-6.8.0-0-g886e009-release
10 10
1.5 8
(aerosol_gravitational_deposition) (aerosol_gravitational_deposition)

Surface Density (kg/m²)

Mass Fraction (kg/kg)

1
Exact Gas 1 Exact Wall Top
Exact Gas 2 Exact Wall Bottom
FDS Gas 1
4 FDS Wall Top
FDS Gas 2 FDS Wall Bottom
0.5
2

0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
Time (s) Time (s)

-5 FDS-6.8.0-0-g886e009-release -8 FDS-6.8.0-0-g886e009-release
10 10
1.5 8
(aerosol_gravitational_deposition_2) (aerosol_gravitational_deposition_2)

6
Surface Density (kg/m²)
Mass Fraction (kg/kg)

1
Exact Gas 1 Exact Wall Top
Exact Gas 2 Exact Wall Bottom
FDS Gas 1
4 FDS Wall Top
FDS Gas 2 FDS Wall Bottom
0.5
2

0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
Time (s) Time (s)

Figure 10.31: Time evolution of soot mass fraction in the gas (left) and soot surface density on the wall (right)
for the aerosol_gravitational_deposition (Top) and aerosol_gravitational_deposition_2
(Bottom) cases.

146
10.11.2 Thermophoretic Settling and Deposition of Aerosols
(aerosol_thermophoretic_deposition)
This verification test consists of two test cases. The second case, aerosol_thermophoretic_deposition_2,
reverses the temperature gradient. The case consists of a box 1 cm on side with adiabatic, free-slip side
walls and a 100 K temperature gradient over the height of the box. The box is filled with two gas species
each having a molecular weight of 28.8 g/mol, a viscosity of 0.00002 kg/(m · s, a thermal conductivity of
0.025 W/(m · K, and specific heat of 1 kJ/(kg · K, and zero diffusivity. One of the gas species is defined
as an aerosol with a diameter of 1 µm, a solid phase density of 2000 kg/m3 , and a solid phase conductiv-
ity of 1 W/(m · K. The initial mass fraction of the aerosol is 0.00001. The gas temperature is initialized
to its steady-state temperature gradient. STRATIFICATION, NOISE, and all aerosol behaviors except for
THERMOPHORETIC_SETTLING and THERMOPHORETIC_DEPOSITION are turned off. Thermophoretic set-
tling rates are weakly dependent on the gas density. Since there is a temperature gradient, the settlings rates
are not uniform over the height of the box. Unlike the gravitational settling case, this means over long
enough time periods the overall settling rate is not linear in time; however, for a short time period a near
linear settling rate is expected and can be determined analytically
-5 FDS-6.8.0-0-g886e009-release -9 FDS-6.8.0-0-g886e009-release
10 10
1 6
(aerosol_thermophoretic_deposition) (aerosol_thermophoretic_deposition)
5
0.8

Surface Density (kg/m²)

Mass Fraction (kg/kg)

4
0.6
Exact Gas Exact Wall
FDS Gas
3 FDS Wall
0.4
2

0.2
1

0 0
0 0.5 1 1.5 2 0 0.5 1 1.5 2
Time (s) Time (s)

-5 FDS-6.8.0-0-g886e009-release -9 FDS-6.8.0-0-g886e009-release
10 10
1 6
(aerosol_thermophoretic_deposition_2) (aerosol_thermophoretic_deposition_2)
5
0.8
Surface Density (kg/m²)
Mass Fraction (kg/kg)

4
0.6
Exact Gas Exact Wall
FDS Gas
3 FDS Wall
0.4
2

0.2
1

0 0
0 0.5 1 1.5 2 0 0.5 1 1.5 2
Time (s) Time (s)

Figure 10.32: Time evolution of soot density in the gas (left) and soot surface density on the wall (right) for
the aerosol_thermophoretic_deposition (Top) and aerosol_thermophoretic_deposition_2
(Bottom) cases.

147
10.11.3 Turbulent Deposition of Aerosols (aerosol_turbulent_deposition)
This verification test consists of three tunnel like geometries 10 cm on side with adiabatic, free-slip walls.
One end of the tunnel is OPEN and the other end has a constant inlet velocity of 0.001 m/s, 0.1 m/s, or 1
m/s. The tunnels are filled with two gas species each having a molecular weight of 28.8 g/mol, a viscosity
of 0.00002 kg/(m · s), a thermal conductivity of 0.025 W/(m · K), and specific heat of 1 kJ/(kg · K), and zero
diffusivity. One of the gas species is defined as an aerosol with a diameter of 100 µm, a solid phase density
of 2000 kg/m3 , and a solid phase conductivity of 1 W/(m · K. The initial mass fraction of the aerosol is
0.00001. STRATIFICATION, NOISE, and all aerosol behaviors except for TURBULENT_DEPOSITION are
turned off. Turbulent deposition is computed using a correlation based open the wall friction velocity and
the wall dimensionless stopping distance. The correlation has three parts, and the selected velocities test
each part. Since the inlet condition is a constant velocity with a constant aerosol mass fraction, the first wall
cell after the inlet will see a uniform settling rate over time.
-14 FDS-6.8.0-0-g886e009-release -8 FDS-6.8.0-0-g886e009-release
10 10
2 4
(aerosol_turbulent_deposition) (aerosol_turbulent_deposition)

1.5 3
Surface Density (kg/m²)

1 Surface Density (kg/m²) 2

0.5 1
Exact Wall 1 Exact Wall 2
FDS Wall 1 FDS Wall 2
0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
Time (s) Time (s)

-7 FDS-6.8.0-0-g886e009-release
10
4
(aerosol_turbulent_deposition)

3
Surface Density (kg/m²)

1
Exact Wall 3
FDS Wall 3
0
0 0.2 0.4 0.6 0.8 1
Time (s)

Figure 10.33: Time evolution of soot surface density on the wall for the aerosol_turbulent_deposition.
Top left is the diffusion regime, top right is the diffusion-impaction regime, and bottom is the inertial regime.

10.11.4 Mass Conservation of Depositing Aerosols (propane_flame_deposition)

The calculations described in this section check the conservation of mass for the aerosol deposition al-
gorithms. The example involves a reaction that invokes the aerosol deposition model in FDS (i.e., soot
deposition). For this example, the fuel will be propane with a peak heat release rate of 100 kW and a 5 %

148
soot yield. For a 100 kW propane fire (HEAT_OF_COMBUSTION of 44,715 kJ/kg), the associated burning
rate is 2.236 g/s. A RAMP is used to specify a burning rate from a propane burner that ramps up linearly to
2.236 g/s from 0 s to 1 s, remains steady at 2.236 g/s from 1 s to 4 s, then linearly decreases to 0 g/s from 4 s
to 5 s. The total fuel mass released is 8.944 g of propane. Therefore, the resulting mass of soot should be
equal to 0.447 g, or 5 % of the total amount of fuel. For a case with no aerosol deposition, we would expect
the mass of soot in the gas phase to be equal to 0.447 g. For a case with aerosol deposition, we would expect
some fraction of soot to exist in the gas phase and the remainder of soot to be deposited on the wall. The
sum of soot in the gas phase and deposited soot should be 0.447 g.
Figure 10.34 shows the time evolution of soot in the gas phase, deposited soot, and total soot (sum of soot
in the gas phase and deposited soot) for five cases. The five cases are (1) all aerosol deposition mechanisms
active, (2) no aerosol deposition mechanisms active, (3) gravitational deposition only, (4) thermophoretic
deposition only, and (5) turbulent deposition only.
FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
0.6 0.012
Gas soot mass (propane_flame_deposition) Wall soot mass (propane_flame_deposition)
0.01
0.55
Expected Mass 0.008
FDS (All Deposition)
FDS (All Deposition)
Mass (g)

FDS (No Deposition)

Mass (g) FDS (No Deposition)
0.5 FDS (Gravitational Deposition)
0.006 FDS (Gravitational Deposition)
FDS (Thermophoretic Deposition)
FDS (Thermophoretic Deposition)
FDS (Turbulent Deposition)
FDS (Turbulent Deposition) 0.004
0.45
0.002

0.4 0
0 2 4 6 8 10 0 2 4 6 8 10
Time (s) Time (s)

FDS-6.8.0-0-g886e009-release
0.6
Total soot mass (propane_flame_deposition)

0.55
Expected Mass
FDS (All Deposition)
Mass (g)

FDS (No Deposition)

0.5 FDS (Gravitational Deposition)
FDS (Thermophoretic Deposition)
FDS (Turbulent Deposition)
0.45

0.4
0 2 4 6 8 10
Time (s)

Figure 10.34: Time evolution of soot mass components for propane_flame_deposition cases. (Top left) Mass
of soot in gas phase. (Top right) Mass of soot deposited on walls. (Bottom) Total mass of soot in gas phase and
deposited on walls.

10.11.5 Agglomeration of Aerosols (aerosol_agglomeration)

This verification case consists of two test cases.
The first case to be discussed consists of a box 10 cm on side with adiabatic, free-slip side walls. The box
is filled with two gas species each having a molecular weight of 28.8 g/mol, a viscosity of 0.00002 kg/(m · s),

149
a thermal conductivity of 0.025 W/(m · K), and specific heat of 1 kJ/(kg · K), and zero diffusivity. One of the
gas species is defined as an aerosol with two particle size bins with a minimum diameter of 1 µm and a maxi-
mum diameter of 10 µm, a solid phase density of 2000 kg/m3 , and a solid phase conductivity of 1 W/(m · K).
The initial mass fraction of the first size bin of the aerosol is 0.00001. STRATIFICATION, NOISE, and all
aerosol behaviors except for AGGLOMERATION are turned off. In this case the only mechanisms operating
are Brownian and gravitational agglomeration. The case is run for 60 s.

10 -5 FDS-6.8.0-0-g886e009-release 10 -9 FDS-6.8.0-0-g886e009-release
1.01 1.5
(aerosol_agglomeration) (aerosol_agglomeration)
Mass Fraction (kg/kg)

Mass Fraction (kg/kg)

1.005 1

Exact Bin 2
FDS Bin 2

1 0.5
Exact Total
Exact Bin 1
FDS Total
FDS Bin 1
0.995 0
0 10 20 30 40 50 60 0 10 20 30 40 50 60
Time (s) Time (s)

Figure 10.35: Time evolution of soot mass fraction in the gas for size bin 1 and both bins (left) and size bin 2 (right)
for the aerosol_agglomeration case.

In the second case three pairs of small (3 mm by 3 mm) vents are placed on opposite corners of the the
walls, floor, and ceiling of the box. Each pair is connected by an HVAC duct that blows air at 1 m/s at the
face of the vents. The average dissipation rate in the box is 0.0204 m2 /s3 . Using the average dissipation
rate, the agglomeration rate is computed and the resulting mass fractions are compared against the volume
averaged mass fractions computed by FDS.
-5 FDS-6.8.0-0-g886e009-release -9 FDS-6.8.0-0-g886e009-release
10 10
1.01 2
(aerosol_agglomeration_2) (aerosol_agglomeration_2)

1.5
Mass Fraction (kg/kg)

Mass Fraction (kg/kg)

1.005

Exact Bin 2
1 FDS Bin 2

1
Exact Total 0.5
Exact Bin 1
FDS Total
FDS Bin 1
0.995 0
0 10 20 30 40 50 60 0 10 20 30 40 50 60
Time (s) Time (s)

Figure 10.36: Time evolution of soot mass fraction in the gas for size bin 1 and both bins (left) and size bin 2 (right)
for the aerosol_agglomeration_2 case.

150
10.11.6 Scrubbing of Aerosols (aerosol_scrubbing)
This verification case tests the scrubbing of aerosols by droplet. The case consists of a box that is 0.4 m on a
side with 4 cells in each direction. The box is initially filled with pure water vapor with a soot mass fraction
of 0.0001. The soot has a mean diameter of 1 µm. The calculation is initialized with a 100 µm droplet in
every cell with an initial downward velocity of 0.1 m/s. Gravity and drag is disabled so the drops do not
change velocity. The drops are given an initial weight so that the total projected area is 25 % of the cell area.
Since every cell always has a single drop of the same diameter and velocity, the result is a constant decay
of the soot mass fraction with a decay constant of 0.0000666. After 5000 s, the case should predict a 28 %
drop in soot mass fraction.

10 -5 FDS-6.8.0-0-g886e009-release
12
(aerosol_scrubbing)
11

10
Mass Fraction

7 Exact
FDS
6
0 1000 2000 3000 4000 5000
Time (s)

Figure 10.37: Time evolution of soot mass fraction in the gas for the aerosol_scrubbing case.

10.11.7 Oxidation of Soot on Surfaces (soot_oxidation_wall)

This verification case checks the routine for oxidizing soot on a surface. The calculation fills a 1 m3 box
with 0.1 g of soot with a mean diameter of 100 microns; the large diameter was selected so the soot would
quickly settle to the floor. The case lets the soot settle for 10 s. This is followed by a one-second ramp
that heats the floor from 20 ◦ C to 1000 ◦ C. This results in rapid oxidation of the soot. Once all the soot is
oxidized, there should have been a heat release of 3.28 kJ and the production of 0.3 mg of CO2 and 8.3 µg
of H2 O.

151
 10 -4 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
1.2 4
Deposited Soot Mass (soot_oxidation_wall) Energy Release (soot_oxidation_wall)
1
3
0.8

Energy (kJ)
Mass (kg)

0.6 2

0.4
1
0.2 Ideal Ideal
FDS FDS
0 0
9 9.5 10 10.5 11 11.5 12 9 9.5 10 10.5 11 11.5 12
Time (s) Time (s)

10 -4 FDS-6.8.0-0-g886e009-release
4
CO Produced (soot_oxidation_wall)
2

3
Mass (kg)

1
Ideal
FDS
0
9 9.5 10 10.5 11 11.5 12
Time (s)

Figure 10.38: Time evolution the accumulated wall mass of soot (upper left), energy release due to soot combustion
(upper right), and mass of CO2 (lower left) and H2 O (lower right) in the gas for the soot_oxidation_wall case.

10.12 Condensation
10.12.1 Condensation and Evaporation in the Gas Phase
(condesnation_1 and condensation_2)
This pair of verification cases test the condensation and evaporation of water vapor. Both cases use 1 m3
box. Case condensation_1 has an initial temperature of 200 ◦ C and contains an initial mass fraction
of condensed water vapor of 0.01 kg/kg. At equilibirum all the condensed vapor will evaporate resulting
in a temperature drop of 27 ◦ C and a pressure drop of 5800 Pa. Case condensation_2 has an initial
temperature of -100 ◦ C and contains an initial mass fraction of water vapor of 0.01 kg/kg. At equilibirum
99.9 % of the vapor will condense resulting in a temperature rise of 36 ◦ C and a pressure rise of 21,000 Pa.

10.12.2 Condensation and Evaporation in the Solid Phase

(wall_cond)
This verification cases tests the condenensation and evaporation of water vapor onto and from solid surfaces.
The case uses 1 m3 box. The case has an initial temperature of 100 ◦ C and contains an initial mass fraction
of water vapor of 0.05 kg/kg. The walls are 1 mm thick with a density of 1000 kg/m3 , a specific heat of
1 kJ/(kg · K), and a conductivity of 100 W/(m · K). At equilibirum the wall and gas temperature will be equal
at 32 ◦ C and the pressure drop will be -20000 Pa. Note that NUCLEATION_SITES on MISC was set to zero

152
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
0 0
Analytical Analytical
Temperature (condensation_1) FDS
Pressure (condensation_1) FDS
-5 -1000

-10 -2000
Temperature (°C)

Pressure (kPa)
-15 -3000

-20 -4000

-25 -5000

-30 -6000
0 0.002 0.004 0.006 0.008 0.01 0 0.002 0.004 0.006 0.008 0.01
Time (s) Time (s)

FDS-6.8.0-0-g886e009-release 10 4 FDS-6.8.0-0-g886e009-release
40 2.5
Temperature (condensation_2)
2 Pressure (condensation_2)
30
Temperature (°C)

Pressure (kPa)
1.5
20
1

10
0.5
Analytical Analytical
FDS FDS
0 0
0 2 4 6 8 10 0 2 4 6 8 10
Time (s) Time (s)

Figure 10.39: Temperature change (left) and pressure change (right) due to the evaporation (top) and condensation
(bottom) of water vapor for the condensation_1 (top) and condensation_2 (bottom) cases.

to turn off condensation in the gas phase.

10.12.3 Radiation for Condensed Phase (condesnation_3)

This verification case tests that the radiation absorption of condensed vapor is equilvalent to the absorption
of an equivalent set of Lagrangian particles. The case contains two 1 m3 boxes filled with dry air at 0 ◦ C.
One box is initialized with an 0.001 mass fraction of condensed water vapor with a droplet size of 1 micron.
The other box is filled with one particle per cell with the same diameter and total mass of dropletsf. One
wall of each box is set to 1000 ◦ C. The radiative heat flux on the opposite face should be the same in both
cases.

153
 FDS-6.8.0-0-g886e009-release 10 4 FDS-6.8.0-0-g886e009-release
100 0
Temperature (wall_cond)
80 -0.5 Pressure (wall_cond)
Temperature (°C)

Pressure (Pa)
60 -1

40 -1.5

20 -2
Analytical Analytical
FDS FDS
0 -2.5
0 0.5 1 1.5 2 0 0.5 1 1.5 2
Time (s) 4 Time (s) 4
10 10

Figure 10.40: Temperature (left) and pressure change (right) for the wall_cond case.

FDS-6.8.0-0-g886e009-release
30

25
Radiative Heat Flux (condensation_3)
Radiative Flux (kW/m²)

20

15

10

5 Analytical
FDS
0
0 0.2 0.4 0.6 0.8 1
Time (s) 10 -5

Figure 10.41: Radiative heat flux for the condensation_3 case

154
Chapter 11

Heat Conduction

This chapter contains examples that test the one-dimensional heat conduction solver in FDS. A one-dimensional
heat conduction equation for the solid phase temperature Ts (x,t) is applied in the direction x pointing into
the solid (the point x = 0 represents the surface)

∂ Ts ∂ ∂ Ts
ρs cs = ks + q̇000
s (11.1)
∂t ∂x ∂x
In cylindrical and spherical coordinates, the heat conduction equation is written
   
∂ Ts 1 ∂ ∂ Ts 000 ∂ Ts 1 ∂ 2 ∂ Ts
ρ s cs = rks + q̇s ; ρs cs = 2 r ks + q̇000
s (11.2)
∂t r ∂r ∂r ∂t r ∂r ∂r

FDS offers the user these options, with the assumption that the obstruction is not actually rectilinear, but
rather cylindrical or spherical in shape. This option is useful in describing the behavior of small, complicated
“targets” like cables or heat detection devices.

11.1 Heat Conduction Through a Solid Slab (heat_conduction)

Analytical solutions of transient, one-dimensional heat conduction through a slab can be found in Refs. [90]
and [91]. Four cases are examined here. In each, a slab of thickness L = 0.1 m is exposed on one face to an
air temperature of Tg = 120 °C. The other face is insulated (adiabatic). The convective heat transfer from
the gas to the slab is q̇00c = h (Tg − Ts ), where h is constant, and Ts is the slab face temperature. No thermal
radiation is included.
Case k ρ c h Bi
(W/(m · K)) (kg/m3 ) (kJ/(kg · K)) (W/(m2 · K)) hL/k
A 0.1 100 1 100 100
B 0.1 100 1 10 10
C 1.0 1000 1 10 1
D 10.0 10000 1 10 0.1

155
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
150 150
Slab Temperature (heat_conduction_a) Slab Temperature (heat_conduction_b)

100 100
Temperature (°C)

Temperature (°C)
Analytical (Front)
Analytical (4 cm)
Analytical (Back)
FDS (Front)
FDS (4 cm) Analytical (Front)
50 FDS (Back) 50 Analytical (4 cm)
Analytical (Back)
FDS (Front)
FDS (4 cm)
FDS (Back)
0 0
0 5 10 15 20 25 30 0 50 100 150 200
Time (min) Time (min)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
150 150
Slab Temperature (heat_conduction_c) Slab Temperature (heat_conduction_d)

100 100
Temperature (°C)

Temperature (°C)
Analytical (Front)
Analytical (4 cm)
Analytical (Back)
FDS (Front)
Analytical (Front) FDS (4 cm)
50 Analytical (4 cm) 50 FDS (Back)
Analytical (Back)
FDS (Front)
FDS (4 cm)
FDS (Back)
0 0
0 100 200 300 400 500 600 0 200 400 600 800
Time (min) Time (min)

Figure 11.1: Comparison of heat conduction calculations with analytical solutions.

11.2 Temperature-Dependent Thermal Properties (heat_conduction_kc)

This example demonstrates the 1-D heat conduction in Cartesian, cylindrical, and spherical geometries with
temperature-dependent thermal properties. The Cartesian solution was computed using HEATING (ver-
sion 7.3), a multi-dimensional, finite-difference, general purpose heat transfer model [92]. The cylindrical
and spherical solutions were computed using a commercial finite-element solver, ABAQUS.
The sample of homogeneous material is initially at 0 °C and at t > 0 exposed to a gas at 700 °C. A
fixed heat transfer coefficient of 10 W/(m2 · K) is assumed. The density of the material is 10,000 kg/m3 . The
conductivity and specific heat are functions of temperature with the following values: k(0) = 0.10 W/(m · K),
k(200) = 0.20 W/(m · K), c(0) = 1.0 kJ/(kg · K), c(100) = 1.2 kJ/(kg · K), c(200) = 1.0 kJ/(kg · K). The
thickness (radius) of the sample is 0.01 m. In the Cartesian case, the back surface of the material is exposed
to a gas at 0 °C. In the figure below, the solid lines are FDS results and the circles are the HEATING results.
An example input with cylindrical geometry looks like:

&MATL ID='MAT_1'
EMISSIVITY = 0.0
CONDUCTIVITY_RAMP='K_RAMP'
SPECIFIC_HEAT_RAMP = 'C_RAMP'
DENSITY=10000. /

&RAMP ID = 'K_RAMP' T=0, F= 0.10 /

&RAMP ID = 'K_RAMP' T=100, F= 0.15 /

156
&RAMP ID = 'K_RAMP' T=200, F= 0.20 /
&RAMP ID = 'C_RAMP' T=0, F= 1.00 /
&RAMP ID = 'C_RAMP' T=100, F= 1.20 /
&RAMP ID = 'C_RAMP' T=200, F= 1.00 /

&SURF ID='SLAB'
STRETCH_FACTOR = 1.0
GEOMETRY = 'CYLINDRICAL'
MATL_ID='MAT_1'
THICKNESS=0.01 /

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
250 50
Surface Temperature (heat_conduction_kc) Back/Inner Temperature (heat_conduction_kc)
200 40
Temperature (°C)

Temperature (°C)
HEATING (cart_back)
150 30 ABAQUS (cyl_back)
ABAQUS (sph_back)
FDS (cart_back)
100 HEATING (cart_surf) 20 FDS (cyl_back)
ABAQUS (cyl_surf) FDS (sph_back)
ABAQUS (sph_surf)
50 FDS (cart_front) 10
FDS (cyl_front)
FDS (sph_front)
0 0
0 5 10 15 0 5 10 15
Time (min) Time (min)

Figure 11.2: Comparison of heat conduction calculations with a finite-element model.

11.3 Convective Cooling (convective_cooling)

This example tests the convective cooling algorithm. A 1 m thick solid slab that is initially at 1000 °C is
suddenly exposed to air at 0 °C. The back of the slab is insulated. Its density is 1000 kg/m3 , its specific heat
is 0.001 kJ/(kg·K), its conductivity is 1 W/(m·K), and its emissivity is zero, meaning there is no radiative
loss from the surface. The convective heat transfer coefficient is 1 W/(m2 ·K). The predicted and exact
temperature values at the back side of the slab are displayed in Fig. 11.3.
To test the second-order accuracy of the solid phase conduction algorithm, this case was run with in-
creasingly finer spatial and temporal resolution. The resulting convergence study is displayed in Fig. 11.4.
The plot shows the relative difference between the computed and exact solution at 1800 s as a function of
grid size.

157
 FDS-6.8.0-0-g886e009-release
1200
Inner Temperature (convective_cooling)
1000

800

Temperature (°C)
600

400

200 Exact (Temp)

FDS (inner temp)
0
0 5 10 15 20 25 30
Time (min)

Figure 11.3: Predicted vs. exact back side temperatures of a thick slab cooled by convection alone.

FDS-6.8.0-0-g886e009-release
10 -2
Convergence (convective_cooling)

10 -4
Relative Error

10 -6
O(?x)
O(?x²)
FDS (Relative Error)
10 -8
-4 -3
10 10
Grid Spacing (m)

Figure 11.4: Convergence study for the convective_cooling test case.

11.4 Simple Thermocouple Model (thermocouples)

This example tests the simple thermocouple model in FDS. It consists of a box whose walls and gas tem-
peratures are fixed at 500 °C. Inside the box are three thermocouples with bead diameters of 1 mm, 2 mm,
and 3 mm. Also included in the box are three “targets” – small solid objects whose surfaces are assumed
to be composed of small spheres of the same diameter as the thermocouples. Figure 11.5 compares the
temperature rise of the objects. The thermocouple model is not compared with an analytical solutions. This
is simply a comparison of the thermally thin thermocouple calculation with the thermally thick “target” cal-
culation. Small differences in temperature are due to slightly different flow conditions in different regions
of the box and numerical error due to node spacing and time step size.

158
 FDS-6.8.0-0-g886e009-release
600
Surface Temperature (thermocouples)
500

400

Temperature (°C)
300
Thick (Target_1)
200 Thick (Target_2)
Thick (Target_3)
Thin (TC_1)
100 Thin (TC_2)
Thin (TC_3)
0
0 20 40 60 80 100 120
Time (s)

Figure 11.5: Comparison of thermally thin and thick heat conduction into a small sphere.

11.5 Heat Conduction through Insulated Steel (insulated_steel_x)

In these examples, a 1 cm thick steel plate and pipe are coated on each side with 2 cm thick insula-
tion material. The conductivity, specific heat, and density of the steel are 50 W/(m · K), 0.5 kJ/(kg · K),
and 7500 kg/m3 , respectively. The corresponding values for the insulation material are 0.2 W/(m · K),
1 kJ/(kg · K), and 200 kg/m3 , respectively. The heat transfer coefficient is 10 W/(m2 · K). The outer sur-
face of the composite is exposed to air at 480 °C while the inner surface is exposed to air at 20 °C. The
assemblies are heated for 10 h and the steady-state temperature profiles are shown in Fig. 11.6. Neglecting
thermal radiation, the heat flux through the plate is given by [77]:
Thot − Tcold
q̇00 = Lins (11.3)
1
h + kins + Lksteel
steel
+ Lkins
ins
+ 1h

where h is the heat transfer coefficient, L is the layer thickness, and k is the thermal conductivity of the layer
material. The heat flux through the inner surface of the pipe, r1 , is given by:
Thot − Tcold
q̇00 = (11.4)
1
h + r1
kins ln r2
r1
r1
+ ksteel ln rr32 + krins1 ln rr34 + rr14 1h

where r1 is the inner radius of the inner layer of insulation, r2 is the inner radius of the steel, r3 is the outer
radius of the steel, and r4 is the outer radius of the outer layer of insulation. The heat flux through each
layer is the same, and the steady-state solution can be obtained by solving for the temperature drop across
each layer. The temperature gradient within each layer of the plate composite is linear. For the pipe, the
temperature profile for each layer is:
Ts,i−1 − Ts,i r
T (r) = Ts,i + ln i = 2, 3, 4 (11.5)
ln(ri−1 /ri ) ri

where Ts,i is the surface temperature corresponding to radial coordinates 1 through 4. Note that the position
of the symbols in Fig. 11.6 indicates the default finite difference nodes of the solid. Also note that the spatial
coordinate refers to the distance from the outer (hot) surface of insulation.

159
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
500 500
Temperature (insulated_steel_plate) Temperature (insulated_steel_pipe)
400 400
Temperature (°C)

Temperature (°C)
300 300

200 200

100 100
Analytical (Temp) Analytical (Temp)
FDS (Temperature) FDS (Temperature)
0 0
0 0.01 0.02 0.03 0.04 0.05 0 0.01 0.02 0.03 0.04 0.05
Depth (m) Depth (m)

Figure 11.6: Steady-state temperature profile within an insulated steel plate and pipe.

11.6 Heat Conduction across Mesh Boundaries (back_wall_test)

In this example, a hollow cube whose walls are either zero, one, or two cells thick contains a solid cube
centered inside with a fixed temperature of 500 ◦ C. The objective of the test is to ensure that the heat
conduction through the steel and across the mesh boundaries is computed properly when each wall has a
specified THICKNESS of 5 cm. In the left hand plot of Fig. 11.7, the net radiative heat flux to the interior
surface of the box matches the net radiative heat flux from the external surface in steady state. In the right
hand plot, the external surface temperature of two sides of the box are compared. One side is a zero cell thick
obstruction; the other side is a two cell thick obstruction. Since the sides all have a specified THICKNESS
of 0.5 cm, the fact that the obstruction is either zero, one, or two cells thick should not matter and the
temperatures should match exactly.
FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
20 500
Heat Flux (back_wall_test) Temperature (back_wall_test)
400
15
Heat Flux (kW/m²)

Temperature (°C)

300
10
200

5
100
Heat Flux In Zero cell thick obstruction
Heat Flux Out Two cell thick obstruction
0 0
0 10 20 30 40 50 60 0 10 20 30 40 50 60
Time (min) Time (min)

Figure 11.7: (Left) The heat flux into and out of a steel plate. (Right) The external temperature of a steel box whose
sides are either zero or two cells thick.

160
11.7 Heat Conduction through Immersed Boundaries (back_wall_test_2)
A 0.8 m by 0.8 m plate with surface temperature 1000 ◦ C and emissivity of 1 exposes a parallel plate sepa-
rated by 0.15 m. The exposed plate is 1 cm thick, with a conductivity of 25 W/(m·K), density 1000 kg/m3 ,
specific heat 0.1 kJ/(kg·K), emissivity 1, and a convective heat transfer coefficient of 0. Three different con-
figurations are considered: (1) the plates are all defined as traditional FDS OBSTructions, (2) the plates are
defined as immersed boundary (GEOM) obstructions, and (3) the same plates as in (2) are rotated 45◦ . Fig-
ure 11.8 displays the front and back surface temperatures for all three cases, compared to the exact solution
where the analytically determined heat flux, Eq. (9.1), is imposed directly on the front plate surface. The
error in all three cases is due mainly to the error in the radiation transport calculation on grids with 2.5 cm
cells.
FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
1000 1000
OBST Surface Temperature (back_wall_test_2) GEOM Surface Temperature (back_wall_test_2)
800 800
Temperature (°C)

Temperature (°C)
600 600

400 400

Exact Front Exact Front

200 Exact Back 200 Exact Back
FDS Front FDS Front
FDS Back FDS Back
0 0
0 5 10 15 20 0 5 10 15 20
Time (s) Time (s)

FDS-6.8.0-0-g886e009-release
1000
Rotated GEOM Surface Temp. (back_wall_test_2)
800
Temperature (°C)

600

400

Exact Front
200 Exact Back
FDS Front
FDS Back
0
0 5 10 15 20
Time (s)

Figure 11.8: Front and back surface temperatures of a plate defined using the OBST namelist parameters (top, left);
GEOM namelist parameters (top, right); GEOM parameters rotated 45◦ .

11.8 Specified Heat Flux Boundary Conditions (adiabatic_xxx_flux)

There are options in FDS to specify either a net (convective and radiative) or convective boundary con-
dition at a solid surface. In these two test cases, Heat_Transfer/adiabatic_con_flux.fds and
Heat_Transfer/adiabatic_net_flux.fds, a sealed cube of side length 2 m has a 0.4 m by 0.4 m
area on the floor with a specified convective and net heat flux of 100 kW/m2 . The case with the specified

161
convective heat flux boundary has the radiation turned off. The case with the net heat flux boundary has
both convection and radiation enabled so that the net flux is 100 kW/m2 . The flux is applied for 30 s and
then turned off. The pressure in the compartment is expected to rise according to the equation:

dp (γ − 1) Q̇
= (11.6)
dt V
where Q̇ = 16 kW, γ = 1.4, and V = 8 m3 . After 30 s, the pressure rise, p, is expected to be 24 kPa (left,
Fig. 11.9). The expected temperature rise, ∆T , is found by solving the energy conservation equation:

Q̇ ∆t = c p ρ V ∆T −V ∆p (11.7)

as shown in the right hand plots of Fig. 11.9.

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
40 120
Pressure (adiabatic_con_flux) Temperature (adiabatic_con_flux)
100
30
80

Temperature (°C)
Pressure (kPa)

20 60

40
10
Analytical (pres)
20 Analytical (temp)
FDS (pres) FDS (temp_gas)
0 0
0 10 20 30 40 50 60 0 10 20 30 40 50 60
Time (s) Time (s)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
40 120
Pressure (adiabatic_net_flux) Temperature (adiabatic_net_flux)
100
30
80
Temperature (°C)
Pressure (kPa)

20 60

40
10
Analytical (pres)
20 Analytical (temp)
FDS (pres) FDS (temp_gas)
0 0
0 10 20 30 40 50 60 0 10 20 30 40 50 60
Time (s) Time (s)

Figure 11.9: Pressure and average temperature of a small compartment with specified heat flux boundary conditions.

162
11.9 Solid Heat Transfer 3-D (Beta)
This section presents test cases for solid phase heat diffusion (conduction) in 3-D. This feature is not yet
connected to the 1-D pyrolysis heat transfer code discussed in the previous sections. Currently, the purpose
of the HT3D functionality is to couple FDS with structural fire calculations.

11.9.1 Energy Conservation in a 3-D Solid (ht3d_energy_conservation)

A 0.8 m by 0.8 m by 2 cm thick metal plate is exposed to a hot radiating surface for 10 s. Figure 11.10
displays the increased internal enthalpy versus the net heat flux integrated over the surface and integrated
over time.

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
100 97
Energy Balance (ht3d_energy_conservation) Energy Balance (ht3d_energy_conservation)
80 96
Enthalpy (kJ)

Enthalpy (kJ)
60 95

40 94

20 93
Enthalpy Enthalpy
Integrated Heat Flux Integrated Heat Flux
0 92
0 2 4 6 8 10 9.5 9.6 9.7 9.8 9.9 10
Time (s) Time (s)

Figure 11.10: Plots of the enthalpy of a metal plate versus the surface heat flux integrated in space and time. At right
is the same plot showing how the 3-D temperature field is updated over three time steps.

Figure 11.11 displays the energy balance for cases where the internal grid cell is 0.25 times the gas phase
grid cell size (left plot), and where the internal grid cells stretch as they would for a 1-D heat conduction
calculation (right). The grid stretching is not recommended for the 3-D solver.

163
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
100 100
Energy Balance (ht3d_energy_conservation_2) Energy Balance (ht3d_energy_conservation_3)
80 80
Enthalpy (kJ)

Enthalpy (kJ)
60 60

40 40

20 20
Enthalpy Enthalpy
Integrated Heat Flux Integrated Heat Flux
0 0
0 2 4 6 8 10 0 2 4 6 8 10
Time (s) Time (s)

Figure 11.11: Plots of the enthalpy of a metal plate versus the surface heat flux integrated in space and time. The left
plot is the case where the internal grid cells are 0.25 the size of the gas phase cells, and the right plot is the case where
the internal cells stretch as is done in the 1-D solver.

11.9.2 Heat Transfer to a Steel Beam (ht3d_beam_heating)

A steel I-beam 1.76 m long, with flanges that are 0.22 m wide and 0.03 m thick and a web that is 0.6 m
deep and 0.02 m thick, is heated within a sealed compartment that is filled with a gas that has a volume Vg =
0.8 m3 , initial temperature Tg,0 = 1073.15 K, pressure p0 = 101.325 kPa, molecular weight W = 0.028 kg/-
mol, specific heat cg = 1 kJ/(kg·K), and density ρg = p0W /(RTg,0 ) ≈ 0.318 kg/m3 . R ≈ 0.008314 kJ/(mol·K)
is the universal gas constant. The steel has a density ρs = 7500 kg/m3 , specific heat cs = 0.5 kJ/(kg·K), and
volume Vs = 0.041536 kg/m3 , and initial temperature Ts,0 = 293.15 K. Initially, the gas and steel have inter-
nal energies given by:

Ug = ρgVg cg Tg,0 − p0Vg ≈ 191.95 kJ ; Us = ρsVs cs Ts,0 ≈ 45661.04 kJ (11.8)

After several minutes, the equilibrium temperature is expected to be:

Ug +Us
T= ≈ 294.045 K ≡ 20.895 ◦ C (11.9)
ρsVs cs + ρgVg cg − nR

where n = p0Vg /(RTg,0 ) ≈ 9.0847 is the number of moles of gas.

The left hand plot in Fig. 11.12 displays the average surface temperature of the steel beam. Its computed
value is approximately 4 % lower than the exact value because the overlap of the web and flanges is not
accounted for with the current 3-D heating algorithm in FDS.
In a second case, a steel plate that is 1.76 m long, 0.58 m wide, and 0.02 m thick is coated with a 1 cm
thick layer of thermal insulation (density ρi = 20 kg/m3 , specific heat ci = 1.0 kJ/(kg·K)) and exposed to
the same thermal environment. The expected equilibrium temperature is 21.808 ◦ C. The right hand plot in
Fig. 11.12 compares the computed and exact temperatures.

164
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
24 24
Heat Transfer (ht3d_beam_heating_1) Heat Transfer (ht3d_beam_heating_2)

23 23
Beam Temperature (°C)

Beam Temperature (°C)

22 22

21 21
Exact Exact
FDS FDS
20 20
0 1 2 3 4 5 0 10 20 30 40 50 60
Time (s) Time (s)

Figure 11.12: (Left) Average surface temperature of a steel beam immersed in hot gas. (Right) Average temperature
of an insulated steel plate exposed to the same conditions.

11.9.3 Mass Conservation in a 3-D Solid (ht3d_mass_conservation)

Three abutting solids, one of which is thin, have a density of ρs = 400 kg/m3 , moisture mass fraction
Ym = 0.05, and total volume V = 0.0106 m3 . When exposed to hot plates of 1000 ◦ C, it is expected that
ρsYmV = 0.212 kg should be evaporated. Figure 11.13 displays the evaporated mass as a function of time.
FDS-6.8.0-0-g886e009-release
0.3
Mass Balance (ht3d_mass_conservation)
0.25

0.2
Mass (kg)

0.15

0.1

0.05 Exact (Mass)

FDS (WOOD MOISTURE)
0
0 20 40 60 80 100
Time (s)

Figure 11.13: Evaporated moisture mass from a collection of solids.

11.9.4 Continuous Heat Flux Boundary (ht3d_slab)

In this test case, we utilize the non-steady state conduction solution provided in Carslaw and Jaeger [91] for
a semi-infinite slab exposed to a constant ambient temperature, T∞ , with a constant convective heat transfer
coefficient, h, and no radiation (see also Drysdale [90], Eq. (2.25)). The initial temperature of the slab, T0 ,
is set to 1000 ◦C. The thermal properties of the slab, k = 1 W/(m · K), ρ = 1000 kg/m3 , and c = 0.001
kJ/(kg · K), are set to give a thermal diffusivity, α, of 0.001 m2 /s. The analytical solution for this problem
for the solid temperature, T (x,t), is given by
     √ 
T − T0 x xh αt x αt
= erfc √ − exp + 2
erfc √ + (11.10)
T∞ − T0 2 αt k (k/h) 2 αt k/h

165
where erfc = 1 − erf(x).
This is a 1-D problem in depth into the solid. In FDS, the problem is solved in 2-D, with the temperature
varying in the x coordinate direction; there is no variation in the vertical, z, direction. While the solution is
provided for a semi-infinite slab, FDS must use a finite thickness. The slab is taken to be 0.5 m in depth,
with x = 0 m marking the interface location between the fluid and the solid. The analytical solution for
T (x = 0.5,t) is used as the “back” boundary condition of the slab. The grid resolution is chosen to be 5
cm (10 cells covering 0.5 m). The stability constraint for the time step in the heat equation is therefore,
δt ≤ δ x2 /(2α) = 1.25 s. This is set as the initial time step in FDS, since no gas phase solution is computed.
The main purpose of this verification case is to check the accuracy of the continuous heat flux boundary
condition used in the HT3D method (see the FDS Tech Guide [48] for details). Figure 11.14 shows the
surface (wall) temperature, Tw , plotted as a function of time (the hot slab is cooling). Note that the exact
solution for the surface temperature is taken from Eq. (11.10) with x = 0 m. That is, Tw (t) ≡ T (0,t).
FDS-6.8.0-0-g886e009-release
1200
Heat Transfer Semi-infinite Slab (ht3d_slab)
1000
Temperature (°C)

800
Exact (Surface Temp)
FDS (Surface Temp)
600

400

200
0 500 1000 1500
Time (s)

Figure 11.14: Test of continuous heat flux boundary condition for 3-D heat transfer.

11.9.5 3-D Heat Diffusion in a Steel I-Beam (ht3d_ibeam)

The purpose of this test case is to compare FDS HT3D with a well-established commercial finite-element
model (ANSYS) for a reasonably practical problem. We consider a steel I-beam cross-section 0.4 m on
each side. The flanges are 6 cm thick and the web is 4 cm thick. The grid resolution for both FEM and
FDS models is ∆x = 1 cm. The thermal properties of the steel are taken to be constant: k = 45 W/(m · K),
ρ = 7850 kg/m3 , and c = 0.60 kJ/(kg · K). The boundary conditions are adiabatic except for a hot patch on
the front half of the bottom flange maintained at 800 ◦C. The initial temperature of the steel is 20 ◦C and the
case is run for 3600 s. Note that the FEM model is run with a time step comparable to the explicit stability
criterion (∆t ≈ 1.7 s) in order to yield time accurate results for comparison with FDS.
Figure 11.15 shows a comparison between the surface temperature contours from ANSYS (left) and
FDS (right). Below that, in Fig. 11.16, we show the time history of the surface temperatures for six locations
on the bottom flange (positions may be identified from the image on the left).

166
Figure 11.15: Three-dimensional heat diffusion in an I-beam, comparison between a commercial finite-element code
(ANSYS, left, courtesy Chao Zhang) and FDS (right). The beam boundary conditions are adiabatic except for a hot
patch maintained at 800 ◦C on the front right of the bottom flange. The initial temperature of the steel is 20 ◦C and the
calculations are run for 3600 s.

FDS-6.8.0-0-g886e009-release
1000
HT3D I-beam Surface Temperature (ht3d_ibeam)
FEM 1
800 FEM 2
FEM 3
FEM 4
Temperature (°C)

600 FEM 5
FEM 6
FDS 1
400 FDS 2
FDS 3
FDS 4
200 FDS 5
FDS 6

0
0 1000 2000 3000
Time (s)

Figure 11.16: (Left) Device locations corresponding to the legend entries in the plot to the right. (Right) Time history
of the surface temperature. Symbols represent the finite-element model (FEM) results and the lines represent the FDS
results.

167
11.9.6 3-D Heat Diffusion in Thin Steel Plates (ht3d_network)
This case tests the 3-D solid phase solver for a network of thin steel plates. Consider the image in Fig. 11.17
which shows a collection of connected steel plates that are 20 cm wide and 1 cm thick. Two hot panels heat
up the single plate to the left and the heat is conducted through the network of plates. Besides the left-most
plate, all other plates are perfectly insulated. The plot on the right simply demonstrates that the temperature
of the outer two of the four right-most extremities are the same, verifying that the distribution of heat is
symmetric given a symmetric arrangement of plates.

FDS-6.8.0-0-g886e009-release
50
HT3D Thin Obstructions (ht3d_network)
40

Temperature (°C)
30

20

10
T-1
T-4
0
0 2 4 6 8
Time (h)

Figure 11.17: (Left) A collection of thin steel plates are heated at the left and heat is conducted laterally along the
plates. (Right) The plot shows the temperature at the outer right extremities after 8 h.

11.9.7 3-D Heat Diffusion in a Sphere (ht3d_sphere)

Salah Benkorichi, Omega Fire Engineering Ltd.

A solid sphere of radius a = 0.1 m with internal heat generated at q̇000 = 200 kW/m3 has an initial
temperature of T0 = 20 ◦ C and its exterior surface temperature is held at this temperature. The 3-D heat
conduction equation can be written as

q̇000
   
∂T 1 ∂ 2∂T
=α 2 2 r + ; T (r, 0) = T0 ; T (a,t) = T0 (11.11)
∂t r ∂r ∂r k

where k is the thermal conductivity and α is the thermal diffusivity. The exact solution to Eq. (11.11) is
from [91] (Sec. 9.8, p. 243):

q̇000 2 q̇000 a3 ∞ (−1)n n2 π 2

 nπr   
2


T = T0 + a −r +2 3 ∑ sin exp −αt 2 (11.12)
6k kπ r n=1 n3 a a

The simulation is run for 180 s. Shown in Fig. 11.18 is a comparison of analytical and numerical radial
temperature profiles at several points in time for the high resolution case. Also shown is the L∞ error as a
function of cell size.

168
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
60 10 0
FDS
O( x)
O( x²)
50
Temperature (°C)

Analytical

error (°C)
FDS t =10 s
FDS t =20 s
40 FDS t =60 s
FDS t =120 s
FDS t =180 s

L
30
10 -1

20
0 0.02 0.04 0.06 0.08 0.1 3 4 5 6 7 8 9
Radial Distance (m) x (m) 10 -3

Figure 11.18: (Left) Comparison of analytical and numerical radial profiles of temperature at various times. (Right)
L∞ error in temperature at t = 180 s for three levels of grid refinement.

169
Chapter 12

Pyrolysis

This chapter includes tests of the routines that calculate the thermal decomposition of materials. Solid
surfaces can consist of multiple layers, and each layer can consist of multiple material components. Each
material component may undergo several competing reactions, and each of these reactions may produce
some other solid component (residue), gaseous fuel, and/or water vapor.

12.1 Conservation of Pyrolyzed Mass (surf_mass_conservation)

The calculations described in this section check the conservation of mass produced by the pyrolysis algo-
rithm. In the tests, four independent modeling options are varied:

1. The SURF line can be associated with either a solid surface as designated by a VENT line, or by solid
particles as described by a PART line.

2. The SURF geometry can be either CARTESIAN, CYLINDRICAL, or SPHERICAL.

3. The MATL can be either charring (non-zero NU_RESIDUE) or non-charring.

4. The pyrolysis product can be either the fuel gas defined by the mixture fraction model, or an additional
gas species, defined by a SPEC line.

In most cases, the wall thickness (or radius for cylindrical and spherical geometries) is 0.01 m. The material
density is 360 kg/m3 and the yield of gaseous products for the charring cases is 0.5, i.e., half of the original
mass. For Cartesian surfaces, the mass loss per unit area is 1.8 kg/m2 for charring, and 3.6 kg/m2 for
non-charring materials. For cylindrical surfaces, the volume per unit surface area is r/2, and thus the mass
loss per unit area is 360(1 − 0.5) × r/2=0.9 kg/m2 for charring, and 1.8 kg/m2 for non-charring materials.
For spherical surfaces, the volume per unit surface area is r/3, and thus the mass loss per unit area is
360(1 − 0.5) × r/3=0.6 kg/m2 for charring, and 1.2 kg/m2 for non-charring materials.

12.1.1 Pyrolysis at a Solid Surface

The analytical mass losses are calculated by multiplying the mass per unit area by the VENT area, which in
all cases is 1 m2 . The expected and computed results for charring material are compared in Fig. 12.1. The
expected and computed results for non-charring material are compared in Fig. 12.2.

171
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
4 4
surf mass vent char cart fuel surf mass vent char cart gas

3 3
Mass (kg)

Mass (kg)
2 2

1 Expected 1
Gaseous Expected
Solid Gaseous
Gaseous Fuel Solid
0 0
0 50 100 150 200 0 50 100 150 200
Time (s) Time (s)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
2
surf mass vent char cyl fuel 2 surf mass vent char cyl gas

1.5
1.5
Mass (kg)

Mass (kg)
1
1

0.5 Expected
Gaseous 0.5 Expected
Solid Gaseous
Gaseous Fuel Solid
0 0
0 20 40 60 80 100 0 20 40 60 80 100
Time (s) Time (s)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
1.5 1.5
surf mass vent char spher fuel surf mass vent char spher gas

1 1
Mass (kg)

Mass (kg)

0.5 0.5
Expected
Gaseous Expected
Solid Gaseous
Gaseous Fuel Solid
0 0
0 20 40 60 80 100 0 20 40 60 80
Time (s) Time (s)

Figure 12.1: Comparison of analytical and predicted mass change for charring surfaces that are introduced as VENTs.

12.1.2 Pyrolysis of Discrete Particles

For Lagrangian particles, the expected values of the mass are obtained by multiplying the material density by
the particle volume by the residue fraction. For Cartesian surfaces (plates), the particle area is two times the
product of the parameters LENGTH and WIDTH on the SURF line, both of which are given a value of 0.05 m.
As a result, the expected masses for plate particles are 360(1 − 0.5) × 2δ LW = 0.009 kg for charring, and
0.018 kg for non-charring materials. Note that the half-thickness, δ = 0.01 m, is specified on the SURF line
as THICKNESS. For cylindrical particles, the LENGTH is 0.1 m, and the radius, r (specified via THICKNESS),
is 0.01 m. The expected masses are thus 360(1 − 0.5) × πr2 L = 0.00565 kg for charring, and 0.0113 kg for

172
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
5 5
surf mass vent nonchar cart fuel surf mass vent nonchar cart gas
4 4

3 Expected 3
Mass (kg)

Mass (kg)
Expected
Gaseous
Gaseous
Solid
Solid
2 Gaseous Fuel 2

1 1

0 0
0 50 100 150 200 0 50 100 150 200
Time (s) Time (s)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release

2 surf mass vent nonchar cyl fuel 2 surf mass vent nonchar cyl gas

1.5 1.5
Expected
Mass (kg)

Mass (kg)
Expected
Gaseous
Gaseous
1 Solid 1 Solid
Gaseous Fuel

0.5 0.5

0 0
0 50 100 150 0 50 100 150
Time (s) Time (s)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
1.5 1.5
surf mass vent nonchar spher fuel surf mass vent nonchar spher gas

1 1
Expected
Mass (kg)

Mass (kg)

Expected
Gaseous
Gaseous
Solid
Solid
Gaseous Fuel
0.5 0.5

0 0
0 20 40 60 80 100 0 20 40 60 80 100
Time (s) Time (s)

Figure 12.2: Comparison of analytical and predicted mass change for non-charring surfaces that are introduced as
VENTs.

173
non-charring materials. For spherical particles, the expected mass is 360(1−0.5)×4πr3 /3 = 7.54×10−4 kg
for charring, and 1.51 × 10−3 kg for non-charring materials.

Charring Particles
The analytical and computed results for charring material are compared in Fig. 12.3. The analytical and
computed results for non-charring material are compared in Fig. 12.4.
FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
0.01 0.01
surf mass part char cart fuel surf mass part char cart gas
0.008 0.008

0.006 0.006
Mass (kg)

Mass (kg)
0.004 0.004

0.002 Expected 0.002 Expected

Gaseous Gaseous
Solid Solid
0 0
0 50 100 150 200 0 50 100 150 200
Time (s) Time (s)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
0.014 0.014
surf mass part char cyl fuel surf mass part char cyl gas
0.012 0.012

0.01 0.01
Mass (kg)

Mass (kg)

0.008 0.008

0.006 0.006

0.004 0.004
Expected Expected
0.002 Gaseous 0.002 Gaseous
Solid Solid
0 0
0 20 40 60 80 100 0 20 40 60 80 100
Time (s) Time (s)

10 -3 FDS-6.8.0-0-g886e009-release 10 -3 FDS-6.8.0-0-g886e009-release
2 2
surf mass part char spher fuel surf mass part char spher gas

1.5 1.5
Mass (kg)

Mass (kg)

1 1

0.5 0.5
Expected Expected
Gaseous Gaseous
Solid Solid
0 0
0 20 40 60 80 100 0 20 40 60 80 100
Time (s) Time (s)

Figure 12.3: Comparison of analytical and predicted mass change for charring particle surfaces.

174
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release

0.01 surf mass part nonchar cart fuel 0.01 surf mass part nonchar cart gas

0.008 0.008
Mass (kg)

Mass (kg)
0.006 Expected 0.006 Expected
Gaseous Gaseous
Solid Solid
0.004 0.004

0.002 0.002

0 0
0 20 40 60 80 100 0 20 40 60 80 100
Time (s) Time (s)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
0.015 0.015
surf mass part nonchar cyl fuel surf mass part nonchar cyl gas

0.01 0.01
Mass (kg)

Mass (kg)
0.005 0.005
Expected Expected
Gaseous Gaseous
Solid Solid
0 0
0 20 40 60 80 100 0 20 40 60 80 100
Time (s) Time (s)

-3 FDS-6.8.0-0-g886e009-release -3 FDS-6.8.0-0-g886e009-release
10 10
2.5 2.5
surf mass part nonchar spher fuel surf mass part nonchar spher gas
2 2

1.5 1.5
Mass (kg)

Mass (kg)

1 1

0.5 Expected 0.5 Expected

Gaseous Gaseous
Solid Solid
0 0
0 20 40 60 80 100 0 20 40 60 80 100
Time (s) Time (s)

Figure 12.4: Comparison of analytical and predicted mass change for non-charring particle surfaces.

175
Particles that Generate Multiple Gas Species
Figure 12.5 shows the results of three test cases involving 0.1 kg of flat, cylindrical, or spherical particles in
a unit cube that simultaneously generate fuel gas and water vapor, like wood or vegetation.
FDS-6.8.0-0-g886e009-release
0.12
Mass conservation (surf_mass_two_species_cart)
0.1

0.08 Expected Fuel Mass

Expected Water Vapor Mass
Mass (kg) Expected Solid Mass
0.06 FDS Fuel Mass
FDS Water Vapor Mass
0.04 FDS Solid Mass

0.02

0
0 200 400 600 800 1000
Time (s)

FDS-6.8.0-0-g886e009-release
0.12
Mass conservation (surf_mass_two_species_cyl)
0.1

0.08 Expected Fuel Mass

Expected Water Vapor Mass
Mass (kg)

Expected Solid Mass

0.06 FDS Fuel Mass
FDS Water Vapor Mass
0.04 FDS Solid Mass

0.02

0
0 200 400 600 800 1000
Time (s)

FDS-6.8.0-0-g886e009-release
0.12
Mass conservation (surf_mass_two_species_spher)
0.1

0.08 Expected Fuel Mass

Expected Water Vapor Mass
Mass (kg)

Expected Solid Mass

0.06 FDS Fuel Mass
FDS Water Vapor Mass
0.04 FDS Solid Mass

0.02

0
0 200 400 600 800 1000
Time (s)

Figure 12.5: Comparison of analytical and predicted mass change for particles that generate water and fuel gas.

176
12.1.3 Rate of Solid Particle Decomposition (part_baking_soda)
Paul Papas (Raytheon Technologies) and Changmin Cao (Collins Aerospace Ireland)

In this set of cases we present a simple analytical solution for the solid phase decomposition of a spher-
ical particle with a constant rate of reaction under isothermal conditions. Let m(t) denote the mass of a
spherical particle with radius r(t), changing with time, t. The initial mass is m0 = m(t = 0). The extent of
reaction is defined as
m0 − m
α≡ (12.1)
m0

First-order Reaction Model

The density of the material, ρ, is a constant. The volume is denoted V (t). The rate of change of α for a
first-order reaction is given by [93]
dα
= k (1 − α) (12.2)
dt
where the rate constant of the reaction is
k = Ae−Ea /(RT ) (12.3)
Here, A is the Arrhenius parameter (1/s) and Ea is the activation energy (J/mol); R is the gas constant (8.3145
J/mol/K) and T is temperature in K.
Integrating with constant density gives
dα
Z Z
=k dt ⇒ [ln(1 − α)]αα0 = −kt (12.4)
1−α

1−α m
= e−kt ; α0 = 0 ; α = 1− (12.5)
1 − α0 m0
m ρV
= e−kt ⇒ = e−kt ⇒ r3 = r03 e−kt (12.6)
m0 ρ0V0

Example
We take A = 3.4e11 1/s, Ea = 103000 J/mol, r0 = 2.5 µm, and compare FDS with the analytical solution
for T = [420, 450, 500] K. Results are presented in Fig. 12.6 below.

Contracting Volume Reaction Model

The rate of change of α for a contracting volume is given by [93]
dα
= k 3(1 − α)2/3 (12.7)
dt
For this model, it is helpful to first convert the extent of reaction back to the particle mass,
dm 1/3
= −k 3 m0 m2/3 (12.8)
dt
Integrating gives, Z Z
1/3
m−2/3 dm = −k 3 m0 dt (12.9)

177
 1/3
m1/3 = m0 (1 − k t) (12.10)
(ρV )1/3 = (ρ0V0 )1/3 (1 − k t) (12.11)
Physically, the density would remain constant and the volume would contract, leading to a linear de-
crease in the particle radius,
r = r0 (1 − k t) (12.12)
However, the FDS reaction equation is not set up to mimic the contracting volume model. Instead, for
the same kinetics parameters, we can achieve the correct mass decay of the particle by holding the volume
constant (ALLOW_SHRINKING=F) and adjusting the preexponential factor to account for the initial mass. By
holding the volume constant, Eq. (12.8) may be written in terms of solid density for material α as follows,
dρs,α 2/3
= −k 3 ρs,α (0)1/3 ρs,α (12.13)
dt | {z }
adjustment factor

Therefore, to match the initial decay of the contracting volume model, we must multiply A by the factor
3 ρs,α (0)1/3 and change the order of the reaction to 2/3.

Example
We consider the contracting volume model applied to a spherical baking soda (NaHCO3 ) particle. We
take A = 3.4e11 × 3(2200)1/3 = 133e11 1/s (note: the initial density of the particle is 2200 kg/m3 ), Ea =
103000 J/mol, r0 = 2.5 µm, and compare FDS with the analytical solution for the particle diameter with the
rate constant evaluated at T = [420, 450, 500] K. We set ALLOW_SHRINKING=F and we report the particle
diameter as the initial diameter times the cube root of the density ratio, d(t) = d0 (ρs [t]/ρs [0])1/3 . The results
are shown in Fig. 12.6 along with a comparison to the rate for the first-order model with the same Arrhenius
parameters. It is worth noting that the contracting volume model leads to a significantly higher decay rate
than the first-order reaction for the same set of parameters.

178
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
6 6
Baking Soda 420 K (part_baking_soda) Baking Soda 450 K (part_baking_soda)
5
5
m)

m)
4
Diameter (

Diameter (
4 3

2
3 Exact first-order Exact first-order
Exact spherical Exact spherical
FDS N=1 A=3.4e11
1 FDS N=1 A=3.4e11
FDS N=2/3 A=133e11 SHRINKING=F FDS N=2/3 A=133e11 SHRINKING=F
2 0
0 2 4 6 8 10 0 2 4 6 8 10
Time (s) Time (s)

FDS-6.8.0-0-g886e009-release
6
Baking Soda 500 K (part_baking_soda)
5
m)

4
Exact first-order
Exact spherical
Diameter (

3 FDS N=1 A=3.4e11

FDS N=2/3 A=133e11 SHRINKING=F
2

0
0 2 4 6 8 10
Time (s)

Figure 12.6: Comparison of analytical and predicted particle diameter for first-order and contracting volume solid
phase reaction under isothermal conditions.

12.2 Conservation of Pyrolyzed Energy (matl_e_cons)

Nine simple test cases are created to verify the process of ensuring the MATL reactions conserve energy.
Each case is a cube 1 m on a side with 4 grid cells in each dimension. One or more materials and gas species
are defined. All materials are defined with a density of 1 kg/m3 . All species are defined with a molecular
weight of 1000 g/mol. This is done for convenience as it means that the heat of formation in kJ/mol is the
same numerical values as the enthalpy in kJ/kg. For each material that is defined, a single wall cell VENT is
defined with a SURF such that the wall cell contains 1 kg of material. The WALL ENTHALPY is output for
each wall cell. The cases are explained below with results shown in Figs. 12.7 and 12.8.
In case 1, there is a single species is given a heat of formation of 0 kJ/mol. A single material is defined
with a specific heat of 0.5 kJ/kg/K and a reaction that produces the gas species with a heat of reaction of
1000 kJ/kg at a reference temperature of 300 ◦ C with a pyrolysis range of 40 K. This gives:

HMATL1 (300 ◦ C) + HMATL1,adj + 1000 kJ/kg = HSPEC1 (300 ◦ C) (12.14)

286.575 kJ/kg + HMATL1,adj + 1000 kJ/kg = 275 kJ/kg (12.15)
HMATL1,adj = −1011.575 kJ/kg (12.16)

At ambient temperature this makes the expected enthalpy of the wall cell 298.15 K × 0.5 kJ/kg/K-
1011.575 kJ/kg=-865 kJ/kg.

179
 Case 2 repeats case 1 with a species heat of formation of 1000 kJ/mol. This makes the expected ambient
enthalpy -1865 kJ/kg.
Case 3 combines case 1 and case 2 into a single input. The case 1 material is changed from using a
reference temperature and pyrolysis range to using the resulting A (7.68 × 1031 1/s) and E (371000 J/mol)
which was taken from the case 1 output file. This case tests the virtual TGA (the case 1 material will only
have the same adjustment if there is the same reference temperature) and that multiple sets of reactions are
correctly handled.
Case 4 takes case 1 and changes the reaction to produce equal amounts of the gas species and a second
material. The second material has a specific heat of 1.5 kJ/kg/K and a reaction that produces the gas species
with a heat of reaction of 500 kJ/kg at a reference temperature of 400 ◦ C. This case has two total reactions
that include two total materials. The solution can be found by hand by first finding the adjustment for the
second material (following the example for case 1) and then using that adjustment to find the value for the
first material.
Case 5 takes case 4 and adds a second reaction for the first material. This reaction creates a second copy
of the first material with a heat of reaction of 500 kJ/kg at a reference temperature of 280 ◦ C. This can be
thought of as a phase change. The second copy also produces equal amounts of the gas species and the other
material expect with a heat of reaction of 500 kJ/kg. This is now an overdetermined system as there are
three materials and four reactions; however, note that the parallel paths to creating the gas species have the
same total enthalpy change. The first material directly pyrolyzing has a heat of reaction of 1000 kJ/kg, and
the path of the phase change also has a total enthalpy change of 1000 kJ/kg split between the phase change
and the pyrolysis. Therefore, even though the system is overdetermined, an exact solution is expected. This
is seen in the results where the first and second material enthalpies in case 5 match those from case 4.
Case 6 takes case 5 and increases the heat of reaction post-phase change to 1000 kJ/kg. Now an exact
solution is no longer expected. It can be seen that the second material has the same enthalpy but the first
material and its phase-change have different ambient enthalpies. The post-phase change material enthalpy
has changed by 333 kJ/kg but the reaction enthalpy changed by 500 kJ/kg so energy is not completely
conserved. As a least squares solution, it is conserved as best as it can be.
Case 7 takes case 4 and removes the reaction from the second material. This is now an underdetermined
system where there are two materials and a single reaction. Following the process for case 1, one gets that

HMATL1,adj − 0.5HMATL2,adj = −719.2125 kJ/kg (12.17)

The minimum matrix solution is HMATL1,adj =-575.37 kJ/kg and HMATL2,adj =287.68 kJ/kg. It can be
shown that this is a minimal solution by taking the norm of the solution vector. This can be expressed in
terms of only HMATL1,adj as
2
HMATL1,adj + (2(HMATL1,adj + 791.2125 kJ/kg))2 (12.18)

The minimum is where the derivative is zero or

10HMATL1,adj + 8 × 791.2125 kJ/kg = −575.37 kJ/kg. (12.19)

Case 8 takes case 1 and deletes the gas production. In this case FDS will assume that a virtual material
is produced with the same specific heat as the original material. This results in a least squares solution where
the reaction enthalpy is split over the original material and the virgin material making the expected ambient
temperature enthalpy of the wall cell 298.15 K × 0.5 kJ/kg/K-500 kJ/kg=-353.425 kJ/kg.
Case 9 takes case 1 and produces a particle of a second material with the same specific heat as the first
material. This is equivalent to case 8 and the same ambient enthalpy is expected.

180
 1000 1000
Exact Exact
(matl_e_cons_1) FDS
(matl_e_cons_2) FDS
500 500

0 0
Enthalpy (kJ/kg)

Enthalpy (kJ/kg)
-500 -500

-1000 -1000

-1500 -1500

-2000 -2000
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
Time (s) Time (s)

1000 1000
Exact 1 Exact 1
(matl_e_cons_3) Exact 2
(matl_e_cons_4) Exact 2
500 FDS 1
500 FDS 1
FDS 2 FDS 2
0 0
Enthalpy (kJ/kg)

Enthalpy (kJ/kg)
-500 -500

-1000 -1000

-1500 -1500

-2000 -2000
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
Time (s) Time (s)

1000 1000
Exact 1 Exact 1
(matl_e_cons_5) Exact 1 PC
(matl_e_cons_6) Exact 1 PC
500 Exact 2
500 Exact 2
FDS 1 FDS 1
0 FDS 1 PC 0 FDS 1 PC
Enthalpy (kJ/kg)

Enthalpy (kJ/kg)

FDS 2 FDS 2

-500 -500

-1000 -1000

-1500 -1500

-2000 -2000
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
Time (s) Time (s)

Figure 12.7: Comparison of analytical and predicted wall enthalpy for the matl_e_cons test cases 1-6.

181
 1000 1000
Exact
(matl_e_cons_7) (matl_e_cons_8) FDS
500 500

0 0
Enthalpy (kJ/kg)

Enthalpy (kJ/kg)
-500 -500

-1000 -1000
Exact 1
Exact 2
-1500 FDS 1
-1500
FDS 2
-2000 -2000
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
Time (s) Time (s)

1000
Exact
(matl_e_cons_9) FDS
500

0
Enthalpy (kJ/kg)

-500

-1000

-1500

-2000
0 0.2 0.4 0.6 0.8 1
Time (s)

Figure 12.8: Comparison of analytical and predicted wall enthalpy for the matl_e_cons test cases 7-9.

182
12.3 Burning Rate of a Small Object (cell_burn_away)
In FDS, it is assumed that heat conduction and pyrolysis within solids are functions of depth only. This can
be problematic in cases where entire solid obstructions burn away. In this example, a single grid cell, 4 cm
on a side, is made up of a material with density ρs = 50 kg/m3 and constant reaction rate, rs = 0.05 s−1 . The
burning rate of the single grid cell, ṁ, is given by:

ṁ = δ ρs rs A (12.20)

where δ is the thickness of the surface layer and A is the area of the six faces of the single grid cell. If we
set the layer thickness so that δ A = V , where V is the volume of the cell, and note that the mass of the cell
is m = δ A ρs , Eq. (12.20) becomes:
ṁ = m rs = m0 rs e−rst (12.21)
Figure 12.9 displays the computed and analytical burning rate. Note that the burning rate is proportional to
the mass and the mass decays exponentially to zero in this case because we have specified the THICKNESS
to be V /A and set the BULK_DENSITY to be 50 kg/m3 on the OBST line defining the single solid grid cell.
-4 FDS-6.8.0-0-g886e009-release
10
2
Burning Rate (cell_burn_away)

1.5
Burning Rate (kg/s)

Ideal
1 FDS

0.5

0
0 20 40 60 80 100
Time (s)

Figure 12.9: Burning rate of a single grid cell with fixed reaction rate.

12.4 Evaporation of a Liquid Surface

12.4.1 Mass Conservation (surf_mass_vent_liquid)
To test the liquid fuel evaporation model, a 1 cm deep pool of heptane is exposed to a 50 kW/m2 heat flux
and the evolution of fuel vapor and surface density are tracked. Results are shown in Fig. 12.10. Note that in
the “nonconforming” case, the specified pool dimensions do not align with the specified mesh, but the same
amount of fuel vapor is expected.

183
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
10 10
surf_mass_vent_liquid_fuel surf_mass_vent_liquid_fuel_nonconforming
8 8

6 6
Mass (kg)

Mass (kg)
Expected Expected
Liquid Mass Liquid Mass
Gaseous Fuel Mass Gaseous Fuel Mass
4 4

2 2

0 0
0 50 100 150 200 0 50 100 150 200
Time (s) Time (s)

Figure 12.10: Comparison of analytical and predicted mass change for a liquid surface.

12.4.2 Evaporation Rate (water_pool)

In this example, air with an RH = 25 % relative humidity and temperature Ta = 305 K flows with velocity
U = 0.15 m/s over an L = 1.2 m long, W = 1 m wide pool of water with a surface temperature of Ts = 293 K.
The evaporation rate of water is given by the following relation:
Ysv −Yg
ṁ = LW hm ρfilm ln (1 + B) ≈ 1.42 × 10−5 kg/s ; B= ≈ 0.00925 (12.22)
1 −Ysv
hm is the mass transfer coefficient, ρfilm is the density within a thin surface layer, B is the Spalding mass
transfer number, Ysv is the “surface vapor” mass fraction of water vapor at the liquid surface, and Yg is the
mass fraction of water vapor in the ambient air. The composition of the surface layer is obtained from the
Clausius-Clapeyron equation for the equilibrium vapor pressure. The volume and mass fraction of water
vapor at the surface is:
  
hvWH2 O 1 1 XsvWH2 O
Xsv = exp − − ≈ 0.0278 ; Ysv = ≈ 0.0175 (12.23)
R Ts Tb XsvWH2 O + (1 − Xsv )Wair
The volume and mass fraction of water vapor in the ambient air is:
  
RH hvWH2 O 1 1 XgWH2 O
Xg = exp − − ≈ 0.0134 ; Yg = ≈ 0.0084 (12.24)
100 R Ta Tb XgWH2 O + (1 − Xg )Wair
Here hv = 2260 kJ/kg is the heat of vaporization of water, WH2 O = 18 kg/kmol is its molecular weight, and
Tb = 373 K is its boiling temperature. R = 8.3145 kJ/kmol/K is the gas constant and Wair = 29 kg/kmol is
the molecular weight of air. The mass transfer coefficient is given by
Sh D
hm = ≈ 1.06 × 10−3 m/s (12.25)
L
Sh is the Sherwood number and D = 2.10 × 10−5 m2 /s is the diffusivity of water vapor into air at T = 293 K.
An empirical correlation for the Sherwood number is
1 4 ρaUL
Sh = 0.037 Sc 3 Re 5 ≈ 60.8 ; Sc = 0.71 ; Re = ≈ 12062 (12.26)
µ
where µ = 1.8 × 10−5 kg/m/s is the viscosity of air at T = 293 K. The gas density at the liquid surface is
given by
p0Wair
ρfilm = ≈ 1.21 kg/m3 (12.27)
1000R Ts
184
where p0 = 101325 Pa.
Note that ADJUST_H=F is set in the input file to make the calculation of the analytical result easier.

10 -5 FDS-6.8.0-0-g886e009-release
2
Evaporation Rate (water_pool)

1.5

Mass Loss Rate (kg/s)

1

0.5
Theory (mdot)
FDS (mdot)
0
0 10 20 30 40 50 60
Time (min)

Figure 12.11: Evaporation rate of a pool of water.

Figure 12.11 compares the FDS predicted versus the theoretical evaporation rate. The prediction and
theory are not exactly the same because the nominal parameters used above do not account for slight changes
in surface temperature and water vapor concentration above the pool. Nevertheless, the two should be
relatively close.

12.5 Change in Surface Emissivity (emissivity)

For thermally thick materials, the surface emissivity is computed as a mass-weighted sum of the individual
values of the emissivity in the first condensed phase grid cell. In this verification test, the initial material,
having emissivity of 1.0, is converted to another material, having emissivity of 0.0, at a constant rate of
0.1 s−1 . As a result, the surface emissivity should change linearly from 1.0 to 0.0 in 10 s (Fig. 12.12).
FDS-6.8.0-0-g886e009-release
1
Surface Emissivity (emissivity)
0.8

0.6
Emissivity

0.4

0.2
Analytical (Emissivity)
FDS (EMISSIVITY)
0
0 2 4 6 8 10
Time (s)

Figure 12.12: Testing the emissivity of solid materials.

185
12.6 Shrinking and swelling materials (shrink_swell)
A single layer of a multi-component material contains reacting material (1) and an inert matrix (2). If the
product material (3) of the reaction is of higher density than the reactant (ρ3 > ρ1 ), the layer thickness ∆
should shrink. If the product density is lower than the reactant’s density (ρ3 < ρ1 ) the layer should swell
(increase) in thickness. If the inert material has a static matrix structure (ALLOW_SHRINKING=.FALSE. or
ALLOW_SWELLING=.FALSE.) the layer thickness should not change. The parameters of the six test cases
are shown in the following table. Figure 12.13 shows the surface densities, which obviously should not
change, and the surface thicknesses over the time of the simulation.

Table 12.1: Parameters used to test the shrinking and swelling of materials.

Case Reactant Inert Product Layer

ρ1 Y1 ρ2 Y2 Allow ρ3 ρs (0) ρs (end) ∆(0) ∆(end)
(kg/m3 ) (kg/m3 ) shrink/swell (kg/m3 ) (kg/m3 ) (kg/m3 ) (m) (m)
Shrink 1 500 1 0 0 T 1000 500 1000 0.001 0.0005
Shrink 2 500 0.9 500 0.1 T 1125 500 1000 0.001 0.0005
Shrink 3 500 0.9 500 0.1 F 1125 500 500 0.001 0.001
Swell 1 1000 1 90 0 T 500 1000 500 0.001 0.002
Swell 2 1000 0.9 1000 0.1 T 450 1000 500 0.001 0.002
Swell 3 1000 0.9 1000 0.1 F 450 1000 1000 0.001 0.001

186
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
1 2
Shrink Mass Conservation (shrink_swell) Swell Mass Conservation (shrink_swell)
0.8
1.5
Surface density (kg/m²)

Surface density (kg/m²)

0.6
1
0.4

Expected 0.5 Expected

0.2 Shrink 1 Swell 1
Shrink 2 Swell 2
Shrink 3 Swell 3
0 0
0 5 10 15 0 5 10 15
Time (s) Time (s)

10 -3 FDS-6.8.0-0-g886e009-release 10 -3 FDS-6.8.0-0-g886e009-release
1.5 2.5
Shrinking Thickness (shrink_swell) Swelling Thickness (shrink_swell)
2

1
Thickness (m)

Thickness (m)
1.5

Expected 1 1 Expected 1
0.5 Expected 2 Expected 2
Expected 3 Expected 3
Shrink 1 0.5 Swell 1
Shrink 2 Swell 2
Shrink 3 Swell 3
0 0
0 5 10 15 0 5 10 15
Time (s) Time (s)

Figure 12.13: Testing the shrinking and swelling surfaces.

12.7 Enthalpy of Solid Materials (enthalpy)

Consider a thin plate of conductive material that is exposed on one side to an elevated temperature heat
source and exposed on the other to an ambient temperature void. In the thermally thin limit, the temperature
of the slab is governed by the following equation

dTs q̇00front + q̇00back

= (12.28)
dt cs ρs δ

In this example, the initial exposure to the front side of the slab is 3 kW/m2 . The original Material A
undergoes a reaction to form Material B. The reaction rate is constant, 6 kg/(m3 ·s), which in this case means
that Material A, with an initial density of 30 kg/m3 , disappears in exactly 5 s. This is achieved by setting ns
and E to 0 and A to 6 kg/(m3 ·s) in the reaction rate term:
 
ns E
r = ρs,A A exp − (12.29)
RTs

The density and conductivity of both materials are 30 kg/m3 and 10 W/(m · K), respectively. The emissivity
of the front and back is 1. The specific heat of Material A changes from 1.0 kJ/(kg · K) to 0.1 kJ/(kg · K)
above 80 °C, while the specific heat of material B is constant at 1.0 kJ/(kg · K). The slab is 1 mm thick. Note
that the “analytical” solution is actually a simple numerical integration of the equations above with a small

187
 FDS-6.8.0-0-g886e009-release
200
Surface Temperature (enthalpy)

150

Temperature (°C)
100

50
Analytical (T_slab)
FDS (T_slab)
0
0 1 2 3 4
Time (s)

Figure 12.14: Testing the enthalpy of solid materials.

time step to ensure accuracy. This example (Fig. 12.14) tests a number of features, including the reaction
rate, mass weighted specific heats, and radiation boundary conditions. Note that the convective heat transfer
has been turned off, and the correct steady-state temperature is calculated by FDS.

12.8 A Simple Two-Step Pyrolysis Example (two_step_solid_reaction)

Before considering actual experimental measurements, it is necessary to check the accuracy of the ordi-
nary differential equation solver within FDS. Consider a simplified set of ordinary differential equations
describing the mass fraction of three components of a solid material undergoing thermal degradation:
dYa
= −KabYa
dt
dYb
= KabYa − KbcYb (12.30)
dt
dYc
= KbcYa
dt
where the mass fraction of component a is 1 initially. The analytical solution is:

Ya (t) = exp(−Kabt)
Kab
Yb (t) = exp(−Kabt) − exp(−Kbct) (12.31)
Kbc − Kab
Yc (t) = [Kab (1 − exp(−Kbct)) + Kbc ∗ (exp(−Kabt) − 1)] /(Kab − Kbc ) (12.32)

The analytical and numerical solution for the parameters Kab = 0.389 and Kbc = 0.262 are shown in Fig. 12.15:

188
 FDS-6.8.0-0-g886e009-release
1.2
Solid Density (two_step_solid_reaction)
1

0.8

Mass Fraction
0.6
Analytical (Ya)
0.4 Analytical (Yb)
Analytical (Yc)
FDS (Ya)
0.2 FDS (Yb)
FDS (Yc)
0
0 5 10 15 20 25 30
Time (s)

Figure 12.15: Comparison of a two-step solid pyrolysis calculation with an analytical solution.

12.9 Interpreting Bench-Scale Measurements

This section describes a method of deriving and applying the values of the kinetic parameters for the ther-
mal decomposition of a solid, following the methodology described by Lyon [94]. This is the basic the-
ory that underlies measurement methods like Thermo-Gravimetric Analysis (TGA) and Micro-Combustion
Calorimetry (MCC).

12.9.1 General Theory

Consider a small sample of solid material that is heated at a relatively slow, constant rate. Assume that the
solid consists of N material components, with each component mass fraction denoted by Yα (t). As the solid
is heated, each component undergoes a reaction to form undetermined gases or a single solid residue whose
mass fraction is denoted Yr (t) and whose yield is denoted by νr . The governing equations for the component
mass fractions are:
 
dYα Eα
= −Aα Yα exp − ; Yα (0) = Yα,0 ; α = 1, N (12.33)
dt RT
dYr dYα
= −νr ∑ ; Yr (0) = 0 (12.34)
dt α dt

In standard test apparatus, the temperature of the sample is increased linearly in time, dT /dt = β . Because
the test results are usually expressed as a function of temperature rather than time, it is convenient to rewrite
Eq. (12.33) as:  
dYα Aα Eα
= − Yα exp − ; Y (T = T0 ) = Yα,0 (12.35)
dT β RT
The decomposition rate for each material component, −dYα /dt, peaks at a temperature denoted by Tα,p with
a value denoted by rα,p . At this temperature, the second derivative of Yα is zero:

d2Yα
   
Aα dYα Eα Aα Eα Eα
= − exp − − Yα exp −
dT 2 β dT RTα,p β 2
RTα,p RTα,p
"   #
dYα Aα Eα Eα
= − exp − − 2
=0 (12.36)
dT β RTα,p RTα,p

189
Equation (12.35) can be integrated from Yα,0 to Yα,p (the value of Yα at the peak), and T0 to Tα,p :
2
dY 0 Aα RTα,p
Z Yα,p   Z Tα,p    
α Aα Eα 0 Eα
=− exp − 0 dT ≈ − exp − (12.37)
Yα,0 Yα0 β T0 RT β (Eα + 2RTα,p ) RTα,p

Using Eq. (12.36) to eliminate Aα yields:

 
Yα,p Eα
ln =− ≈ −1 (Eα  2RTα,p ) (12.38)
Yα,0 Eα + 2RTα,p

or more simply, Yα,p ≈ Yα,0 /e. Now, the activation energy can be evaluated using Eqs. (12.33) and (12.36):
  2
RTα,p 2
RTα,p
2 Aα Eα rα,p e rα,p
Eα = RTα,p exp − = ≈ (12.39)
β RTα,p β Yα,p β Yα,0
Then Aα can be evaluated directly from Eq. (12.33):
   
rα,p Eα e rα,p Eα
Aα = exp ≈ exp (12.40)
Yα,p RTα,p Yα,0 RTα,p
Note that the formulae for Aα and Eα can be evaluated with parameters that are obtained directly by inspec-
tion of the plot of mass loss rate versus temperature. For each peak, the values of Tα,p are easy to obtain
by inspection. The initial mass fractions, Yα,0 , can be estimated based on the relative area underneath each
peak. The values of Yα,0 should sum to 1.
The reaction rates, rα,p , require a slight interpretation. Typically, tests like TGA produce plots of the
normalized sample mass, m/m0 , and its first derivative with time, − d(m/m0 )/ dt. However, the reaction
rate in the analysis above is interpreted as the change in the component mass fraction with time:
dYα
rα = − (12.41)
dt
The relationship between these two rates is
dYα 1 d(m/m0 )
∑ = (12.42)
α dt 1 − νr dt

Thus, the reaction rates we need for Eqs. (12.39) and (12.40), rα,p , are related to the normalized mass loss
rate curve as follows:
1 d(m/m0 )
rα,p = − (12.43)
1 − νr dt T =Tα,p

The examples in the following section describe how to interpret output from bench-scale material tests.

12.9.2 Interpreting TGA Data (tga_sample)

The solid curve of the left hand plot of Fig. 12.16 shows the measured normalized mass of a small solid
sample undergoing heating at a rate of 5 ◦ C/min in the TGA apparatus. The solid curve of the right hand
plot is the mass loss rate, or minus the first derivative of the plot on the left. The goal of this example is
to obtain the kinetic parameters, Eα and Aα , from Eqs. (12.39) and (12.40) to solve numerically the ODE,
Eqs. (12.33) and (12.34). The first decision to make is the number of reactions. Usually, this corresponds
to the number of peaks in the mass loss rate curve, but for realistic cases, some judgment is needed. In
this example, it is assumed that there are two material components, each of which is undergoing a single

190
 FDS-6.8.0-0-g886e009-release 10 -4 FDS-6.8.0-0-g886e009-release
1.2 6
TGA Mass (tga_sample) TGA Mass Loss Rate (tga_sample)
1 5

Mass Loss Rate (1/s)

0.8 4
Mass Fraction

Measured
0.6 3 FDS (mlrpua)

0.4 2

0.2 Measured
1
FDS (mpua)
0 0
0 200 400 600 800 0 200 400 600 800
Temperature (°C) Temperature (°C)

Figure 12.16: Results of a TGA analysis. At left is the normalized sample mass as a function of the temperature. At
right is the corresponding mass loss rate.

reaction in the neighborhood of a certain temperature. By inspection, T1,p = 300 ◦ C and T2,p = 500 ◦ C. The
next piece of information from the measurement is the residue fraction, which is obtained from the plot on
the left, νr = 0.53. From Eq. (12.43), the reaction rates are obtained by converting the peak mass loss rates
from the plot on the right: r1,p = 0.00045/0.47 s−1 and r2,p = 0.0001/0.47 s−1 . Finally, the initial mass
fractions of the two components are estimated from the relative areas under the peaks in the right hand plot.
In this case, Y1,0 = 0.6 and Y2,0 = 0.4.
In the FDS input file, tga_sample.fds, this information is conveyed via the following lines:

&SURF ID = 'SAMPLE'
...
MATL_ID(1,1:2) = 'component 1','component 2'
MATL_MASS_FRACTION(1,1:2) = 0.6,0.4 /

&MATL ID = 'component 1'

...
N_REACTIONS = 1
REFERENCE_TEMPERATURE = 300.
REFERENCE_RATE = 0.0016
HEATING_RATE = 5.
NU_SPEC = 0.47
SPEC_ID = 'OFF-GAS'
NU_MATL = 0.53
MATL_ID = 'residue' /

&MATL ID = 'component 2'

...
N_REACTIONS = 1
REFERENCE_TEMPERATURE = 500.
REFERENCE_RATE = 0.0005
HEATING_RATE = 5.
NU_SPEC = 0.47
SPEC_ID = 'OFF-GAS'
NU_MATL = 0.53
MATL_ID = 'residue' /

Note that REFERENCE_RATE, the FDS input parameter, represents the quantity, rα,p /Yα,0 . Note also that
both material components have the same residue yield, NU_MATL. Solely from TGA data, it is not possible

191
to determine how many unique material components there are, or what the reaction sequence is. The point
of this exercise is simply to get FDS to mimic the total mass loss rate curve.

12.9.3 Effect of the TGA Heating Rate (birch_tga)

This is an example of a comparison of a candidate solid phase model with TGA data that shows the effect
of the sample heating rate. The sample cases called birch_tga_1step_2 and birch_tga_1step_20
simulate two standard TGA experiments in which small samples of birch wood are heated up slowly at
constant rates of 2 °C/min and 20 °C/min, respectively. The model of the wood consists of only one reaction
that converts virgin wood to char and fuel gases. There is also a reaction in the simulation that does nothing
more than evaporate the small amount of moisture in the wood. This evaporation is evident in Fig. 12.17
near the temperature of 100 °C.
FDS-6.8.0-0-g886e009-release

1 Exp. 2 K/min
Exp. 20 K/min
FDS 2 K/min
0.8 FDS 20 K/min
Mass Fraction

0.6

0.4

0.2

0
0 200 400 600 800
Temperature (°C)

Figure 12.17: Comparison of a solid phase model of birch wood with TGA data.

12.9.4 TGA for a Charring Sample (Needle_TGA)

In this section, kinetic parameters for a longleaf pine needle are derived based on a set of TGA measure-
ments1 . Cases A and B listed in Table 12.2 have been used to derive the kinetic parameters for all eight
cases. The mass and mass loss rate curves for all eight cases are shown in Figs. 12.19 through 12.22 on the
following pages.
It is assumed that the pine needle consists of moisture and three solid phase components making up
0.25, 0.50, and 0.25 of the total dry mass, and reaching peak decomposition rates at 250 ◦ C, 350 ◦ C, and
425 ◦ C, respectively. These estimates are obtained by inspection of the experimental data shown Fig. 12.18.
It is further assumed that 0.25 of the dry mass is converted to char, and the kinetics of the oxidation reaction
are estimated based on Case B.

1 TGA data provided by the Forest Products Laboratory of the U.S. Forest Service

192
Table 12.2: Parameters for pine needle TGA experiments. Note that the Heating Rate refers to the linear
ramp up in temperature after the sample has been dried out at approximately 100 ◦ C.

Case Heating Rate (K/min) Moisture Fraction Atmosphere

A 10 0.075 Inert
B 10 0.053 Air
C 10 1.22 Inert
D 10 1.63 Air
E 3.5 0.100 Inert
F 60 0.064 Inert
G 60 1.56 Air
H 60 1.33 Air

-3 FDS-6.8.0-0-g886e009-release -3 FDS-6.8.0-0-g886e009-release
10 10
2 2
Case A; 10 °C/min; M=0.075; Inert Case B; 10 °C/min; M=0.053; Air
Normalized Mass Loss Rate (1/s)

Normalized Mass Loss Rate (1/s)

1.5 1.5

Exp (Mdot) Exp (Mdot)

1 FDS (mlr)
1 FDS (mlr)

0.5 0.5

0 0
0 200 400 600 800 0 200 400 600 800
Temperature (°C) Temperature (°C)

Figure 12.18: Mass loss rates for Cases A (left) and B (right). These cases are used to determine the kinetic parameters
for all eight cases, A-H.

193
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
800 800
Case A; 10 °C/min; M=0.075; Inert Case B; 10 °C/min; M=0.053; Air

600 600
Temperature (°C)

Temperature (°C)
Exp (Temperature) Exp (Temperature)
400 FDS (T)
400 FDS (T)

200 200

0 0
0 10 20 30 40 50 60 70 0 10 20 30 40 50 60 70
Time (min) Time (min)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
1.2 1.2
Case A; 10 °C/min; M=0.075; Inert Case B; 10 °C/min; M=0.053; Air
1 1

0.8 0.8
Normalized Mass

Normalized Mass
Exp (Mass) Exp (Mass)
0.6 FDS (mass)
0.6 FDS (mass)

0.4 0.4

0.2 0.2

0 0
0 10 20 30 40 50 60 70 0 10 20 30 40 50 60 70
Time (min) Time (min)

-3 FDS-6.8.0-0-g886e009-release -3 FDS-6.8.0-0-g886e009-release
10 10
2 2
Case A; 10 °C/min; M=0.075; Inert Case B; 10 °C/min; M=0.053; Air
Normalized Mass Loss Rate (1/s)

Normalized Mass Loss Rate (1/s)

1.5 1.5

Exp (Mdot) Exp (Mdot)

1 FDS (mlr)
1 FDS (mlr)

0.5 0.5

0 0
0 10 20 30 40 50 60 70 0 10 20 30 40 50 60 70
Time (min) Time (min)

Figure 12.19: Results of TGA calibration exercise for a pine needle, Cases A (left) and B (right). (Top) Sample
temperature vs time. (Middle) Sample mass fraction vs time. (Bottom) Sample mass loss rate vs time.

194
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
800 800
Case C; 10 °C/min; M=1.22; Inert Case D; 10 °C/min; M=1.63; Air

600 600
Temperature (°C)

Temperature (°C)
Exp (Temperature) Exp (Temperature)
400 FDS (T)
400 FDS (T)

200 200

0 0
0 20 40 60 80 0 20 40 60 80
Time (min) Time (min)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
1.2 1.2
Case C; 10 °C/min; M=1.22; Inert Case D; 10 °C/min; M=1.63; Air
1 1

0.8 0.8
Normalized Mass

Normalized Mass
Exp (Mass) Exp (Mass)
0.6 FDS (mass)
0.6 FDS (mass)

0.4 0.4

0.2 0.2

0 0
0 20 40 60 80 0 20 40 60 80
Time (min) Time (min)

-3 FDS-6.8.0-0-g886e009-release -3 FDS-6.8.0-0-g886e009-release
10 10
2 2
Case C; 10 °C/min; M=1.22; Inert Case D; 10 °C/min; M=1.63; Air
Normalized Mass Loss Rate (1/s)

Normalized Mass Loss Rate (1/s)

1.5 1.5

Exp (Mdot) Exp (Mdot)

1 FDS (mlr)
1 FDS (mlr)

0.5 0.5

0 0
0 20 40 60 80 0 20 40 60 80
Time (min) Time (min)

Figure 12.20: Results of TGA calibration exercise for a pine needle, Cases C (left) and D (right). (Top) Sample
temperature vs time. (Middle) Sample mass fraction vs time. (Bottom) Sample mass loss rate vs time.

195
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
800 800
Case E; 3.5 °C/min; M=0.010; Inert Case F; 60 °C/min; M=0.064; Inert

600 600
Temperature (°C)

Temperature (°C)
Exp (Temperature) Exp (Temperature)
400 FDS (T)
400 FDS (T)

200 200

0 0
0 50 100 150 0 5 10 15 20
Time (min) Time (min)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
1.2 1.2
Case E; 3.5 °C/min; M=0.010; Inert Case F; 60 °C/min; M=0.064; Inert
1 1

0.8 0.8
Normalized Mass

Normalized Mass
Exp (Mass) Exp (Mass)
0.6 FDS (mass)
0.6 FDS (mass)

0.4 0.4

0.2 0.2

0 0
0 50 100 150 0 5 10 15 20
Time (min) Time (min)

-3 FDS-6.8.0-0-g886e009-release -3 FDS-6.8.0-0-g886e009-release
10 10
2 5
Case E; 3.5 °C/min; M=0.010; Inert Case F; 60 °C/min; M=0.064; Inert
Normalized Mass Loss Rate (1/s)

Normalized Mass Loss Rate (1/s)

4
1.5

3
Exp (Mdot) Exp (Mdot)
1 FDS (mlr) FDS (mlr)
2

0.5
1

0 0
0 50 100 150 0 5 10 15 20
Time (min) Time (min)

Figure 12.21: Results of TGA calibration exercise for a pine needle, Cases E (left) and F (right). (Top) Sample
temperature vs time. (Middle) Sample mass fraction vs time. (Bottom) Sample mass loss rate vs time.

196
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
800 800
Case G; 60 °C/min; M=1.56; Air Case H; 60 °C/min; M=1.33; Air

600 600
Temperature (°C)

Temperature (°C)
Exp (Temperature) Exp (Temperature)
400 FDS (T)
400 FDS (T)

200 200

0 0
0 10 20 30 40 0 5 10 15 20 25 30
Time (min) Time (min)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
1.2 1.2
Case G; 60 °C/min; M=1.56; Air Case H; 60 °C/min; M=1.33; Air
1 1

0.8 0.8
Normalized Mass

Normalized Mass
Exp (Mass) Exp (Mass)
0.6 FDS (mass)
0.6 FDS (mass)

0.4 0.4

0.2 0.2

0 0
0 10 20 30 40 0 5 10 15 20 25 30
Time (min) Time (min)

-3 FDS-6.8.0-0-g886e009-release -3 FDS-6.8.0-0-g886e009-release
10 10
5 5
Case G; 60 °C/min; M=1.56; Air Case H; 60 °C/min; M=1.33; Air
Normalized Mass Loss Rate (1/s)

Normalized Mass Loss Rate (1/s)

4 4

3 3
Exp (Mdot) Exp (Mdot)
FDS (mlr) FDS (mlr)
2 2

1 1

0 0
0 10 20 30 40 0 5 10 15 20 25 30
Time (min) Time (min)

Figure 12.22: Results of TGA calibration exercise for a pine needle, Cases G (left) and H (right). (Top) Sample
temperature vs time. (Middle) Sample mass fraction vs time. (Bottom) Sample mass loss rate vs time.

197
12.9.5 TGA of various Mediterranean vegetation (Morvan_TGA)
In this section we analyze the TGA data for various Mediterranean vegetation species presented in Morvan
and Dupuy, 2004 [95]. The FDS feature TGA_ANALYSIS=T was used with the model set up with 10 %
moisture for the vegetation model. The heating rate is 1.6 ◦C/min, as reported in the paper. The reference
temperature and pyrolysis range approach was used to tune the model to the data. The parameters can be
found in the input file Verification/WUI/Morvan_TGA.fds. Note that an unusually high char fraction
of 50 % is needed to match the TGA data. The results are shown in Fig. 12.23.
FDS-6.8.0-0-g886e009-release 10 -4 FDS-6.8.0-0-g886e009-release
1.2 1.5
Morvan TGA; 1.6 °C/min Morvan TGA; 1.6 °C/min
1

Mass Loss Rate (1/s)

0.8 1
Normalized Mass

Exp (Morvan 2004) SG deriv (Morvan 2004)

0.6 FDS TGA (Total Mass) FDS TGA (Total MLR)

0.4 0.5

0.2

0 0
0 100 200 300 400 500 600 700 0 100 200 300 400 500 600 700
Temperature (°C) Temperature (°C)

Figure 12.23: Results of performing the FDS TGA analysis on the data for various Mediterranean vegetation presented
in Morvan and Dupuy, 2004 [95]. (Left) Total mass vs time. (Right) Mass loss rate vs time—the Morvan and Dupuy
data were Savitsky-Golay filtered and the derivative of the resulting curve was normalized to give the dashed line.

198
12.9.6 Interpreting MCC Data (cable_XX_mcc)
This section describes a method for interpreting micro-combustion calorimeter (MCC) measurements. The
pyrolysis combustion flow calorimeter (PCFC), developed by Lyon and Walters [96] at the U.S. Federal Avi-
ation Administration (FAA), is a device used to measure the heat generated from the combustion of small
(4 mg to 6 mg) material samples by oxygen depletion calorimetry. Samples are pyrolyzed at a specified heat-
ing rate in an anaerobic atmosphere (typically N2 ) and the resulting gases are mixed with excess oxygen and
combusted in a separate chamber. The heat release rate from the specimen is obtained from measurements
of the concentration of oxygen in the effluent exiting the combustor as a function of time. The methodology
is the basis for the standard test ASTM D 7309 [97].
The results of PCFC measurements for the jacket and insulation materials of several electrical cables are
shown in Fig. 12.24. These samples were pyrolyzed in the PCFC at a rate of 1 °C/s from 100 °C to 600 °C
in a nitrogen atmosphere and the effluent combusted at 900 °C in a mixture consisting of 20 % O2 and 80 %
N2 . The resulting curve shows the heat release rate of the sample as it was heated, normalized by the mass
of the original sample. There are usually one, two or three noticeable peaks in the curve, corresponding
to temperatures where a significant decomposition reaction occurs. Each peak can be characterized by the
maximum value of the heat release rate per unit mass, q̇α,p , the temperature, Tα,p , and the relative fraction
of the original sample mass that undergoes this particular reaction (Yα,0 ). The area under the curve
Z ∞
q̇(T ) dT = β ∆h0 (12.44)
0

is the sample heating rate, β , multiplied by the energy released per unit mass of the original sample, ∆h0 .
This latter quantity is related to the more conventional2 heat of combustion via the relation

∆h0
∆h = (12.45)
1 − νr
where νr is the fraction of the original mass that remains as residue. Sometimes this is referred to as the
“char yield.” Note that it is assumed to be the same for all reactions.
The MCC measurement is similar to TGA in that it is possible to derive the kinetic parameters, Aα
and Eα , for the various reactions from the heat release rate curve. As an example of how to work with
MCC data, consider the two plots shown in Fig. 12.24. The solid curves in the figures display the results
of micro-calorimetry measurements for the insulation and jacket material of multi-conductor control cables
(the numbers have no particular meaning other than to distinguish them from other cables being studied).
Taking Cable 11 as an example, the insulation material exhibits two fairly well-defined peaks, whereas
the jacket material exhibits three. Thus, the insulation material is modeled using two solid components,
each undergoing a single-step reaction that produces fuel gas and a solid residue. The jacket material is
modeled using three solid components. The residue yield for the insulation material is 6 %; for the jacket
49 %, obtained simply by weighing the sample before and after the measurement. It is not known which
reaction produces what fraction of the residue. Rather, it is assumed that each reaction yields the same
residue in the same relative amount. The dashed curves in Fig. 12.24 are the results of FDS simulations of
the MCC measurements. To mimic the sample heating, a very thin sheet comprised of a mixture of the solid
components with an insulated backing is heated at the rate specified in the experiment (1 °C/s or 60 °C/min,
the units needed in FDS). For each reaction, the kinetic parameters are calculated using the formulae (12.39)
and (12.40). The values of Tα,p are obtained directly from the figures. The peak reaction rate for reaction α,
2 In fire protection engineering, it is typically assumed that the heat of combustion is the energy released per unit mass of

vaporized fuel.

199
 FDS-6.8.0-0-g886e009-release
FDS-6.8.0-0-g886e009-release
800 800
Heat Release Rate (cable_11_insulation_mcc) Heat Release Rate (cable_11_jacket_mcc)

600 600

HRR (W/g)
HRR (W/g)

Exp (HRR) Exp (HRR)

400 400 FDS (hrrpum)
FDS (hrrpum)

200 200

0 0
0 100 200 300 400 500 600 0 100 200 300 400 500 600
Temperature (°C) Temperature (°C)

FDS-6.8.0-0-g886e009-release
FDS-6.8.0-0-g886e009-release
800 800
Heat Release Rate (cable_23_insulation_mcc) Heat Release Rate (cable_23_jacket_mcc)

600 600

HRR (W/g)
HRR (W/g)

Exp (HRR) Exp (HRR)

400 400 FDS (hrrpum)
FDS (hrrpum)

200 200

0 0
0 100 200 300 400 500 600 0 100 200 300 400 500 600
Temperature (°C) Temperature (°C)

FDS-6.8.0-0-g886e009-release
FDS-6.8.0-0-g886e009-release
800 800
Heat Release Rate (cable_701_insulation_mcc) Heat Release Rate (cable_701_jacket_mcc)

600 600
HRR (W/g)
HRR (W/g)

Exp (HRR) Exp (HRR)

400 400 FDS (hrrpum)
FDS (hrrpum)

200 200

0 0
0 100 200 300 400 500 600 0 100 200 300 400 500 600
Temperature (°C) Temperature (°C)

Figure 12.24: Results of a micro-calorimetry analyses of cable insulation (left) and jacket materials (right).

200
rα,p , is obtained from:
q̇α,p 1
Z ∞
0
rα,p = ; ∆h = q̇(T ) dT (12.46)
∆h0 β 0

where q̇α,p is the peak heat release rate corresponding to reaction α. The values, Yα,0 , can be estimated from
the relative area under the curve. Their sum ought to be 1. It is important to check the units of all of these
quantities because the results of these experiments are often presented in different ways depending on the
particular application. A mistake in units can result in values of Aα and/or Eα that will invariably cause
spurious results. For this reason, the direct calculation of the kinetic parameters is avoided by inputting the
reaction parameters instead.
The dashed curves in Fig. 12.24 are the results of numerically integrating Eq. (12.33) within FDS for
each material component. A typical input line for FDS that describes a single material component undergo-
ing a single reaction is given by:

&MATL ID = 'Cable 11 Jacket Component A'

...
N_REACTIONS = 1
REFERENCE_TEMPERATURE = 300.
REFERENCE_RATE = 0.0064
HEATING_RATE = 60.
NU_MATL = 0.49
MATL_ID = 'char'
NU_SPEC = 0.51 /

Only the relevant parameters are shown. The other parameters are not relevant in this exercise. Note that
REFERENCE_TEMPERATURE is Tα,p in units of ◦C. REFERENCE_RATE is actually rα,p /Yα,0 , in units of s−1 .
HEATING_RATE is β in units of °C/min. NU_MATL is νr and NU_SPEC is (1 − νr ). Table 12.3 lists all of the
kinetic parameters for the cable insulation and jacket materials. The peak temperatures are easy to estimate,
and the values of r p,i /Y0,i can be fine-tuned to closely match the data. Note that it is possible to compute
values of Ai and Ei and input them directly into FDS, rather than inputting those listed in the table. However,
the values of Ai and Ei are fairly large numbers and have little meaning in their own.

Table 12.3: Parameters used to derive the kinetic constants for Cable 11 materials. The heating rate for both
is 60 °C/min.

Parameter Insulation, νr = 0.06 Jacket, νr = 0.49

α =1 α =2 α =1 α =2 α =3
Tα,p (°C) 355 485 300 345 450
rα,p /Yα,0 (s−1 ) 0.0384 0.2426 0.0064 0.3500 0.0156

201
12.10 Cone Data Extrapolation Model (cone_demo_2)
This section tests a model for extrapolating heat release rate data from a cone calorimeter to heat fluxes
other than the one used during the cone test. In the case a surface is assigned an HRRPUA and RAMP_Q from
a cone test performed at 50 kW/m2 . The surface is given an IGNITION_TEMPERATURE of 0 ◦ C so it ignites
immediately. The surface is exposed to a sawtooth EXTERNAL_FLUX that varies between 10 kW/m2 and
110 kW/m2 . The results are shown below in Fig 12.25.
FDS-6.8.0-0-g886e009-release
500
Cone Demo 2
400
HRRPUA (kW/m²)

300
Cone
FDS
200

100

0
0 100 200 300 400 500 600
Time (s)

Figure 12.25: Check of the cone data extrapolation model.

12.11 Melting Ice Cube (ice_cube)

This section tests the ability for a MATL reaction to produce particles. A 10 cm cube (1 kg) block of ice is
defined with a 10 kW/m2 EXTERNAL_FLUX. The mass of water drops produced is tracked. The results are
shown below in Fig 12.26.
FDS-6.8.0-0-g886e009-release
1.2
ice_cube
1

0.8
Mass (kg)

Expected
0.6 Predicted

0.4

0.2

0
0 100 200 300 400 500
Time (s)

Figure 12.26: Check of the particle production by surface reaction feature.

202
Chapter 13

Lagrangian Particles

This chapter contains verification cases that test all aspects of particles, droplets, sprays, and so on. Note
that in FDS, Lagrangian particles are used for a variety of purposes, not just water droplets.

13.1 Particle and Droplet Drag

The cases in this section test accuracy, stability, and momentum conservation for the general case of mo-
mentum transfer between solid particles or liquid droplets and air. Details of the FDS integration scheme
are given in the Technical Reference Guide [1]. In short, we find an analytical solution for the case of
binary fluid-particle interaction and apply this solution to the case of multiple particles by enforcing total
momentum conservation within a fluid cell. The resulting scheme is guaranteed to be stable because the
fluid velocity cannot overshoot the cell equilibrium velocity (total momentum of fluid and particles divided
by the total mass).

13.1.1 Dragless Particle (cannon_ball)

This simple case tests the particle transport scheme in the absence of drag. A 1 cm diameter sphere is
shot from ground level with an initial velocity of (u0 , v0 , w0 ) = (10, 10, 10) m/s. The ball is expected to
fall backqto the ground in t = 2 w0 /g ≈ 2.04 s at which time it will have travelled a horizontal distance
of d = t u20 + v20 ≈ 28.86 m. The grid has a resolution of 1 m, and the default time step is 0.5 s. Thus,
the particle trajectory is computed via only 5 calls to particle transport routine. However, the integration is
performed using sub-steps such that the particle cannot cross farther than a single grid cell in a sub-step.
This ensures an accurate computation of the trajectory, as shown in Fig. 13.1.

13.1.2 Particles Swept Up in a Free Stream (fluid_part_mom)

Consider a 1 m square duct that is 10 m long. The boundary conditions are periodic in the streamwise
direction and free slip along the inside walls. The fluid velocity is initialized to be 10 m/s. A monodisperse
distribution of 1000 particles is initialized with zero velocity and random positions in the first cubic meter
of the duct. Gravity is set to zero. As the case proceeds momentum is exchanged between the fluid and the
particles. The results for all three tests are shown in Fig. 13.2. On the top, the green symbols represent the
total momentum of the system, which, as can be seen, is conserved. The fluid (blue) and the particle (red)
momenta are closely following the analytical solution (solid lines) which is the basis of the scheme. The
plot on the bottom shows the fluid (blue) and particle (red) velocities compared to the system equilibrium

203
 FDS-6.8.0-0-g886e009-release
30
Particle Trajectory (cannon_ball)
25

20

Distance (m)
15

10

5 Exact
FDS (d)
0
0 0.5 1 1.5 2 2.5
Time (s)

Figure 13.1: Horizontal distance travelled by a dragless sphere as a function of time.

velocity (green dashed line). By construction of the method, both the fluid and the particle velocities are
approaching the equilibrium velocity at long times.

13.1.3 Particles Slowing Down a Gas Flow (particle_drag)

The particle drag test cases consider a 1 m by 1 m by 1 m channel with periodic boundary conditions on the
x-faces and FREE_SLIP walls on y- and z-faces. Static droplets are placed in the center of the channel, one
particle per computational cell, so that they form a surface perpendicular to the flow direction. Gravity is set
to zero. Due to the symmetry of the problem, the flow is one dimensional. Assuming that the droplets are
of uniform diameter and the drag coefficient and gas density are constant, the velocity in the channel decays
according to
u0 1 ∑ CD π rd2
u= ; B= (13.1)
1 + B u0 t 2 V
In the above, V is the volume of the channel, rd is the droplet radius, CD is the droplet drag coefficient, and
u is the gas velocity in the x-direction. The summation is over all N particles. The common parameters used
in all the simulations are: Cd = 10, rd = 0.005 m. The initial velocity, u0 , for Cases A-F are 10, 50, 100, 50,
100, and 150 m/s, respectively. Comparisons of computed and analytical results are shown in Fig. 13.3.
Following the above notation, the total drag force can also be written as:
 2
u0
Fd = ρV B (13.2)
1 + B u0 t

Checking this quantity is useful for confirming the PARTICLE DRAG FORCE X/Y/Z output quantities and
SPATIAL_STATISTIC = ’SUM’ for a volume of particles. The computed results for Case A are compared
to the analytical solution in Fig. 13.4.

13.1.4 Stationary Spherical Particles in a Duct (sphere_drag)

Consider a 2 m long duct with a 1 m square cross section, fixed air velocity of u0 = 2 m/s, and free-slip
walls. Three ducts are stacked vertically, each with its own mesh and a plane of particles spanning the duct

204
 FDS-6.8.0-0-g886e009-release
150

FDS fluid U
FDS fluid V

Momentum (kg m/s)

FDS fluid W
100
FDS particle U
FDS particle V
FDS particle W
FDS total U
FDS total V
50 FDS total W
Analytical fluid
Analytical particle

0
0 0.2 0.4 0.6 0.8 1
Time (s)

FDS-6.8.0-0-g886e009-release
10

8
FDS fluid U
Velocity (m/s)

FDS fluid V
6 FDS fluid W
FDS particle U
FDS particle V
4 FDS particle W
Equilibrium velocity

0
0 0.2 0.4 0.6 0.8 1
Time (s)

Figure 13.2: Outputs of the fluid_part_mom cases. (Top) The fluid momentum (blue), the particle momentum
(red), and the total momentum (green) are plotted for each coordinate direction, x (circles), y (inverted triangles), and
z (plus symbols). The analytical solutions for the fluid and particle equations are shown by the solid blue and red
lines, respectively. The total momentum should be constant, as indicated by the green symbols. (Bottom) The fluid
(blue) and particle (red) velocities are plotted for each direction. Also, the equilibrium velocity (total momentum
[fluid plus particles] divided by the total mass) is shown as a dashed green line. Stability of the momentum exchange
is guaranteed because, by construction, the fluid velocity cannot overshoot the equilibrium velocity; the fluid velocity
relaxes toward it.

at its center point. The particles are 1 cm in diameter and 10 particles are specified in each grid cell. The
expected pressure drop is given by the formula:

1 ∑ Cd π rp2 u20
∆p = ρ (13.3)
2 A
where A is the 1 m2 cross-sectional area, ρ = 1.2 kg/m3 is the density of air, and the summation is over 4000
particles. For specified drag coefficients of 5, 10, and 20 in the three ducts, the pressure drops are expected
to be 3.77 Pa, 7.54 Pa, and 15.1 Pa. Comparisons of computed and analytical results are shown in the left
hand plot of Fig. 13.5.
In a second test case, consider a single 10 m long, 1 m square duct. Spherical particles 2 mm in diameter
with a density of 514 kg/m3 are randomly distributed in the section of the duct between 4 m and 5 m from
the upstream end. The particle mass per unit volume is set to 1.66 kg/m3 . The number of particles included
in the simulation is 10000, which means that each particle actually represents 77.1 real particles. The drag
coefficient is approximately 1.6, based on the local Reynolds number, which is about 40. The free stream
velocity in the duct is 0.3 m/s, but the speed varies slightly within the cloud of particles. The pressure is

205
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
12 60
Gas phase velocity (particle_drag_A) Gas phase velocity (particle_drag_D)
10 50

8 40
Velocity (m/s)

Velocity (m/s)
Analytical
6 FDS
30

4 20

2 10 Analytical
FDS
0 0
0 20 40 60 80 100 0 0.5 1 1.5 2
Time (s) Time (s)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
60 120
Gas phase velocity (particle_drag_B) Gas phase velocity (particle_drag_E)
50 100

40 80
Velocity (m/s)

Velocity (m/s)
Analytical
30 FDS
60

20 40

10 20 Analytical
FDS
0 0
0 10 20 30 40 50 0 0.5 1 1.5 2
Time (s) Time (s)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
120 200
Gas phase velocity (particle_drag_C) Gas phase velocity (particle_drag_F)
100
150
80
Velocity (m/s)

Velocity (m/s)

Analytical
60 FDS
100

40
50
20 Analytical
FDS
0 0
0 5 10 15 20 25 0 0.5 1 1.5 2
Time (s) Time (s)

Figure 13.3: Outputs of the particle_drag test cases compared with analytical solutions.

206
 FDS-6.8.0-0-g886e009-release
1
Total drag force (particle_drag_A)
0.8

0.6

Force (N)
Analytical
FDS
0.4

0.2

0
0 20 40 60 80 100
Time (s)

Figure 13.4: Output of total drag from the particle_drag_U10_N16 test case compared with analytical solution.

expected to drop linearly from approximately 0.21 Pa to 0 Pa over the 1 m of duct filled by particles, as
shown in the right hand plot of Fig. 13.5.
FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
0.3
Pressure Drop (sphere_drag_1) Pressure Drop (sphere_drag_2)
15 0.25
Pressure Drop (Pa)

Exact 5 0.2
Pressure (Pa)

Exact 10
10 Exact 20 Exact
FDS 5
0.15 FDS
FDS 10
FDS 20 0.1
5
0.05

0 0
0 2 4 6 8 10 0 2 4 6 8 10
Time (s) Length (m)

Figure 13.5: Outputs of the sphere_drag test cases compared with analytical solutions.

13.1.5 Drag Caused by Vegetation (vegetation_drag)

Vegetation (leaves, grasses, etc.) is sometimes modeled using Lagrangian particles with flat, cylindrical, or
spherical geometries. In this test case. cylindrical particles represent pine needles. Consider a 4 m long duct
with a 1 m square cross section, fixed air velocity of u0 = 2 m/s, and free-slip walls. Three ducts are stacked
vertically, each with a 10 cm uniform mesh and a 1 m by 1 m by 1 m volume filled with particles spanning
the duct near its center. The cylindrical particles are D = 1 mm in diameter and L = 2 cm long. In the lowest
duct, 5000 particles are inserted randomly throughout the volume. In the middle duct, 1 particle is inserted
at the center of each grid cell within the volume containing the pine needles. In the top duct, 5 particles are
inserted randomly in each grid cell, for a total of 5000 particles. The drag coefficient for each cylindrical
particle is given by the empirical relation:

0.6 + 0.4 Re0.8 ρ u0 D

Cd = 10 ≈ 1.55 ; Re = ≈ 132.9 (13.4)
Re µair

207
The viscosity of air at 20 ◦ C is µair = 1.8 × 10−5 kg/(m·s).

Conventional Notation
The pressure drop in each duct is expected to the be same, regardless of how the particles are inserted:

1 np Cd Ac u20 m000 V
∆p = − ρ ≈ −18.9 Pa ; np = ≈ 254648 (13.5)
2 A π r2 L ρp

where A is the 1 m2 cross-sectional area of the duct, Ac = LD is the cross sectional area of the cylindrical
particle, and ρ = 1.196 kg/m3 is the density of air. The number of actual pine needles within the volume, np ,
is the mass per unit volume, m000 = 2 kg/m3 , times the volume, V = 1 m3 , divided by the mass of an individual
pine needle. The density of the needle is ρp = 500 kg/m3 . Comparisons of computed and analytical pressure
drops are shown in the left plot of Fig. 13.6. Differences between the ideal and computed results are due to
the fact that the air velocity is not exactly 2 m/s within the volume filled with the particles. In particular, the
duct filled randomly with particles exhibits the greatest deviation from the ideal because the velocity does
not remain uniform within the volume.

Notation used by the Wildland Fire Community

The notation used in the previous section is inconvenient when characterizing the drag force exerted by
actual vegetation that cannot be assumed to be perfect spheres or cylinders. Instead, Eq. (13.5) can be
rewritten as follows:
L
∆p = − ρ Cd Cs β σ u20 ≈ −18.9 Pa (13.6)
2
Here, Cs is the ratio of the particle’s cross sectional area to surface area, which is approximately 1/π for
objects that are cylindrical in shape. β is the packing ratio, which is the ratio of the volume occupied by the
solid particles to the overall volume of particle-laden region; 0.004 in this case. σ is the surface to volume
ratio of an individual particle; 2/r for a cylinder.

Pressure Drop due to Mass-Generating Particles

Using this same basic configuration, another test is performed where the cylindrical particles now exert no
drag force on the air flow, but instead generate additional air at a rate of 0.1 kg/(m2 ·s). The surface area of
each of the 254648 particles is 6.28 × 10−5 m2 , resulting in a total air injection rate of ṁ = 1.6 kg/s. The
momentum equation becomes one-dimensional and time-independent:

d p u2
 
1 ṁ
+ =− u (13.7)
dx ρ 2 ρV

Using the subscript 1 to denote the upstream region of the duct, and 2 the downstream, the air velocity
increases as it passes through the particles: u2 = u1 + ṁ/(ρ A). Multiplying Eq. (13.7) by the constant
density, ρ, and integrating both sides over the length of the duct containing the particles:
u2 + u1 ṁ u2 + u1
∆p ≡ p2 − p1 = −ρ (u2 − u1 ) − = −8.534 Pa (13.8)
2 A 2
The results are shown in the right hand plot of Fig. 13.6 for the various different methods of representing
the particles in the model.

208
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
30 12
Vegetation Drag (vegetation_drag_1) Vegetation Drag (vegetation_drag_2)
25 10
Pressure Drop (Pa)

Pressure Drop (Pa)

20 8

15 6
Ideal (Pressure) Ideal (Pressure)
10 Ideal (Pressure) 4 Ideal (Pressure)
Ideal (Pressure) Ideal (Pressure)
FDS (p1) FDS (p1)
5 FDS (p2)
2 FDS (p2)
FDS (p3) FDS (p3)
0 0
0 2 4 6 8 10 0 2 4 6 8 10
Time (s) Time (s)

Figure 13.6: Pressure drop in duct filled with pine needles for three different particle insertion methods.

13.1.6 Mesh Screen in a Duct (screen_drag)

These two test cases consider a 1 m wide by 1 m tall by 2 m long channel with either a 2 m/s or a 20 m/s fixed
velocity boundary condition fixed velocity boundary condition on one end, an OPEN boundary condition on
the other end, and FREE_SLIP on the y- and z-faces. Three channels are stacked vertically (each with its
own MESH, and a plane of particles is placed across each channel at 1 m. The plane of particles are defined
as a screen with a wire diameter of 0.0003 m. A FREE_AREA_FRACTION of 0.1, 0.4, and 0.8 is set for the
particles in the channels. According to the screen drag formula
 
µ Y 2
∆p = u + ρ √ u ∆xscreen (13.9)
K K
where K and Y are empirical constants defined in the FDS User’s Guide, the pressure drops for the 2 m/s
inlet should be 126 Pa, 14 Pa, and 5.3 Pa, and for the 20 m/s inlet should be 1260 Pa, 168 Pa, and 124 Pa.
Comparisons of computed and analytical results are shown in Fig. 13.7.
FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
150 1500
Screen Pressure Drop Screen Pressure Drop
Pressure Drop (Pa)

Pressure Drop (Pa)

100 Exact 0.1 1000 Exact 0.1

Exact 0.4 Exact 0.4
Exact 0.8 Exact 0.8
FDS 0.1 FDS 0.1
FDS 0.4 FDS 0.4
50 FDS 0.8 500 FDS 0.8

0 0
0 1 2 3 4 0 2 4 6 8 10
Time (s) Time (s)

Figure 13.7: Outputs of the screen_drag test cases compared with analytical solutions. Left 2 m/s inlet, right
20 m/s inlet. The three cases represent different values of FREE_AREA_FRACTION (0.1, 0.4, and 0.8) within the
channel.

209
13.1.7 Flow through Porous Media (porous_media)
This case involves a 2 m long duct with a 0.5 m by 0.5 m cross section and 1 m/s flow in the x direction. An
8.5 cm thick block of rigid aluminum foam is positioned in the duct. The pressure drop induced by the foam
is expected to be approximately:  
µ Y 2
∆p ≈ δ u+ρ √ u (13.10)
K K
where K = 1 × 10−7 m2 and Y = 0.1 in all three coordinate directions. The block thickness δ = 0.085 m, the
velocity u = 1 m/s, the air density ρ = 1.2 kg/m3 , and the air viscosity µ = 1.8 × 10−5 kg/m/s. Comparisons
of computed and analytical results for three different grid resolutions (1 corresponds to low resolution, 2
medium, 3 fine) are shown in Fig. 13.8.
FDS-6.8.0-0-g886e009-release
60
Pressure (porous_media)
50
Pressure Drop (Pa)

40

30
Expected (Pres)
20 Expected (Pres)
Expected (Pres)
FDS (pres_1)
10 FDS (pres_2)
FDS (pres_3)
0
0 2 4 6 8 10
Time (s)

Figure 13.8: Pressure drop caused by a block of aluminum foam.

13.1.8 Particle Terminal Velocity (terminal_velocity)

If a single particle with a constant drag coefficient falls through a stationary fluid, its terminal velocity can
be found analytically. The equation governing this motion is
du ρaCd A
= Ku2 − g ; K≡ (13.11)
dt 2m
where ρa is the air density, Cd = 1 is the drag coefficient, A is the cross-sectional area of the particle, and m
is the mass. Defining ρd = 1000 kg/m3 as the particle density and D = 0.01 m as the particle diameter, the
solution of the ODE is:
r  √

dx g p  ln cosh gKt 3ρaCd
u= =− tanh gKt ; x=− ; K≡ (13.12)
dt K K 4ρd D

The L∞ (maximum) error of the FDS computed terminal velocity and position are computed and plotted for
each time step. As shown in Fig. 13.9, FDS exhibits first-order convergence for the particle position and the
terminal velocity error is near machine precision.

210
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
10 0 10
0
Terminal Velocity Error (m/s)

-5

Position Error (m)

10

10 -5

10 -10

10 -15 10 -10
10 -4 10 -3 10 -2 10 -1 10 0 10 -4 10 -3 10 -2 10 -1 10 0
Time Step (s) Time Step (s)

Figure 13.9: (Left) FDS convergence for the particle velocity in the terminal_velocity case. (Right) FDS
convergence for the particle position in the terminal_velocity case.

13.1.9 Particle Trajectory (flat_fire)

There are no known solutions of the particle trajectory equations in multiple dimensions. However, ballis-
ticians have developed an approximate solution useful for computing trajectories of projectiles when guns
are shot within a certain range of angles. This solution makes use of the “flat fire” assumption. In short, if
a particle is launched with its velocity in only the horizontal direction and its vertical velocity is relatively
small compared to the horizontal velocity, then throughout the entire trajectory the horizontal velocity com-
ponent, up , is a good approximation for the velocity magnitude, |up |. In 2-D, this approximation decouples
the horizontal (x) equation of motion from the vertical (z) and reduces the vertical equation of motion to a
linear ODE:
dup 3ρaCd
= −K u2p ; K ≡ (13.13)
dt 8ρp rp

dwp
= −K up wp − g (13.14)
dt
Given the initial conditions, xp (0) = 0, up (0) = V0 , zp (0) = h, wp (0) = 0, the exact solution to the ODE is:
ln(V0 Kt + 1) g ln(V0 Kt + 1) gt 2 gt
xp = ; zp = h + 2
− − (13.15)
K 2(V0 K) 4 2V0 K
V0 g gt g
up = ; wp = − − (13.16)
V0 Kt + 1 2V0 K (V0 Kt + 1) 2 2V0 K
The following case was run in FDS and compared against the approximate solution:
Drag coefficient Cd 0.2
Particle diameter 2 rp 5 mm
Gravitational acceleration g 9.8 m/s2
Particle density ρp 1000 kg/m3
Air density ρa 1.2 kg/m3
Initial horizontal velocity V0 400 m/s
Initial height h 8m
Figure 13.10 compares the computed solution to the “flat fire” solution given above.

211
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
10 100
Particle Trajectory (flat_fire) Horizontal Position (flat_fire)
8 80

6 60
z p (m)

x p (m)
4 40

2 20
Exact Exact
FDS FDS
0 0
0 20 40 60 80 100 0 0.5 1 1.5 2
x p (m) Time (s)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
10 400
Vertical Position (flat_fire) Horizontal Velocity (flat_fire)
8
300

u p (m/s)
z p (m)

Exact Exact
FDS
200 FDS
4

100
2

0 0
0 0.5 1 1.5 2 0 0.5 1 1.5 2
Time (s) Time (s)

FDS-6.8.0-0-g886e009-release
0
Vertical Velocity (flat_fire)
-2

-4
w p (m/s)

Exact
FDS
-6

-8

-10
0 0.5 1 1.5 2
Time (s)

Figure 13.10: Results of the flat_fire test case.

212
13.2 Solid Particle Mass and Energy Transfer
13.2.1 Radiation Absorption by Vegetation
In this example, six channels, each measuring 4 m by 1 m by 1 m, are filled with various amounts of
particles that represent vegetation. The vegetation occupies exactly one cubic meter within each channel.
One end of each channel is a solid wall with temperature Thot = 1273 K, and the opposite end is a wall with
temperature Tcold = 0 K. Each wall has an emissivity of 1 and a heat transfer coefficient of 0. The side walls
of the channels are “mirror” boundaries which essentially means that the end walls and vegetation layer
are infinitely wide, rendering the configuration one-dimensional. With no particles present, the radiative
4 /π ≈ 47.4 kW/(m2 ·sr). With particles
intensity, or radiance, in the axial direction at the hot wall is I0 = σ Thot
present, the radiance is expected to attenuate according to the following:
I = I0 e−0.25 β σp L (13.17)
β = m000 /ρp is the “packing ratio,” the vegetation mass per unit volume, or “bulk density,” divided by the
density of the vegetation itself, ρp = 500 kg/m3 . σp = 4000 m−1 is the particle surface area to volume ratio,
and L = 1 m is the depth of the vegetation layer. Figure 13.11 displays the comparison of the expected
radiative heat flux versus that calculated by FDS.
FDS-6.8.0-0-g886e009-release
60
Radiation Absorption (vegetation_absorb)
50
Radiance (kW/m²/sr)

40

Ideal (rad)
30 FDS (rad)

20

10

0
0 0.5 1 1.5
Bulk Density (kg/m³)

Figure 13.11: Decrease in radiance caused by layers of vegetation of various densities.

13.2.2 Heating a Metal Sphere via Radiation and Convection

A small metal sphere with mass, ms = 0.005 kg, is suspended in a 1 m cube filled with mg = 0.318 kg
nitrogen with a specified specific heat, c p = 1 kJ/(kg·K), and initial temperature, Tg,i = 1073.15 K. The metal
has a specified specific heat, cs = 1 kJ/(kg·K), and initial temperature, Ts,i = 293.15 K. The walls of the box
are adiabatic. In the first case, particle_heating_convection, the sphere is heated via convection only
and there is no radiation heat transfer. In the second case, particle_heating_radiation, the sphere
is heated via radiation only. The initial heat flux in both cases is approximately 75 kW/m2 , and the final
temperature in both cases, Tf , is found from solving an equation that equates the internal energy gained by
the solid with the internal energy lost by the gas:
R 8.3145
ms cs (Tf − Ts,i ) = mg cv (Tg,i − Tf ) ; = 1− cv = c p −
≈ 0.703 kJ/(kg · K) (13.18)
W 28
The final temperature, Tf = 1056.1 K or 782.9 ◦ C, as shown in Fig. 13.12.

213
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
1000 1000
Temperature (particle_heating_convection) Temperature (particle_heating_radiation)
800 800
Temperature (°C)

Temperature (°C)
600 600

400 400

200 Analytical 200 Analytical

FDS (T_gas) FDS (T_gas)
FDS (T_ball) FDS (T_ball)
0 0
0 200 400 600 800 0 200 400 600 800
Time (s) Time (s)

Figure 13.12: Temperature of a small cold sphere and surrounding hot gas when the sphere is heated by convection
only (left) and radiation only (right).

214
13.3 Water Droplet Evaporation (water_evaporation)
The cases called water_evaporation_x test the water evaporation routine in FDS. The geometry consists
either of a sealed box or a tunnel with a cross flow. The walls are assumed adiabatic, meaning that there are
no leaks or heat losses. For the box cases, the air within the box is stirred to maintain uniform conditions.
In each case, the numerical solutions are compared to analytical solutions.
For Cases 1-3 where the water droplets are confined to a sealed box, the change in the enthalpy of the
droplets should equal the change in the enthalpy of the gas minus the work performed due to pressure. It
is the internal energy of the system that is conserved. The internal energy can be expressed in terms of the
enthalpy, pressure and density:
p
e = h− (13.19)
ρ
In differential form:  
1 1
de = dh − dp − p d = dh − v dp − p dv (13.20)
ρ ρ
Multiplying by mass and noting that the volume, V , is constant yields:
dE = dH −V dp (13.21)
The enthalpy decrease of the liquid water droplets is equal to the enthalpy gain of the gas (both expressed in
kJ) minus the pressure increase times the volume in units of kPa and m3 , respectively.

13.3.1 Case 1
In the first case, the dimensions of the box are 1 m on a side, the initial air temperature is 200 °C, the
median volumetric diameter of the droplets is 200 µm, the water temperature is 20 °C, and the total mass of
water droplets is 0.01 kg. The initial mass fraction of water vapor is 0. It is expected that the droplets will
all evaporate within about 10 s. Figure 13.13 displays the average enthalpy, humidity, density, temperature,
pressure, and mass of water vapor within the box. The solid horizontal lines denote the expected steady-state
values.

13.3.2 Case 2
In the second case, the dimensions of the box are 40 cm on a side, and the initial air temperature is 500 °C.
There is only one fictitiously large water droplet in the box whose diameter is 10 cm and whose temperature
is 20 °C. The initial mass fraction of water vapor is 0. It is expected that the water will evaporate until
the air in the box is saturated. The analytical solution for the stationary state is derived from the first law
of thermodynamics and the Clausius-Clapeyron equation by assuming that the liquid water temperature is
equal to air temperature at the steady state. Results are shown in Fig. 13.14.

13.3.3 Case 3
In the third case, the dimensions of the box are 1 m on a side, the initial air temperature is 500 °C, the
initial diameter of all the droplets is 200 µm, the water temperature is 20 °C, and the total mass of water
droplets is 0.1 kg. The initial mass fraction of water vapor is 0. It is expected to reach the saturation
state. As in the second case, the analytical solution for the stationary state derives from the first principle
of thermodynamics and the Clausius-Clapeyron equation by assuming that the liquid water temperature is
equal to air temperature at the stationary state. Results are shown in Fig. 13.15.

215
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
2.5
200 Enthalpy Change (water_evaporation_1) Relative Humidity (water_evaporation_1)
2

100

Humidity (%)
Enthalpy (kJ)

Expected (h_gas) 1.5

Exact (h_water)
FDS (h_gas)
0
FDS (h_water) 1

-100 0.5
Exact (Rel. Hum)
FDS (humid)
-200 0
0 2 4 6 8 10 0 2 4 6 8 10
Time (s) Time (s)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
0.015 220
Density Change (water_evaporation_1) Temperature (water_evaporation_1)
200

0.01

Temperature (°C)
Density (kg/m³)

180

160
0.005

140
Exact (dens) Expected (temp)
FDS (dens) FDS (Temp)
0 120
0 2 4 6 8 10 0 2 4 6 8 10
Time (s) Time (s)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
0 0.015
Pressure Change (water_evaporation_1) Evaporated Mass (water_evaporation_1)
-2

0.01
Pressure (kPa)

-4
Mass (kg)

Expected (pres)
FDS (pres)
-6
0.005

-8
Exact (vapor)
FDS (WATER VAPOR)
-10 0
0 2 4 6 8 10 0 2 4 6 8 10
Time (s) Time (s)

Figure 13.13: Output of the water_evaporation_1 test case.

216
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
40 120
Enthalpy Change (water_evaporation_2) Relative Humidity (water_evaporation_2)
100
20
80

Humidity (%)
Enthalpy (kJ)

Expected (h_gas)
Exact (h_water) Exact (humid)
0 FDS (h_gas)
60 FDS (humid)
FDS (h_water)
40
-20
20

-40 0
0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14
Time (h) Time (h)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
0.025
Density Change (water_evaporation_2) 500 Temperature (water_evaporation_2)
0.02
400

Temperature (°C)
Density (kg/m³)

0.015 Expected (Temp)

300 Expected (temp_w)
FDS (Temp)
0.01 FDS (temp_w)
200

0.005 100
Expected (dens)
FDS (dens)
0 0
0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14
Time (h) Time (h)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
0 0.05
Pressure Change (water_evaporation_2) Water Concentration (water_evaporation_2)
-10
0.04
-20
Density (kg/m³)
Pressure (kPa)

-30 0.03
Expected (pres))
FDS (pres)
-40 0.02
-50
0.01
-60 Expected (H2O)
FDS (H2O)
-70 0
0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14
Time (h) Time (h)

Figure 13.14: Output of the water_evaporation_2 test case.

217
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
1500 120
Enthalpy Change (water_evaporation_3) Relative Humidity (water_evaporation_3)
1000 100

500 80

Humidity (%)
Enthalpy (kJ)

Expected (h_gas)
Exact (h_water) Exact (humid)
0 FDS (h_gas)
60 FDS (humid)
FDS (h_water)
-500 40

-1000 20

-1500 0
0 5 10 15 20 0 5 10 15 20
Time (s) Time (s)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
0.1
Density Change (water_evaporation_3) 500 Temperature (water_evaporation_3)
0.08
400

Temperature (°C)
Density (kg/m³)

0.06 Expected (Temp)

300 Expected (temp_w)
FDS (Temp)
0.04 FDS (temp_w)
200

0.02 100
Expected (dens)
FDS (dens)
0 0
0 5 10 15 20 0 5 10 15 20
Time (s) Time (s)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
0 0.15
Pressure Change (water_evaporation_3) Water Concentration (water_evaporation_3)
-10

-20
0.1
Mass Fraction
Pressure (kPa)

-30
Expected (pres)
FDS (pres)
-40
0.05
-50

-60 Expected (H2O)

FDS (H2O)
-70 0
0 5 10 15 20 0 5 10 15 20
Time (s) Time (s)

Figure 13.15: Output of the water_evaporation_3 test case.

218
13.3.4 Case 4
In the fourth case, air at 500 °C is forced through a 3 m long, 1 m wide, 1 m tall tunnel at 1 m/s. Water
droplets with a median volumetric diameter of 20 µm are introduced in the middle of the tunnel at a rate of
0.05 kg/s, starting at 10 s. Two cases are run with different values of DT_INSERT and N_PARTICLES
while keeping the total rate of mass injected the same. The first case uses DT_INSERT=0.001 s and
N_PARTICLES=1000; the second case, water_evaporation_4_np100, uses DT_INSERT=0.01 s and
N_PARTICLES=100, reducing the total number of particles by a factor of 100. The water temperature is
20 °C. The mass fraction of water vapor in the hot air flowing into the tunnel is 0. The water evaporates
at approximately the same rate at which it is introduced. Figure 13.16, left, displays the time history of the
enthalpy flow out of the tunnel. The enthalpy flow rate of the dry hot air is 229 kW above its ambient value.
The steady-state temperature is 211 °C. FDS reports heat flow as the difference in enthalpy of the flowing
gas at its current temperature from that at ambient temperature. For the steady-state condition of 0.05 kg/s
in the air flow at 211 °C the heat flow is expected to be 107 kW.
FDS-6.8.0-0-g886e009-release
300
Enthalpy Flow (water_evaporation_4)
250
Enthalpy Flow (kW)

200

150

100

Expected (Heat_Flow)
50 FDS np 1000 (heat)
FDS np 100 (heat)
0
0 10 20 30 40
Time (s)

Figure 13.16: Output of water_evaporation_4 test case. The black line indicates the FDS results for the test
case with 1000 particles injected every 0.001 s. The dashed red line shows results for FDS with 100 particles injected
every 0.01 s. Note that the same total mass is injected in both cases.

13.3.5 Case 5
This case is based upon one of the Ranz and Marshall experiments [98] contained in the FDS Validation
Guide [20] in which a single, stationary water droplet of diameter 1043 µm in relatively dry air evapo-
rates in approximately 15 min. The droplet is modeled using the thermally-thin liquid droplet model and
the thermally-thick model usually used for solid particles. For the thermally-thick droplet, the evaporation
model used for liquid pools is applied at the surface of the sphere. A temperature-dependent heat of va-
porization is applied. Fig. 13.17 compares the droplet diameter of the thermally-thick and thermally-thin
models compared to the experimental measurement.

13.3.6 Case 6
This pair of cases tests the conservation of energy between droplets, surfaces, and the gas. It also demon-
strates that one can specify thermophysical properties rather than rely on tabulated data. In both cases, a box
with volume, V = 1 m3 , is filled with gas with the following properties: mg,0 = 1 kg, c p,g = 1 kJ/(kg · K),

219
 10 -3 FDS-6.8.0-0-g886e009-release
1.2
(water_evaporation_5)
1

0.8

Diameter (m)
0.6

0.4
Measured
0.2 FDS thermally thin
FDS thermally thick
0
0 200 400 600
Time (s)

Figure 13.17: Output of water_evaporation_5 test case.

W = 28 g/mol, Tg,0 = 373.15 K. The box is filled with droplets with total mass, md = 0.1 kg, specific heat, cl
= 2 kJ/(kg · K), and heat of vaporization, hv = 173.15 kJ/kg. The droplets evaporate into the same gas as the
background. The droplets have an initial temperature, Td,0 = 293.15 K, and a boiling point, Tb = 373.15 K.
In case 6, all walls of the box except the floor are adiabatic. The floor is thermally thin (e.g., high
conductivity compared to its thickness), with a perfectly insulated backing. It has a total mass, mw = 1 kg,
a specific heat, cw = 1 kJ/(kg · K), and an initial temperature, Tw,0 = 523.15 K. After approximately 100 s,
all the droplets evaporate, the wall and the gas temperatures converge to a single final temperature, Tg,f , and
the pressure rises to pf (see Fig. 13.18). The final gas temperature can be found by solving the equation,
∆E = ∆H − V ∆p, where ∆E is the energy transferred from the hot floor to the gas, ∆H is the increased
enthalpy of the gas due to the convective heating of the floor and the evaporation of the water droplets, and
∆p is the change in gas pressure.

∆E = mw cw (Tw,0 − Tg,f )
 
∆H = mg,0 c p,g (Tg,f − Tg,0 ) + md cl (Tb − Td,0 ) + c p,g (Tg,f − Tb ) + hv

R

∆p = mg,f Tg,f − mg,0 Tg,0 ; R = 8.314 J/mol/K
WV
The final temperature is calculated to be 445.2 K (172.0 ◦ C) and the pressure increase is 34.6 kPa.
In case 6a, the floor is also set to adiabatic. For a fixed heat flux surface, FDS does not compute heat
transfer between droplets and the surface as this can lead to instabilities. In this case the drops fall to the
floor and begin evaporating in the hot gas without transferring heat to the walls. The internal energy of the
box should change an equal and opposite amount as the enthalpy change to the droplets. The drops form a
cold layer at the bottom as heat is transferred from the gas to the droplets. This temperature stratification
means this case would take a very long time to reach equilibrium. For verification purposes the case is
run long enough to capture the initial high rates of heat transfer until the gas around the droplets is in near
equilibrium with the droplets.

13.3.7 Case 7
In this case, a gas burner lies near the bottom of a 1 m by 1 m by 5 m high vertical channel whose walls are
made of a thin sheet of insulated steel. Air at ambient temperature is forced into the bottom of the channel at

220
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
200 40
Temperature (water_evaporation_6) Pressure Rise (water_evaporation_6)
190

180 35
Temperature (°C)

Pressure (kPa)
170
30
160

150
Exact (T_gas) 25
140 Exact (T_wall)
FDS (T_gas) Exact (Pressure)
130 FDS (T_wall) FDS (pres)
20
0 50 100 150 200 0 50 100 150 200
Time (s) Time (s)

Figure 13.18: Output of water_evaporation case 6.

FDS-6.8.0-0-g886e009-release
10
Change in Energy (water_evaporation_6a)

5
Change in Energy (kJ)

Exact (sum)
FDS (u_gas)
0 FDS (h_liquid)
FDS (sum)

-5

-10
0 5 10 15 20
Time (s)

Figure 13.19: Output of water_evaporation_6a test case.

1 m/s. Monodisperse water droplets with a diameter of 2,000 µm are introduced via a nozzle in the middle
of the channel at a rate of 1 L/min, starting at 60 s. The water temperature is 20 °C, and the spray is directed
at the walls with an initial velocity of 5 m/s. The water completely evaporates before it drips down to the
bottom of the channel. Figure 13.20 displays the energy balance for this case. The heat release rate, HRR,
of the fire is expected to be 384 kW. The rate at which the water droplets extract energy from the system,
Q_PART, is expected to be

− (4.189 kJ/(kg · K) × 80 K + 2269 kJ/kg) × 1/60 kg/s = −43.4 kW (13.22)

The sum of all the terms, Q_TOTAL, is expected to be zero. The three other quantities, Q_COND, Q_RADI,
Q_CONV, all have plausible values, but there is no way to determine the exact values.

13.3.8 Case 8
This case is similar to Case 5 except that here the droplet has a diameter of 100 µm and it is dropped with an
initial velocity of 3 m/s. The measurements are described in Li and Chow [99]. The specific case is shown
in Fig. 6 of the paper. The simulation consists of an injection of a 100 µm droplet at 10 ◦ C into 60 ◦ C air

221
 FDS-6.8.0-0-g886e009-release
500
400 Energy Budget (water_evaporation_7)

300 Expected (HRR)

Heat Loss/Gain (kW)

Expected (Q_TOTAL)
200 Expected (Q_PART)
FDS (HRR)
100 FDS (Q_COND)
FDS (Q_TOTAL)
0 FDS (Q_PART)
FDS (Q_RADI)
-100 FDS (Q_CONV)

-200
-300
0 50 100 150 200 250 300
Time (s)

Figure 13.20: Output of water_evaporation case 7.

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
4 120
(water_evaporation_8) (water_evaporation_8)
3.5 100
3
80
Diameter (µm)
Position (m)

2.5 Measured
FDS thermally thin 60
2 FDS thermally thick

40
1.5
Measured
1 20 FDS thermally thin
FDS thermally thick
0.5 0
0 5 10 15 0 5 10 15
Time (s) Time (s)

Figure 13.21: Output of water_evaporation case 8.

with an initial water vapor mass fraction of 10 %. The initial speed of the droplet is 3 m/s in the negative z
direction (downward with gravity turned on).

13.4 Water Droplets on Solid Surfaces

13.4.1 Basic Flow and Mass Conservation (geom_sprk_mass)
In the test case called Sprinklers_and_Sprays/geom_sprk_mass.fds, two identical sprinklers pour
water on two identical obstructions, one of which is created using the traditional OBST input parameters,
the other is created using the immersed boundary GEOM parameters. The objective is simply to ensure that
the droplets behave the same way in both cases. The sprinklers operate for 4 s, each discharging 10.5 L of
water. The HORIZONTAL_VELOOCITY of the droplets on the solid surfaces is set to a low value of 0.05 m/s
to allow for all of the water to pool before dripping over the edge. Figure 13.22 displays the accumulated
mass of water on each obstruction as a function of time, demonstrating equivalent behavior by both types of
solid obstructions.

222
 FDS-6.8.0-0-g886e009-release
15
Accumulated Water (geom_sprk_mass)

10

Mass (kg)
Expected
FDS (m OBST)
FDS (m GEOM)

0
0 10 20 30 40 50
Time (s)

Figure 13.22: Mass of water pooled on two solid obstructions.

13.4.2 Obstructions of Different Orientations (geom_particle_cascade_2)

In the test case called Complex_Geometry/geom_particle_cascade_2.fds, four identical sprinklers
pour water on four obstructions, one of which is created using the traditional OBST input parameters, the
other three are created using the immersed boundary GEOM parameters. The objective is simply to ensure
that the droplets drip off the obstructions appropriately. The sprinklers operate for 10 s, each discharging
10 L of water. Figure 13.23 displays the accumulated mass of water on the floor as a function of time,
demonstrating equivalent behavior by different types of solid obstructions.
FDS-6.8.0-0-g886e009-release
12
Droplet Mass (geom_particle_cascade_2)
10

8
Mass (kg)

4 Exact
Case 1
Case 2
2 Case 3
Case 4
0
0 5 10 15
Time (s)

Figure 13.23: Mass of water flowing over four obstructions.

13.4.3 Adiabatic Surface Temperature (adiabatic_surf_temp_spray)

This test case is a modified version of Radiation/adiabatic_surface_temperature. A thin metal
plate with insulated backing is heated for 100 s. Adiabatic surface temperature (AST) is measured from the
surface of the plate. Figure 13.24 shows that after about 40 s, the plate temperature and AST match each
other. At 50 s, a water spray is turned on and it starts to cool down the surface. At 100 s, the heater is turned
off, and the plate cools down even more. AST and plate temperatures should match each other in the end of
every 50 s period.

223
 FDS-6.8.0-0-g886e009-release
400
Plate Temp vs AST
(adiabatic_surf_temp_spray)
300

Temperature (°C)
Plate (temp)
200 AST

100

0
0 50 100 150
Time (s)

Figure 13.24: Adiabatic and true surface temperatures under radiative heating (0 - 50 s), heating and spray cooling
(50-100 s) and only spray cooling (100-150 s).

13.5 Flow Rate and Pipe Networks (flow_rate_2)

In this test case four sprinklers are placed in a room. The first three sprinklers are on one pipe network
(PIPE_INDEX = 1) and the fourth sprinkler is on a second pipe network (PIPE_INDEX = 2). The sprin-
klers are defined with a pressure ramp such that all of the sprinklers within a pipe network flow at a combined
rate of 10 L/min regardless of the number of open heads. The sprinklers are set to open every 15 s and to
close at 60 s. After 45 s, the first three sprinklers (black line) have opened for a combined flow rate of
10 L/min and an accumulated water mass of 7.5 kg. At 60 s, the fourth sprinkler (red line) flowing at 10
L/min plus the first three sprinklers flowing at 10 L/min have produced 12.5 kg of water. The results are
shown in Fig. 13.25.
FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
5 15
Open Nozzles (flow_rate) Accumulated Water (flow_rate)
4

10
Open Nozzles

3
Mass (kg)

2
5

1
Analytical (Water)
FDS (Water)
0 0
0 10 20 30 40 50 60 70 0 10 20 30 40 50 60 70
Time (s) Time (s)

Figure 13.25: Output of the flow_rate_2 test case.

13.6 Turbulent Dispersion (random_walk)

Two cases are set up to test the turbulent dispersion of tracer particles. Tracer particles are injected into
a channel with a steady laminar flow with mean velocity 1 m/s. In the first case (random_walk_1),
the diffusivity is set to D = 0.0834 m2 /s (this value results from setting the dynamic viscosity to µ =

224
0.1 kg/(m·s) with an ambient density of ρ = 1.199 kg/m3 and a Schmidt number of unity). In the second
case (random_walk_2), the diffusivity is set to D = 0.00834 m2 /s. The results are compared with Einstein’s
classic result for Brownian motion [100]:

n exp −z2 /4Dt

f (z,t) = √ (13.23)
4πD t

where f (z,t) is the particle number density at lateral distance z from the source at time t, n is the total
number of particles in the system. A virtual phase doppler particle analyzer (PDPA) is set up to measure
the FDS number concentration at the outlet of the channel as a function of height, z. The outlet is 15 m
downstream of the point source of tracer particles. With a mean velocity of 1 m/s, the time is taken as t = 15
s. The FDS results are compared with Eq. (13.23) in Fig. 13.26.
FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
0.4 1
Turbulent Dispersion (random_walk_1) Turbulent Dispersion (random_walk_2)
0.8
Normalized Number Density

Normalized Number Density

0.3

0.6
Expected Expected
0.2 FDS FDS
0.4

0.1
0.2

0 0
0 2 4 6 8 10 0 2 4 6 8 10
Position (m) Position (m)

Figure 13.26: Turbulent dispersion of tracer particles random_walk test cases.

225
Chapter 14

Heating, Ventilating, and Air Conditioning

(HVAC)

14.1 Duct Flow Losses (HVAC_flow_loss)

The pressure drop due to steady flow in a duct is given by
Kρ u2
∆p = (14.1)
2
where K is the loss coefficient for the duct, ρ is the gas density, and u is the mean velocity. In this example,
a duct is split at a tee joint into two ducts. All ducts have the same cross sectional area. The primary duct
is assigned a constant flow rate of 0.3 m3 /s. The other two ducts are assigned loss coefficients of 4 and 16
and outlets at the same elevation. From Eq. (14.1), it can be shown that since the two branch ducts have the
same pressure at their outlets, the steady-state velocities in the three ducts should be 3 m/s, 2 m/s, and 1 m/s.
FDS results are shown in Fig. 14.1.
FDS-6.8.0-0-g886e009-release
5
Ideal 1 m/s
Ideal 2 m/s
Velocities (HVAC_flow_loss)
4 Ideal 3 m/s
FDS 1 m/s
FDS 2 m/s
Velocity (m/s)

3 FDS 3 m/s

0
0 2 4 6 8 10
Time (s)

Figure 14.1: Duct velocities in three ducts with different loss coefficients.

14.2 Node Flow Losses (HVAC_tee_loss_1, HVAC_tee_loss_2)

These two cases use the same geometry as in Sec. 14.1. However, instead of losses assigned to the ducts, the
losses are assigned to the duct nodes. In the first case, HVAC_tee_loss_1, the flow goes into the primary

227
leg of the tee. The secondary branches are assigned loss coefficients of 15 and 3 and the discharge nodes are
assigned losses of 1. This gives total losses of 16 and 4. In the second case, HVAC_tee_loss_2, the flow
goes out the primary leg of the tee. Duct directions are reversed so the sign on velocity remains the same as
in the first case. The same flow losses are assigned. In the first case, the two exhaust ducts are expected to
have the same losses as in Sec. 14.1 and the duct velocities should be the same (3 m/s, 2 m/s, and 1 m/s). In
the second case, the exhaust ducts now function as supply ducts, and the supply duct now functions as an
exhaust duct. This means that only the duct with the specified sees the pressure drop imposed by the tee as
the other two ducts are now upstream of the tee. Since those ducts do not have any losses other than the node
losses of 1, both ducts have the same total loss and see the same pressure drop. In this case the expected
flows are 3 m/s, 1.5 m/s, and 1.5 m/s. FDS results are shown in Fig. 14.2.
FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
5 5
Ideal 1 m/s Ideal 1.5 m/s A
Ideal 2 m/s
Velocities (HVAC_tee_loss_1) Ideal 1.5 m/s B
Velocities (HVAC_tee_loss_2)
4 Ideal 3 m/s 4 Ideal 3 m/s
FDS 1 m/s FDS 1.5 m/s A
FDS 2 m/s FDS 1.5 m/s B
Velocity (m/s)

Velocity (m/s)
3 FDS 3 m/s 3 FDS 3 m/s

2 2

1 1

0 0
0 2 4 6 8 10 0 2 4 6 8 10
Time (s) Time (s)

Figure 14.2: Duct velocities with varied node loss. HVAC_tee_loss_1 (left) and HVAC_tee_loss_2 (right).

14.3 HVAC Mass Conservation (HVAC_mass_conservation)

A 1 m by 1 m by 1 m compartment is initialized with the lower half of the compartment being filled with
a species called SPEC1 having the same molecular weight as air, see Fig. 14.3. An HVAC system of two
ducts is created with one duct taking suction from the lower half (red vent) and discharging into the upper
half (blue vent), and the second duct taking suction from the upper half (green vent) and discharging into
the lower half (yellow vent). The two ducts are assigned the same volume flow rate. It is expected that the
total mass in the compartment will be conserved, that the mass of the second species will be conserved (at
50 % of the total mass), and that at the start of the computation a slice file of species will show pure ambient
being discharged in the lower half and pure SPEC1 being discharged in the upper half.

228
 FDS-6.8.0-0-g886e009-release
1.5
Total Mass (HVAC_mass_conservation)

Mass (kg)
0.5
Ideal Species 1
Ideal Total
FDS Species 1
FDS Total
0
0 2 4 6 8 10
Time (s)

Figure 14.3: SPEC1 mass fraction at 1 s (left) and SPEC1 and total mass (right).

14.4 HVAC Energy Conservation and Pressure (HVAC_energy_pressure)

A 1 m by 1 m by 1 m box is divided into two equal compartments by a vertical wall, see Fig. 14.4. The
ambient species is defined with a molecular weight of 28 g/mol, a specific heat of 1 kJ/(kg · K) and a specific
enthalpy of 293.15 kJ/kg at the reference temperature of 293.15 K. The left side of the compartment is
initialized to twice the ambient temperature. A simple HVAC system of ducts is created. The first duct
takes suction from the lower left half and discharges into the lower right half and is given an area of 0.1 m2
and fixed volume flow rate 0.1 m3 /s. The second duct is placed in the upper half of the compartment and
is assigned a flow loss of 1 and an area of 0.01 m2 . It is expected that the top and bottom duct velocities
will be 10 m/s and 1 m/s respectively, the total energy will be conserved, that at the start of the computation
hot gas will be seen entering the lower right and cold gas entering the upper left, and that at the end of the
computation the pressure difference between the two sides will be 56.3 Pa based on Eq. (14.1)

229
 FDS-6.8.0-0-g886e009-release
15
Velocities (HVAC_energy_pressure)

10

Velocity (m/s)
Ideal Top
Ideal Bottom
FDS Top
FDS Bottom
5

0
0 5 10 15 20
Time (s)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
400 70
Total Enthalpy (HVAC_energy_pressure) Pressures (HVAC_energy_pressure)
65
380

60
Enthalpy (kJ)

Pressure (Pa)
360
55
340
50

320
Ideal Enthalpy
45 Ideal Delta P
FDS Enthalpy FDS Delta P
300 40
0 5 10 15 20 0 5 10 15 20
Time (s) Time (s)

Figure 14.4: Clockwise from top left: Temperature at 1 s, duct velocities, total enthalpy in the domain, and duct node
pressures for the top duct.

14.5 HVAC Dampers (HVAC_damper)

The HVAC_damper example case is a simple duct network of three ducts connected by a tee. A damper is
present in one of the ducts and linked to a DEVC that will change state at 10 s. With the damper open, the
respective duct flow speeds should be 3 m/s, 1 m/s, and 2 m/s. When the damper closes, the respective duct
flow speeds should be 3 m/s, 3 m/s, and 0 m/s. Results of this are shown in Fig. 14.5.

230
 FDS-6.8.0-0-g886e009-release
5
Ideal Duct 1
Ideal Duct 2
Velocities (HVAC_damper)
4 Ideal Duct 3
FDS Duct 1
FDS Duct 2

Velocity (m/s)
3 FDS Duct 3

0
0 5 10 15
Time (s)

Figure 14.5: Simple duct network with a damper, case HVAC_damper.

14.6 HVAC Filters (HVAC_filter)

The sample case HVAC_filter demonstrates the use of the filter input. A 1 m3 compartment is initialized
with a particulate species with a mass fraction of 0.001. A 100 % efficient filter with a clean loss of 1 and
a loading loss of 7732.446 kg−1 (gives a total loss of 10 when all the soot in the compartment is filtered).
A quadratic fan in a 0.01 m2 duct with a maximum pressure of 20 Pa and a maximum flow of 0.2 m3 /s
takes suction from one side of the compartment and discharges into the other. Over time, the filter removes
the particulate from the compartment. However, since the filter loss increases with the mass of particulate
filtered, the rate of removal will decrease over time. Applying the conservation of energy to the compartment
and the steady state duct momentum equation to the duct, we can solve for the compartment temperature and
pressure and the duct velocity. These results along with the loading of the filter and the mass of particulate
in the compartment are shown in Fig. 14.6.

231
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
0 22
Pressure Change (HVAC_filter) Temperature (HVAC_filter)
21
Pressure Change (Pa)

-50

Temperature (°C)
20
Ideal Pressure
FDS Pressure
19
-100

18
Ideal Temperature
FDS Temperature
-150 17
0 100 200 300 400 500 0 100 200 300 400 500
Time (s) Time (s)

FDS-6.8.0-0-g886e009-release 10 -3 FDS-6.8.0-0-g886e009-release
2 1.4
Duct Velocity (HVAC_filter) Particulate Mass (HVAC_filter)
1.2
1.9
1

Particulate Mass (kg)

Duct Velocity (m/s)

Ideal Filter
0.8
Ideal Compartment
1.8 FDS Filter
0.6 FDS Compartment

0.4
1.7
Ideal Velocity 0.2
FDS Velocity
1.6 0
0 100 200 300 400 500 0 100 200 300 400 500
Time (s) Time (s)

Figure 14.6: Compartment pressure (upper left), compartment temperature (upper right), duct velocity (lower left),
and particulate mass in the compartment and on the filter (lower right).

14.7 HVAC Fans (fan_test)

This case tests a quadratic fan using the default approach to finding the fan operation point. Two compart-
ments share a common wall. Both compartments are considered as separate “pressure zones.” Two, 0.04 m2
HVAC ducts are defined. One duct has a quadratic fan with a maximum volumetric flow of 0.16 m3 /s and a
stall pressure of 10 Pa. The second is an open duct with a flow loss of 10. The volume flow through the fans
is obtained from the equation below:
s
|∆p − ∆pmax |
V̇fan = V̇max sign(∆pmax − ∆p) (14.2)
∆pmax

V̇max = 0.16 m3 /s is the maximum output of the fan from Zone 1 to Zone 2 and 0.2 m/s is the flow speed
from Zone 2 to Zone 1. ∆pmax = 10 Pa is the maximum pressure difference at which the fan can operate.
In steady state, the volume flow from compartment to compartment should be equal and opposite in sign.
This occurs when the positive pressure added by the fan equals the pressure drop due to the flow loss, K,
through the return duct. The volume flow through the duct for a given pressure drop can be found from the
duct momentum equation ignoring all terms but the flow loss:
s
2∆p
V̇fan = Aduct (14.3)
ρK

232
Setting the above equation equal to Eq. (14.2), the pressure drop can be determined:
s r
2|p2 − p1 | 2 |p1 − p2 − 10 Pa|
3
(0.04 m ) = (0.16 m3 /s) (14.4)
(1.2 kg/m )10 10 Pa

The solution is p2 = 4.5 Pa and p1 = −4.5 Pa (see Fig. 14.7). Note that the sign of the volume flow in the
duct indicates whether the flow is moving from the first node to the second (positive) or the second node to
the first (negative). This convention can make these types of calculations a bit tricky.
FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
8
6 Pressure (fan_test) 0.06 Volume Flow (fan_test)

4 0.04

Volume Flow (m³/s)

2 0.02
Pressure (Pa)

Ideal (pres_1) Ideal (vflow_1)

Ideal (pres_2) Ideal (vflow_2)
0 FDS (pres_1)
0 FDS (vflow_1)
FDS (pres_2) FDS (vflow_2)
-2 -0.02
-4 -0.04
-6 -0.06
-8 -0.08
0 10 20 30 40 50 60 0 10 20 30 40 50 60
Time (s) Time (s)

Figure 14.7: Compartment pressure and volume flow for the fan_test case.

14.8 HVAC Operation Point Single Fan (qfan_test)

This case tests a quadratic fan using the system curve approach to finding the fan operation point. Two
compartments share a common wall. Both compartments are considered as separate “pressure zones.” Two,
0.04 m2 HVAC ducts are defined. One duct has a quadratic fan with a maximum volumetric flow of 0.16 m3 /s
and a stall pressure of 10 Pa. Both ducts have a flow loss of 5. At each time step system curve is generated
by fitting a quadratic to two points. The first point is using the current compartment pressure difference and
solving for the system flow with the fan off (V̇system 0 ).
s
2|∆p|
V̇system 0 = Aduct sign(∆p) (14.5)
ρK

The second point applies the maximum fan pressure and determines the volume flow with the fan ap-
plying its full force on the flow (V̇system max ).
s
2|∆p − ∆pmax |
V̇system max = Aduct sign(∆p − ∆pmax ) (14.6)
ρK

The operation point for the fan at each timestep will be where the fan curve intersects a quadratic
between the two points just defined. At steady-state the operation point flow rate through the duct with
the fan must match the flow rate through the duct without the fan. If it doesn’t, then pressures in the
compartments will change and the condition won’t be steady-state. This sets up two functions. One function

233
is the operation point as a function of the pressure difference between the two compartments. The second
function is the flow through the duct without a fan as a function the pressure diffrence between the two
compartments. These two functions will equal each other at a pressure difference of 4.52 Pa split equally
between the compartments and a flow rate of 0.0491 m3 /s.
FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
4
Pressure (qfan_test) 0.06 Volume Flow (qfan_test)

2 0.04

Volume Flow (m³/s)

0.02
Pressure (Pa)

Ideal (pres_1) Ideal (vflow_1)

Ideal (pres_2) Ideal (vflow_2)
0 FDS (pres_1)
0 FDS (vflow_1)
FDS (pres_2) FDS (vflow_2)
-0.02
-2 -0.04
-0.06
-4 -0.08
0 5 10 15 20 0 5 10 15 20
Time (s) Time (s)

Figure 14.8: Compartment pressure and volume flow for the qfan_test case.

14.9 HVAC Operation Point Multiple Fans (qfan_multi)

This case tests the system curve approach for the cases a multiple fans in a duct run. Two duct runs are
define. In one case two fans take parallel suction and discharge through a combined exhaust. In the second
case there are two fans in series. All ducts are defined with an area of 0.04 m2 and a loss of 5. In each case
two different quadratic fans are used. The first fan has a maximum volumetric flow of 0.2 m3 /s and a stall
pressure of 10 Pa. The first fan has a maximum volumetric flow of 0.6 m3 /s and a stall pressure of 30 Pa.
Since the domain is open, for all fans the zero fan pressure flow will be 0.0 m3 /s. With two fans in each
duct run, two full fan pressure points are created, one for each fan. When the first fan is being solved for,
the pressure of the second fan is set to zero and vice versa. This means in the parallel case that the system
solution will mean some air exhausting through the duct branch with the fan that is off.
In the parallel case the two full fan system points are (10 Pa, 0.0653 m3 /s) and (30.0 Pa, 0.113 m3 /s).
In the series case the values are (10 Pa, 0.0517 m3 /s) and (30 Pa, 0.0895 m3 /s). The parallel case is higher
since the effective exhaust duct area is 0.2 m2 rather than 0.1 m2 .
The fan pressures that intersect the system curves are 9.03 Pa and 29.0 Pa for the parallel case and
9.37 Pa and 29.4 Pa for the serial case. For the serial case this results in a system flow of 0.102 m3 /s. In the
case of the parallel fans, the larger fan actually pushes flow out the smaller fans inlet. The predicted flows
are -0.0258 m3 /s for the small fan, 0.0999 m3 /s for the large fan, and 0.0741 m3 /s out the exhaust duct.

234
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
0.15 0.15
Parallel Volume Flow (fan_multi) Series Volume Flow (fan_multi)
0.1
Volume Flow (m³/s)

Volume Flow (m³/s)

Ideal (vflow_1) 0.1
0.05 Ideal (vflow_2)
Ideal (vflow_e) Ideal (vflow)
FDS (vflow_1) FDS (vflow)
0 FDS (vflow_2)
FDS (vflow_e) 0.05

-0.05

-0.1 0
0 5 10 15 20 0 5 10 15 20
Time (s) Time (s)

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
40 40
Parallel Fan Pressure (fan_multi) Series Fan Pressure (fan_multi)

30 30
Pressure (Pa)

Pressure (Pa)
Ideal (fan pres_1) Ideal (fan pres_1)
Ideal (fan pres_2) Ideal (fan pres_2)
20 FDS (fan pres_1)
20 FDS (fan pres_1)
FDS (fan pres_2) FDS (fan pres_2)

10 10

0 0
0 5 10 15 20 0 5 10 15 20
Time (s) Time (s)

Figure 14.9: Volume flow and fan pressure for the qfan_multi. Left is two fans in parallel, and right is two fans in
series. Top is volume flow, and bottom is fan pressure.

14.10 Leakage
In the following examples, both leaks and fans are demonstrated.

14.10.1 Single Zone Leakage (leak_test)

Two compartments (each 0.9 m by 1 m by 1 m) are connected by a small fan and a leak at the floor. The
volume flow rate of the fan is given by the “fan curve”:
s
|∆p − ∆pmax |
V̇fan = AductUmax sign(∆pmax − ∆p) (14.7)
∆pmax

where ∆p is the difference in pressure and Aduct = 0.1 m2 , Umax = 0.6 m/s, and ∆pmax = 300 Pa. The volume
flow due to the leak is given by: s
2∆p
V̇leak = Aleak (14.8)
ρ

where Aleak = 0.002 m2 and ρ is the density at the discharge side. Equating the fan and leak volume flow
rates, we can solve for ∆p. The results are shown in Fig. 14.10.

235
 FDS-6.8.0-0-g886e009-release
200
Pressure (leak_test)

195

Pressure Difference (Pa)

190

185
Ideal
FDS
180
0 5 10 15 20 25 30
Time (s)

Figure 14.10: Pressure difference between two compartments with leakage connected by a fan.

14.10.2 Multiple Zone Leakage (leak_test_2)

This case starts with the same input as leak_test. A second leakage vent is added to the floor of the
discharge side of the fan. This vent leaks to the ambient. When the simulation starts the fan pushes air from
one compartment to the other. As the pressure rises, the leakage increases until the leakage flow rate is equal
to the fan flow rate. Since some air from the interior compartment is being discharged to the ambient, the
average pressure of the two compartments decreases until the pressure of the interior compartment is equal
to that of the ambient. At this point in time, there is no more leakage to the outside and the fan flow rate
equals the leakage flow rate between the two compartments. Given the leakage area and the fan curve, this
occurs when the pressure difference between the two compartments is the same as the pressure difference
from leak_test, only now the discharge side should be at 0 Pa.
FDS-6.8.0-0-g886e009-release
50
Pressure (leak_test_2)
0
Pressure (Pa)

-50 Ideal (pres_1)

Ideal (pres_2)
FDS (pres_1)
-100 FDS (pres_2)

-150

-200
0 5 10 15 20 25 30
Time (s)

Figure 14.11: Vector plot of pressure at 10 s with geometry clipped to show compartment interiors (left) and a plot of
zone pressures (right).

14.10.3 Leakage through Immersed Boundary Wall (leak_test_3)

This case is intended to ensure that an immersed boundary obstruction (GEOM) handles leakage in the same
way as a simple grid-conforming obstruction (OBST). Two rooms are separated by a wall specified via either
an OBST line or a GEOM line. Air is pumped into the room on one side of the wall at a rate of V̇ = 0.02 m3 /s,

236
and the wall is assigned a leakage area AL = 0.001 m2 . The room on the other side of the wall leaks to the
outside through a hole of equal area. The pressure rise in rooms 1 and 2, ∆p1 and ∆p2 , are found by solving:
s s
2|∆p1 − ∆p2 | 2|∆p2 |
V̇ = AL ; V̇ = AL (14.9)
ρ0 ρ0

The solution is that ∆p1 = 480 Pa and ∆p2 = 240 Pa, as shown in Fig. 14.12.
FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
600 600
Pressure (leak_test_3) (OBST) Pressure (leak_test_3) (GEOM)
500 500

400 400
Pressure (Pa)

Pressure (Pa)
300 300

200 200
Ideal (p1) Ideal (p3)
Ideal (p2) Ideal (p4)
100 FDS (p1)
100 FDS (p3)
FDS (p2) FDS (p4)
0 0
0 5 10 15 20 25 30 0 5 10 15 20 25 30
Time (s) Time (s)

Figure 14.12: Pressure rise in adjacent compartments separated by an OBST (left) and GEOM (right) wall.

14.10.4 Leakage through Immersed Boundary Compartment (leak_test_4)

This case is intended to ensure that a simple compartment constructed of immersed boundary obstructions
(GEOM) handles leakage in the same way as a simple grid-conforming compartment. Air is pumped into the
compartment via a small rotated cube inside at a rate of V̇ = 0.024 m3 /s, and the walls are assigned a total
leakage area AL = 0.001 m2 . The pressure rise, ∆p is found by solving:
s
2|∆p|
V̇ = AL ; ρ0 = 1.195 kg/m3 (14.10)
ρ0

The solution is that ∆p = 344.16 Pa, as shown in Fig. 14.13.

14.10.5 Localized Leakage (leak_enthalpy)

A domain that is 2 m by 1 m by 1 m with 10 cm grid resolution is divided into two halves by a 20 cm thick
wall. In the left side is given a 100 kW/m3 heat source in its entire volume that lasts for 4 s. A localized
leakage path that is 0.001 m2 is defined for one grid cell at the center on either side of the dividing wall. The
parameter LEAK_PATH is set to true. All the boundaries are defined as adiabatic. The heat source creates
pressure on the left side which force heated on into the right side. The localized leakage should conserve
mass, and the energy source should add 360 kJ to the domain. Results are shown in Fig. 14.14

237
 FDS-6.8.0-0-g886e009-release
500
Pressure (leak_test_4)
400

Pressure (Pa)
300

200

100
Ideal
FDS
0
0 5 10 15 20 25 30
Time (s)

Figure 14.13: Pressure rise in a compartment constructed of immersed boundary obstructions.

-4
10
1
Ideal Ideal
FDS 400 FDS

Change in Energy (kJ)

Total Mass (HVAC_leak_enthalpy)
Change in Mass (kg)

300

0
200

100
Internal Energy (HVAC_leak_enthalpy)
-1 0
0 1 2 3 4 5 0 1 2 3 4 5
Time (s) Time (s)

Figure 14.14: Change in mass (left) and energy (right) in the domain.

14.10.6 Multiple Zone, Multiple Mesh Leakage (zone_shape_2)

A duct with a square 1 m cross section winds its way through a 2 by 2 by 2 array of meshes that span a
sealed 6.2 m cubic domain. A fan forces air into one end of the duct at a rate of V̇ = 0.1 m3 /s. The other end
of the duct has a leak with effective area AL = 0.007746 m2 . The expected pressure difference between the
interior and exterior of the duct is  2
ρ V̇
∆p = ≈ 100 Pa (14.11)
2 AL
The overall mass in the domain should not change. Figure 14.15 displays the pressure difference and overall
mass.

238
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
150 400
Pressure Difference (zone_shape_2) Mass (zone_shape_2)

300
100
Pressure (Pa)

Mass (kg)
200

50
100
Ideal (Delta p) Ideal (Mass)
FDS (Delta p) FDS (Total)
0 0
0 50 100 150 200 250 300 0 50 100 150 200 250 300
Time (s) Time (s)

Figure 14.15: (Left) Pressure difference between interior and exterior of the duct. (Right) Overall mass.

14.10.7 Leak Pressure Exponent (HVAC_leak_exponent)

As the interior pressure rises in a typical building, the leakage area will grow as small gaps, cracks, and other
leakage paths open up. Leakage tests performed according to test standards such as ASTM E779 provide
two additional data points to quantify this behavior. These are the LEAK_PRESSURE_EXPONENT and the
LEAK_REFERENCE_PRESSURE. The use of these additional inputs are shown in the equation below as n and
∆pref respectively where AL,ref is given by LEAK_AREA. Additionally, a discharge coefficient, Cd , is often
applied in practice. s
|∆p| n−0.5 2|∆p|
 
V̇L = Cd AL,ref (14.12)
∆pref ρ0
By default, Cd = 1, n = 0.5 and ∆pref = 4 Pa, meaning that the leak area will not change with pressure unless
you specify an exponent other than 0.5.
The first verification case involves three 1000 m3 compartments that have 0.01 m2 of leakage to the
ambient. Each compartment is supplied with 0.16 m3 /s of inlet flow. The first compartment has the default
values for the exponent and reference pressure, the second changes the exponent to 0.6, and the third third
changes exponent and the reference pressure respectively to 0.6 and 10 Pa. The expected leakage velocities
for these conditions are 160 m/s, 11.8 m/s, and 12.8 m/s. Note in the left plot of Fig. 14.16 that the velocities
are negative because the FDS output for the leakage velocity is from the lower numbered pressure zone (in
this case ambient) to the higher number (inside the compartment). Negative velocity indicates flow from
inside to outside as expected.
The second case involves a steel enclosure with four localized leaks, two near the bottom and two near
the top. The air flow into the compartment is ramped up slowly, and the right hand plot of Fig. 14.16
shows the ideal and predicted relationship between the volume flow and compartment pressure. In this case,
Cd = 0.61, n = 0.6, ∆pref = 1 Pa, and AL,ref = 0.18 at each of the four “cracks.”

239
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
0 0.12
Ideal 1
Ideal 2
Velocities (HVAC_leak_exponent) Leakage (HVAC_leak_exponent_2)
Ideal 3
0.1
-5 FDS 1

Volume Flow (m³/s)

FDS 2 0.08
Velocity (m/s)

FDS 3

-10 0.06

0.04
-15
0.02 Ideal
FDS
-20 0
0 20 40 60 80 100 0 2 4 6 8
Time (s) Pressure Rise (Pa)

Figure 14.16: (Left) Leakage velocities resulting from varying LEAK_PRESSURE_EXPONENT and
LEAK_REFERENCE_PRESSURE. (Right) Volume flow versus pressure for a leaking steel enclosure.

14.11 ASHRAE Sample Problem 7

(ashrae_7_fixed_flow, ashrae_7_quadratic, ashrae_7_table)
ASHRAE Sample Problem 7 [101] is a simple HVAC system that represents a metal working exhaust system
for a machine shop where particulates from three pieces of equipment are removed by a dust collector. The
network, shown in Fig. 14.17, consists of three inlets, two tees, the dust collector, and a fan. The lengths,
diameters, and friction losses for all the components are specified, as well as the fan flow rate. The pressure
drops and flow rates can be computed by hand. Also listed in Fig. 14.17 are the predicted pressure drops for a
fan with a fixed flow rate, a fan with a quadratic fan curve (MAX_FLOW=2.4094, MAX_PRESSURE=3000.)
that includes the fixed flow state (i.e., the flow rate of 1.45 m3 /s and pressure drop of 1913.5 Pa from the
fixed flow simulation sits on the quadratic curve), and a fan with a RAMP to define the same quadratic fan
curve at 500 Pa intervals. The hand calculations assume a constant density, whereas the FDS computation
will have a slightly non-uniform density in the network due to pressure drops. It is expected that there will
be small differences that will be more noticeable at higher pressure changes. Additionally since specifying
a fan curve via a table results in a piecewise linear approximation, it is expected that the user-defined fan
curve results will vary slightly from the quadratic and fixed flow results.

240
 7

Dust
1 5 6 Fan
Collector
4

2 3

Duct Number 1 2 3 4 5 56 6 7
Experiment 716.200 441.200 441.200 275.000 121.300 731.100 32.200 313.200
Fixed Flow 714.326 440.088 440.088 274.255 120.952 728.779 32.111 312.023
Quadratic 715.572 440.854 440.854 274.734 121.164 730.060 32.167 312.564
Table 713.734 439.723 439.724 274.027 120.851 728.170 32.084 311.765

Figure 14.17: Schematic of ASHRAE Sample Problem 7, and tabulated pressure drops (Pa) for a fixed flow fan curve,
quadratic fan curve, and a table defined fan curve.

14.12 HVAC transient mass transport

This section documents verification of the mass transport subroutine of the HVAC module. Unless otherwise
stated, the general test case has a 1 m3 domain split into two equally sized compartments (being their own
pressure zones) connected in some manner by an HVAC network. The left and right compartments are
initialized with species 1 and species 2 respectively. Both species have an initial density of 1.0 kg/m3 . The
ducts have a cross sectional area of 0.01 m2 and a flow rate of 0.01 m3 /s (velocity of 1 m/s).

14.12.1 Convergence study for HVAC mass transport (HVAC_convergence_study_x)

The time-dependent mass transport subroutine is an explicit Euler method solver which uses a Godunov
upwinding scheme. This method is formally first-order accurate in space and should exhibit L1 and L2
norm error decay rates of O(∆x) for a continuous solution.
We undertake a convergence study to demonstrate that the adopted numerical method exhibits the ex-
pected spatial error decay. The two compartments are connected with a single 1 m long duct which is
initialized using linear interpolation between the upstream and downstream duct node. In this manner the
mass fraction of species 1 linearly decreases from 1 to 0 kg/kg. If we examine the species 1 mass fraction at
the downstream node we expect to see it’s value increase linearly from 0 to 1 kg/kg between 0 s to 1 s and
then to remain at 1 kg/kg until the end of the simulation. The mass fraction of species 1 at the downstream
duct node is examined and compared to the analytical result for a range of duct cell numbers (20, 40, 80,
160 and 320). The time step is fixed to satisfy the CFL of the case cell size (CFL = ∆x/2u); this ensures

241
that the ∆x/∆t ratio is fixed for each case. Fig. 14.18 shows the mass fraction at the downstream duct node
for all cases and the analytical solution. Fig. 14.19 gives the L1 and L2 norm errors.
FDS-6.8.0-0-g886e009-release
1

0.8

Mass fraction (kg/kg)

0.6

0.4 Exact solution

FDS N_CELLS = 20
FDS N_CELLS = 40
0.2 FDS N_CELLS = 80
FDS N_CELLS = 160
FDS N_CELLS = 320
0
0 0.5 1 1.5 2
Time (s)

Figure 14.18: Species 1 mass fraction in the downstream duct node for the cell size convergence study. The mass
fraction starts at zero and then linearly increases due to the initial conditions in the duct. At 1 s the mass fraction
reaches its maximum and remains at this value until the end of the simulation. The figure illustrates that at lower cell
counts (blue line) there is increased numerical diffusivity and that, for higher cell counts (cyan line), the numerical
solution converges towards the analytical solution (black dashed line).

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release

-2
10 10
-2
L1 error (kg/kg)

L2 error (kg/kg)

-3
10 10 -3
FDS FDS
O( x) O( x)

-2 -2
10 10
x (m) x (m)

Figure 14.19: L1 (left) and L2 (right) norm errors in mass fraction for the convergence study using various cell
numbers (320, 160, 80, 40, 20 from left to right). The plot demonstrates errors decreasing for increasing cell numbers
(i.e. decreasing cell size). L1 and L2 norm errors decay at O(∆x).

14.12.2 Transport time (HVAC_mass_transport)

A single 10 m long duct connects the two compartments. The duct is initialized using data from the right
duct node (i.e. species 2). As HVAC does not account for diffusion in the duct (only advection) we expect
to see species 1 mass fraction at the downstream duct node start and remain at 0 kg/kg until 10 s and then
instantaneously increase to 1 kg/kg. Ideal result and FDS output is presented in Fig. 14.20.

242
 FDS-6.8.0-0-g886e009-release
1.2
Mass fraction (HVAC_mass_transport)
1

0.8

Mass Fraction
0.6

0.4

0.2 Ideal Species 1

FDS Species 1
0
0 5 10 15 20
Time (s)

Figure 14.20: Species 1 mass fraction in the downstream duct node.

14.12.3 Duct definition sensitivity analysis (HVAC_mass_transport_x)

A sensitivity analysis is carried out to verify that results are not sensitive to differing methods of defining the
same duct network. The general arrangement is identical to that of Sec. 14.12.2. Test cases 1 to 4 incorporate
the use of increasing numbers of DUCTs (one, two, four and five respectively) and affiliated DUCTNODEs but
the number of cells in the discretized duct (N_CELLS), and hence the cell size, remains identical. We look
at species 1 mass fraction at the downstream DUCTNODE and expect to see no variation in the results for the
differing number of DUCTs. Ideal result and FDS output is presented in Fig. 14.21.
FDS-6.8.0-0-g886e009-release
1.2
Mass fraction (HVAC_mass_transport_x)
1

0.8
Mass Fraction

0.6

0.4 Ideal Species 1

FDS case 1 (1 duct)
FDS case 2 (2 ducts)
0.2 FDS case 3 (4 ducts)
FDS case 4 (5 ducts)
0
0 5 10 15 20
Time (s)

Figure 14.21: Species 1 mass fraction in the downstream duct node for different duct definitions. Although different
numbers of ducts are used, the number of cells is identical for all cases. Therefore the output for all cases should also
be identical.

14.12.4 Mass conservation 1 (HVAC_mass_transport_conservation_1)

The two compartments are connected by two ducts, A and B, each being 10 m long with areas of 0.01 m2 .
The ducts are initialized with species 1 with Duct A having flow from left to right and duct B having flow
from right to left. The total mass in the FDS domain and in both compartments is expected to remain the
same. The mass of species 1 in the left compartment is expected to decrease from 0.5 kg to 0.292 kg and in
the right compartment it is expected to increase from 0 kg to 0.292 kg. The spatially mirrored behavior is

243
expected of species 2 with ending values of 0.208 kg. Fig. 14.22 shows mass in the left compartment, right
compartment and in the whole FDS domain. Fig. 14.23 presents the species 1 and 2 masses in the left and
right compartments.
FDS-6.8.0-0-g886e009-release
1.2
Mass (HVAC_mass_transport_conservation_1)
1

0.8
Mass (kg)
0.6
Ideal Mass_total
0.4 Ideal Mass_left
Ideal Mass_right
FDS Mass_total
0.2 FDS Mass_left
FDS Mass_right
0
0 20 40 60 80 100
Time (s)

Figure 14.22: Total mass in the FDS domain and total mass in the left and right compartments.

FDS-6.8.0-0-g886e009-release
0.5
Mass (HVAC_mass_transport_conservation_1)
0.4
Ideal Species 1_left
Ideal Species 1_right
0.3 Ideal Species 2_left
Mass (kg)

Ideal Species 2_right

FDS Species 1_left
0.2 FDS Species 1_right
FDS Species 2_left
FDS Species 2_right
0.1

0
0 20 40 60 80 100
Time (s)

Figure 14.23: Species 1 and 2 masses in the left and right compartments.

14.12.5 Mass conservation 2 (HVAC_mass_transport_conservation_2)

The case setup is identical to that of Sec. 14.12.4 with the exception of the length of duct A is reduced to 5 m
and the length of duct B is increased to 20 m. This gives a total FDS and HVAC domain species 1 : species 2
mass ratio of 0.6 : 0.4. We expect that the species 1 and 2 masses within the FDS domain will come to
equilibrium at this ratio. Fig. 14.24 shows the total masses of species 1 and 2.

14.12.6 Mass conservation 3 (HVAC_mass_transport_conservation_3)

The case setup is identical to that of Sec. 14.12.4 except that both ducts A and B have a length of 12.5 m and
are initialized with species 2. This gives a total domain (FDS and HVAC) species 1 : species 2 mass ratio of

244
 FDS-6.8.0-0-g886e009-release
1
Mass (HVAC_mass_transport_conservation_2)
0.8

0.6

Mass (kg)
0.4

Ideal Species 1
0.2 Ideal Species 2
FDS Species 1
FDS Species 2
0
0 20 40 60 80 100
Time (s)

Figure 14.24: Total species 1 and 2 masses in the FDS domain. The initial equilibrium is prior to species 1 entering
duct B and species 2 entering duct A.

0.4 : 0.6. We expect that the species 1 and 2 masses within the FDS domain will come to equilibrium at this
ratio. Fig. 14.25 presents the ideal and FDS total masses of species 1 and 2.
FDS-6.8.0-0-g886e009-release
1
Mass (HVAC_mass_transport_conservation_3)
0.8

0.6
Mass (kg)

0.4

Ideal Species 1
0.2 Ideal Species 2
FDS Species 1
FDS Species 2
0
0 20 40 60 80 100
Time (s)

Figure 14.25: Total species 1 and 2 masses in the FDS domain. Due to the initial contents of the two ducts (species 2)
the mass ratio of species 1 and 2 is 0.4 : 0.6; therefore at equilibrium this ratio should be observed throughout the
domain.

14.12.7 Transient transport branching ducts (HVAC_mass_transport_branch)

The two compartments are connected by an HVAC network comprising a single 5 m feed duct which
branches into two ducts (ducts A and B) at a tee with both ducts discharging to the right compartment.
Ducts A and B each have a cross-sectional area of 0.005 m2 , half that of the feed duct, and will therefore
both have a velocity of 1 m/s. Duct A has a length of 5 m and duct B has a length of 10 m. The duct net-
work is initialized using data from the right compartment. Based on the duct lengths and a constant velocity
we expect to see the species 1 mass fraction at outlet A and B increase from 0 to 1 kg/kg at 10 s and 15 s
respectively. Ideal result and FDS output is presented in Fig. 14.26.

245
 FDS-6.8.0-0-g886e009-release
1.2
Mass fraction (HVAC_mass_transport_branch)
1

0.8

Mass Fraction
0.6

0.4
Ideal Outlet A
Ideal Outlet B
0.2 FDS Outlet A
FDS Outlet B
0
0 5 10 15 20
Time (s)

Figure 14.26: Species 1 mass fraction in the downstream duct nodes. The distance from the inlet to outlet A and
from the inlet to outlet B is 10 m and 15 m respectively. Given a constant velocity of 1 m/s through the network, the
transport time from inlet to outlets A and B should be 10 s and 15 s respectively.

14.12.8 Transient transport combining ducts (HVAC_mass_transport_combine)

The left compartment is separated into two sub-compartments, A and B. These sub-compartments are ini-
tialized with species 1 and 2 respectively. An inlet vent is located in each sub-compartment and these vents
are connected to separate ducts, of lengths 1 m and 2 m respectively, which combine at an internal tee. A
single duct, with a length of 1 m, connects this tee to the right compartment where it discharges to a single
outlet into the right compartment. The right compartment and the duct network is initialized with species 3.
If we examine species mass fractions at the downstream outlet duct node we expect to see a species 3 mass
fraction of 1 kg/kg for 2 s, a species 1 and 3 mass fraction of 0.5 kg/kg each from 2 s to 3 s and a species 1
and 2 mass fraction of 0.5 kg/kg each from 3 s until the end of the simulation. Ideal result and FDS output
is presented in Fig. 14.27.
FDS-6.8.0-0-g886e009-release
1.2
Outlet mass fraction (HVAC_mass_transport_combine)
1

0.8
Mass Fraction

0.6
Ideal Species 1
0.4 Ideal Species 2
Ideal Species 3
FDS Species 1
0.2 FDS Species 2
FDS Species 3
0
0 1 2 3 4 5
Time (s)

Figure 14.27: Species mass fractions at the outlet duct node. Species 3 starts and remains at 1 kg/kg for 2 s as the
initial contents of the duct flows out of the outlet. At 2 s the species from the sub-compartment connected via the
shorter duct and the remaining initial species from the longer duct flows out of the outlet. At 3 s the species from the
sub-compartment connected by the longer duct reaches the outlet and the mass fraction is split between the species
from the two upstream sub-compartments - no initial species remains in the ducts.

246
14.12.9 Energy conservation and pressure (HVAC_mass_transport_energy)
The left compartment is initialized with a temperature of 100 ◦C, the right compartment remains at ambient
(20 ◦C). Both compartments and the duct network are initialized with a single species. The species has a
specific heat of 1 kJ/(kg · K). The compartments are connected by two separate ducts, duct A and duct B,
both with a length of 1 m. Duct A has a specified flow from left to right with a velocity of 1 m/s, duct B has
no defined flow and a minor losses coefficient of 2. Duct A and B are initialized using gas data from the left
and right compartment duct nodes respectively. Based on Eq. 14.13 enthalpy in the FDS domain should be
293.15 kJ and should remain constant. Based upon Eq. 14.1 the pressure difference between the upstream
and downstream duct node of duct B should be 0.8928 Pa. Duct node pressure difference and total enthalpy
are presented in Fig. 14.28.

V (Tleft
hCFD = c p ρleft + Tright ρright ) (14.13)
2

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
1.2 350
Pressure difference (HVAC_mass_transport_energy) Total enthalpy (HVAC_mass_transport_energy)
1.1

1 Enthalpy (kJ)
Pressure (Pa)

0.9 300

0.8

0.7 Ideal p Ideal Enthalpy

FDS p FDS Enthalpy
0.6 250
0 10 20 30 40 0 10 20 30 40
Time (s) Time (s)

Figure 14.28: Pressure difference between the inlet and outlet of duct B (left). Total enthalpy in the CFD domain
(right).

247
Chapter 15

Wildland Fire Spread

This chapter deals with verification cases which are specific to solving problems related to wildland and
Wildland-Urban Interface (WUI) fires. Currently, some relevant verification tests can also be found in other
sections (light blue links refer to the User’s Guide):

• Convective heating in vegetation: Figure 17.4.4;

• Radiation attenuation by vegetation: Figures 9.11 and 17.4.3;

• Mass and energy conservation in burning vegetation: Figures 17.3 and 17.4.

15.1 Level Set Model

The level set model refers to an approach for solving parameterized fire spread over a surface using a Eule-
rian approach to front tracking [102]. FDS allows for several levels of complexity in level set simulations,
and the sections here deal with verification of relevant aspects.

15.1.1 Elliptical Fire Spread

One possible approach for advancing the fire front, and the default for level set in FDS, is to assume that point
ignitions grow in an elliptical shape. Thus, the spread rate at any point along the front can be determined
by assuming elliptical growth from this point (Huygens’ principle) [103, 104]. The shape of the ellipse is
determined by the Rothermel fire spread equation [105] and empirical assumptions regarding the length-to-
breadth ratio [106] and the head-to-backing fire spread ratio [104]. The orientation of the ellipse at a given
point along the fireline is determined by the angle between the normal vector to the fireline and the vector
summation of the wind and slope direction (with slope treated as an effective wind). Further details of the
relevant equations and the solution technique employed in FDS can be found in Bova et al. [107].
In the simplest configuration of a static uniform wind field (LEVEL_SET_MODE = 1), a uniform slope,
and a uniform fuel it is possible to solve the equations for elliptical spread directly and to calculate the
spread in all directions from a point ignition. This allows verification of the algorithm in FDS. Four cases
are configured to test the influence of wind and slope:

• LS_ellipse_0ms_00deg.fds: no wind and no slope;

• LS_ellipse_5ms_00deg.fds: a 5 m/s wind from the west and no slope;

• LS_ellipse_0ms_30deg.fds: no wind and a 30° slope oriented to the northeast;

249
 • LS_ellipse_5ms_30deg.fds: a 5 m/s wind from the west and a 30° slope oriented to the northeast.

The cases involve a point ignition at (x, y) = (0, 0) and a fuel with an arbitrary zero-wind zero-slope
spread rate (VEG_LSET_ROS_00) of 0.05 m/s. The fuel layer has the default packing ratio (VEG_LSET_BETA)
of 0.01 and surface-to-volume ratio (VEG_LSET_SIGMA) of 5000 m−1 , which are used to calculate the effect
of wind and slope on fire spread [105]. The default fuel height (VEG_LSET_HT) of 1 m is also applied,
which is used to calculate the midflame wind speed from the value of the wind aloft [108].
The shapes of the fire fronts are shown in Figure 15.1. The relative errors between the FDS cases and
the empirical model are shown in Figure 15.2.

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
4 4

3 3

2 2

1 1
y (m)

y (m)
0 0

-1 -1

-2 FDS 0 m/s -2 FDS 0 m/s

Expected 0 m/s Expected 0 m/s
-3 FDS 5 m/s -3 FDS 5 m/s
wind Expected 5 m/s wind slope Expected 5 m/s
-4 -4
-2 0 2 4 6 8 10 -2 0 2 4 6 8 10
x (m) x (m)

Figure 15.1: Expected versus predicted shape of fire front after a period of 10 s. The cases on the left are for zero
slope, and those on the right have 30° slope oriented to the northeast.

FDS-6.8.0-0-g886e009-release
0.16
0 m/s; 00°
0.14 5 m/s; 00°
0 m/s; 30°
0.12 5 m/s; 30°
relative error (-)

0.1

0.08

0.06

0.04

0.02

0
0 50 100 150 200 250 300
compass angle (°)

Figure 15.2: Relative error between FDS and empirical model. FDS spread rates are obtained from arrival times at
devices placed 4 m from the ignition point at compass angle increments of 45°.

250
Chapter 16

Unstructured Geometry (Beta)

This section describes some of the test cases used to verify that FDS and Smokeview are computing and
visualizing unstructured geometric objects correctly. Also here, the behavior of the different algorithms
used to include unstructured geometries whithin FDS is tested. Several test cases that showcase and test the
use of the GEOM line are presented, followed by accuracy and conservation tests.

16.1 Verification of setup checks on broken geometries

This subsection describes some of the test cases used to verify that FDS correctly enforces on geometries
the quality rules.
The following verification cases contain specific broken geometries in an otherwise correctly formatted
FDS input file:

• geom_bad_inconsistent_normals.fds: a cube with only one face normal inverted;

• geom_bad_inverted_normals.fds: a cube with all face normals inverted;

• geom_bad_non_manifold_edge.fds: a cube and a box sharing the same non manifold edge;

• geom_bad_open_surface.fds: a cube missing one face, that is an open surface;

To allow Firebot to consider the test successful if it catches the error condition, each of the above input
files contain the following parameter in the MISC line:

&MISC POSITIVE_ERROR_TEST = T

The effect of this parameter is replacing the ERROR label of the message with a SUCCESS label.
After Firebot is run, the geom_positive_errors.m Matlab script checks the existence of the planned
error condition in the CHID.err file, reporting in cases where the condition is not met.

251
16.2 Simple geometry and properties definitions
16.2.1 Simple Geometry
Figure 16.1 was created using the following input. The VERTS keyword is used to specify 3 vertices and the
FACES keyword is used to specify the indices of the triangular face and the SURF_ID the face has.

&GEOM ID='geom1', VERTS=0.0,0.0,0.0, 1.0,0.5,0.0, 1.0,1.0,1.0,

FACES=1,2,3,1 SURF_ID='surf1' /

Figure 16.1: A triangle created using 3 vertices and 1 face. Case: geom_simple.fds

16.2.2 Cuboid Obstacle

Figure 16.2 was created using the following MOVE and GEOM features. The XB keyword is used the same way
to specify a block as on a OBST or VENT line.

&MOVE ID='RotOBST', AXIS=0.,0.5,-0.866, ROTATION_ANGLE=90 /

&GEOM ID='obst',SURF_ID='surf1', XB=-0.6,0.6,-0.6,0.6,-0.2,0.2, MOVE_ID='RotOBST' /

252
Figure 16.2: A block generated using the XB keyword. The block is refined automatically to be consistent with the
underlying grid resolution. Case: geom_obst.fds

16.2.3 Self-generated Sphere Geometry

Figure 16.3 was created using LEVEL=n in the following GEOM namelist where n ranges from 0 to 5.

&GEOM ID='sphere',SURF_ID='surf1',SPHERE_RADIUS=0.5,
N_LEVELS=n,SPHERE_ORIGIN=0.0,0.0,0.0 /

The N_LEVELS keyword is used to specify the resolution of the sphere, larger N_LEVELS values result
in a more highly resolved sphere.LEVEL=0 produces a 20 sided sphere approximation, an icosahedron .
LEVEL=n produces a sphere approximation with four times as many triangles as the sphere produced with
LEVEL=n-1. The SPHERE_RADIUS and SPHERE_ORIGIN keywords are used to specify the size and lo-
cation of the sphere. This discretization technique should result in equilateral triangles at each recursion
level.
Figure 16.4 was created by using N_LAT and N_LONG keywords in a GEOM namelist to split a sphere in
latitudinal (north/south) and longitudinal (east/west) directions respectively. The minimum values of N_LAT
and N_LONG permitted are 3 and 6. This discretization technique results in triangles with high aspect ratios
near the poles at higher discretization levels.

253
 LEVEL=0 LEVEL=1

LEVEL=2 LEVEL=3

LEVEL=4 LEVEL=5

Figure 16.3: Recursive sphere discretization. A sphere at a given level is obtained by split-
ting each triangle from the previous level into four parts and renormalizing added vertices. Cases:
geom_sphere1a.fds,...,geom_sphere1f.fds

254
 N_LAT=3,N_LONG=6 N_LAT=6,N_LONG=12

N_LAT=12,N_LONG=24 N_LAT=24,N_LONG=48

N_LAT=48,N_LONG=96 N_LAT=96,N_LONG=192

Figure 16.4: Latitude/longitude sphere discretization. Spheres are split in longitudinal (east-
/west) and latitudinal (north/south) directions using the N_LAT and N_LONG keywords. Cases:
geom_sphere3a.fds,...,geom_sphere3f.fds

255
16.2.4 Terrain from Digital Elevation Map
The terrain of Fig. 16.5 was created using the ZVALS keyword for &GEOM.

Figure 16.5: Elevations are defined on a rectangular array of grid points using the ZVALS keyword. Case:
gridgeom_terrain.fds

16.2.5 Adding Texture Map

Figure 16.6 was created using the following SURF and GEOM namelists. The TEXTURE_MAP keyword on the
SURF is used to specify the name of the image file used to texture map the geometric object.

&SURF ID='surf1',TEXTURE_MAP='nistleft.jpg',TEXTURE_WIDTH=0.6,
TEXTURE_HEIGHT=0.2,COLOR='BLUE' /

&GEOM ID='texture',
VERTS=0.0,0.0,0.0, 1.0,0.0,0.0, 1.0,1.0,0.0,
FACES=1,2,3,1, SURF_ID='surf1'/

256
 Figure 16.6: A texture is applied to a triangle. Case: geom_texture.fds

16.2.6 Adding Multiple Texture Maps

Figure 16.7 was created using the following SURF and GEOM namelists. The TEXTURE_MAP keyword on the
SURF is used to specify the name of the image file used to texture map the geometric object. This example
has two textures.

&SURF ID='surf1' TEXTURE_MAP='nistleft.jpg',

TEXTURE_WIDTH=0.6,TEXTURE_HEIGHT=0.2,COLOR='BLUE' /
&SURF ID='surf2',TEXTURE_MAP='grass.jpg',
TEXTURE_WIDTH=0.6,TEXTURE_HEIGHT=0.2,COLOR='GREEN' /

&GEOM ID='texture',
VERTS=0.0,0.0,0.0, 1.0,0.0,0.0, 1.0,1.0,0.0,
FACES=1,2,3,1 SURF_ID='surf1'/

&GEOM ID='texture2',
VERTS=0.0,0.0,0.0, 1.0,1.0,0.0, 0.0,1.0,0.0,
FACES=1,2,3,1 SURF_ID='surf2'/

257
 Figure 16.7: Two texture maps are applied to two separate triangles. Case: geom_texture2.fds

16.2.7 Adding Texture Maps with Different Methods

The spheres in Figure 16.8 were created using SURF and GEOM namelists similar to the following.

&SURF ID='surf1',TEXTURE_MAP='sphere_cover_03.png'COLOR='BLUE',
TEXTURE_WIDTH=1.0,TEXTURE_HEIGHT=1.0/

&GEOM ID='sphere1',SURF_ID='surf1',
N_LAT=50,N_LONG=50,SPHERE_RADIUS=0.25,SPHERE_ORIGIN=0.25,0.5,0.5,
TEXTURE_ORIGIN=0.25,0.5,0.5,TEXTURE_MAPPING='SPHERICAL' /

&GEOM ID='sphere2',SURF_ID='surf3',
N_LEVELS=3,SPHERE_RADIUS=0.25,SPHERE_ORIGIN=0.75,0.5,0.5,
TEXTURE_ORIGIN=0.75,0.5,0.5,TEXTURE_MAPPING='SPHERICAL' /

The TEXTURE_MAP keyword is used to specify the name of the image file applied to the geometric object.
The SPHERICAL setting indicates that that the texture map image is applied to the object using spherical
coordinates.

258
 lat/long discretization recursive discretization

Figure 16.8: Texture maps applied to spheres discretized using two different methods. The two spheres on the left
are discretized by splitting the sphere along longitudinal (east/west) and latitudinal (north/south) directions. The
two spheres on the right are discretized recursively starting with an icosahedron (20 sided polyhedron). Texture maps
applied to the top two sphere have 8 segments. Texture map applied to the bottom two spheres have 5 segments. Cases:
geom_texture3a.fds, geom_texture3b.fds, geom_texture4a.fds, geom_texture4b.fds

16.3 Computational Geometry

16.3.1 Cut-cell Definition Around Immersed Cone: Split Cells, Piercing and Alignment
We test the computational geometry algorithm on its treatment of degenerate cases [109], namely, split cells
within a cartesian cell, cartesian cell piercing, and body surface-grid alignment. A cone of base diameter DB
and height H = 4DB is placed centered in a Cartesian domain of size x = [−1.5DB : 1.5DB ], y = [−1.5DB :
1.5DB ], z = [−2DB : 6DB ], as shown in figure 16.9a. The cone is composed of 316 vertices and 628 triangles.
A single Cartesian mesh is used with nx = 5, ny = 5 and nz = 12 cells in each corresponding direction.
A center column of cartesian cells and corresponding cut-cells is shown in Fig. 16.9b. FDS correctly
defines cut-faces and polyhedrons in three situations, namely: Solid piercing a cartesian cell, split cut-cells,

259
 (a) (b)

← pierced cell

← split cells

← aligned face

Figure 16.9: Polyhedra computation around immersed cone: (a) problem setup and grid on a y plane. (b) Cut-cells
on mid cell column along z direction: (pierced cell) cut-cell formed by cartesian cell pierced by object; (split cells)
cartesian cell complement of body defines 4 split cells of which one is kept in numerical simulations, other are blocked;
and (aligned face) cut-cell of same size of regular cell contains boundary cut-faces aligned with z = 0 grid plane.

and cells defined from the cone base aligned with grid plane z = 0.
Three integral conservation tests are defined for a completely immersed geometry problem. These are
namely, relative difference of areas computed from the surface triangulation or from the computed boundary
cut-faces |AB,τ − AB,CF |/AB,τ ; relative difference of G region volumes computed from the body triangulation
and from the cut-cells and G region regular cells |VGB −VGRC |/VGB ; and finally, difference on the G region
centroid computed from body centroid or cut-cells and G cells centroids |xGB − xGRC |∞ . See Appendix C for
a formal definition of these tests.
The results of the three tests on this cone problem can be seen on table 16.1. We note that cone area and
G region volume computed either through the solid triangulation or through the polyhedra produced by the
method match up to relative values of 10−15 , commensurate with the floating point tolerance ε = 10−12 Lz
used. The centroid of the G region computed through the two different avenues matches up to 10−16 DB ,
and is an excellent integral test of the proposed method to compute both cut-face and cut-cell centroids.
Note that centroids along x and y direction are not zero, due to the asymmetry of the geometry triangulation
employed. Given that the numerical scheme for momentum and scalars integration is defined for one cut-cell
per cartesian cell, the cut-cell with largest volume is kept and others are blocked as solid.

260
 |AB,τ −AB,CF |
AB,τ AB,CF AB,τ
0.7104E+01 0.7104E+01 0.6252E-15
|VGB −VGRC |
VGB VGRC VGB
0.7096E+02 0.7096E+02 0.2212E-14
xGB xGRC |xGB − xGRC |∞
(-0.306E-6,0.586E-6,0.202E+1) (-0.306E-6,0.586E-6,0.202E+1) 0.1856E-15

Table 16.1: Results for conservation tests on polyhedra defined on immersed cone problem.

16.4 Cut-cell Method for Scalar Transport - Stress Method

16.4.1 Temporal Error Analysis for Variable Density Projection
This case is a variation of problem 3.4 for a domain that includes an imposed cut-cell region, where the
cut-cells have same geometry as regular cells. The problem allows us to verify the temporal accuracy
of our scalar transport implementation for cut-cells. The problem is defined by the advection of a non-
reacting mixture of two fluids with different densities (10:1 ratio) in one cartesian direction. Zero diffusivity
is prescribed. For the problem definition and temporal error computation details, we refer the reader to
problem 3.4. The domain is defined by a mesh containing Nx = 512 cells in the x direction, where the
domain x = [−1, 1]. Note that the problem defined is one-dimensional in x. On this domain, cut-cells with
same shape and size of regular cells are defined in the space x = [−0.5, 0.5]. The test verifies that the
implementation details of this cut-cell region does not affect the temporal order of accuracy seen on the
default verification case of Sec. 3.4.
Three cases are defined corresponding to default FDS DNS mode in cells not covered by cut-cells, and
in the cut-cell region, for CFL numbers 0.25, 0.125, 0.0625. These cases are defined in
/Complex_Geometry/saad_CC_explicit_*.fds.
Same procedure as explained in Sec. 3.4 is used to define the temporal order of accuracy p. As noted
there, this procedure filters out the spatial error from the computed temporal convergence. Initial and final
fields for temperature and mixture fraction for the smallest integration time step are shown in the top figures
of Fig. 16.10. Fields advect to the right with unitary velocity. The bottom figure shows the order p computed
point wise in the x direction for explicit time integration on the cut-cell region. Deviations from p ' 2 are
dependent on the formula used to compute p, as explained in Sec. 3.4. The l2 norm of p computed pointwise
for explicit integration in cut-cells is 2.0053.

16.4.2 Variable-Density Manufactured Solution

We test the manufactured solution test case 3.5 for scalar transport, including an internal region of the two di-
mensional domain where cut-cells with same shape and size as regular Cartesian cells are artificially defined.
This problem allows us to check the overall accuracy (spatial+temporal) of the cut-cell implementation for
scalar transport on the cut-cell region. The parameters of the test follow the ones defined by Shunn et al. [43]
and described in Sec. 3.5.
The treatment of sources for scalar transport and momentum equations due to the manufactured solution
is essentially the same as defined in Sec. 3.5. Therefore, source terms from manufactured solution are treated
following the time integration method defined in FDS. The domain is defined on a square of length (L = 2
m), with periodicity prescribed in its boundaries. A smaller square of length (L = 1 m) centered on the
previous is filled with cut-cells of regular cell geometry.

261
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
1 5
initial field initial field
final field final field
0.8 4
mixture fraction

density (kg/m³)
0.6 3

0.4 2

0.2 1

0 0
-1 -0.5 0 0.5 1 -1 -0.5 0 0.5 1
x (m) x (m)
FDS-6.8.0-0-g886e009-release
4

3
p order density

0
-1 -0.5 0 0.5 1
x (m)

Figure 16.10: Temporal order for a variable-density projection on domain including cut-cell region. (Upper-left) Initial
and final field for mixture fraction, explicit time integration on cut-cells. (Upper-right) Initial and final field for density,
explicit time integration on cut-cells. (Bottom) p computed pointwise for density at final time, explicit integration on
cut-cells. Fluctuations are due to degenerate points in the formula for p.

Simulations were performed for N = {32, 64, 128, 256, 320}, where N is the number of cells in each
direction, using an adaptive time step satisfying a CFL number of 0.25. The scalar transport equations for
both species defined on Sec. 3.5, are solved using the GODUNOV flux limiter and the DNS default CHARM
limiter. Figure 16.11, presents the L2 error norm defined on the whole domain (regular+cut-cells) at time
t = 0.9 as a function of grid spacing, ∆x = L/N. Time integration in the cut-cell region produces second-
order accurate solutions for density, mixture fraction, and velocity, when CHARM interpolation is used to
define the advective mass flux. As expected, these variables exhibit asymptotic first order accuracy when
the GODUNOV flux limiter is used. Note that central differences are used on the diffusive terms of these
equations, rendering a second order accurate approximation of diffusion on these regular shaped cells. The
pressure H is, in same fashion as seen on Sec. 3.5, first order accurate. This is consistent with the projection
scheme used for time integration of velocities used by FDS.

16.4.3 Stress Method on Manufactured Solution Around Rotated Cube

Consider the solution of the following constant density, incompressible flow, momentum and scalar transport
equations:

∂ ρu(x,t)
= − [Fv (u, x,t) + ∇p(x,t) + fev (u, p)] (16.1)
∂t
262
 FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
10 0
FDS FDS
-2 FDS Z FDS Z
10 FDS u FDS u
FDS H -2 FDS H
10
O( x) O( x)
L Error

L Error
O( x2) O( x2)
2

2
10 -4
10 -4

10 -6 10 -6
10 -3 10 -2 10 -1 10 -3 10 -2 10 -1
x (m) x (m)

Figure 16.11: Overall accuracy for manufactured solution variable-density projection on domain including cut-cell
region. The L2 error defined on the whole domain at time t = 0.9 s is plotted as a function of grid spacing for
N = {32, 64, 128, 256, 320} points in each direction. Two flux limiting options for scalar advection are used: CHARM
and GODUNOV. (Left) Convergence for explicit scalar time integration and CHARM flux limiter for advection. (Right)
Convergence for explicit scalar time integration and GODUNOV flux limiter.

∂ ρZ
= − [Fz (u, x,t) + f eZ (u, Z)] (16.2)
∂t
which are extended by the manufactured solution (MMS) sources fev (u, p) and f eZ (u, Z) respectively. The
vectors Fv = −µ∇2 (u) + ∇ · (ρu ⊗ u) and Fz = ∇ · (ρZu − ρDz ∇Z) contain the corresponding advection and
diffusion operators.
Now, consider in Fig. 16.12, a local set of fixed axes x0 , y0 and a manufactured solution defined on the
fluid region Ω − Ω1 of the form:

u0 (x0Bp ,t) = − sin(t) sin(kx0 )2 sin(2ky0 ) (16.3)

0
v (x0Bp ,t)
= sin(t) sin(2kx ) sin(ky ) 0 0 2
(16.4)
sin(t)
p(x0Bp ,t) = 2 + cos(2kx0 ) 2 + cos(2ky0 ) − sin(t)



(16.5)
4
A Z sin(t) π  π 
h i
Z(x0Bp ,t) = 1 − cos(2k(x0 − )) 1 − cos(2k(y0 − )) − 1 + Zmean (16.6)
3 2 2
which is similar on velocities and pressure to the functions given in Ref. [110]. These functions imply zero
initial conditions, a divergence free velocity field and the following
∂ p ∂Z
u0 (x0 ,t) = v0 (x0 ,t) = 0, = = 0, f or kx0 = ky0 = mπ ; m ∈ N (16.7)
∂n ∂n
which can be used as no-slip wall conditions in square boundaries defined by the integer m. The MMS
sources fe0v and f e0Z are computed analytically as functions of (x0 ,t) replacing definitions (16.3)-(16.6)
in the governing equations (16.1)-(16.2) and computing symbolically the corresponding terms. See the
Matlab script Utilities/Matlab/scripts/source_mms_rotcube.m in the FDS repository for
said sources derivations.
We take values m = 0, 1 and wavenumber k = 1, and consider the external flow to the resulting square
immersed body, as seen in figure 16.12. The velocity field function inside the immersed object Ω1 (i.e. solid
region) is assumed zero at all times, consistent with the fixed rigid body hypothesis. An angular parameter
α defines the orientation of the local axes x0 , y0 respect to the Eulerian frame coordinates x, y. The center of

263
Figure 16.12: Reference frames for Eulerian (E) and immersed body local system (B), and vectors relating a generic
point p on the fluid region Ω − Ω1 .

Ω1 in the Eulerian frame is situated in position xEc = (π, π). The rotation matrix from Eulerian E to local
frame B coordinates is  
cos(α) sin(α)
[TBE ] = (16.8)
− sin(α) cos(α)
In the local system, the center of the square in local coordinates is given by x0Bc = (π/2, π/2). Then, the
location of a general point p in the local system is:
x0Bp = x0Bc + x0cp (16.9)
The same point p location can be defined in terms of the E frame as xE p = xEc + xcp (Eulerian coordinates).
Then
x0cp = [TBE ]xcp = [TBE ] (xE p − xEc ) (16.10)
and finally
x0Bp = x0Bc + [TBE ] (xE p − xEc ) (16.11)
This last equation is employed to translate a given point xE p in the Eulerian frame (with Eulerian frame co-
ordinates x, y) to the local frame coordinates where equations (16.3)-(16.6) are defined. This transformation
is used to define the initial conditions in the local frame, for points collocated in the Eulerian frame aligned
fluid grid. It is also used to compute the source terms fe0v and f e0Z in local coordinates at different time
levels. To transform fe0v and f e0Z to their Eulerian counterparts used in the numerical integration we use:
fev = [TBE ]T fe0v (16.12)
f eZ = f e0Z (16.13)
Similarly, the manufactured variables can be defined at the final time to define error norms. The velocity
components of interest are the ones defined in the Eulerian frame, and can be readily computed as
   0 
u T u
= [TBE ] (16.14)
v p v0 p
264
We assume periodic boundary conditions on ∂ Ω. Then, the size of the domain Ω depends on the periodicity
lengths for a rotated flow field. When α = 0, a domain Ω of size Lx = Ly = 2π centered in xEc is sufficient
√ α = arctan(1/r), r = 1, 2, 3, ... the periodicity
for application of periodic boundary conditions. Then, for
2
√ frame are Lx = Ly = 2π 1 + r . That
lengths on the Eulerian reference √ is, if α = arctan(1), the periodicity
length increases to Lx = Ly = 8π. If α = arctan(1/2), Lx = Ly = 20π.
FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release

FDS Z FDS Z
10 -2 FDS u 10 -2 FDS u
FDS H FDS H
O( x) O( x)
x2) x2)
L Error

L Error
O( O(
10 -4 10 -4
2

2
10 -6 10 -6

10 -2 10 -1 10 -2 10 -1
x (m) x (m)

Figure 16.13: Zero degree rotation. The L2 error defined on the whole domain at time t = 0.95 s is plotted as a function
of grid spacing for N = {32, 64, 128, 256, 320} points in each direction. (Left) Stress method, (Right) square defined
by &OBST.

FDS-6.8.0-0-g886e009-release

FDS Z
FDS u
FDS H
10 -2
O( x)
x2)
L Error

O(

10 -4
2

10 -6

10 -2 10 -1
x (m)

Figure 16.14: 27 degree rotation. The L2 error defined on the whole domain at time t = 0.95 s is plotted as a function
of grid spacing for N = {32, 64, 128, 256, 320} points in each direction. Stress method.

Although the different orientation angles correspond to different periodicity lengths and domain sizes all
cases were run with N = {32, 64, 128, 256, 320} points in each direction for accuracy testing. The different
inclination angles and grids were run with the stress method for shear stress imposition. The cases with zero
inclination were also run with the &OBST functionality for comparison. Calculations were run at constant
CFL = 0.1 and errors were computed on the solution at the first time level after t = 0.95. The cases input
files are found on /Complex_Geometry/rotated_cube_*deg_*_*.fds.
Overall the error norm behaves as O(∆x) for the scalar transport equation in cases where the cube is
inclined. This is expected as the diffusive flux differences used are of this spatial accuracy. Note that for

265
the non inclined cases of figure 16.13 the scalar Z presents better accuracy, simply because the two point
centroid to centroid difference coincides with a central difference. The O(∆x2 ) accuracy in Z is also seen
for OBST in the right figure of 16.13, which use central differences for diffusive fluxes. For OBST and grid
aligned stress method where the momentum equations are solved on velocity points next to the surface,
velocities behave as O(∆x2 ), in accordance to the discretization schemes used (right figure 16.13). For
the stress method and inclined cases the spatial accuracy is higher than O(∆x). Note that pressure spatial
accuracy is O(∆x) across all tests, consistent with the fractional step projection method used by FDS.
FDS-6.8.0-0-g886e009-release

FDS Z
10 -2 FDS u
FDS H
O( x)
2
L Error

O( x )
10 -4
2

10 -6

-2 -1
10 10
x (m)

Figure 16.15: 45 degree rotation. The L2 error defined on the whole domain at time t = 0.95 s is plotted as a function
of grid spacing for N = {32, 64, 128, 256, 320} points in each direction. Stress method.

16.4.4 Poiseuille Flow in Channel Created by Two GEOMs

Consider a gas with density ρ = 1.165 kg/m3 , and molecular viscosity µ = 0.025 kg/m/s, being advected
through a 2-D channel of height H = 1 m and length L = 10 m, by means of a constant force field along
the channel direction Fx = 1 N/m3 . This force field is equivalent to a pressure drop d p/dx = −1 Pa/m. The
analytical solution for the friction factor in this laminar flow is:

fan = 24/ReH (16.15)

where in our case ReH = ρuH/µ ' 155, being the analytical value for the bulk velocity u = 3.3̂ m/s. The
(Moody) friction factor formula for pressure losses is:
L1 2
∆p/L = f ρu (16.16)
H2
which can be used to compute the simulation error in f , as a function of the mean velocity evaluated on the
numerical mesh. To evaluate the solution quality of the cut-cell scheme we consider three sets of test cases.
See figure 16.16. We define as h the distance from the lower geometry wall to the closest grid line within it.
The we consider discretizations with 10, 20, 40, 80 cells along the z direction and cases: grid aligned h = 0,
grid dependent h = ∆z/3, and fixed h = ∆z10 /11, where ∆z10 is the wall normal cell size in the coarsest
discretization used.
For stress method for cross velocities (STM) and unlinked cut-cells, the friction factor error | f − fan |
convergence can be seen in Figs. 16.17a,b.

266
Figure 16.16: Poiseuille flow created by two geometries Ω1 , Ω2 : Location of channel walls respect to Cartesian grid
lines (in blue).

FDS-6.8.0-0-g886e009-release FDS-6.8.0-0-g886e009-release
10 -2 10 -2
Friction Factor Error

Friction Factor Error

10 -4 10 -4

10 -6 10 -6
-2 -1 -2 -1
10 10 10 10

Figure 16.17: Poiseuille flow created by two geometries, STM for cross velocities: Convergence trends on friction
factor (Moody) for h = 0, grid dependent h = ∆z/3 and fixed h = ∆z10 /11.

16.5 Geometry Tests Involving Heat and Mass Transfer

16.5.1 Radiating Sphere
A sphere with a radius of 1 m and a surface temperature of T = 773.15 K is centered in a box that is 4 m on
a side with a surface temperature of 0 K. The emissivity of all surfaces is 1, and there is no convective heat
transfer to or from any surface. The heat flux from the hot sphere to the nearest cold wall surface 1 m away
is expected to be σ T 4 /4 = 5.065 kW/m2 . The results are shown in Fig. 16.18.

267
 FDS-6.8.0-0-g886e009-release
8
Heat Flux (sphere_radiate)

Heat Flux (kW/m²)

4

2
Exact
FDS
0
0 0.002 0.004 0.006 0.008 0.01
Time (s)

Figure 16.18: Left: a radiating hot sphere. Right: heat flux to the nearest wall surface.

16.5.2 Leaking Sphere

A hollow sphere with an inner radius of 1 m has a small cube within it that is blowing air into the sphere at
a rate of V̇ = 0.008 m3 /s. The sphere is assigned a leakage area of A = 0.0001 m2 . It is expected that the
pressure rise within the sphere shall be
 2
ρ∞ V̇
∆p = ≈ 3840 Pa (16.17)
2 A

as shown in Fig. 16.19.

FDS-6.8.0-0-g886e009-release
5000
Pressure Rise (sphere_leak)
4000
Pressure (Pa)

3000

2000

1000
Exact
FDS
0
0 20 40 60 80 100
Time (s)

Figure 16.19: Pressure rise in a leaking sphere.

16.6 Restart Function

In this test, a GEOM cube is heated from below by a surface at a fixed temperature. The internal temperature
inside the geometry is monitored. The full calculation runs for 1,200 s. At 700 s, a RESTART is performed
and compared with the base case. The results are shown in Fig. 16.20.

268
 FDS-6.8.0-0-g886e009-release
80
GEOM Restart (geom_restart)

60

Internal Temperature (°C)

40

20
without restart
restarted at 700 s
0
0 200 400 600 800 1000 1200
Time (s)

Figure 16.20: Test of restart functionality for complex geometry.

269
Chapter 17

Outputs

17.1 Statistical Quantities

17.1.1 RMS, Co-Variance, and Cross-Correlation (rms_cov_corr)
FDS can output the root mean square (RMS), co-variance, and cross-correlation for both point and line DEVC
outputs. To test these outputs a 1 m3 box with open sides and a 10 cm grid size is defined with two inlet
vents centered on adjacent faces. This results in two orthogonal flow streams that collide at the center of
the box and exit diagonally. Within the diagonal portion of the flow, are placed point measurements for the
FDS outputs of the u-velocity RMS, the u-velocity/w-velocity co-variance, and the u-velocity/w-velocity
cross-correlation.
Computation of the these quantities by their typical statistical definition would require having the entire
time history for the velocities. This is obviously not possible during run-time. Instead FDS computes the
values using logarithmic averaging and a user-defined averaging window. One would expect, once steady-
state conditions are reached, that the FDS predicted values should approach the analytic values. Results are
shown in Fig. 17.1. In each figure the red line represents the value computed using the entire time history
after steady-state is reached. As seen, the FDS predicted values approach the analytic values.

271
 FDS-6.8.0-0-g886e009-release
0.2

0.15

u rms (m/s)
Analytic
0.1
FDS

0.05

0
-500 0 500 1000
Time (s)

10 -3 FDS-6.8.0-0-g886e009-release
5

0
uw covariance (m²/s²)

-5
Analytic
FDS
-10

-15

-20
-500 0 500 1000
Time (s)

FDS-6.8.0-0-g886e009-release
0.2
Analytic
0 FDS
uw cross correlation

-0.2

-0.4

-0.6

-0.8

-1
-500 0 500 1000
Time (s)

Figure 17.1: Output of the rms_cov_corr test case. By construction, the statistic matches the analytical result at
time t = T_END.

272
17.1.2 RMS as a Running Average (rms_example)
In this example, we want to compute the mean and rms of a hypothetical velocity time series
u(t) = tanh(t) + U (−0.5, 0.5) (17.1)
where U (−0.5, 0.5) is a uniformly distributed random variable. The expected RMS of the time series
beyond
p its initial transient is equal to the standard deviation of the uniform distribution over the unit interval,
σ = 1/12 ≈ 0.2887. The expected mean is 1. This exercise is as much a check of the intrinsic random
number generator in Fortran as it is the computation within FDS. Figure 17.2 displays the result. Note that
FDS-6.8.0-0-g886e009-release
2.5
Expected Mean
rms_example Expected RMS
2 FDS (u)
FDS (u_rms)
Velocity (m/s)

1.5

0.5

0
0 5 10 15 20
Time (s)

Figure 17.2: Output of the rms_example test case in which the mean and RMS of the time series is compared to
expected values.

this example is also a useful demonstration of how to use the math functions within FDS. Although there is
not a tanh function included, it can be constructed from the expression:
et − e−t
tanh(t) = (17.2)
et + e−t
using the base functions as shown below.

&DEVC ID='rn', QUANTITY='RANDOM NUMBER', XYZ=0,0,0 /

&DEVC ID='t', QUANTITY='TIME', XYZ=0,0,0 /
&DEVC ID='u', QUANTITY='CONTROL VALUE', CTRL_ID='sum', XYZ=0,0,1, UNITS='m/s' /
&DEVC ID='u_rms', QUANTITY='CONTROL VALUE', CTRL_ID='sum', XYZ=0,0,1, UNITS='m/s',
STATISTICS='RMS', STATISTICS_START=5. /

&CTRL ID='exp(t)', FUNCTION_TYPE='EXP', INPUT_ID='t' /

&CTRL ID='-t', FUNCTION_TYPE='MULTIPLY', INPUT_ID='t','CONSTANT', CONSTANT=-1. /
&CTRL ID='exp(-t)', FUNCTION_TYPE='EXP', INPUT_ID='-t' /
&CTRL ID='num', FUNCTION_TYPE='SUBTRACT', INPUT_ID='exp(t)','exp(-t)' /
&CTRL ID='den', FUNCTION_TYPE='SUM', INPUT_ID='exp(t)','exp(-t)' /
&CTRL ID='tanh', FUNCTION_TYPE='DIVIDE', INPUT_ID='num','den' /
&CTRL ID='rn2', FUNCTION_TYPE='SUBTRACT', INPUT_ID='rn','CONSTANT', CONSTANT=0.5 /
&CTRL ID='sum', FUNCTION_TYPE='SUM', INPUT_ID='tanh','rn2' /

17.1.3 Favre Average (shunn3_FavreZ)

The Shunn manufactured solution (see Sec. 3.5) provides a 2-D analytical solution for both the density and
mixture fraction. In this section, to test the statistical output TEMPORAL_STATISTIC=’FAVRE AVERAGE’

273
we compare FDS results at two grid resolutions, δ x = [L/32, L/64], with L = 2. The case is run to t = 10 s.
Both the running average and Favre average are output at (x = 0, y = 0), the center of the domain. Note that
this imparts an O(δ x) error to the output. The higher grid resolution is therefore more accurate.
FDS-6.8.0-0-g886e009-release
0.15

0.1
Mass Fraction
Analytical Favre
Analytical Average
0.05 FDS 32 Favre
FDS 32 Average
FDS 64 Favre
FDS 64 Average
0
0 2 4 6 8 10
Time (s)

Figure 17.3: Output of the shunn3_FavreZ test cases in which the Favre average (continuous line) and running
average (dashed line) at (x = 0, y = 0) are compared against expected values from the Shunn manufactured solution
(see Sec. 3.5).

17.2 Interpolation
For line devices (POINTS=...) used together with SPATIAL_STATISTIC=’INTERPOLATION’ the near
wall values are usually truncated to the cell value. However, with temperature and velocity components the
wall value is used to provide an interpolation between the first cell value and the wall value. In the test cases
below, a linear field of velocity and temperature are generated with 100 points along the device and only
5 cells across the domain. Three linear point arrays are created, one at each end along the periodic mesh
boundary, and one in the center of the domain, which is also a cell center.

1 25
Velocity (devc_interpolation) Temperature (devc_interpolation)
0.8 24
Temperature (°C)
Velocity (m/s)

0.6 23

0.4 22

u prescribed T prescribed
0.2 U-VEL-1 21 TMP-1
U-VEL-2 TMP-2
U-VEL-3 TMP-3
0 20
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
z (m) z (m)

Figure 17.4: Output of the devc_interpolation test cases.

274
Bibliography

[1] K. McGrattan, S. Hostikka, R. McDermott, J. Floyd, C. Weinschenk, and K. Overholt. Fire Dynam-
ics Simulator, Technical Reference Guide. National Institute of Standards and Technology, Gaithers-
burg, Maryland, USA, and VTT Technical Research Centre of Finland, Espoo, Finland, sixth edition,
September 2013. Vol. 1: Mathematical Model; Vol. 2: Verification Guide; Vol. 3: Validation Guide;
Vol. 4: Software Quality Assurance. v, 55, 61, 124, 145, 203

[2] K. McGrattan, S. Hostikka, R. McDermott, J. Floyd, C. Weinschenk, and K. Overholt. Fire Dynamics
Simulator, User’s Guide. National Institute of Standards and Technology, Gaithersburg, Maryland,
USA, and VTT Technical Research Centre of Finland, Espoo, Finland, sixth edition, September 2013.
v, 73, 74

[3] American Society for Testing and Materials, West Conshohocken, Pennsylvania. ASTM E1355-12,
Standard Guide for Evaluating the Predictive Capabilities of Deterministic Fire Models, 2012. v, 1,
9

[4] W. Mell, K.B. McGrattan, and H. Baum. Numerical Simulation of Combustion in Fire Plumes. In
Twenty-Sixth Symposium (International) on Combustion, pages 1523–1530. Combustion Institute,
Pittsburgh, Pennsylvania, 1996. 3

[5] K.B. McGrattan, H.R. Baum, and R.G. Rehm. Large Eddy Simulations of Smoke Movement. Fire
Safety Journal, 30:161–178, 1998. 3

[6] H.R. Baum, R.G. Rehm, P.D. Barnett, and D.M. Corley. Finite Difference Calculations of Buoyant
Convection in an Enclosure, Part I: The Basic Algorithm. SIAM Journal of Scientific and Statistical
Computing, 4(1):117–135, March 1983. 3

[7] H.R. Baum and R.G. Rehm. Finite Difference Solutions for Internal Waves in Enclosures. SIAM
Journal of Scientific and Statistical Computing, 5(4):958–977, December 1984. 3

[8] H.R. Baum and R.G. Rehm. Calculations of Three Dimensional Buoyant Plumes in Enclosures.
Combustion Science and Technology, 40:55–77, 1984. 3

[9] R.G. Rehm, P.D. Barnett, H.R. Baum, and D.M. Corley. Finite Difference Calculations of Buoyant
Convection in an Enclosure: Verification of the Nonlinear Algorithm. Applied Numerical Mathemat-
ics, 1:515–529, 1985. 3

[10] K.B. McGrattan, T. Kashiwagi, H.R. Baum, and S.L. Olson. Effects of Ignition and Wind on the
Transition to Flame Spread in a Microgravity Environment. Combustion and Flame, 106:377–391,
1996. 4

275
[11] T. Kashiwagi, K.B. McGrattan, S.L. Olson, O. Fujita, M. Kikuchi, and K. Ito. Effects of Slow
Wind on Localized Radiative Ignition and Transition to Flame Spread in Microgravity. In Twenty-
Sixth Symposium (International) on Combustion, pages 1345–1352. Combustion Institute, Pittsburgh,
Pennsylvania, 1996. 4

[12] W. Mell and T. Kashiwagi. Dimensional Effects on the Transition from Ignition to Flame Spread
in Microgravity. In Twenty-Seventh Symposium (International) on Combustion, pages 2635–2641.
Combustion Institute, Pittsburgh, Pennsylvania, 1998. 4

[13] W. Mell, S.L. Olson, and T. Kashiwagi. Flame Spread Along Free Edges of Thermally-Thin Samples
in Microgravity. In Twenty-Eighth Symposium (International) on Combustion, pages 2843–2849.
Combustion Institute, Pittsburgh, Pennsylvania, 2000. 4

[14] K. Prasad, Y. Nakamura, S.L. Olson, O. Fujita, K. Nishizawa, K. Ito, and T. Kashiwagi. Effect of
Wind Velocity on Flame Spread in Microgravity. In Twenty-Ninth Symposium (International) on
Combustion, pages 2553–2560. Combustion Institute, Pittsburgh, Pennsylvania, 2002. 4

[15] Y. Nakamura, T. Kashiwagi, K.B. McGrattan, and H.R. Baum. Enclosure Effects on Flame Spread
over Solid Fuels in Microgravity. Combustion and Flame, 130:307–321, 2002. 4

[16] W.E. Mell, K.B. McGrattan, and H.R. Baum. g-Jitter Effects on Spherical Diffusion Flames. Micro-
gravity Science and Technology, 15(4):12–30, 2004. 4

[17] A. Mukhopadhyay and I.K. Puri. An Assessment of Stretch Effects on Flame Tip Using the Thin
Flame and Thick Formulations. Combustion and Flame, 133:499–502, 2003. 4

[18] A. Hamins, M. Bundy, I.K. Puri, K.B. McGrattan, and W.C. Park. Suppression of Low Strain Rate
Non-Premixed Flames by an Agent. In Proceedings of the 6th International Microgravity Combustion
Workshop, NASA/CP-2001-210826, pages 101–104. National Aeronautics and Space Administration,
Lewis Research Center, Cleveland, Ohio, May 2001. 4

[19] K.B. McGrattan, R.G. Rehm, and H.R. Baum. Fire-Driven Flows in Enclosures. Journal of Compu-
tational Physics, 110(2):285–291, 1994. 4

[20] K. McGrattan, S. Hostikka, R. McDermott, J. Floyd, C. Weinschenk, and K. Overholt. Fire Dynamics
Simulator, Technical Reference Guide, Volume 3: Validation. National Institute of Standards and
Technology, Gaithersburg, Maryland, USA, and VTT Technical Research Centre of Finland, Espoo,
Finland, sixth edition, September 2013. 5, 9, 219

[21] P. Friday and F. W. Mowrer. Comparison of FDS Model Predictions with FM/SNL Fire Test Data.
NIST GCR 01-810, National Institute of Standards and Technology, Gaithersburg, Maryland, April
2001. 6

[22] A. Bounagui, N. Benichou, C. McCartney, and A. Kashef. Optimizing the Grid Size Used in CFD
Simulations to Evaluate Fire Safety in Houses. In 3rd NRC Symposium on Computational Fluid
Dynamics, High Performance Computing and Virtual Reality, pages 1–8, Ottawa, Ontario, Canada,
December 2003. National Research Council, Canada. 6

[23] R.L. Alpert. SFPE Handbook of Fire Protection Engineering, chapter Ceiling Jet Flows. Springer,
New York, 5th edition, 2016. 6

276
[24] A. Bounagui, A. Kashef, and N. Benichou. Simulation of the Dynamics of the Fire for a Section of
the L.H.-La Fontaine Tunnel. IRC-RR- 140, National Research Council Canada, Ottawa, Canada,
K1A0R, September 2003. 6

[25] Y. Xin. Assessment of Fire Dynamics Simulation for Engineering Applications: Grid and Domain
Size Effects. In Proceedings of the Fire Suppression and Detection Research Application Symposium,
Orlando, Florida. National Fire Protection Association, Quincy, Massachusetts, 2004. 6

[26] J.A. Ierardi and J.R. Barnett. A Quantitative Method for Calibrating CFD Model Calculations. In
Proceedings of the CIB-CTBUH International Conference on Tall Buildings, pages 507–514. Inter-
national Council for Research and Innovation in Building and Construction (CIB), 2003. 6

[27] G. Heskestad. SFPE Handbook of Fire Protection Engineering, chapter Fire Plumes, Flame Height
and Air Entrainment. Springer, New York, 5th edition, 2016. 6

[28] N.M. Petterson. Assessing the Feasibility of Reducing the Grid Resolution in FDS Field Modeling.
Fire Engineering Research Report 2002/6, University of Canterbury, Christchurch, New Zealand,
March 2002. 6

[29] A. Musser, K. B. McGrattan, and J. Palmer. Evaluation of a Fast, Simplified Computational Fluid
Dynamics Model for Solving Room Airflow Problems. NISTIR 6760, National Institute of Standards
and Technology, Gaithersburg, Maryland, June 2001. 7

[30] W. Zhang, A. Hamer, M. Klassen, D. Carpenter, and R. Roby. Turbulence Statistics in a Fire Room
Model by Large Eddy Simulation. Fire Safety Journal, 37:721–752, 2002. 7

[31] J. Smagorinsky. General Circulation Experiments with the Primitive Equations. I. The Basic Experi-
ment. Monthly Weather Review, 91(3):99–164, March 1963. 7

[32] M. Germano, U. Piomelli, P. Moin, and W.H. Cabot. A Dynamic Subgrid-Scale Eddy Viscosity
Model. Physics of Fluids A, 3(7):1760–1765, 1991. 7

[33] D.K. Lilly. A Proposed Modification of the Germano Subgrid-Scale Closure Method. Physics of
Fluids A, 4(3):633–635, 1992. 7

[34] P. Moin, K. Squires, W. Cabot, and S. Lee. A dynamic subgrid-scale model for compressible turbu-
lence and scalar transport. Phys. Fluids A, 3(11):2746–2757, 1991. 7, 47

[35] S. Hostikka. Development of fire simulation models for radiative heat transfer and probabilistic risk
assessment. VTT Publications 683, VTT Technical Research Centre of Finland, Espoo, Finland,
2008. 8

[36] J. Hietaniemi, S. Hostikka, and J. Vaari. FDS Simulation of Fire Spread – Comparison of Model Re-
sults with Experimental Data. VTT Working Papers 4, VTT Building and Transport, Espoo, Finland,
2004. 8

[37] C. Lautenberger, G. Rein, and C. Fernandez-Pello. The Application of a Genetic Algorithm to Esti-
mate the Material Properties for Fire Modeling from Bench-Scale Fire Test Data. Fire Safety Journal,
41:204–214, 2006. 9

[38] J.C. Adams, W.S. Brainerd, J.T. Martin, B.T. Smith, and J.L. Wagener. Fortran 95 Handbook: Com-
plete ISO/ANSI Reference. MIT Press, Cambridge, Massachusetts, 1997. 9

277
[39] R. McDermott. A nontrivial analytical solution to the 2-d incompressible Navier-Stokes equations.
https://sites.google.com/site/randymcdermott/NS_exact_soln.pdf, 2003. 11

[40] R. J. Leveque. Finite Volume Methods for Hyperbolic Problems. Cambridge University Press, New
York, 2002. 20

[41] A. Biglari, T. Saad, and J. Sutherland. An efficient and explicit pressure projection method for react-
ing flow simulations. In 8th US National Combustion Meeting, 2013. 23

[42] P. Moin. Fundamentals of Engineering Numerical Analysis. Cambridge University Press, 2001. 23

[43] L. Shunn, F. Ham, and P. Moin. Verification of variable-density flow solvers using manufactured
solutions. Journal of Computational Physics, 231:3801–3827, 2012. 24, 25, 26, 27, 261

[44] https://github.com/firemodels/fds. 25, 39, 58

[45] S. Armfield and R. Street. An analysis and comparison of the time accuracy of fractional-step methods
for Navier-Stokes equations on staggered grids. Int. J. Num. Meth. Fluids, 38:255–282, 2001. 26, 28

[46] J.-C. Jouhaud. Benchmark on the vortex preservation.

http://elearning.cerfacs.fr/pdfs/numerical/TestCaseVortex2D.pdf, 2010. 30

[47] D.K. Kahaner, W.F. Lawkins, and S. Thompson. On the use of rootfinding ODE software for the so-
lution of a common problem in nonlinear dynamical systems. Journal of Computational and Applied
Mathematics, 28:219–230, 1989. 33

[48] K. McGrattan, S. Hostikka, R. McDermott, J. Floyd, C. Weinschenk, and K. Overholt. Fire Dy-
namics Simulator, Technical Reference Guide, Volume 1: Mathematical Model. National Institute of
Standards and Technology, Gaithersburg, Maryland, USA, and VTT Technical Research Centre of
Finland, Espoo, Finland, sixth edition, September 2013. 33, 34, 37, 47, 49, 55, 74, 134, 166

[49] R.L. Curl. Disperse phase mixing: I. theory and effects in simple reactors. AIChE Journal, 9:175,
1963. 34

[50] G. Comte-Bellot and S. Corrsin. Simple Eulerian time correlation of full- and narrow-band velocity
signals in grid-generated, ‘isotropic’ turbulence. J. Fluid Mech., 48:273–337, 1971. 45

[51] Stephen M. de Bruyn Kops. Numerical simulation of non-premixed turbulent combustion. PhD thesis,
The University of Washington, 1999. 45

[52] R. McDermott, A. Kerstein, R. Schmidt, and P. Smith. Characteristics of 1D spectra in finite-volume

large-eddy simulations with one-dimensional turbulence subgrid closure. In 58th Annual Meeting of
the American Physical Society, Division of Fluid Dynamics, Chicago, Illinois, November 2005. 45,
46

[53] Y. Morinishi, T. S. Lund, O. V. Vasilyev, and P. Moin. Fully conservative high order finite difference
schemes for incompressible flow. Journal of Computational Physics, 143:90–124, 1998. 46

[54] F. E. Ham, F. S. Lien, and A. B. Strong. A fully conservative second-order finite difference scheme for
incompressible flow on non-uniform grids. Journal of Computational Physics, 177:117–133, 2002.
46

278
[55] R. McDermott. Discrete kinetic energy conservation for variable-density flows on staggered grids. In
60th Annual Meeting of the American Physical Society, Division of Fluid Dynamics, Salt Lake City,
Utah, November 2007. 46

[56] M. Germano, U. Piomelli, P. Moin, and W. Cabot. A dynamic subgrid-scale eddy viscosity model.
Physics of Fluids A, 3(7):1760–1765, 1991. 47

[57] M. Pino Martin, U. Piomelli, and G. Candler. Subgrid-scale models for compressible large-eddy
simulation. Theoret. Comput. Fluid Dynamics, 13:361–376, 2000. 47

[58] T. S. Lund. On the use of discrete filters for large eddy simulation. Center for Turbulence Research
Annual Research Briefs, 1997. 47

[59] Stephen B. Pope. Turbulent Flows. Cambridge University Press, 2000. 48, 55, 58

[60] S.B. Pope. Ten questions concerning the large-eddy simulation of turbulent flows. New Journal of
Physics, 6:1–24, 2004. 48

[61] J.W. Deardorff. Stratocumulus-capped mixed layers derived from a three-dimensional model.
Boundary-Layer Meteorol., 18:495–527, 1980. 48

[62] J. Bardina, J. H. Ferziger, and W. C. Reynolds. Improved Subgrid Scale Models for Large Eddy Sim-
ulation. In AIAA 13th Fluid & Plasma Dynamics Conference, AIAA-80-1357, Snowmass, Colorado,
July 1980. American Institute of Aeronautics and Astronautics. 49

[63] B. Vreman. An eddy-viscosity subgrid-scale model for turbulent shear flow: Algebraic theory and
applications. Physics of Fluids, 16(10):3670–3681, 2004. 49

[64] F. Nicoud and F. Ducros. Subgrid-Scale Stress Modelling Based on the Square of the Velocity Gra-
dient Tensor. Flow, Turbulence, and Combustion, 62:183–200, 1999. 49, 50

[65] D.C. Wilcox. Turbulence Modeling for CFD. DCW Industries, Inc., 2nd edition, 1998. 49

[66] W. Kümmel. Technische strömungsmechanik. Technical report, B.G. Teubner, 3. Auflage, 2007. 53

[67] J. S. Baggett. Some modeling requirements for wall models in large eddy simulation. Stanford Center
for Turbulence Research Annual Research Briefs, 1997. 55

[68] J. S. Baggett. On the feasibility of merging LES with RANS for the near-wall region of attached
turbulent flows. Stanford Center for Turbulence Research Annual Research Briefs, 1998. 55

[69] W. Cabot. Large-eddy simulations with wall models. Stanford Center for Turbulence Research
Annual Research Briefs, 1995. 55

[70] Pierre Sagaut. Large Eddy Simulation for Incompressible Flows. Springer, 2001. 55

[71] L. F. Moody. Friction factors for pipe flow. Transactions of the ASME, 66, 1944. 55

[72] Bruce R. Munson, Donald F. Young, and Theodore H. Okiishi. Fundamentals of Fluid Mechanics.
John Wiley and Sons, 1990. 56

[73] R.L. Panton. Incompressible Flow. Wiley, New Jersey, 2005. 58, 59

[74] J.P. Holman. Heat Transfer. McGraw-Hill, New York, 7th edition, 1990. 60, 65, 66, 67, 68

279
[75] J. Kim and P. Moin. Transport of Passive Scalars in Turbulent Channel Flow. Technical Memorandum
89463, National Aeronautics and Space Administration, Moffet Field, California, September 1987.
62

[76] R. Moser, J. Kim, and N. Mansour. Direct numerical simulation of turbulent channel flow up to
Reτ = 590. Phys. Fluids, 11:943–945, 1999. 62

[77] T.L. Bergman, A.S. Lavine, F.P. Incropera, and D.P. DeWitt. Fundamentals of Heat and Mass Trans-
fer. John Wiley and Sons, New York, 7th edition, 2011. 63, 118, 159

[78] T. Yuge. Experiments on Heat Transfer from Spheres Including Combined Natural and Forced Con-
vection. Journal of Heat Transfer, 82:214, 1960. 68

[79] W.S. Amato and C.L. Tien. Free Convection Heat Transfer from Isothermal Spheres in Water. Int. J.
Heat Mass Transfer, 15:327, 1972. 68

[80] L. Casara and T. Arts. Aerodynamic Performance of a Rib Roughened Cooling Channel Flow with
High Blockage Ratio. In Eleventh International Symposium on Application of Laser Techniques to
Fluid Mechanics, Lisbon, Portugal, 2002. 69, 70, 72

[81] L. Casara and T. Arts. Experimental Investigation of the Aerothermal Performance of a High Block-
age Rib-Roughened Cooling Channel. Journal of Turbomachinary, 127:580–588, 2005. 69, 70, 72

[82] M. Lohász, P. Rambaud, and C. Benocci. LES Simulation of Ribbed Square Duct Flow with Fluent
and Comparison with PIV Data. In The 12th International Conference on Fluid Flow Technologies,
Budapest, Hungary, 2003. 69

[83] P. von Schoenberg. Development of a Large-Eddy Simulation code for the atmospheric boundary
layer. FOI-R–1533–SE, Swedish Defense Research Agency, 2004. 74, 75

[84] R. Siegel and J. R. Howell. Thermal Radiation Heat Transfer. Taylor & Francis, New York, 4th
edition, 2002. 109, 110

[85] Y.B. Zel’dovich and Y.P. Raizer. Physics of shock waves and high-temperature hydrodynamic phe-
nomena. Dover Publications, New York, 2002. Translated from the Russian and then edited by
W.D.Hayes and R.F. Probstein. 111

[86] C.K. Westbrook and F.L. Dryer. Simplified Reaction Mechanisms for the Oxidation of Hydrocarbon
Fuels in Flames. Combustion Science and Technology, 27:31–43, 1981. 131

[87] S. Gordon and B.J. McBride. Computer program for calculation of complex chemical equilibrium
compositions and applications. NASA Reference Publication 1311, National Aeronautics and Space
Administration, Office of Management, Scientific and Technical Information Program, Cleveland,
Ohio, 1994. 131

[88] S.R. Turns. An Introduction to Combustion. McGraw-Hill, New York, 2nd edition, 1996. 137

[89] D.A. Purser. SFPE Handbook of Fire Protection Engineering, chapter Combustion Toxicity. Springer,
New York, 5th edition, 2016. 139, 140

[90] D. Drysdale. An Introduction to Fire Dynamics. John Wiley and Sons, New York, 2nd edition, 2002.
155, 165

280
 [91] H.S. Carslaw and J.C. Jaegar. Conduction of Heat in Solids. Oxford University Press, 2nd edition,
1959. 155, 165, 168

[92] K.W. Childs. HEATING 7: Multidimensional, Finite-Difference Heat Conduction Analysis Code
System. Technical Report PSR-199, Oak Ridge National Laboratory, Oak Ridge, TN, 1998. 156

[93] A. Khawam and D. R. Flanagan. Solid-State Kinetic Models: Basics and Mathematical Fundamen-
tals. J. Phys. Chem. B, 110:17315–17328, 2006. 177

[94] R.E. Lyon. Heat Release Kinetics. Fire and Materials, 24:179–186, 2000. 189

[95] D. Morvan and J.L. Dupuy. Modeling the propagation of a wildfire through a Mediterranean shrub
using a multiphase formulation. Combustion and Flame, 138(3):199–210, 2004. 198

[96] R.E. Lyon and R.N. Walters. Pyrolysis Combustion Flow Calorimetry. Journal of Analytical and
Applied Pyrolysis, 71(1):27–46, March 2004. 199

[97] American Society for Testing and Materials, West Conshohocken, Pennsylvania. ASTM D 7309-
11, Standard Test Method for Determining Flammability Characteristics of Plastics and Other Solid
Materials Using Microscale Combustion Calorimetry, 2011. 199

[98] W. E. Ranz and W. R. Marshall. Evaporation from drops - Part II. Chemical Engineering Progress,
48:173–180, March 1952. 219

[99] Y.F. Li. and W.K. Chow. Study of water droplet behavior in hot air layer in fire extinguishment. Fire
Technology, 44:351–381, 2008. 221

[100] C.W. Gardiner. Handbook of Stochastic Methods for Physics, Chemistry and the Natural Sciences.
Springer-Verlag, Berlin, 3rd edition, 2004. 225

[101] M.S. Owen. 2009 ASHRAE Handbook: Fundamentals. American Society of Heating, Refrigerating,
and Air-Conditioning Engineers, Inc., Atlanta, GA, 2009. 240

[102] R.G. Rehm and R. McDermott. Fire-Front Propagation Using the Level Set Method. NISTTN 1611,
National Institute of Standards and Technology, Gaithersburg, Maryland, April 2009. 249

[103] G.D. Richards. An elliptical growth model of forest fire fronts and its numerical solution. Interna-
tional Journal for Numerical Methods in Engineering, 30:1163–1179, 1990. 249

[104] M.A. Finney. Fire Area Simulator – Model, Development and Evaluation. Research Paper RMRS-
RP-4 Revised, United States Forest Service, Rocky Mountain Research Station, Missoula, Montana,
February 2004. http://www.fs.fed.us/rm/pubs/rmrs_rp004.pdf. 249

[105] R.C. Rothermel. A Mathematical Model for Predicting Fire Spread in Wildland Fuels. Research
Paper INT-115, Intermountain Forest and Range Experiment Station, USDA Forest Service, Ogden,
Utah, January 1972. http://www.treesearch.fs.fed.us/pubs/32533. 249, 250

[106] H.E. Anderson. Predicting wind-driven wild land fire size and shape. Research Paper INT-305,
Intermountain Forest and Range Experiment Station, USDA Forest Service, Ogden, Utah, 1983. 249

[107] A.S. Bova, W.E. Mell, and C.M. Hoffman. A comparison of level set and marker methods for the
simulation of wildland fire front propagation. International Journal of Wildland Fire, 25(2):229–241,
2015. http://dx.doi.org/10.1071/WF13178. 249

281
[108] P.L. Andrews. Modeling wind adjustment factor and midflame wind speed for Rothermel’s surface
fire spread model. General Technical Report RMRS-GTR-266, Rocky Mountain Research Station,
USDA Forest Service, Fort Collins, Colorado, 2012. 250

[109] M. Berger. Chapter 1 - cut cells: Meshes and solvers. In Rémi Abgrall and Chi-Wang Shu, editors,
Handbook of Numerical Methods for Hyperbolic Problems, volume 18 of Handbook of Numerical
Analysis, pages 1–22. Elsevier, 2017. 259

[110] Weinan E and Jian-Guo Liu. Gauge Method for Viscous Incompressible Flows. Communications in
Mathematical Sciences, 1(2):317–332, 2003. 263

[111] G.E. Danielson and R.H. Lagdon. Safety Software Guide for Use with 10 CFR 830 Subpart A,
Qualiity Assurance Requirements, and DOE O 414.1C, Quality Assurance. DOE G 414.1-4, U.S.
Department of Energy, Washington, DC, June 2005. 303

[112] F. Odar. Software Quality Assurance Procedures for NRC Thermal Hydraulic Codes. NUREG-1737,
U.S. Nuclear Regulatory Commission, Washington, DC, December 2000. 303

282
Appendix A

Summary of Verification Results

This appendix summarizes the accuracy of cases in the FDS verification suite. These cases are an integral
part of the continuous integration process for FDS in which daily source code changes are checked each
night against this growing list of cases to verify the overall accuracy of FDS and to quickly identify errors
introduced into the source code.
The verification cases that are listed in Table A.1 are described either in the FDS Verification Guide or
the FDS User’s Guide. Those cases that are in the User’s Guide help to explain various input parameters.
Dark blue links refer to sections in the Verification Guide; light blue links refer to the User’s Guide. For
each case, there are one or more point to point comparisons between an FDS predicted value and an exact
or expected value. If the relative or absolute difference between the predicted and exact/expected values is
within the specified tolerance, the case passes.

283
 Table A.1: Summary of verification cases.

Case Name Section Expected Expected Predicted Predicted Type of Error Error Within
Quantity Value Quantity Value Error Tolerance Tol.
1_step_2_step 10.8 Q1S 1.61e+02 Q2S 1.61e+02 Relative 3.05e-04 5.00e-03 Yes
activate_vents 18.5.11 controller 1 1.00e+00 controller 1 1.00e+00 Relative 0.00e+00 0.00e+00 Yes
activate_vents 18.5.11 controller 3 1.00e+00 controller 3 1.00e+00 Relative 0.00e+00 0.00e+00 Yes
activate_vents 18.5.11 controller 4 1.00e+00 controller 4 1.00e+00 Relative 0.00e+00 0.00e+00 Yes
adiabatic_con_flux 11.8 pres 2.40e+04 pres 2.39e+04 Relative 6.01e-03 1.00e-02 Yes
adiabatic_con_flux 11.8 temp 6.92e+01 temp_gas 6.91e+01 Relative 1.97e-03 1.00e-02 Yes
adiabatic_net_flux 11.8 pres 2.40e+04 pres 2.39e+04 Relative 2.90e-03 1.00e-02 Yes
adiabatic_net_flux 11.8 temp 6.92e+01 temp_gas 6.93e+01 Relative 7.31e-04 1.00e-02 Yes
adiabatic_surface_temperature 22.10.13 temp 3.36e+02 AST 3.36e+02 Relative 1.32e-04 1.00e-02 Yes
adiabatic_surface_temperature 22.10.13 AST 3.36e+02 AST-Gas 3.36e+02 Relative 0.00e+00 1.00e-04 Yes
aerosol_agglomeration 10.11.5 Exact Total 1.00e-05 FDS Total 1.00e-05 Relative 0.00e+00 2.00e-02 Yes
aerosol_agglomeration 10.11.5 Exact Bin 1 1.00e-05 FDS Bin 1 1.00e-05 Relative 4.00e-06 2.00e-02 Yes
aerosol_agglomeration 10.11.5 Exact Bin 2 1.06e-09 FDS Bin 2 1.06e-09 Relative 2.17e-06 2.00e-02 Yes
aerosol_agglomeration 10.11.5 Exact Total 1.00e-05 FDS Total 1.00e-05 Relative 0.00e+00 2.00e-02 Yes
aerosol_agglomeration 10.11.5 Exact Bin 1 1.00e-05 FDS Bin 1 1.00e-05 Relative 2.63e-06 2.00e-02 Yes
aerosol_agglomeration 10.11.5 Exact Bin 2 1.08e-09 FDS Bin 2 1.07e-09 Relative 5.79e-03 2.00e-02 Yes
aerosol_gravitational_deposition 10.11.1 Exact Gas 1 7.21e-06 FDS Gas 1 7.28e-06 Relative 9.73e-03 2.50e-02 Yes
aerosol_gravitational_deposition 10.11.1 Exact Gas 2 1.20e-05 FDS Gas 2 1.20e-05 Relative 2.67e-03 2.50e-02 Yes

284
aerosol_gravitational_deposition 10.11.1 Exact Wall Top 1.00e-12 FDS Wall Top 1.00e-12 Relative 0.00e+00 2.00e-02 Yes
aerosol_gravitational_deposition 10.11.1 Exact Wall Bottom 5.26e-08 FDS Wall Bottom 5.27e-08 Relative 1.03e-03 2.00e-02 Yes
aerosol_gravitational_deposition 10.11.1 Exact Gas 1 1.20e-05 FDS Gas 1 1.20e-05 Relative 2.67e-03 2.50e-02 Yes
aerosol_gravitational_deposition 10.11.1 Exact Gas 2 7.21e-06 FDS Gas 2 7.28e-06 Relative 9.73e-03 2.50e-02 Yes
aerosol_gravitational_deposition 10.11.1 Exact Wall Top 5.26e-08 FDS Wall Top 5.27e-08 Relative 1.03e-03 2.00e-02 Yes
aerosol_gravitational_deposition 10.11.1 Exact Wall Bottom 1.00e-12 FDS Wall Bottom 1.00e-12 Relative 0.00e+00 2.00e-02 Yes
aerosol_thermophoretic_deposition 10.11.2 Exact Gas 6.01e-06 FDS Gas 6.06e-06 Relative 8.28e-03 2.00e-02 Yes
aerosol_thermophoretic_deposition 10.11.2 Exact Wall 3.69e-09 FDS Wall 3.70e-09 Relative 2.57e-03 2.00e-02 Yes
aerosol_thermophoretic_deposition 10.11.2 Exact Gas 6.01e-06 FDS Gas 6.06e-06 Relative 8.28e-03 2.00e-02 Yes
aerosol_thermophoretic_deposition 10.11.2 Exact Wall 3.69e-09 FDS Wall 3.70e-09 Relative 2.57e-03 2.00e-02 Yes
aerosol_turbulent_deposition 10.11.3 Exact Wall 1 1.78e-14 FDS Wall 1 1.78e-14 Relative 9.86e-05 2.00e-02 Yes
aerosol_turbulent_deposition 10.11.3 Exact Wall 2 2.92e-08 FDS Wall 2 2.92e-08 Relative 1.00e-04 2.00e-02 Yes
aerosol_turbulent_deposition 10.11.3 Exact Wall 3 3.61e-07 FDS Wall 3 3.61e-07 Relative 1.00e-04 2.00e-02 Yes
aerosol_scrubbing 10.11.6 MF 1.00e-04 MF 1.00e-04 Relative 0.00e+00 2.00e-02 Yes
ashrae7_fixed_flow 14.11 Duct 1 7.16e+02 Duct 1 7.14e+02 Relative 2.62e-03 1.00e-02 Yes
ashrae7_fixed_flow 14.11 Duct 2 4.41e+02 Duct 2 4.40e+02 Relative 2.52e-03 1.00e-02 Yes
ashrae7_fixed_flow 14.11 Duct 3 4.41e+02 Duct 3 4.40e+02 Relative 2.52e-03 1.00e-02 Yes
ashrae7_fixed_flow 14.11 Duct 4 2.75e+02 Duct 4 2.74e+02 Relative 2.71e-03 1.00e-02 Yes
ashrae7_fixed_flow 14.11 Duct 5 1.21e+02 Duct 5 1.21e+02 Relative 2.87e-03 1.00e-02 Yes
ashrae7_fixed_flow 14.11 Duct 56 7.31e+02 Duct 56 7.29e+02 Relative 3.17e-03 1.00e-02 Yes
ashrae7_fixed_flow 14.11 Duct 6 3.22e+01 Duct 6 3.21e+01 Relative 2.78e-03 1.00e-02 Yes
ashrae7_fixed_flow 14.11 Duct 7 3.13e+02 Duct 7 3.12e+02 Relative 3.76e-03 1.00e-02 Yes
ashrae7_quadratic 14.11 Duct 1 7.16e+02 Duct 1 7.16e+02 Relative 8.77e-04 1.00e-02 Yes
ashrae7_quadratic 14.11 Duct 2 4.41e+02 Duct 2 4.41e+02 Relative 7.85e-04 1.00e-02 Yes
ashrae7_quadratic 14.11 Duct 3 4.41e+02 Duct 3 4.41e+02 Relative 7.84e-04 1.00e-02 Yes
 Case Name Section Expected Expected Predicted Predicted Type of Error Error Within
Quantity Value Quantity Value Error Tolerance Tol.
ashrae7_quadratic 14.11 Duct 4 2.75e+02 Duct 4 2.75e+02 Relative 9.67e-04 1.00e-02 Yes
ashrae7_quadratic 14.11 Duct 5 1.21e+02 Duct 5 1.21e+02 Relative 1.12e-03 1.00e-02 Yes
ashrae7_quadratic 14.11 Duct 56 7.31e+02 Duct 56 7.30e+02 Relative 1.42e-03 1.00e-02 Yes
ashrae7_quadratic 14.11 Duct 6 3.22e+01 Duct 6 3.22e+01 Relative 1.01e-03 1.00e-02 Yes
ashrae7_quadratic 14.11 Duct 7 3.13e+02 Duct 7 3.13e+02 Relative 2.03e-03 1.00e-02 Yes
ashrae7_table 14.11 Duct 1 7.16e+02 Duct 1 7.14e+02 Relative 3.44e-03 1.00e-02 Yes
ashrae7_table 14.11 Duct 2 4.41e+02 Duct 2 4.40e+02 Relative 3.35e-03 1.00e-02 Yes
ashrae7_table 14.11 Duct 3 4.41e+02 Duct 3 4.40e+02 Relative 3.35e-03 1.00e-02 Yes
ashrae7_table 14.11 Duct 4 2.75e+02 Duct 4 2.74e+02 Relative 3.54e-03 1.00e-02 Yes
ashrae7_table 14.11 Duct 5 1.21e+02 Duct 5 1.21e+02 Relative 3.70e-03 1.00e-02 Yes
ashrae7_table 14.11 Duct 56 7.31e+02 Duct 56 7.28e+02 Relative 4.01e-03 1.00e-02 Yes
ashrae7_table 14.11 Duct 6 3.22e+01 Duct 6 3.21e+01 Relative 3.62e-03 1.00e-02 Yes
ashrae7_table 14.11 Duct 7 3.13e+02 Duct 7 3.12e+02 Relative 4.58e-03 1.00e-02 Yes
aspiration_detector 18.3.7 ideal 4.57e+01 asp1 4.62e+01 Relative 1.19e-02 2.00e-02 Yes
back_wall_test 11.6 Flux_in 9.33e+00 Flux_out 9.33e+00 Relative 3.61e-06 5.00e-03 Yes
back_wall_test 11.6 Temp_thin 3.71e+02 Temp_thicker 3.71e+02 Relative 0.00e+00 5.00e-03 Yes
back_wall_test_2 11.7 T_F_Target 7.82e+02 T_F1 7.67e+02 Relative 1.83e-02 2.00e-02 Yes
back_wall_test_2 11.7 T_B_Target 7.56e+02 T_B1 7.43e+02 Relative 1.73e-02 2.00e-02 Yes
back_wall_test_2 11.7 T_F_Target 7.82e+02 T_F2 7.68e+02 Relative 1.83e-02 2.00e-02 Yes
back_wall_test_2 11.7 T_B_Target 7.56e+02 T_B2 7.43e+02 Relative 1.72e-02 2.00e-02 Yes
back_wall_test_2 11.7 T_F_Target 7.82e+02 T_F3 7.89e+02 Relative 9.65e-03 2.00e-02 Yes
back_wall_test_2 11.7 T_B_Target 7.56e+02 T_B3 7.64e+02 Relative 9.43e-03 2.00e-02 Yes
beam_detector 18.3.6 Obs 9.56e+01 beam_1 9.56e+01 Relative 8.61e-06 1.00e-02 Yes

285
beam_detector 18.3.6 Obs 9.56e+01 beam_2 9.56e+01 Relative 1.03e-05 1.00e-02 Yes
beam_detector 18.3.6 Obs 9.56e+01 beam_3 9.56e+01 Relative 5.27e-06 1.00e-02 Yes
bound_test_1 3.8.1 F1 1.00e-01 XF1 1.00e-01 Absolute 0.00e+00 1.00e-03 Yes
bound_test_1 3.8.1 A1 1.00e-12 XA1 1.19e-17 Absolute 1.00e-12 1.00e-03 Yes
bound_test_1 3.8.1 P1 4.00e-01 XP1 4.00e-01 Absolute 0.00e+00 1.00e-03 Yes
bound_test_1 3.8.1 F2 1.00e-12 XF2 1.00e-12 Absolute 0.00e+00 1.00e-03 Yes
bound_test_1 3.8.1 A2 3.00e-01 XA2 3.00e-01 Absolute 0.00e+00 1.00e-03 Yes
bound_test_1 3.8.1 P2 2.00e-01 XP2 2.00e-01 Absolute 0.00e+00 1.00e-03 Yes
bound_test_1 3.8.1 Sum 1.00e+00 SUM 1.00e+00 Absolute 0.00e+00 1.00e-03 Yes
bound_test_2 3.8.1 F1 1.00e-12 XF1 1.00e-12 Absolute 0.00e+00 1.00e-03 Yes
bound_test_2 3.8.1 F2 2.40e-01 XF2 2.40e-01 Absolute 0.00e+00 1.00e-03 Yes
bound_test_2 3.8.1 A1 2.40e-01 XA1 2.40e-01 Absolute 0.00e+00 1.00e-03 Yes
bound_test_2 3.8.1 A2 1.00e-12 XA2 2.98e-17 Absolute 1.00e-12 1.00e-03 Yes
bound_test_2 3.8.1 P1 1.20e-01 XP1 1.20e-01 Absolute 0.00e+00 1.00e-03 Yes
bound_test_2 3.8.1 P2 3.20e-01 XP2 3.20e-01 Absolute 0.00e+00 1.00e-03 Yes
bound_test_2 3.8.1 P3 8.00e-02 XP3 8.00e-02 Absolute 0.00e+00 1.00e-03 Yes
bound_test_2 3.8.1 Sum 1.00e+00 SUM 1.00e+00 Absolute 0.00e+00 1.00e-03 Yes
box_burn_away1 9.2.7 Mass (kg) 1.28e+00 Mass fuel 1.28e+00 Relative 8.85e-04 2.00e-02 Yes
box_burn_away2 9.2.7 Mass (kg) 1.28e+00 Mass GAS 1.28e+00 Relative 2.16e-04 2.00e-02 Yes
box_burn_away3 9.2.7 Mass (kg) 1.28e+00 Mass fuel 1.28e+00 Relative 3.16e-03 2.00e-02 Yes
box_burn_away4 9.2.7 Mass4 (kg) 9.60e-01 Mass fuel 9.61e-01 Relative 1.18e-03 2.00e-02 Yes
box_burn_away5 9.2.7 Mass (kg) 1.28e+00 Mass fuel 1.28e+00 Relative 1.43e-04 2.00e-02 Yes
box_burn_away6 9.2.7 Mass_residue (kg) 6.40e-01 Mass fuel 1 6.40e-01 Relative 1.07e-04 2.00e-02 Yes
box_burn_away6 9.2.7 Mass (kg) 1.28e+00 Mass fuel 2 1.28e+00 Relative 1.07e-04 2.00e-02 Yes
 Case Name Section Expected Expected Predicted Predicted Type of Error Error Within
Quantity Value Quantity Value Error Tolerance Tol.
box_burn_away7 9.2.7 Mass (kg) 3.20e-02 Mass fuel 3.19e-02 Relative 3.22e-03 2.00e-02 Yes
box_burn_away_2D 9.2.7 Mass (kg) 1.28e+00 Mass fuel 1.28e+00 Relative 2.24e-04 2.00e-02 Yes
box_burn_away_2D 9.2.7 Mass_end (kg) 1.00e-12 Mass solid 1.00e-12 Relative 0.00e+00 2.00e-02 Yes
box_burn_away8 9.2.7 Mass4 (kg) 9.60e-01 Mass fuel 9.61e-01 Relative 1.18e-03 2.00e-02 Yes
box_burn_away_2D_residue 9.2.7 Mass_residue (kg) 6.40e-01 Mass fuel 6.40e-01 Relative 9.64e-05 2.00e-02 Yes
box_burn_away_2D_residue 9.2.7 Mass_end (kg) 1.00e-12 Mass solid 1.00e-12 Relative 0.00e+00 2.00e-02 Yes
bucket_test_1 22.9.1 Mass (kg) 9.00e+01 Mass 8.87e+01 Relative 1.49e-02 2.00e-02 Yes
bucket_test_2 18.3.1 Mass (kg) 5.00e+00 Mass 4.95e+00 Relative 1.06e-02 3.00e-02 Yes
bucket_test_3 15.5.3 Mass (kg) 8.00e-03 Mass 7.98e-03 Relative 2.84e-03 2.00e-02 Yes
bucket_test_4 22.9.3 Mass (kg) 2.75e-02 total 2.75e-02 Relative 0.00e+00 2.00e-02 Yes
bucket_test_4 22.9.3 Mass Flux (kg/s) 5.00e-04 flux 5.01e-04 Relative 2.02e-03 5.00e-02 Yes
bulk_density_file 17.2.2 dry canopy mass 3.50e+02 dry canopy mass 3.50e+02 Absolute 3.00e-05 1.00e-03 Yes
burke_schumann_spec 10.6 Fuel 1.00e+00 FDS_Fuel 1.00e+00 Absolute 0.00e+00 1.00e-03 Yes
burke_schumann_spec 10.6 Ox 1.00e-12 FDS_Ox 1.00e-12 Absolute 0.00e+00 1.00e-03 Yes
burke_schumann_spec 10.6 Prod 1.00e-12 FDS_Prod 1.00e-12 Absolute 0.00e+00 1.00e-03 Yes
burke_schumann_temp 10.6 Temperature 1.00e-12 FDS_Temperature 1.00e-12 Absolute 0.00e+00 1.00e-03 Yes
cable_11_insulation 12.9.6 HRR (W/g) 2.44e+04 hrrpum 2.51e+04 Relative 2.88e-02 3.00e-02 Yes
cable_23_insulation 12.9.6 HRR (W/g) 2.34e+04 hrrpum 2.35e+04 Relative 4.61e-03 1.00e-02 Yes
cable_701_insulation 12.9.6 HRR (W/g) 1.27e+04 hrrpum 1.26e+04 Relative 5.64e-03 1.00e-02 Yes
cable_11_jacket 12.9.6 HRR (W/g) 8.10e+03 hrrpum 7.89e+03 Relative 2.54e-02 5.00e-02 Yes
cable_23_jacket 12.9.6 HRR (W/g) 7.34e+03 hrrpum 7.14e+03 Relative 2.76e-02 5.00e-02 Yes
cable_701_jacket 12.9.6 HRR (W/g) 1.41e+04 hrrpum 1.41e+04 Relative 4.49e-03 1.00e-02 Yes
cannon_ball 13.1.1 Distance 2.89e+01 d 2.89e+01 Relative 1.94e-04 1.00e-02 Yes

286
cascade 15.7.1 Mass 1.00e+00 water mass 9.95e-01 Relative 4.92e-03 2.00e-02 Yes
clocks_restart 8.10 HRR 1.26e+03 HRR 1.26e+03 Relative 2.75e-06 1.00e-02 Yes
cell_burn_away 12.3 m_dot 3.25e-03 MLR_PROPANE 3.19e-03 Relative 2.01e-02 4.00e-02 Yes
char_oxidation_1 17.1.3 HR 1.01e+03 HRR 9.94e+02 Relative 1.74e-02 3.00e-02 Yes
char_oxidation_1 17.1.3 mass 5.31e-04 mass foliage 5.31e-04 Relative 3.77e-04 1.00e-02 Yes
char_oxidation_2 17.1.3 Delta e_part -2.41e-01 Delta e_part dry needle -2.41e-01 Relative 1.19e-04 1.00e-01 Yes
char_oxidation_2 17.1.3 Delta e_gas 2.41e-01 Delta e_gas dry needle 2.41e-01 Relative 8.69e-05 1.00e-01 Yes
char_oxidation_2 17.1.3 mass fuel gas 2.95e-05 mass fuel gas dry needle 2.95e-05 Relative 1.93e-05 1.00e-02 Yes
char_oxidation_2 17.1.3 mass CO2 2.58e-05 mass CO2 dry needle 2.58e-05 Relative 1.36e-06 1.00e-02 Yes
char_oxidation_2 17.1.3 mass foliage 1.96e-07 dry needle mass 1.96e-07 Relative 1.78e-03 1.00e-02 Yes
char_oxidation_2 17.1.3 mass O2 -1.62e-05 mass O2 dry needle -1.62e-05 Relative 7.18e-05 1.00e-02 Yes
circular_burner 7.4.2 Mdot 1.57e-02 MLR_PROPANE 1.57e-02 Relative 7.13e-05 1.00e-02 Yes
condensation_evaporation 10.12.1 temp -2.71e+01 Temp -2.70e+01 Relative 5.20e-03 2.00e-02 Yes
condensation_evaporation 10.12.1 pres -5.80e+03 pres -5.78e+03 Relative 5.09e-03 2.00e-02 Yes
condensation_evaporation 10.12.1 temp 3.64e+01 Temp 3.65e+01 Relative 2.22e-03 2.00e-02 Yes
condensation_evaporation 10.12.1 pres 2.10e+04 pres 2.14e+04 Relative 1.82e-02 2.00e-02 Yes
condensation_radiation 10.12.3 PART 2.80e+01 COND 2.80e+01 Relative 5.52e-04 1.00e-02 Yes
condensation_evaporation 10.12.1 Tgas 3.20e+01 Temp 3.21e+01 Relative 1.90e-03 5.00e-02 Yes
condensation_evaporation 10.12.1 dP -2.02e+04 pres -2.00e+04 Relative 6.70e-03 5.00e-02 Yes
control_test_2 18.5.9 Add 3.00e+01 CF Add 3.00e+01 Relative 6.13e-06 1.00e-02 Yes
control_test_2 18.5.9 Multiply 1.00e+01 CF Multiply 1.00e+01 Relative 5.70e-06 1.00e-02 Yes
control_test_2 18.5.9 Subtract -1.00e+01 CF Subtract -1.00e+01 Relative 4.10e-05 1.00e-02 Yes
control_test_2 18.5.9 Divide 2.00e+00 CF Divide 2.00e+00 Relative 3.54e-05 1.00e-02 Yes
control_test_2 18.5.9 Power 3.16e+00 CF Power 3.16e+00 Relative 7.06e-04 1.00e-02 Yes
 Case Name Section Expected Expected Predicted Predicted Type of Error Error Within
Quantity Value Quantity Value Error Tolerance Tol.
control_test_2 18.5.9 PID -1.50e+00 CF PID -1.50e+00 Relative 3.33e-04 1.00e-02 Yes
convective_cooling 11.3 Temp 2.95e+02 inner temp 2.96e+02 Relative 1.82e-03 1.00e-02 Yes
couch 9.2.2 HRR 3.60e+03 HRR 4.10e+03 Relative 1.38e-01 2.00e-01 Yes
cyl_test 8.1.3 temp1 9.03e+01 temp2 9.03e+01 Relative 0.00e+00 1.00e-02 Yes
cyl_test 8.1.3 temp1 7.12e+02 temp1 7.12e+02 Relative 0.00e+00 1.00e-02 Yes
device_restart 8.10 temp 6.70e+01 temp 6.71e+01 Absolute 1.38e-01 5.00e+00 Yes
devc_interpolation 17.2 u 1.00e+00 U-VEL-3 1.00e+00 Absolute 0.00e+00 1.00e-02 Yes
devc_interpolation 17.2 T 2.50e+01 TMP-3 2.50e+01 Absolute 0.00e+00 1.00e-02 Yes
divergence_test_1 8.3.1 div_min 1.00e-12 div_min -3.53e-14 Absolute 1.04e-12 1.00e-10 Yes
divergence_test_1 8.3.1 div_max 1.00e-12 div_max 2.59e-14 Absolute 9.74e-13 1.00e-10 Yes
divergence_test_2 8.3.2 div_min -5.00e-16 Vdot -1.07e-17 Absolute 4.89e-16 1.00e-10 Yes
divergence_test_2 8.3.2 div_max 5.00e-16 Vdot -1.07e-17 Absolute 5.11e-16 1.00e-10 Yes
divergence_test_3 8.3.3 div_min 1.00e-12 div_min -4.92e-13 Absolute 1.49e-12 1.00e-10 Yes
divergence_test_3 8.3.3 div_max 1.00e-12 div_max 4.96e-13 Absolute 5.04e-13 1.00e-10 Yes
door_crack 10.3.2 Pressure 9.94e+02 Pressure 9.89e+02 Relative 4.29e-03 2.00e-02 Yes
door_crack 10.3.2 HRR 1.60e+02 HRR 1.53e+02 Relative 4.52e-02 8.00e-02 Yes
dragon_5a 17.2.3 Mass 2.40e-02 Mass 2.39e-02 Relative 4.74e-03 5.00e-02 Yes
droplet_absorption_cart 9.7 Absorption 8.88e+00 ADT 8.82e+00 Relative 6.80e-03 1.00e-02 Yes
droplet_absorption_cyl 9.7 Absorption 7.11e+00 ADT 7.06e+00 Relative 6.97e-03 1.00e-02 Yes
droplet_distributions 15.3.3 Mass 5.78e-02 LN_M 5.78e-02 Relative 1.91e-04 2.00e-02 Yes
droplet_distributions 15.3.3 Mass 5.78e-02 RR-LN_M 5.86e-02 Relative 1.40e-02 2.00e-02 Yes
droplet_distributions 15.3.3 Mass 5.78e-02 RR_M 5.81e-02 Relative 4.68e-03 2.00e-02 Yes
droplet_distributions 15.3.3 Mass 5.78e-02 USER_M 5.83e-02 Relative 8.12e-03 2.00e-02 Yes

287
duct_flow 21.1.2 Flow 1.00e+00 flow_out 9.87e-01 Relative 1.34e-02 5.00e-02 Yes
duct_flow 21.1.2 Flow 1.00e+00 flow_out 1.00e+00 Relative 0.00e+00 5.00e-02 Yes
e_coefficient 15.7.2 HRR 1.00e+01 HRR_1 1.11e+01 Absolute 1.12e+00 3.00e+00 Yes
e_coefficient 15.7.2 HRR 1.00e+01 HRR_2 1.08e+01 Absolute 7.78e-01 3.00e+00 Yes
emissivity 12.12 Emissivity 1.00e-12 EMISSIVITY 9.00e-03 Absolute 9.00e-03 1.00e-01 Yes
energy_budget_adiabatic_walls 7.2.1 HRR 1.20e+03 HRR 1.20e+03 Absolute 1.13e-01 1.20e+01 Yes
energy_budget_adiabatic_walls 7.2.1 Net Enthalpy Change 1.00e-12 Q_TOTAL 8.35e+00 Absolute 8.35e+00 1.20e+01 Yes
energy_budget_adiabatic_walls 7.2.1 Q CONV 1.20e+03 Q_CONV 1.20e+03 Absolute 2.18e+00 1.20e+01 Yes
energy_budget_adiabatic_two_fuels 13.2.4 HRR 1.20e+03 HRR 1.20e+03 Absolute 9.22e-01 1.20e+01 Yes
energy_budget_cold_walls 7.2.1 HRR 1.20e+03 HRR 1.20e+03 Absolute 1.89e-02 2.00e+01 Yes
energy_budget_cold_walls 7.2.1 Net Enthalpy Change 1.00e-12 Q_TOTAL 2.30e+00 Absolute 2.30e+00 2.00e+01 Yes
energy_budget_cold_walls 7.2.1 Losses 1.20e+03 Q_RADI+Q_CONV+Q_COND+Q_DIFF 1.20e+03 Absolute 2.28e+00 2.00e+01 Yes
energy_budget_combustion 7.2.6 dp 3.69e+02 dh 3.75e+02 Relative 1.55e-02 5.00e-02 Yes
energy_budget_dns_100 7.2.4 div -1.36e-08 div -1.36e-08 Relative 2.97e-04 1.00e-02 Yes
energy_budget_particles 7.2.9 Q_ENTH -7.50e+01 Q_TOTAL -7.50e+01 Relative 1.64e-04 1.00e-02 Yes
energy_budget_tmix 7.2.5 Temp 1.00e+02 Temp 1.00e+02 Relative 2.37e-04 1.00e-02 Yes
energy_budget_solid 7.2.7 T_g Exact 9.63e+01 T_g FDS 9.65e+01 Relative 1.80e-03 1.00e-02 Yes
energy_budget_solid 7.2.7 T_w Exact 9.63e+01 T_w FDS 9.64e+01 Relative 6.87e-04 1.00e-02 Yes
energy_budget_solid 7.2.7 P Exact 2.27e+04 P FDS 2.27e+04 Relative 1.69e-03 1.00e-02 Yes
enthalpy 12.7 T_slab 1.56e+02 T_slab 1.56e+02 Relative 7.81e-04 1.00e-02 Yes
fan_test 14.7 pres_1 4.52e+00 pres_1 4.51e+00 Relative 1.20e-03 4.00e-02 Yes
fan_test 14.7 pres_2 -4.52e+00 pres_2 -4.52e+00 Relative 1.12e-03 4.00e-02 Yes
fan_test 14.7 vflow1 -4.98e-02 vflow1 -4.90e-02 Relative 1.66e-02 2.00e-02 Yes
fan_test 14.7 vflow2 4.98e-02 vflow2 4.90e-02 Relative 1.67e-02 2.00e-02 Yes
 Case Name Section Expected Expected Predicted Predicted Type of Error Error Within
Quantity Value Quantity Value Error Tolerance Tol.
qfan_test 14.8 pres_1 2.26e+00 pres_1 2.26e+00 Relative 1.99e-03 4.00e-02 Yes
qfan_test 14.8 pres_2 -2.26e+00 pres_2 -2.27e+00 Relative 4.27e-03 4.00e-02 Yes
qfan_test 14.8 vflow1 -4.91e-02 vflow1 -4.90e-02 Relative 1.69e-03 2.00e-02 Yes
qfan_test 14.8 vflow2 4.91e-02 vflow2 4.90e-02 Relative 1.73e-03 2.00e-02 Yes
qfan_multi 14.9 fpp1 9.03e+00 FT1 9.05e+00 Relative 2.75e-03 5.00e-02 Yes
qfan_multi 14.9 fpp2 2.90e+01 FT2 2.90e+01 Relative 1.09e-03 5.00e-02 Yes
qfan_multi 14.9 vflowp1 -2.58e-02 vflowp1 -2.57e-02 Relative 4.17e-03 2.00e-02 Yes
qfan_multi 14.9 vflowp2 9.99e-02 vflowp2 9.98e-02 Relative 9.14e-04 2.00e-02 Yes
qfan_multi 14.9 vflowpe 7.41e-02 vflowpe 7.41e-02 Relative 8.38e-05 2.00e-02 Yes
qfan_multi 14.9 fsp1 9.37e+00 FSI 9.36e+00 Relative 5.70e-04 2.00e-02 Yes
qfan_multi 14.9 fsp2 2.94e+01 FSO 2.93e+01 Relative 1.84e-03 2.00e-02 Yes
qfan_multi 14.9 vflows 1.02e-01 vflows 1.02e-01 Relative 3.59e-03 2.00e-02 Yes
FED_FIC 10.9.1 FED (O2 CO2 CO) 5.99e-01 FED (O2 CO2 CO) 5.99e-01 Relative 6.73e-04 1.00e-02 Yes
FED_FIC 10.9.1 FED (Asphyxiants) 9.74e-01 FED (Asphyxiants) 9.73e-01 Relative 1.51e-03 1.00e-02 Yes
FED_FIC 10.9.1 FED (Irritants) 8.26e-03 FED (Irritants) 8.25e-03 Relative 8.75e-04 1.00e-02 Yes
FED_FIC 10.9.1 FED (All) 5.14e-01 FED (All) 5.13e-01 Relative 1.38e-03 1.00e-02 Yes
FED_FIC 10.9.1 FIC (O2 CO2 CO) 1.00e-12 FIC (O2 CO2 CO) 1.00e-12 Absolute 0.00e+00 1.00e-02 Yes
FED_FIC 10.9.1 FIC (Asphyxiants) 1.00e-12 FIC (Asphyxiants) 1.00e-12 Absolute 0.00e+00 1.00e-02 Yes
FED_FIC 10.9.1 FIC (Irritants) 8.57e-01 FIC (Irritants) 8.56e-01 Relative 1.26e-03 1.00e-02 Yes
FED_FIC 10.9.1 FIC (All) 4.29e-01 FIC (All) 4.28e-01 Relative 1.06e-03 1.00e-02 Yes
FED_FIC_SMIX 10.9.1 FED (O2 CO2 CO) 5.99e-01 FED (O2 CO2 CO) 5.99e-01 Relative 6.73e-04 1.00e-02 Yes
FED_FIC_SMIX 10.9.1 FED (Asphyxiants) 9.74e-01 FED (Asphyxiants) 9.73e-01 Relative 1.51e-03 1.00e-02 Yes
FED_FIC_SMIX 10.9.1 FED (Irritants) 8.26e-03 FED (Irritants) 8.25e-03 Relative 8.75e-04 1.00e-02 Yes

288
FED_FIC_SMIX 10.9.1 FED (All) 5.14e-01 FED (All) 5.13e-01 Relative 1.38e-03 1.00e-02 Yes
FED_FIC_SMIX 10.9.1 FIC (O2 CO2 CO) 1.00e-12 FIC (O2 CO2 CO) 1.00e-12 Absolute 0.00e+00 1.00e-02 Yes
FED_FIC_SMIX 10.9.1 FIC (Asphyxiants) 1.00e-12 FIC (Asphyxiants) 1.00e-12 Absolute 0.00e+00 1.00e-02 Yes
FED_FIC_SMIX 10.9.1 FIC (Irritants) 8.57e-01 FIC (Irritants) 8.56e-01 Relative 1.26e-03 1.00e-02 Yes
FED_FIC_SMIX 10.9.1 FIC (All) 4.29e-01 FIC (All) 4.28e-01 Relative 1.06e-03 1.00e-02 Yes
FED_CO_HCN 10.9.2 FED 2.90e-01 FED 2.90e-01 Relative 8.16e-04 1.00e-02 Yes
species_props 10.1 MU_0 1.00e-05 MU_0 1.00e-05 Absolute 0.00e+00 1.00e-02 Yes
species_props 10.1 MU_1 3.00e-05 MU_1 3.00e-05 Absolute 0.00e+00 1.00e-02 Yes
species_props 10.1 MU_2 1.59e-05 MU_2 1.59e-05 Absolute 4.21e-08 1.00e-02 Yes
species_props 10.1 MU_3 2.18e-05 MU_3 2.18e-05 Absolute 4.29e-08 1.00e-02 Yes
species_props 10.1 MU_4 1.64e-05 MU_4 1.64e-05 Absolute 3.59e-08 1.00e-02 Yes
species_props 10.1 K_0 1.00e-02 K_0 1.00e-02 Relative 0.00e+00 1.00e-02 Yes
species_props 10.1 K_1 3.00e-02 K_1 3.00e-02 Relative 0.00e+00 1.00e-02 Yes
species_props 10.1 K_2 1.59e-02 K_2 1.59e-02 Relative 2.65e-03 1.00e-02 Yes
species_props 10.1 K_3 2.18e-02 K_3 2.18e-02 Relative 1.97e-03 1.00e-02 Yes
species_props 10.1 K_4 1.64e-02 K_4 1.64e-02 Relative 2.19e-03 1.00e-02 Yes
species_props 10.1 CP_0 1.00e+00 CP_0 1.00e+00 Relative 0.00e+00 1.00e-02 Yes
species_props 10.1 CP_1 3.00e+00 CP_1 3.00e+00 Relative 0.00e+00 1.00e-02 Yes
species_props 10.1 CP_2 1.67e+00 CP_2 1.67e+00 Relative 2.00e-03 1.00e-02 Yes
species_props 10.1 CP_3 2.33e+00 CP_3 2.33e+00 Relative 1.14e-03 1.00e-02 Yes
species_props 10.1 CP_4 1.80e+00 CP_4 1.80e+00 Relative 0.00e+00 1.00e-02 Yes
fire_const_gamma 3.9.2 dH_exact 2.81e+04 dH_FDS 2.81e+04 Relative 5.14e-06 1.00e-05 Yes
fire_const_gamma 3.9.2 dP_exact 6.28e+04 dP_FDS 6.28e+04 Relative 6.20e-06 1.00e-05 Yes
flat_fire 13.1.9 x 8.91e+01 x1 8.77e+01 Relative 1.60e-02 2.00e-02 Yes
 Case Name Section Expected Expected Predicted Predicted Type of Error Error Within
Quantity Value Quantity Value Error Tolerance Tol.
flat_fire 13.1.9 z -7.16e+00 z1 -7.11e+00 Relative 6.51e-03 1.00e-02 Yes
flat_fire 13.1.9 u -3.84e+02 u1 -3.85e+02 Relative 1.77e-03 1.00e-02 Yes
flat_fire 13.1.9 w -8.41e+00 w1 -8.28e+00 Relative 1.59e-02 2.00e-02 Yes
flat_fire 13.1.9 z -7.16e+00 z1 -7.11e+00 Relative 6.51e-03 1.00e-02 Yes
flow_rate 18.3.2 Water 1.00e+01 Water mass 9.99e+00 Relative 1.35e-03 1.00e-02 Yes
flow_rate 18.3.2 Nozzles 4.00e+00 Nozzles 4.00e+00 Relative 0.00e+00 0.00e+00 Yes
flow_rate_2 13.5 Water 1.25e+01 Water mass 1.25e+01 Relative 8.07e-05 1.00e-02 Yes
flow_rate_2 13.5 Nozzles 1 3.00e+00 Nozzles 1 3.00e+00 Relative 0.00e+00 0.00e+00 Yes
flow_rate_2 13.5 Nozzles 2 1.00e+00 Nozzles 2 1.00e+00 Relative 0.00e+00 0.00e+00 Yes
hot_rods 15.4.4 PDF 5.00e-04 PDF_rods 5.00e-04 Relative 2.17e-16 1.00e-02 Yes
hot_rods 15.4.4 Mass 7.00e-01 Mass 2 7.02e-01 Relative 3.01e-03 1.00e-02 Yes
gas_filling 12.1.1 Mass 3.00e+00 HYDROGEN 3.00e+00 Relative 5.92e-05 1.00e-02 Yes
geom_hrrpuv_init 7.2.9 Q 8.75e+01 HRR 8.75e+01 Relative 0.00e+00 1.00e-03 Yes
geom_mass_file_test 7.1.8 Total 5.69e-01 Total 5.70e-01 Relative 3.81e-04 1.00e-03 Yes
geom_mass_file_test 7.1.8 NITROGEN 4.34e-01 NITROGEN 4.34e-01 Relative 3.78e-04 1.00e-03 Yes
geom_mass_file_test 7.1.8 OXYGEN 1.32e-01 OXYGEN 1.32e-01 Relative 3.81e-04 1.00e-03 Yes
geom_mass_file_test 7.1.8 CARBON DIOXIDE 3.37e-04 CARBON DIOXIDE 3.37e-04 Relative 3.79e-04 1.00e-03 Yes
geom_mass_file_test 7.1.8 WATER VAPOR 3.38e-03 WATER VAPOR 3.38e-03 Relative 4.18e-04 1.00e-03 Yes
geom_particle_cascade_2 13.4.2 Mass 1.00e+01 m1 1.00e+01 Relative 2.08e-03 1.00e-02 Yes
geom_particle_cascade_2 13.4.2 Mass 1.00e+01 m2 1.00e+01 Relative 2.08e-03 1.00e-02 Yes
geom_particle_cascade_2 13.4.2 Mass 1.00e+01 m3 1.00e+01 Relative 2.08e-03 1.00e-02 Yes
geom_particle_cascade_2 13.4.2 Mass 1.00e+01 m4 1.00e+01 Relative 2.08e-03 1.00e-02 Yes
geom_rad 9.1.2 E_OB 2.88e+01 E_GE 2.88e+01 Relative 7.05e-04 1.00e-03 Yes

289
geom_rad 9.1.2 HF_OB 2.84e+01 HF_GE 2.84e+01 Relative 1.15e-05 1.00e-03 Yes
geom_rad_2 9.1.3 Energy Flow 2.86e+01 E 2.84e+01 Relative 7.05e-03 1.00e-02 Yes
geom_rad_2 9.1.3 Heat Flux 2.84e+01 HF 2.85e+01 Relative 5.04e-03 1.00e-02 Yes
geom_restart 16.6 S_15 7.49e+01 S_15 7.49e+01 Relative 4.82e-05 1.00e-03 Yes
geom_sprk_mass 13.4.1 Mass 1.05e+01 m OBST 1.05e+01 Relative 4.72e-03 2.00e-02 Yes
geom_sprk_mass 13.4.1 Mass 1.05e+01 m GEOM 1.05e+01 Relative 4.70e-03 2.00e-02 Yes
geom_stretched_grid 6.2 psolver mesh 1 3.00e+00 psolver mesh 1 3.00e+00 Relative 0.00e+00 1.00e-03 Yes
geom_stretched_grid 6.2 psolver mesh 2 1.00e-12 psolver mesh 2 1.00e-12 Relative 0.00e+00 1.00e-03 Yes
ground_vegetation_conv 17.4.4 T_A 2.71e+02 temp 1A 2.62e+02 Relative 3.29e-02 5.00e-02 Yes
ground_vegetation_conv 17.4.4 T_A 2.71e+02 temp 2A 2.67e+02 Relative 1.79e-02 5.00e-02 Yes
ground_vegetation_conv 17.4.4 T_B 2.43e+02 temp 1B 2.32e+02 Relative 4.47e-02 5.00e-02 Yes
ground_vegetation_conv 17.4.4 T_B 2.43e+02 temp 2B 2.39e+02 Relative 1.87e-02 5.00e-02 Yes
ground_vegetation_drag 17.4.2 pres 1 6.07e-01 pres 2 6.10e-01 Relative 5.22e-03 5.00e-02 Yes
ground_vegetation_load 17.4.1 Mass 2.25e-01 Mass 1 2.18e-01 Relative 3.24e-02 1.50e-01 Yes
ground_vegetation_load 17.4.1 Mass 2.25e-01 Mass 2 2.22e-01 Relative 1.55e-02 1.50e-01 Yes
ground_vegetation_radi 17.4.3 HF 1.07e+01 hf 1 1.16e+01 Relative 8.80e-02 1.00e-01 Yes
hallways 22.10.1 Q_ENTH 4.34e-01 Q_TOTAL 6.07e-01 Absolute 1.73e-01 3.00e+00 Yes
heat_conduction_a 11.1 Front 1.19e+02 Front 1.19e+02 Relative 3.18e-05 1.00e-02 Yes
heat_conduction_a 11.1 4 cm 7.28e+01 4 cm 7.27e+01 Relative 1.57e-03 1.00e-02 Yes
heat_conduction_a 11.1 Back 4.21e+01 Back 4.20e+01 Relative 3.05e-04 1.00e-02 Yes
heat_conduction_b 11.1 Front 1.19e+02 Front 1.19e+02 Relative 4.61e-05 1.00e-02 Yes
heat_conduction_b 11.1 4 cm 1.15e+02 4 cm 1.15e+02 Relative 1.09e-04 1.00e-02 Yes
heat_conduction_b 11.1 Back 1.13e+02 Back 1.13e+02 Relative 8.65e-05 1.00e-02 Yes
heat_conduction_c 11.1 Front 1.18e+02 Front 1.18e+02 Relative 5.37e-05 1.00e-02 Yes
 Case Name Section Expected Expected Predicted Predicted Type of Error Error Within
Quantity Value Quantity Value Error Tolerance Tol.
heat_conduction_c 11.1 4 cm 1.18e+02 4 cm 1.18e+02 Relative 8.49e-05 1.00e-02 Yes
heat_conduction_c 11.1 Back 1.17e+02 Back 1.17e+02 Relative 8.26e-05 1.00e-02 Yes
heat_conduction_d 11.1 Front 6.04e+01 Front 6.04e+01 Relative 2.68e-04 1.00e-02 Yes
heat_conduction_d 11.1 4 cm 5.85e+01 4 cm 5.84e+01 Relative 3.16e-04 1.00e-02 Yes
heat_conduction_d 11.1 Back 5.74e+01 Back 5.73e+01 Relative 2.23e-04 1.00e-02 Yes
heat_conduction_kc 11.2 cart_surf 1.59e+02 cart_front 1.59e+02 Relative 5.92e-04 1.00e-02 Yes
heat_conduction_kc 11.2 cyl_surf 1.82e+02 cyl_front 1.82e+02 Relative 1.17e-03 1.00e-02 Yes
heat_conduction_kc 11.2 sph_surf 2.10e+02 sph_front 2.10e+02 Relative 1.80e-03 1.00e-02 Yes
heat_conduction_kc 11.2 cart_back 4.99e+00 cart_back 4.94e+00 Relative 1.08e-02 2.00e-02 Yes
heat_conduction_kc 11.2 cyl_back 2.00e+01 cyl_back 1.97e+01 Relative 1.40e-02 2.00e-02 Yes
heat_conduction_kc 11.2 sph_back 4.54e+01 sph_back 4.48e+01 Relative 1.23e-02 2.00e-02 Yes
helium_1d_isothermal 3.8.2 Temp 2.50e+01 maxT 2.50e+01 Absolute 0.00e+00 1.00e-06 Yes
helium_1d_isothermal 3.8.2 Temp 2.50e+01 minT 2.50e+01 Absolute 0.00e+00 1.00e-06 Yes
helium_1d_isothermal 3.8.2 H_S 1.00e-12 maxHS 6.83e-13 Absolute 3.17e-13 1.00e-06 Yes
helium_1d_isothermal 3.8.2 H_S 1.00e-12 minHS 1.72e-14 Absolute 9.83e-13 1.00e-06 Yes
helium_1d_isothermal 3.8.2 W 1.00e+00 maxW 1.00e+00 Absolute 0.00e+00 1.00e-06 Yes
helium_1d_isothermal 3.8.2 W 1.00e+00 minW 1.00e+00 Absolute 1.00e-08 1.00e-06 Yes
helium_1d_const_gamma 3.9.1 Temp 2.50e+01 maxT 2.50e+01 Absolute 0.00e+00 1.00e-06 Yes
helium_1d_const_gamma 3.9.1 Temp 2.50e+01 minT 2.50e+01 Absolute 0.00e+00 1.00e-06 Yes
helium_1d_const_gamma 3.9.1 H_S 3.55e+02 maxHS 3.55e+02 Absolute 0.00e+00 1.00e-06 Yes
helium_1d_const_gamma 3.9.1 H_S 3.55e+02 minHS 3.55e+02 Absolute 0.00e+00 1.00e-06 Yes
helium_1d_const_gamma 3.9.1 W 1.00e+00 maxW 1.00e+00 Absolute 0.00e+00 1.00e-06 Yes
helium_1d_const_gamma 3.9.1 W 1.00e+00 minW 1.00e+00 Absolute 1.00e-08 1.00e-06 Yes

290
helium_2d_isothermal 3.8.2 Isothermal 2.00e+01 maxT 2.00e+01 Absolute 1.20e-04 1.00e-02 Yes
helium_2d_isothermal 3.8.2 Isothermal -2.00e+01 minT -2.00e+01 Absolute 1.90e-03 1.00e-02 Yes
helium_air_jet_floor 3.8.3 Isothermal 2.00e+01 maxT 2.00e+01 Absolute 2.55e-02 5.00e-02 Yes
helium_air_jet_floor 3.8.3 Isothermal -2.00e+01 minT -2.00e+01 Absolute 1.00e-06 5.00e-02 Yes
HoC_Ideal 10.7 Ideal 8.75e-04 MLR_METHANE 8.75e-04 Relative 3.74e-06 1.00e-04 Yes
HoC_NonIdeal 10.7 Non-Ideal 8.00e-04 MLR_METHANE 8.00e-04 Relative 0.00e+00 1.00e-04 Yes
hole 8.7 HRR 1.02e+02 HRR 1.02e+02 Absolute 0.00e+00 1.00e-04 Yes
hot_spheres 9.6 Rad Loss 1 -6.23e-02 rad loss 1 -6.10e-02 Relative 2.11e-02 3.00e-02 Yes
hot_spheres 9.6 Rad Loss 2 -2.98e+01 rad loss 2 -2.82e+01 Relative 5.34e-02 6.00e-02 Yes
hot_spheres 9.6 Rad Flux 2.03e+01 rad flux 1.97e+01 Relative 2.56e-02 4.00e-02 Yes
hot_spheres 9.6 Intensity 8.10e+01 UII 8.07e+01 Relative 3.84e-03 1.00e-02 Yes
hrrpuv_reac_arrhenius 10.5 hrrpuv 2.98e+00 q1 2.98e+00 Relative 0.00e+00 1.00e-03 Yes
hrrpuv_reac_extinction 10.5 hrrpuv 3.50e+03 q1 3.50e+03 Relative 1.43e-07 1.00e-05 Yes
hrrpuv_reac_parallel 10.5 hrrpuv 6.63e+03 q1+q2 6.63e+03 Relative 1.36e-07 1.00e-05 Yes
hrrpuv_reac_parallel_2 10.5 hrrpuv 6.68e+03 q1+q2 6.68e+03 Relative 1.35e-07 1.00e-05 Yes
hrrpuv_reac_series 10.5 hrrpuv 4.93e+03 q1+q2+q3 4.93e+03 Relative 0.00e+00 1.00e-05 Yes
hrrpuv_reac_simple 10.5 hrrpuv 3.50e+03 q1 3.50e+03 Relative 0.00e+00 1.00e-05 Yes
hrrpuv_reac_single 10.5 hrrpuv 6.49e+03 q1 6.49e+03 Relative 1.39e-07 1.00e-05 Yes
hrrpuv_reac_soot 10.5 hrrpuv 2.24e+03 q1+q2 2.24e+03 Relative 2.68e-08 1.00e-05 Yes
ht3d_beam_heating 11.9.2 T 8.95e-01 wall temp 8.65e-01 Relative 3.28e-02 5.00e-02 Yes
ht3d_beam_heating 11.9.2 T 1.81e+00 wall temp 1.80e+00 Relative 2.94e-03 5.00e-02 Yes
ht3d_demo 8.3.10 H 8.55e+00 Q_net 8.56e+00 Relative 1.15e-03 1.00e-02 Yes
ht3d_energy_conservation 11.9.1 E3D 9.57e+01 Q_net 9.60e+01 Relative 2.53e-03 1.00e-02 Yes
ht3d_energy_conservation 11.9.1 E3D 9.57e+01 Q_net 9.60e+01 Relative 2.53e-03 1.00e-02 Yes
 Case Name Section Expected Expected Predicted Predicted Type of Error Error Within
Quantity Value Quantity Value Error Tolerance Tol.
ht3d_energy_conservation 11.9.1 E3D 9.61e+01 Q_net 9.61e+01 Relative 3.75e-04 1.00e-02 Yes
ht3d_energy_conservation 11.9.1 E3D 9.87e+01 Q_net 9.61e+01 Relative 2.64e-02 3.00e-02 Yes
ht3d_ibeam 11.9.5 Ts_x195_40 7.64e+02 TS_x195-40 7.64e+02 Relative 9.41e-05 8.00e-02 Yes
ht3d_ibeam 11.9.5 Ts_x145_30 5.93e+02 TS_x145-30 5.91e+02 Relative 4.19e-03 8.00e-02 Yes
ht3d_ibeam 11.9.5 Ts_x095_20 3.62e+02 TS_x095-20 3.55e+02 Relative 2.00e-02 8.00e-02 Yes
ht3d_ibeam 11.9.5 Ts_x025_40 2.89e+02 TS_x025-40 2.77e+02 Relative 4.29e-02 8.00e-02 Yes
ht3d_ibeam 11.9.5 Ts_x195_1 2.53e+02 TS_x195-01 2.43e+02 Relative 3.83e-02 8.00e-02 Yes
ht3d_ibeam 11.9.5 Ts_x025_1 1.52e+02 TS_x025-01 1.42e+02 Relative 6.12e-02 8.00e-02 Yes
ht3d_mass_conservation 11.9.3 Mass 2.12e-01 WOOD MOISTURE 2.12e-01 Relative 5.19e-04 1.00e-02 Yes
ht3d_network 11.9.6 T-1 2.93e+01 T-4 2.93e+01 Relative 0.00e+00 1.00e-03 Yes
ht3d_slab 11.9.4 Ts 3.50e+02 TS 3.50e+02 Relative 2.62e-04 1.00e-03 Yes
HVAC_aircoil 10.2.6 Ideal Q 4.52e+01 FDS Q 4.52e+01 Relative 0.00e+00 1.00e-02 Yes
HVAC_aircoil 10.2.6 Ideal T 5.89e+01 FDS T 5.89e+01 Relative 3.40e-08 1.00e-02 Yes
HVAC_damper 14.5 Ideal Duct 1 3.00e+00 FDS Duct 1 3.00e+00 Relative 9.13e-04 5.00e-03 Yes
HVAC_damper 14.5 Ideal Duct 2 1.00e-12 FDS Duct 2 1.00e-12 Relative 0.00e+00 5.00e-03 Yes
HVAC_damper 14.5 Ideal Duct 3 3.00e+00 FDS Duct 3 3.00e+00 Relative 0.00e+00 5.00e-03 Yes
HVAC_energy_pressure 14.4 Ideal Top 1.00e+01 FDS Top 1.00e+01 Relative 4.34e-04 1.00e-02 Yes
HVAC_energy_pressure 14.4 Ideal Bottom 1.00e+00 FDS Bottom 1.00e+00 Relative 0.00e+00 1.00e-02 Yes
HVAC_energy_pressure 14.4 Ideal Delta P 4.36e+01 FDS Delta P 4.36e+01 Relative 8.05e-04 1.00e-02 Yes
HVAC_energy_pressure 14.4 Ideal Enthalpy 3.07e+02 FDS Energy 3.07e+02 Relative 1.92e-06 1.00e-02 Yes
HVAC_filter 14.6 Ideal Filter 1.16e-03 FDS Filter 1.17e-03 Absolute 1.84e-06 1.00e-05 Yes
HVAC_filter 14.6 Ideal Compartment 1.00e-12 FDS Compartment 6.88e-08 Absolute 6.88e-08 1.00e-05 Yes
HVAC_filter 14.6 Ideal Velocity 1.85e+00 FDS Velocity 1.86e+00 Relative 8.90e-03 7.00e-02 Yes

291
HVAC_filter 14.6 Ideal Temperature 1.99e+01 FDS Temperature 1.99e+01 Relative 1.97e-05 1.00e-02 Yes
HVAC_filter 14.6 Ideal Pressure -1.44e+02 FDS Pressure -1.44e+02 Relative 1.57e-03 1.00e-02 Yes
HVAC_flow_loss 14.1 Ideal 1 m/s 1.00e+00 FDS 1 m/s duct 1.01e+00 Relative 9.14e-03 2.00e-02 Yes
HVAC_flow_loss 14.1 Ideal 2 m/s 2.00e+00 FDS 2 m/s duct 1.99e+00 Relative 4.69e-03 2.00e-02 Yes
HVAC_flow_loss 14.1 Ideal 3 m/s 3.00e+00 FDS 3 m/s duct 3.00e+00 Relative 0.00e+00 2.00e-02 Yes
HVAC_leak_exponent 14.10.7 Ideal 1 -1.60e+01 L1 -1.60e+01 Relative 1.33e-03 1.00e-02 Yes
HVAC_leak_exponent 14.10.7 Ideal 2 -1.18e+01 L2 -1.18e+01 Relative 7.31e-04 1.00e-02 Yes
HVAC_leak_exponent 14.10.7 Ideal 3 -1.27e+01 L3 -1.27e+01 Relative 1.09e-03 1.00e-02 Yes
HVAC_leak_exponent_2 14.10.7 Vdot (m3/s) 1.09e-01 Vdot 1.09e-01 Relative 4.35e-04 1.00e-02 Yes
leak_enthalpy 14.10.5 Mass 1.00e-12 TM 2.43e-07 Absolute 2.43e-07 1.00e-05 Yes
leak_enthalpy 14.10.5 Energy 3.60e+02 TE 3.62e+02 Relative 5.71e-03 1.00e-02 Yes
HVAC_mass_conservation 14.3 Ideal Species 1 5.82e-01 FDS Species 1 5.82e-01 Relative 1.74e-05 1.00e-02 Yes
HVAC_mass_conservation 14.3 Ideal Total 1.16e+00 FDS Total 1.16e+00 Relative 0.00e+00 1.00e-02 Yes
HVAC_mass_transport 14.12.2 Ideal Species 1 1.00e+00 FDS Species 1 1.00e+00 Relative 2.62e-05 1.00e-02 Yes
HVAC_mass_transport_1 14.12.3 Ideal Species 1 1.00e+00 FDS Species 1 1.00e+00 Relative 2.62e-05 1.00e-02 Yes
HVAC_mass_transport_2 14.12.3 Ideal Species 1 1.00e+00 FDS Species 1 1.00e+00 Relative 2.61e-05 1.00e-02 Yes
HVAC_mass_transport_3 14.12.3 Ideal Species 1 1.00e+00 FDS Species 1 1.00e+00 Relative 2.58e-05 1.00e-02 Yes
HVAC_mass_transport_4 14.12.3 Ideal Species 1 1.00e+00 FDS Species 1 1.00e+00 Relative 2.57e-05 1.00e-02 Yes
HVAC_mass_transport_branch 14.12.7 Ideal Outlet A 1.00e+00 FDS Outlet A 1.00e+00 Relative 4.31e-05 1.00e-02 Yes
HVAC_mass_transport_branch 14.12.7 Ideal Outlet B 1.00e+00 FDS Outlet B 1.00e+00 Relative 0.00e+00 1.00e-02 Yes
HVAC_mass_transport_conservation_1 14.12.4 Ideal Mass_total 1.00e+00 FDS Mass_total 1.00e+00 Relative 0.00e+00 1.00e-02 Yes
HVAC_mass_transport_conservation_1 14.12.4 Ideal Mass_left 5.00e-01 FDS Mass_left 5.00e-01 Relative 0.00e+00 1.00e-02 Yes
HVAC_mass_transport_conservation_1 14.12.4 Ideal Mass_right 5.00e-01 FDS Mass_right 5.00e-01 Relative 0.00e+00 1.00e-02 Yes
HVAC_mass_transport_conservation_1 14.12.4 Ideal Species 1_left 2.92e-01 FDS Species 1_left 2.92e-01 Relative 3.56e-04 1.00e-02 Yes
 Case Name Section Expected Expected Predicted Predicted Type of Error Error Within
Quantity Value Quantity Value Error Tolerance Tol.
HVAC_mass_transport_conservation_1 14.12.4 Ideal Species 1_right 2.92e-01 FDS Species 1_right 2.92e-01 Relative 4.17e-04 1.00e-02 Yes
HVAC_mass_transport_conservation_1 14.12.4 Ideal Species 2_left 2.08e-01 FDS Species 2_left 2.08e-01 Relative 4.99e-04 1.00e-02 Yes
HVAC_mass_transport_conservation_1 14.12.4 Ideal Species 2_right 2.08e-01 FDS Species 2_right 2.08e-01 Relative 5.84e-04 1.00e-02 Yes
HVAC_mass_transport_conservation_2 14.12.5 Ideal Species 1 6.00e-01 FDS Species 1 6.00e-01 Relative 4.64e-04 1.00e-02 Yes
HVAC_mass_transport_conservation_2 14.12.5 Ideal Species 2 4.00e-01 FDS Species 2 4.00e-01 Relative 6.96e-04 1.00e-02 Yes
HVAC_mass_transport_conservation_3 14.12.6 Ideal Species 1 6.00e-01 FDS Species 1 6.00e-01 Relative 4.01e-04 1.00e-02 Yes
HVAC_mass_transport_conservation_3 14.12.6 Ideal Species 2 4.00e-01 FDS Species 2 4.00e-01 Relative 6.01e-04 1.00e-02 Yes
HVAC_mass_transport_combine 14.12.8 Ideal Species 1 5.00e-01 FDS Species 1 5.00e-01 Absolute 0.00e+00 1.00e-02 Yes
HVAC_mass_transport_combine 14.12.8 Ideal Species 2 5.00e-01 FDS Species 2 5.00e-01 Absolute 0.00e+00 1.00e-02 Yes
HVAC_mass_transport_combine 14.12.8 Ideal Species 3 1.00e-12 FDS Species 3 2.34e-15 Absolute 9.98e-13 1.00e-02 Yes
HVAC_mass_transport_energy 14.12.9 Ideal Enthalpy 2.93e+02 FDS Enthalpy 2.93e+02 Relative 4.16e-07 1.00e-02 Yes
HVAC_mass_transport_energy 14.12.9 Ideal Delta P 8.93e-01 FDS Delta P 8.93e-01 Relative 4.25e-05 1.00e-02 Yes
HVAC_tee_loss_1 14.2 Ideal 1 m/s 1.00e+00 FDS 1 m/s duct 1.00e+00 Relative 7.73e-05 1.00e-02 Yes
HVAC_tee_loss_1 14.2 Ideal 2 m/s 2.00e+00 FDS 2 m/s duct 2.00e+00 Relative 2.02e-04 1.00e-02 Yes
HVAC_tee_loss_1 14.2 Ideal 3 m/s 3.00e+00 FDS 3 m/s duct 3.00e+00 Relative 0.00e+00 1.00e-02 Yes
HVAC_tee_loss_2 14.2 Ideal 1.5 m/s A 1.50e+00 E1 1.49e+00 Relative 7.16e-03 1.00e-02 Yes
HVAC_tee_loss_2 14.2 Ideal 1.5 m/s B 1.50e+00 E2 1.51e+00 Relative 7.23e-03 1.00e-02 Yes
HVAC_tee_loss_2 14.2 Ideal 3 m/s 3.00e+00 S1 3.00e+00 Relative 0.00e+00 1.00e-02 Yes
insulated_steel_pipe 11.5 Temp 2.60e+02 Temperature 2.59e+02 Relative 3.37e-03 1.00e-02 Yes
insulated_steel_plate 11.5 Temp 1.35e+02 Temperature 1.35e+02 Relative 1.80e-04 1.00e-02 Yes
isentropic 7.2.2 Density_1 1.32e+00 density_1 1.32e+00 Relative 8.32e-07 1.00e-02 Yes
isentropic 7.2.2 Density_2 1.32e+00 density_2 1.32e+00 Relative 8.32e-07 1.00e-02 Yes
isentropic 7.2.2 Pressure_1 1.96e+04 pressure_1 1.96e+04 Relative 1.17e-04 1.00e-02 Yes

292
isentropic 7.2.2 Pressure_2 1.96e+04 pressure_2 1.96e+04 Relative 2.18e-05 1.00e-02 Yes
isentropic 7.2.2 Temperature_1 3.52e+01 temperature_1 3.52e+01 Relative 1.63e-04 1.00e-02 Yes
isentropic 7.2.2 Temperature_2 3.52e+01 temperature_2 3.52e+01 Relative 2.77e-05 1.00e-02 Yes
isentropic 7.2.2 Enthalpy_1 4.23e+02 enthalpy_1 4.23e+02 Relative 1.88e-05 1.00e-02 Yes
isentropic 7.2.2 Enthalpy_2 4.23e+02 enthalpy_2 4.23e+02 Relative 3.47e-06 1.00e-02 Yes
isentropic2 7.2.3 Density_1 1.37e+00 density_1 1.36e+00 Relative 6.90e-04 1.00e-02 Yes
isentropic2 7.2.3 Density_2 1.37e+00 density_2 1.36e+00 Relative 6.89e-04 1.00e-02 Yes
isentropic2 7.2.3 Pressure_1 6.43e+04 pressure_1 6.29e+04 Relative 2.16e-02 3.00e-02 Yes
isentropic2 7.2.3 Pressure_2 6.43e+04 pressure_2 6.40e+04 Relative 4.33e-03 3.00e-02 Yes
isentropic2 7.2.3 Temperature_1 1.36e+02 temperature_1 1.33e+02 Relative 1.79e-02 3.00e-02 Yes
isentropic2 7.2.3 Temperature_2 1.36e+02 temperature_2 1.35e+02 Relative 3.56e-03 3.00e-02 Yes
isentropic2 7.2.3 Enthalpy_1 5.80e+02 enthalpy_1 5.75e+02 Relative 8.40e-03 1.00e-02 Yes
isentropic2 7.2.3 Enthalpy_2 5.80e+02 enthalpy_2 5.79e+02 Relative 1.68e-03 1.00e-02 Yes
lapse_rate 8.4 T 1.96e+01 T 1.96e+01 Relative 1.49e-05 1.00e-03 Yes
lapse_rate 8.4 P 1.01e+05 P 1.01e+05 Relative 5.82e-07 1.00e-03 Yes
layer 8.9 z_int 6.42e-01 z_int 6.16e-01 Relative 4.08e-02 1.00e-01 Yes
layer 8.9 T_low 3.04e+01 T_low 3.01e+01 Relative 7.32e-03 1.00e-01 Yes
layer 8.9 T_up 7.20e+01 T_up 7.36e+01 Relative 2.20e-02 1.00e-01 Yes
leak_test 14.10.1 pres_diff 1.91e+02 pres_diff 1.90e+02 Relative 1.71e-03 1.00e-02 Yes
leak_test_2 14.10.2 pres_1 -1.91e+02 pres_1 -1.90e+02 Absolute 2.47e-01 1.90e+00 Yes
leak_test_2 14.10.2 pres_2 1.00e-12 pres_2 -1.01e-02 Absolute 1.01e-02 1.90e+00 Yes
leak_test_3 14.10.3 p1 4.80e+02 p1 4.80e+02 Relative 3.32e-04 1.00e-02 Yes
leak_test_3 14.10.3 p2 2.40e+02 p2 2.40e+02 Relative 8.90e-04 1.00e-02 Yes
leak_test_3 14.10.3 p1 4.80e+02 p3 4.80e+02 Relative 3.38e-04 1.00e-02 Yes
 Case Name Section Expected Expected Predicted Predicted Type of Error Error Within
Quantity Value Quantity Value Error Tolerance Tol.
leak_test_3 14.10.3 p2 2.40e+02 p4 2.40e+02 Relative 9.12e-04 1.00e-02 Yes
leak_test_4 14.10.4 p 3.44e+02 p1 3.59e+02 Relative 4.29e-02 5.00e-02 Yes
level_set_fuel_model_1 17.5.1 mass 6.29e+04 mass 6.26e+04 Relative 3.96e-03 1.00e-02 Yes
liquid_mixture 9.2.6 Fuel Mass 1.27e+00 N-HEXANE 1.27e+00 Relative 2.74e-04 1.00e-02 Yes
liquid_mixture 9.2.6 Water Mass 2.05e-01 WATER VAPOR 2.04e-01 Relative 3.11e-03 1.00e-02 Yes
mass_flux_comparison 22.10.10 mdot 1.60e-03 MF-1 1.60e-03 Relative 6.04e-04 1.00e-02 Yes
mass_flux_comparison 22.10.10 mdot 1.60e-03 MF-1 1.60e-03 Relative 7.61e-04 1.00e-02 Yes
mass_heat_wall_device_test 7.2.8 MF vent -4.00e+00 MF vent -4.00e+00 Relative 0.00e+00 1.00e-02 Yes
mass_heat_wall_device_test 7.2.8 MF open 4.00e+00 MF open 4.00e+00 Relative 5.28e-04 1.00e-02 Yes
mass_heat_wall_device_test 7.2.8 HF vent -3.00e+02 HF vent -3.00e+02 Relative 1.71e-04 1.00e-02 Yes
mass_heat_wall_device_test 7.2.8 HF open 3.00e+02 HF open 3.00e+02 Relative 1.98e-04 1.00e-02 Yes
mass_heat_wall_device_test_2 7.2.8 MF vent -4.00e+00 MF vent -4.00e+00 Relative 0.00e+00 1.00e-02 Yes
mass_heat_wall_device_test_2 7.2.8 MF open 4.00e+00 MF open 3.99e+00 Relative 2.10e-03 1.00e-02 Yes
mass_heat_wall_device_test_2 7.2.8 HF vent -3.00e+02 HF vent -3.00e+02 Relative 1.00e-05 1.00e-02 Yes
mass_heat_wall_device_test_2 7.2.8 HF open 3.00e+02 HF open 2.99e+02 Relative 2.22e-03 1.00e-02 Yes
methane_flame 10.3 CO2 2.78e-02 pCO2 2.78e-02 Relative 3.85e-06 1.00e-02 Yes
methane_flame 10.3 pCO2 1.24e-01 lfCO2 1.24e-01 Relative 7.13e-04 1.00e-02 Yes
methane_flame 10.3 CO2 2.78e-02 CO2 2.78e-02 Relative 3.88e-04 1.00e-02 Yes
methane_flame 10.3 H2O 3.70e-02 H2O 3.70e-02 Relative 2.52e-04 1.00e-02 Yes
methane_flame 10.3 CO 3.06e-02 CO 3.06e-02 Relative 4.08e-04 1.00e-02 Yes
multiple_reac_hrrpua 9.1 CH4 2.00e-04 MLR_METHANE 2.00e-04 Relative 6.60e-04 1.00e-02 Yes
multiple_reac_hrrpua 9.1 C2H6 1.94e-04 MLR_ETHANE 1.94e-04 Relative 8.36e-04 1.00e-02 Yes
multiple_reac_hrrpua 9.1 C3H8 4.85e-05 MLR_PROPANE 4.85e-05 Relative 8.36e-04 1.00e-02 Yes

293
multiple_reac_hrrpua 9.1 CHO 9.23e-04 MLR_MYFUEL 9.23e-04 Relative 1.13e-04 1.00e-02 Yes
multiple_reac_hrrpua 9.1 H2+C7H8 1.24e-04 MLR_MYFUEL2 1.24e-04 Relative 2.32e-03 1.00e-02 Yes
multiple_reac_n_simple 13.1.3 CH4_CO 4.46e-03 CH4_CO 4.46e-03 Relative 4.29e-05 1.00e-02 Yes
multiple_reac_n_simple 13.1.3 CH4_H2 6.41e-04 CH4_H2 6.41e-04 Relative 6.58e-04 1.00e-02 Yes
multiple_reac_n_simple 13.1.3 C3H8_CO 1.98e-03 C3H8_CO 1.98e-03 Relative 1.50e-04 1.00e-02 Yes
multiple_reac_n_simple 13.1.3 C3H8_H2O 1.76e-02 C3H8_H2O 1.76e-02 Relative 2.41e-05 1.00e-02 Yes
multiple_reac_n_simple 13.1.3 C2H6_CO 2.92e-02 C2H6_CO 2.92e-02 Relative 1.54e-05 1.00e-02 Yes
multiple_reac_n_simple 13.1.3 C2H6_H2 3.15e-03 C2H6_H2 3.15e-03 Relative 9.04e-06 1.00e-02 Yes
methanol_evaporation 9.2.6 mdot 1.60e-02 mdot2 1.62e-02 Relative 9.61e-03 1.50e-02 Yes
methanol_evaporation 9.2.6 Tb 6.47e+01 Tsurf 6.44e+01 Relative 3.49e-03 1.50e-02 Yes
MO_velocity_profile_stable 6.1 u (m/s) 2.24e+02 u 2.32e+02 Relative 3.29e-02 5.00e-02 Yes
MO_velocity_profile_unstable 6.1 u (m/s) 2.88e+02 u 2.96e+02 Relative 2.78e-02 5.00e-02 Yes
mwtest_cfl 3.8.4 Isothermal 2.00e+01 maxT 2.00e+01 Absolute 0.00e+00 1.00e-06 Yes
mwtest_cfl 3.8.4 Isothermal -2.00e+01 minT -2.00e+01 Absolute 0.00e+00 1.00e-06 Yes
ns2d_64_nupt1 3.1 u-vel 1.09e+00 UVEL 1.10e+00 Relative 1.70e-04 1.00e-02 Yes
obst_coarse_fine_interface 10.3.1 DP 1.00e-12 DP -1.47e+00 Absolute 1.47e+00 2.00e+00 Yes
opening_ulmat 10.3.1 Pressure Tolerance 1.00e-12 perr-max 9.45e-11 Absolute 9.35e-11 1.00e-10 Yes
parabolic_profile 10.5 Pressure 2.15e+03 pres 2.15e+03 Relative 1.12e-04 1.00e-02 Yes
particle_drag_U10_N16 13.1.3 U 1.37e+00 U-VEL 1.37e+00 Absolute 5.39e-05 1.00e-02 Yes
particle_drag_U50_N16 13.1.3 U 2.99e+00 U-VEL 2.99e+00 Absolute 5.58e-04 1.00e-02 Yes
particle_drag_U100_N16 13.1.3 U 5.99e+00 U-VEL 5.99e+00 Absolute 1.12e-03 1.00e-02 Yes
particle_drag_U50_N1600 13.1.3 U 3.95e-01 U-VEL 3.93e-01 Absolute 2.23e-03 1.00e-02 Yes
particle_drag_U100_N1600 13.1.3 U 3.17e-01 U-VEL 3.17e-01 Absolute 7.92e-04 1.00e-02 Yes
particle_drag_U150_N1600 13.1.3 U 2.65e-01 U-VEL 2.64e-01 Absolute 3.56e-04 1.00e-02 Yes
 Case Name Section Expected Expected Predicted Predicted Type of Error Error Within
Quantity Value Quantity Value Error Tolerance Tol.
particle_drag_U10_N16 13.1.3 F -1.42e-02 drag force -1.42e-02 Relative 3.82e-03 1.00e-02 Yes
particle_flux 15.5.1 mass 6.40e-01 mass 6.41e-01 Relative 1.00e-03 1.00e-02 Yes
particle_heating_convection 13.2.2 Temp 7.83e+02 T_gas 7.83e+02 Relative 7.86e-05 1.00e-02 Yes
particle_heating_convection 13.2.2 Temp 7.83e+02 T_ball 7.82e+02 Relative 1.19e-03 1.00e-02 Yes
particle_heating_radiation 13.2.2 Temp 7.83e+02 T_gas 7.85e+02 Relative 2.95e-03 1.00e-02 Yes
particle_heating_radiation 13.2.2 Temp 7.83e+02 T_ball 7.85e+02 Relative 2.95e-03 1.00e-02 Yes
particle_isotropic_radi 9.8 Delta e_ball 1.18e+00 Q_rad ball 1.20e+00 Relative 1.18e-02 1.50e-02 Yes
particle_isotropic_radi 9.8 Delta e_cyl 4.44e+00 Q_rad cyl 4.48e+00 Relative 7.92e-03 1.50e-02 Yes
particle_isotropic_radi 9.8 Delta e_plate 7.07e+00 Q_rad plate 7.12e+00 Relative 7.93e-03 1.50e-02 Yes
particle_anisotropic_radi 9.8 Delta e_ball 9.20e-01 Q_rad ball 9.27e-01 Relative 7.65e-03 1.00e-02 Yes
particle_anisotropic_radi 9.8 Delta e_cyl 3.45e+00 Q_rad cyl 3.47e+00 Relative 5.11e-03 1.00e-02 Yes
particle_anisotropic_radi 9.8 Delta e_plate 5.49e+00 Q_rad plate 5.52e+00 Relative 5.11e-03 1.00e-02 Yes
particle_offgas_1 7.1.6 T_min 1.00e-12 T_min 1.00e-12 Absolute 0.00e+00 1.00e-03 Yes
particle_offgas_1 7.1.6 T_max 1.00e-12 T_max 1.00e-12 Absolute 0.00e+00 1.00e-03 Yes
particle_offgas_2 7.1.6 T_min 1.00e-12 T_min 1.00e-12 Absolute 0.00e+00 1.00e-03 Yes
particle_offgas_2 7.1.6 T_max 1.00e-12 T_max 1.00e-12 Absolute 0.00e+00 1.00e-03 Yes
particle_offgas_3 7.1.6 T_min 1.00e-12 T_min 1.00e-12 Absolute 0.00e+00 1.00e-03 Yes
particle_offgas_4 7.1.6 T_min 1.00e-12 T_min 1.00e-12 Absolute 0.00e+00 1.00e-03 Yes
part_path_ramp_jog 15.5.3 Z 8.00e-01 ZH 8.00e-01 Relative 0.00e+00 1.00e-03 Yes
part_baking_soda 12.1.3 Dia (mu-m) First-order 420 K 4.59e+01 diam first-order 4.58e+01 Relative 8.11e-04 1.00e-02 Yes
part_baking_soda 12.1.3 Dia (mu-m) Spherical 420 K 3.68e+01 SPHERE DIA 3.66e+01 Relative 5.43e-03 1.00e-02 Yes
part_baking_soda 12.1.3 Dia (mu-m) First-order 450 K 2.85e+01 diam first-order 2.84e+01 Relative 3.57e-03 5.00e-02 Yes
part_baking_soda 12.1.3 Dia (mu-m) Spherical 450 K 6.64e+00 SPHERE DIA 6.58e+00 Relative 9.50e-03 5.00e-02 Yes

294
part_baking_soda 12.1.3 Dia (mu-m) First-order 500 K 2.55e+00 diam first-order 2.59e+00 Relative 1.82e-02 1.00e-01 Yes
part_baking_soda 12.1.3 Dia (mu-m) Spherical 500 K 4.55e-01 SPHERE DIA 4.57e-01 Relative 4.59e-03 1.00e-01 Yes
pine_needles 17.2 Fuel Gas 6.40e-01 fuel gas mass 6.40e-01 Relative 0.00e+00 1.00e-02 Yes
pine_needles 17.2 Water Vapor 2.00e-01 water vapor mass 2.00e-01 Relative 0.00e+00 1.00e-02 Yes
pine_needles 17.2 Char 1.60e-01 solid mass 1.60e-01 Relative 0.00e+00 1.00e-02 Yes
porous_media 13.1.7 Pres 4.76e+01 pres_1 4.74e+01 Relative 4.42e-03 1.00e-02 Yes
porous_media 13.1.7 Pres 4.76e+01 pres_2 4.74e+01 Relative 4.42e-03 1.00e-02 Yes
porous_media 13.1.7 Pres 4.76e+01 pres_3 4.74e+01 Relative 4.42e-03 1.00e-02 Yes
pressure_rise 10.3.1 Pres 5.00e+04 Pres 3 4.99e+04 Relative 2.87e-03 1.00e-02 Yes
pressure_rise 10.3.1 Temp 5.50e+01 Temp 3 5.46e+01 Relative 7.69e-03 2.00e-02 Yes
pressure_rise 10.3.1 Dens 1.60e+00 Dens 3 1.60e+00 Relative 2.99e-03 1.00e-02 Yes
propane_flame_2reac 13.1.3 HRR 4.81e+01 HRR 4.81e+01 Relative 5.41e-07 1.00e-02 Yes
propane_flame_2reac 13.1.3 Q_RADI -1.06e+01 Q_RADI -1.06e+01 Relative 1.88e-07 1.00e-02 Yes
propane_flame_deposition 10.11.4 Mass 4.48e-04 depo_none 4.48e-04 Relative 2.23e-04 1.00e-02 Yes
propane_flame_deposition 10.11.4 Mass 4.48e-04 depo_all 4.48e-04 Relative 2.23e-04 1.00e-02 Yes
pvc_combustion 13.2.2 CO2 1.55e-01 CO2 1.55e-01 Absolute 3.96e-04 1.00e-03 Yes
pvc_combustion 13.2.2 CO 1.44e-02 CO 1.44e-02 Absolute 3.25e-05 1.00e-03 Yes
pvc_combustion 13.2.2 H2O 6.59e-02 H2O 6.60e-02 Absolute 1.28e-04 1.00e-03 Yes
pvc_combustion 13.2.2 S 3.93e-02 S 3.96e-02 Absolute 2.94e-04 1.00e-03 Yes
pvc_combustion 13.2.2 HCL 1.33e-01 HCL 1.34e-01 Absolute 2.60e-04 1.00e-03 Yes
race_test 8.8 TMP 8.72e+02 TMP 8.72e+02 Absolute 0.00e+00 1.00e-02 Yes
race_test 8.8 VEL 9.24e-01 VEL 9.24e-01 Absolute 0.00e+00 1.00e-02 Yes
radiation_box 9.2 Phi_HdA 2.01e-01 Flux_100_50 2.05e-01 Relative 1.66e-02 2.00e-02 Yes
radiation_gas_panel 9.15 HF_cen 3.08e+01 HF_cen 3.05e+01 Relative 9.66e-03 2.00e-02 Yes
 Case Name Section Expected Expected Predicted Predicted Type of Error Error Within
Quantity Value Quantity Value Error Tolerance Tol.
radiation_gas_panel 9.15 HF_off 1.74e+01 HF_off 1.75e+01 Relative 3.04e-03 2.00e-02 Yes
radiation_gas-veg_consistency_gas 9.10 radx0 8.04e+01 radx0 8.04e+01 Relative 5.85e-07 1.00e-02 Yes
radiation_gas-veg_consistency_gas 9.10 radx0.5 5.76e+00 radx0.5 5.76e+00 Relative 1.61e-05 1.00e-02 Yes
radiation_gas-veg_consistency_gas 9.10 radx1 8.35e-01 radx1 8.35e-01 Relative 3.49e-05 1.00e-02 Yes
radiation_gas-veg_consistency_veg 9.10 radx0 8.04e+01 radx0 8.04e+01 Relative 1.02e-04 1.00e-02 Yes
radiation_gas-veg_consistency_veg 9.10 radx0.5 5.76e+00 radx0.5 5.76e+00 Relative 3.98e-04 1.00e-02 Yes
radiation_gas-veg_consistency_veg 9.10 radx1 8.35e-01 radx1 8.35e-01 Relative 5.40e-04 1.00e-02 Yes
radiation_shield 9.14 Flux 4.76e+00 Flux 4.76e+00 Relative 4.19e-04 1.00e-02 Yes
radiation_shield 9.14 Temp 3.80e+02 Temp 3.80e+02 Relative 2.21e-04 1.00e-02 Yes
ramp_chi_r 14.5 CHI_R 1.40e-01 chi_r_max 1.40e-01 Absolute 0.00e+00 1.00e-03 Yes
random_meshes 8.5 Vdot 1.20e-02 Vdot 1.20e-02 Relative 3.75e-04 2.00e-02 Yes
random_walk 13.6 f_1 2.52e-01 npv 2.61e-01 Relative 3.26e-02 5.00e-02 Yes
random_walk 13.6 f_2 7.91e-01 npv 7.39e-01 Relative 6.62e-02 1.00e-01 Yes
reactionrate_arrhenius_0order_1step 10.4.2 O2 6.57e-05 O2 6.57e-05 Absolute 4.47e-08 1.00e-03 Yes
reactionrate_arrhenius_0order_1step 10.4.2 CO 5.40e-11 CO 7.64e-11 Absolute 2.24e-11 1.00e-03 Yes
reactionrate_arrhenius_0order_1step 10.4.2 CO2 4.55e-01 CO2 4.55e-01 Absolute 3.74e-06 1.00e-03 Yes
reactionrate_arrhenius_2order_1step 10.4.2 O2 1.30e-02 O2 1.32e-02 Absolute 1.34e-04 1.00e-03 Yes
reactionrate_arrhenius_2order_1step 10.4.2 C3H8 3.36e-03 C3H8 3.27e-03 Absolute 8.84e-05 1.00e-03 Yes
reactionrate_arrhenius_2order_1step 10.4.2 CO2 1.70e-01 CO2 1.70e-01 Absolute 7.04e-05 1.00e-03 Yes
reactionrate_arrhenius_2order_1step 10.4.2 H2O 9.27e-02 H2O 9.27e-02 Absolute 2.23e-05 1.00e-03 Yes
reactionrate_arrhenius_1p75order_2step 10.4.2 O2 3.84e-02 O2 3.80e-02 Absolute 3.67e-04 1.00e-03 Yes
reactionrate_arrhenius_1p75order_2step 10.4.2 C3H8 9.72e-03 C3H8 9.52e-03 Absolute 2.01e-04 1.00e-03 Yes
reactionrate_arrhenius_1p75order_2step 10.4.2 CO 3.90e-03 CO 3.84e-03 Absolute 6.78e-05 1.00e-03 Yes

295
reactionrate_arrhenius_1p75order_2step 10.4.2 CO2 1.45e-01 CO2 1.45e-01 Absolute 4.18e-04 1.00e-03 Yes
reactionrate_arrhenius_1p75order_2step 10.4.2 H2O 8.23e-02 H2O 8.25e-02 Absolute 2.20e-04 1.00e-03 Yes
reactionrate_arrhenius_1p75order_2stepr 10.4.2 O2 3.68e-02 O2 3.64e-02 Absolute 3.76e-04 1.00e-03 Yes
reactionrate_arrhenius_1p75order_2stepr 10.4.2 C3H8 8.01e-03 C3H8 7.83e-03 Absolute 1.81e-04 1.00e-03 Yes
reactionrate_arrhenius_1p75order_2stepr 10.4.2 CO 1.19e-02 CO 1.17e-02 Absolute 2.32e-04 1.00e-03 Yes
reactionrate_arrhenius_1p75order_2stepr 10.4.2 CO2 1.37e-01 CO2 1.38e-01 Absolute 6.07e-04 1.00e-03 Yes
reactionrate_arrhenius_1p75order_2stepr 10.4.2 H2O 8.51e-02 H2O 8.53e-02 Absolute 1.84e-04 1.00e-03 Yes
reactionrate_arrhenius_equilibrium 10.4.2 O2 2.27e-02 O2 2.27e-02 Absolute 4.09e-05 1.00e-03 Yes
reactionrate_arrhenius_equilibrium 10.4.2 C3H8 1.00e-12 C3H8 7.25e-14 Absolute 9.28e-13 1.00e-03 Yes
reactionrate_arrhenius_equilibrium 10.4.2 CO 3.98e-02 CO 3.98e-02 Absolute 8.74e-05 1.00e-03 Yes
reactionrate_arrhenius_equilibrium 10.4.2 CO2 1.18e-01 CO2 1.18e-01 Absolute 1.43e-04 1.00e-03 Yes
reactionrate_arrhenius_equilibrium 10.4.2 H2O 9.86e-02 H2O 9.86e-02 Absolute 3.34e-06 1.00e-03 Yes
reactionrate_arrhenius_equilibrium 10.4.2 Temperature 2.55e+03 TEMP 2.55e+03 Relative 1.04e-03 1.00e-02 Yes
reactionrate_arrhenius_equilibrium 10.4.2 Pressure 3.85e+05 PRES 3.86e+05 Relative 1.37e-03 1.00e-02 Yes
reactionrate_EDC_O2lim_1step 10.4.1 O2 3.40e-05 O2 7.47e-06 Absolute 2.65e-05 1.00e-03 Yes
reactionrate_EDC_O2lim_1step 10.4.1 CH4 3.75e-01 CH4 3.75e-01 Absolute 6.41e-06 1.00e-03 Yes
reactionrate_EDC_O2lim_1step 10.4.1 CO2 6.87e-02 CO2 6.88e-02 Absolute 1.83e-05 1.00e-03 Yes
reactionrate_EDC_O2lim_1step 10.4.1 H2O 5.63e-02 H2O 5.63e-02 Absolute 1.47e-05 1.00e-03 Yes
reactionrate_EDC_O2lim_2fuel 10.4.1 CH4 2.66e-01 CH4 2.66e-01 Absolute 0.00e+00 1.00e-03 Yes
reactionrate_EDC_O2lim_2fuel 10.4.1 C3H8 3.55e-01 C3H8 3.55e-01 Absolute 0.00e+00 1.00e-03 Yes
reactionrate_EDC_O2lim_2fuel 10.4.1 O2 1.51e-05 O2 1.51e-05 Absolute 3.30e-11 1.00e-03 Yes
reactionrate_EDC_O2lim_2fuel 10.4.1 CO2 2.29e-01 CO2 2.29e-01 Absolute 0.00e+00 1.00e-03 Yes
reactionrate_EDC_O2lim_2fuel 10.4.1 H2O 1.50e-01 H2O 1.50e-01 Absolute 0.00e+00 1.00e-03 Yes
reactionrate_EDC_flim_1step_CH4 10.4.1 CH4 1.00e-06 O2 1.50e-06 Absolute 4.98e-07 1.00e-03 Yes
 Case Name Section Expected Expected Predicted Predicted Type of Error Error Within
Quantity Value Quantity Value Error Tolerance Tol.
reactionrate_EDC_flim_1step_CH4 10.4.1 O2 1.00e-12 CH4 6.72e-09 Absolute 6.72e-09 1.00e-03 Yes
reactionrate_EDC_flim_1step_CH4 10.4.1 CO2 1.51e-01 CO2 1.51e-01 Absolute 3.10e-07 1.00e-03 Yes
reactionrate_EDC_flim_1step_CH4 10.4.1 H2O 1.24e-01 H2O 1.24e-01 Absolute 1.00e-08 1.00e-03 Yes
reactionrate_EDC_flim_1step_CH4 10.4.1 TEMP 2.54e+03 TEMP 2.54e+03 Relative 8.58e-05 1.00e-02 Yes
reactionrate_EDC_flim_1step_CH4 10.4.1 PRES 8.56e+05 PRES 8.56e+05 Relative 8.32e-05 1.00e-02 Yes
reactionrate_EDC_flim_1step_C3H8 10.4.1 C3H8 1.00e-12 O2 2.67e-08 Absolute 2.67e-08 1.00e-03 Yes
reactionrate_EDC_flim_1step_C3H8 10.4.1 O2 4.00e-06 C3H8 3.93e-06 Absolute 7.04e-08 1.00e-03 Yes
reactionrate_EDC_flim_1step_C3H8 10.4.1 CO2 1.81e-01 CO2 1.81e-01 Absolute 1.90e-07 1.00e-03 Yes
reactionrate_EDC_flim_1step_C3H8 10.4.1 H2O 9.86e-02 H2O 9.86e-02 Absolute 2.35e-07 1.00e-03 Yes
reactionrate_EDC_flim_1step_C3H8 10.4.1 TEMP 2.63e+03 TEMP 2.63e+03 Relative 8.85e-04 1.00e-02 Yes
reactionrate_EDC_flim_1step_C3H8 10.4.1 PRES 9.24e+05 PRES 9.25e+05 Relative 8.86e-04 1.00e-02 Yes
reactionrate_EDC_flim_2step 10.4.1 O2 4.01e-01 O2 4.01e-01 Absolute 3.50e-07 1.00e-03 Yes
reactionrate_EDC_flim_2step 10.4.1 CH4 1.00e-12 CH4 2.52e-10 Absolute 2.51e-10 1.00e-03 Yes
reactionrate_EDC_flim_2step 10.4.1 CO 1.00e-12 CO 3.23e-26 Absolute 1.00e-12 1.00e-03 Yes
reactionrate_EDC_flim_2step 10.4.1 CO2 2.74e-01 CO2 2.74e-01 Absolute 3.70e-07 1.00e-03 Yes
reactionrate_EDC_flim_2step 10.4.1 H2O 2.25e-01 H2O 2.25e-01 Absolute 2.00e-08 1.00e-03 Yes
reactionrate_EDC_1step_CH4_nonmix 10.4.1 O2 1.00e-12 O2 9.67e-11 Absolute 9.57e-11 1.00e-03 Yes
reactionrate_EDC_1step_CH4_nonmix 10.4.1 CH4 3.90e-03 CH4 3.91e-03 Absolute 6.89e-06 1.00e-03 Yes
reactionrate_EDC_1step_CH4_nonmix 10.4.1 CO2 1.49e-01 CO2 1.49e-01 Absolute 8.35e-06 1.00e-03 Yes
reactionrate_EDC_1step_CH4_nonmix 10.4.1 H2O 1.22e-01 H2O 1.22e-01 Absolute 7.14e-06 1.00e-03 Yes
reactionrate_EDC_1step_CH4_nonmix 10.4.1 TEMP 2.47e+03 TEMP 2.49e+03 Relative 5.02e-03 1.00e-02 Yes
reactionrate_EDC_1step_CH4_nonmix 10.4.1 PRES 8.34e+05 PRES 8.36e+05 Relative 2.40e-03 1.00e-02 Yes
reactionrate_fast_slow 10.4.3 CH3OH 1.00e-12 CH3OH 5.88e-18 Absolute 1.00e-12 1.00e-03 Yes

296
reactionrate_fast_slow 10.4.3 O2 1.00e-12 O2 1.13e-05 Absolute 1.13e-05 1.00e-03 Yes
reactionrate_fast_slow 10.4.3 CO2 1.92e-02 CO2 1.93e-02 Absolute 6.00e-05 1.00e-03 Yes
reactionrate_fast_slow 10.4.3 CO 2.88e-02 CO 2.88e-02 Absolute 1.37e-05 1.00e-03 Yes
reactionrate_lumped_two_air 10.4.1 C3H8 1.00e-12 C3H8 3.83e-05 Absolute 3.83e-05 1.00e-03 Yes
reactionrate_lumped_two_air 10.4.1 AIR1 1.00e-12 AIR1 1.94e-11 Absolute 1.84e-11 1.00e-03 Yes
reactionrate_lumped_two_air 10.4.1 AIR2 1.00e-12 AIR2 1.94e-11 Absolute 1.84e-11 1.00e-03 Yes
reactionrate_lumped_two_air 10.4.1 P1 1.00e+00 P1 1.00e+00 Absolute 3.83e-05 1.00e-03 Yes
reactionrate_series_reaction 10.4.1 O2 7.37e-02 O2 7.37e-02 Absolute 6.32e-06 1.00e-03 Yes
reactionrate_series_reaction 10.4.1 C3H8 2.00e-06 C3H8 1.94e-08 Absolute 1.98e-06 1.00e-03 Yes
reactionrate_series_reaction 10.4.1 CO2 1.20e-01 CO2 1.20e-01 Absolute 5.34e-06 1.00e-03 Yes
reactionrate_series_reaction 10.4.1 H2O 6.54e-02 H2O 6.54e-02 Absolute 2.89e-06 1.00e-03 Yes
realizable_mass_fractions 7.1.7 Sum 1.00e+00 YF 1.00e+00 Absolute 0.00e+00 1.00e-06 Yes
rms_example 17.1.2 mean 1.00e+00 u 9.96e-01 Relative 4.44e-03 4.00e-02 Yes
rms_example 17.1.2 sd 2.89e-01 u_rms 2.80e-01 Relative 3.17e-02 4.00e-02 Yes
screen_drag_1 13.1.6 Exact 0.1 1.26e+02 FDS 0.1 1.25e+02 Relative 8.34e-03 3.00e-02 Yes
screen_drag_1 13.1.6 Exact 0.4 1.40e+01 FDS 0.4 1.39e+01 Relative 7.98e-03 3.00e-02 Yes
screen_drag_1 13.1.6 Exact 0.8 5.29e+00 FDS 0.8 5.26e+00 Relative 6.99e-03 3.00e-02 Yes
screen_drag_2 13.1.6 Exact 0.1 1.26e+03 FDS 0.1 1.25e+03 Relative 7.69e-03 8.00e-02 Yes
screen_drag_2 13.1.6 Exact 0.4 1.68e+02 FDS 0.4 1.67e+02 Relative 6.91e-03 8.00e-02 Yes
screen_drag_2 13.1.6 Exact 0.8 1.24e+02 FDS 0.8 1.23e+02 Relative 5.11e-03 8.00e-02 Yes
shrink_swell 12.6 M_1 (kg/m2) 5.00e-01 M_1 5.00e-01 Relative 0.00e+00 1.00e-02 Yes
shrink_swell 12.6 M_4 (kg/m2) 1.00e+00 M_4 1.00e+00 Relative 0.00e+00 1.00e-02 Yes
shrink_swell 12.6 L_1 (m) 5.00e-04 L_1 5.00e-04 Relative 0.00e+00 1.00e-02 Yes
shrink_swell 12.6 L_4 (m) 2.00e-03 L_4 2.00e-03 Relative 0.00e+00 1.00e-02 Yes
 Case Name Section Expected Expected Predicted Predicted Type of Error Error Within
Quantity Value Quantity Value Error Tolerance Tol.
simple_duct 7.1.1 Mass 5.00e-01 CARBON DIOXIDE 5.00e-01 Relative 2.71e-04 1.00e-02 Yes
smoke_detector 10.10 Obscuration 6.47e+01 SD 6.46e+01 Relative 2.68e-04 1.00e-02 Yes
soot_oxidation_wall 10.11.7 Mass 9.18e-16 SOOT W 9.03e-11 Absolute 9.03e-11 1.00e-06 Yes
soot_oxidation_wall 10.11.7 Energy 3.28e+00 E 3.28e+00 Relative 3.93e-05 1.00e-03 Yes
soot_oxidation_wall 10.11.7 CO2 3.66e-04 CO2 G 3.66e-04 Relative 3.91e-05 1.00e-03 Yes
species_conservation_1 7.1.2 Mass 3.60e-02 PROPANE 3.57e-02 Relative 8.54e-03 1.00e-02 Yes
species_conservation_2 7.1.2 M1 1.00e-12 M1 1.00e-12 Absolute 0.00e+00 1.00e-02 Yes
species_conservation_2 7.1.2 M2 1.00e+00 M2 9.96e-01 Absolute 3.78e-03 1.00e-02 Yes
species_conservation_3 7.1.2 Mass N2 9.31e-03 Mass N2 9.34e-03 Relative 2.72e-03 1.00e-01 Yes
species_conservation_3 7.1.2 Total 8.00e-01 Total 8.00e-01 Relative 2.03e-05 1.00e-01 Yes
species_conservation_4 7.1.2 Mass Ar 1.00e-12 Mass Ar 4.10e-08 Absolute 4.10e-08 5.00e-03 Yes
species_conservation_4 7.1.2 Total 4.00e-01 Total 3.98e-01 Absolute 1.51e-03 5.00e-03 Yes
sphere_drag_1 13.1.4 Exact 5 3.77e+00 FDS 5 3.75e+00 Relative 5.11e-03 5.00e-02 Yes
sphere_drag_1 13.1.4 Exact 10 7.54e+00 FDS 10 7.42e+00 Relative 1.57e-02 5.00e-02 Yes
sphere_drag_1 13.1.4 Exact 20 1.51e+01 FDS 20 1.48e+01 Relative 1.52e-02 5.00e-02 Yes
sphere_drag_2 13.1.4 Pres 2.10e-01 pres 2.13e-01 Relative 1.53e-02 2.00e-02 Yes
spray_burner 15.3.1 HRR 8.00e+04 HRR 8.00e+04 Relative 1.22e-04 5.00e-02 Yes
stack_effect 16.5.1 Ideal Upper 4.15e+00 FDS Upper 4.10e+00 Relative 1.20e-02 5.00e-02 Yes
stack_effect 16.5.1 Ideal Lower 3.95e+00 FDS Lower 3.94e+00 Relative 2.27e-03 5.00e-02 Yes
stack_effect 16.5.1 Ideal Upper Exterior 1.23e+00 FDS Upper Exterior 1.23e+00 Relative 1.69e-03 1.00e-02 Yes
stack_effect 16.5.1 Ideal Lower Exterior 1.24e+00 FDS Lower Exterior 1.24e+00 Relative 1.64e-03 1.00e-02 Yes
stack_effect 16.5.1 Ideal Upper Interior 1.19e+00 FDS Upper Interior 1.18e+00 Relative 2.08e-03 1.00e-02 Yes
stack_effect 16.5.1 Ideal Lower Interior 1.20e+00 FDS Lower Interior 1.20e+00 Relative 1.66e-03 1.00e-02 Yes

297
cone_demo_2 12.10 Exact 1.68e+02 HRRPUA 1.69e+02 Relative 6.29e-03 1.00e-02 Yes
surf_mass_vent_nonchar_cart_fuel 12.1.1 Mass (kg) 3.60e+00 density 3.60e+00 Relative 3.57e-04 1.00e-02 Yes
surf_mass_vent_char_cart_fuel 12.1.1 Mass (kg) 1.80e+00 density 1.80e+00 Relative 2.33e-06 1.00e-02 Yes
surf_mass_vent_nonchar_cyl_fuel 12.1.1 Mass (kg) 1.80e+00 density 1.80e+00 Relative 3.89e-07 1.00e-02 Yes
surf_mass_vent_char_cyl_fuel 12.1.1 Mass (kg) 9.00e-01 density 9.00e-01 Relative 1.43e-06 1.00e-02 Yes
surf_mass_vent_nonchar_spher_fuel 12.1.1 Mass (kg) 1.20e+00 density 1.20e+00 Relative 3.33e-07 1.00e-02 Yes
surf_mass_vent_char_spher_fuel 12.1.1 Mass (kg) 6.00e-01 density 6.00e-01 Relative 8.33e-08 1.00e-02 Yes
surf_mass_part_nonchar_cart_fuel 12.1.2 Mass (kg) 9.00e-03 density 9.00e-03 Relative 4.51e-04 1.00e-02 Yes
surf_mass_part_char_cart_fuel 12.1.2 Mass (kg) 4.50e-03 density 4.50e-03 Relative 0.00e+00 1.00e-02 Yes
surf_mass_part_nonchar_cyl_fuel 12.1.2 Mass (kg) 1.13e-02 density 1.13e-02 Relative 8.61e-04 1.00e-02 Yes
surf_mass_part_char_cyl_fuel 12.1.2 Mass (kg) 5.65e-03 density 5.65e-03 Relative 8.61e-04 1.00e-02 Yes
surf_mass_part_nonchar_spher_fuel 12.1.2 Mass (kg) 1.51e-03 density 1.51e-03 Relative 1.35e-03 1.00e-02 Yes
surf_mass_part_char_spher_fuel 12.1.2 Mass (kg) 7.54e-04 density 7.54e-04 Relative 2.36e-05 1.00e-02 Yes
surf_mass_vent_nonchar_cart_gas 12.1.1 Mass (kg) 3.60e+00 density 3.60e+00 Relative 2.16e-04 1.00e-02 Yes
surf_mass_vent_char_cart_gas 12.1.1 Mass (kg) 1.80e+00 density 1.80e+00 Relative 3.33e-07 1.00e-02 Yes
surf_mass_vent_nonchar_cyl_gas 12.1.1 Mass (kg) 1.80e+00 density 1.80e+00 Relative 3.89e-07 1.00e-02 Yes
surf_mass_vent_char_cyl_gas 12.1.1 Mass (kg) 9.00e-01 density 9.00e-01 Relative 4.33e-07 1.00e-02 Yes
surf_mass_vent_nonchar_spher_gas 12.1.1 Mass (kg) 1.20e+00 density 1.20e+00 Relative 2.50e-07 1.00e-02 Yes
surf_mass_vent_char_spher_gas 12.1.1 Mass (kg) 6.00e-01 density 6.00e-01 Relative 5.00e-07 1.00e-02 Yes
surf_mass_part_nonchar_cart_gas 12.1.2 Mass (kg) 9.00e-03 density 9.00e-03 Relative 3.44e-04 1.00e-02 Yes
surf_mass_part_char_cart_gas 12.1.2 Mass (kg) 4.50e-03 density 4.50e-03 Relative 0.00e+00 1.00e-02 Yes
surf_mass_part_nonchar_cyl_gas 12.1.2 Mass (kg) 1.13e-02 density 1.13e-02 Relative 8.61e-04 1.00e-02 Yes
surf_mass_part_char_cyl_gas 12.1.2 Mass (kg) 5.65e-03 density 5.65e-03 Relative 8.61e-04 1.00e-02 Yes
surf_mass_part_nonchar_spher_gas 12.1.2 Mass (kg) 1.51e-03 density 1.51e-03 Relative 1.35e-03 1.00e-02 Yes
 Case Name Section Expected Expected Predicted Predicted Type of Error Error Within
Quantity Value Quantity Value Error Tolerance Tol.
surf_mass_part_char_spher_gas 12.1.2 Mass (kg) 7.54e-04 density 7.54e-04 Relative 2.36e-05 1.00e-02 Yes
surf_mass_part_specified 15.4.7 Ar mass 5.00e-03 Ar mass 4.99e-03 Relative 2.75e-03 1.00e-02 Yes
surf_mass_part_specified 15.4.7 SO2 mass 6.28e-03 SO2 mass 6.27e-03 Relative 2.72e-03 1.00e-02 Yes
surf_mass_part_specified 15.4.7 He mass 2.51e-03 He mass 2.51e-03 Relative 2.64e-03 1.00e-02 Yes
surf_mass_two_species_cart 12.1.2 Mass (fuel) 3.20e-02 fuel gas mass 3.20e-02 Relative 2.69e-06 2.00e-02 Yes
surf_mass_two_species_cart 12.1.2 Mass (H2O) 2.00e-02 water vapor mass 2.00e-02 Relative 2.75e-06 2.00e-02 Yes
surf_mass_two_species_cart 12.1.2 Mass (total) 4.80e-02 solid mass 4.80e-02 Relative 0.00e+00 2.00e-02 Yes
surf_mass_two_species_cyl 12.1.2 Mass (fuel) 3.20e-02 fuel gas mass 3.20e-02 Relative 2.69e-06 4.00e-02 Yes
surf_mass_two_species_cyl 12.1.2 Mass (H2O) 2.00e-02 water vapor mass 2.00e-02 Relative 2.75e-06 4.00e-02 Yes
surf_mass_two_species_cyl 12.1.2 Mass (total) 4.80e-02 solid mass 4.80e-02 Relative 0.00e+00 4.00e-02 Yes
surf_mass_two_species_spher 12.1.2 Mass (fuel) 3.20e-02 fuel gas mass 3.20e-02 Relative 2.69e-06 4.00e-02 Yes
surf_mass_two_species_spher 12.1.2 Mass (H2O) 2.00e-02 water vapor mass 2.00e-02 Relative 2.75e-06 4.00e-02 Yes
surf_mass_two_species_spher 12.1.2 Mass (total) 4.80e-02 solid mass 4.80e-02 Relative 0.00e+00 4.00e-02 Yes
surf_mass_vent_liquid_fuel 12.4.1 Mass (kg) 6.84e+00 Burning rate 6.84e+00 Relative 1.33e-04 1.00e-02 Yes
surf_mass_vent_liquid_fuel_nonconforming 12.4.1 Mass (kg) 6.84e+00 Burning rate 6.84e+00 Relative 1.53e-04 1.00e-02 Yes
ice_cube 12.11 Mass (kg) 1.00e+00 AMPUA 1.00e+00 Relative 5.54e-04 1.00e-02 Yes
symmetry_test 8.1.1 u_1 3.85e-03 u_2 3.85e-03 Absolute 0.00e+00 1.00e-02 Yes
symmetry_test 8.1.1 v_1 3.85e-03 v_2 3.85e-03 Absolute 0.00e+00 1.00e-02 Yes
symmetry_test 8.1.1 w_1 3.85e-03 w_2 3.85e-03 Absolute 0.00e+00 1.00e-02 Yes
symmetry_test_mpi 8.1.1 u_1 5.66e-04 u_2 5.80e-04 Absolute 1.41e-05 1.00e-02 Yes
symmetry_test_mpi 8.1.1 v_1 5.66e-04 v_2 5.80e-04 Absolute 1.41e-05 1.00e-02 Yes
symmetry_test_mpi 8.1.1 w_1 5.66e-04 w_2 5.80e-04 Absolute 1.41e-05 1.00e-02 Yes
symmetry_test_2 8.1.2 T_left 1.46e+02 T_right 1.46e+02 Relative 0.00e+00 1.00e-03 Yes

298
tangential_velocity 10.1.4 u 2.00e+00 u 2.00e+00 Relative 8.12e-04 4.00e-02 Yes
tangential_velocity 10.1.4 v 3.00e+00 v 2.92e+00 Relative 2.79e-02 4.00e-02 Yes
target_test 9.13 flux 1 8.07e+01 flux 2 8.07e+01 Relative 2.90e-06 1.00e-02 Yes
TC_heating 9.11 Temp 6.00e+02 TC 6.00e+02 Relative 0.00e+00 1.00e-02 Yes
TC_view_factor 9.12 TC 3.80e+02 TC_1 3.90e+02 Relative 2.47e-02 3.00e-02 Yes
tga_sample 12.9.2 Mass 5.33e-01 mpua 5.30e-01 Relative 6.19e-03 1.00e-02 Yes
tga_sample 12.9.2 MLR 4.61e-04 mlrpua 4.83e-04 Relative 4.68e-02 5.00e-02 Yes
thermocouples 11.4 Target_1 5.00e+02 TC_1 5.00e+02 Relative 0.00e+00 1.00e-02 Yes
thermocouples 11.4 Target_2 5.00e+02 TC_2 5.00e+02 Relative 0.00e+00 1.00e-02 Yes
thermocouples 11.4 Target_3 4.95e+02 TC_3 4.95e+02 Relative 4.04e-05 1.00e-02 Yes
thermocouple_time_constant 22.10.8 T_0.5 2.63e+01 T_0.5 2.64e+01 Relative 2.01e-03 1.00e-02 Yes
thermocouple_time_constant 22.10.8 T_3.0 2.63e+01 T_3.0 2.64e+01 Relative 1.98e-03 1.00e-02 Yes
thermocouple_time_constant 22.10.8 T_8.0 2.63e+01 T_8.0 2.65e+01 Relative 8.58e-03 1.00e-02 Yes
test_hrr_2d_cyl 22.10.1 HRR 1.56e+02 HRR 1.55e+02 Relative 4.00e-03 1.00e-02 Yes
tmp_lower_limit 3.8.5 T 2.00e+01 minT 2.00e+01 Absolute 1.98e-02 2.00e-01 Yes
tmp_lower_limit 3.8.5 T 2.00e+01 minT 2.00e+01 Absolute 1.10e-03 1.00e-02 Yes
tmp_lower_limit 3.8.5 T 2.00e+01 minT 2.00e+01 Absolute 1.51e-03 1.00e-02 Yes
tunnel_const_gamma 3.10.1 p 7.78e-01 p 7.78e-01 Relative 3.66e-04 1.00e-02 Yes
tunnel_const_gamma 3.10.1 u 1.71e+00 u 1.71e+00 Relative 4.89e-05 1.00e-02 Yes
tunnel_const_gamma 3.10.1 rho 7.02e-01 rho 7.02e-01 Relative 4.34e-04 1.00e-02 Yes
tunnel_const_gamma 3.10.1 T 2.27e+02 T 2.28e+02 Relative 1.35e-03 1.00e-02 Yes
tunnel_linear_cp 3.10.2 p 7.72e-01 p 7.72e-01 Relative 8.55e-05 1.00e-02 Yes
tunnel_linear_cp 3.10.2 u 1.70e+00 u 1.70e+00 Relative 5.08e-05 1.00e-02 Yes
tunnel_linear_cp 3.10.2 rho 7.05e-01 rho 7.05e-01 Relative 4.92e-05 1.00e-02 Yes
 Case Name Section Expected Expected Predicted Predicted Type of Error Error Within
Quantity Value Quantity Value Error Tolerance Tol.
tunnel_linear_cp 3.10.2 T 2.26e+02 T 2.26e+02 Relative 5.18e-05 1.00e-02 Yes
tunnel_demo 21.3 v_target -1.00e-01 Velocity Error -5.10e-05 Absolute 9.99e-02 1.00e+02 Yes
tunnel_demo 21.3 p_target -5.00e+02 Pressure Error -2.00e-06 Absolute 5.00e+02 1.00e+05 Yes
tunnel_demo_glmat 21.3 v_target -1.00e-01 Velocity Error -5.29e-19 Absolute 1.00e-01 1.00e+02 Yes
tunnel_demo_glmat 21.3 p_target -5.00e+02 Pressure Error -2.00e-06 Absolute 5.00e+02 1.00e+05 Yes
two_step_solid_reaction 12.8 Ya 3.57e-09 Ya 3.60e-09 Absolute 2.85e-11 1.00e-02 Yes
two_step_solid_reaction 12.8 Yb 6.25e-06 Yb 6.34e-06 Absolute 8.64e-08 1.00e-02 Yes
two_step_solid_reaction 12.8 Yc 1.00e+00 Yc 1.00e+00 Absolute 6.34e-06 1.00e-02 Yes
vegetation_absorb 13.2.1 rad 2.13e+01 rad 2.14e+01 Relative 7.04e-03 3.00e-02 Yes
vegetation_drag_1 13.1.5 Pressure 1.89e+01 p1 1.85e+01 Relative 2.26e-02 3.00e-02 Yes
vegetation_drag_1 13.1.5 Pressure 1.89e+01 p2 1.86e+01 Relative 1.24e-02 3.00e-02 Yes
vegetation_drag_1 13.1.5 Pressure 1.89e+01 p3 1.85e+01 Relative 1.85e-02 3.00e-02 Yes
vegetation_drag_2 13.1.5 Pressure 8.53e+00 p1 8.54e+00 Relative 1.05e-03 2.00e-02 Yes
vegetation_drag_2 13.1.5 Pressure 8.53e+00 p2 8.44e+00 Relative 1.10e-02 2.00e-02 Yes
vegetation_drag_2 13.1.5 Pressure 8.53e+00 p3 8.43e+00 Relative 1.21e-02 2.00e-02 Yes
velocity_bc_test 8.2 Vel1 1.64e+00 Vel2 1.61e+00 Relative 2.19e-02 3.00e-02 Yes
velocity_bc_test 8.2 Pres1 1.83e-02 Pres2 2.06e-02 Absolute 2.33e-03 5.00e-02 Yes
volume_flow_1 10.1.1 vflow 1.00e-02 vflow 1.00e-02 Relative 6.79e-05 1.00e-02 Yes
volume_flow_2 10.1.1 vflow 1.00e-02 vflow 1.00e-02 Relative 2.51e-04 1.00e-02 Yes
water_evap_1_const_gamma 3.9.3 Rel. Hum 2.11e+00 humid 2.10e+00 Relative 4.55e-03 1.00e-02 Yes
water_evap_1_const_gamma 3.9.3 h_gas -2.77e+01 h_gas -2.76e+01 Relative 3.27e-03 1.00e-02 Yes
water_evap_1_const_gamma 3.9.3 h_water 1.98e+01 h_water 1.98e+01 Relative 0.00e+00 1.00e-02 Yes
water_evap_1_const_gamma 3.9.3 dens 1.00e-02 dens 9.95e-03 Relative 4.66e-03 1.00e-02 Yes

299
water_evap_1_const_gamma 3.9.3 temp 1.54e+02 Temp 1.54e+02 Relative 1.04e-03 1.00e-02 Yes
water_evap_1_const_gamma 3.9.3 pres -7.90e+03 pres -7.88e+03 Relative 3.27e-03 1.00e-02 Yes
water_evap_1_const_gamma 3.9.3 vapor 1.00e-02 WATER VAPOR 9.95e-03 Relative 4.66e-03 1.00e-02 Yes
water_evaporation_1 13.3.1 Rel. Hum 2.11e+00 humid 2.10e+00 Relative 4.55e-03 1.00e-02 Yes
water_evaporation_1 13.3.1 h_gas -1.67e+02 h_gas -1.66e+02 Relative 4.42e-03 1.00e-02 Yes
water_evaporation_1 13.3.1 h_water 1.59e+02 h_water 1.59e+02 Relative 1.58e-06 1.00e-02 Yes
water_evaporation_1 13.3.1 dens 1.00e-02 dens 9.95e-03 Relative 4.65e-03 1.00e-02 Yes
water_evaporation_1 13.3.1 temp 1.54e+02 Temp 1.54e+02 Relative 1.04e-03 1.00e-02 Yes
water_evaporation_1 13.3.1 pres -7.81e+03 pres -7.79e+03 Relative 3.30e-03 1.00e-02 Yes
water_evaporation_1 13.3.1 vapor 1.00e-02 WATER VAPOR 9.95e-03 Relative 4.65e-03 1.00e-02 Yes
water_evaporation_2 13.3.2 humid 1.00e+02 humid 9.97e+01 Relative 3.07e-03 1.00e-02 Yes
water_evaporation_2 13.3.2 h_gas -3.32e+01 h_gas -3.30e+01 Relative 4.66e-03 1.00e-02 Yes
water_evaporation_2 13.3.2 h_water 2.94e+01 h_water 2.92e+01 Relative 5.77e-03 1.00e-02 Yes
water_evaporation_2 13.3.2 dens 2.16e-02 dens 2.15e-02 Relative 4.92e-03 1.00e-02 Yes
water_evaporation_2 13.3.2 Temp 2.34e+01 Temp 2.34e+01 Relative 8.53e-04 1.00e-02 Yes
water_evaporation_2 13.3.2 temp_w 2.34e+01 temp_w 2.34e+01 Relative 9.63e-04 1.00e-02 Yes
water_evaporation_2 13.3.2 pres -5.95e+04 pres -5.95e+04 Relative 2.95e-04 1.00e-02 Yes
water_evaporation_2 13.3.2 H2O 4.54e-02 H2O 4.52e-02 Relative 4.70e-03 1.00e-02 Yes
water_evaporation_3 13.3.3 humid 1.00e+02 humid 9.98e+01 Relative 2.16e-03 1.00e-02 Yes
water_evaporation_3 13.3.3 h_gas -1.09e+03 h_gas -1.09e+03 Relative 6.89e-04 7.00e-02 Yes
water_evaporation_3 13.3.3 dens 6.55e-02 dens 6.54e-02 Relative 8.40e-04 2.00e-02 Yes
water_evaporation_3 13.3.3 Temp 4.46e+01 Temp 4.47e+01 Relative 7.64e-04 1.00e-02 Yes
water_evaporation_3 13.3.3 temp_w 4.46e+01 temp_w 4.47e+01 Relative 7.64e-04 1.00e-02 Yes
water_evaporation_3 13.3.3 pres -5.01e+04 pres -5.01e+04 Relative 5.64e-05 1.00e-02 Yes
 Case Name Section Expected Expected Predicted Predicted Type of Error Error Within
Quantity Value Quantity Value Error Tolerance Tol.
water_evaporation_3 13.3.3 H2O 1.26e-01 H2O 1.26e-01 Relative 7.35e-04 2.00e-02 Yes
water_evaporation_4 13.3.4 Heat_Flow 1.07e+02 heat 1.04e+02 Relative 2.62e-02 6.00e-02 Yes
water_evaporation_4 13.3.4 Heat_Flow 1.07e+02 heat 1.04e+02 Relative 2.28e-02 6.00e-02 Yes
water_evaporation_5 13.3.5 Diameter 3.04e-04 D_thin 3.53e-04 Absolute 4.93e-05 1.00e-04 Yes
water_evaporation_5 13.3.5 Diameter 3.04e-04 D_thick 2.97e-04 Absolute 6.04e-06 1.00e-04 Yes
water_evaporation_6 13.3.6 T_gas 1.72e+02 T_gas 1.72e+02 Relative 1.99e-04 1.00e-02 Yes
water_evaporation_6 13.3.6 T_wall 1.72e+02 T_wall 1.72e+02 Relative 8.27e-04 1.00e-02 Yes
water_evaporation_6 13.3.6 Pressure 3.46e+04 pres 3.46e+04 Relative 3.23e-04 1.50e-02 Yes
water_evaporation_6 13.3.6 sum 1.00e-12 u_gas 1.35e-03 Absolute 1.35e-03 1.00e-02 Yes
water_evaporation_7 13.3.7 Q_PART -4.34e+01 Q_PART -4.15e+01 Relative 4.49e-02 1.00e-01 Yes
water_evaporation_8 13.3.8 z 5.41e-01 z_thin 5.49e-01 Absolute 7.69e-03 5.00e-02 Yes
water_evaporation_8 13.3.8 z 5.41e-01 z_thick 5.57e-01 Absolute 1.61e-02 5.00e-02 Yes
water_evaporation_8 13.3.8 d 2.72e+01 D_thin 2.54e+01 Absolute 1.76e+00 5.00e+00 Yes
water_evaporation_8 13.3.8 d 2.72e+01 D_thick 2.39e+01 Absolute 3.32e+00 5.00e+00 Yes
water_pool 12.4.2 mdot 1.42e-05 mdot 1.46e-05 Relative 2.80e-02 6.00e-02 Yes
wind_example_5 16.1 u_ideal 1.00e-12 u_mean 1.87e-02 Absolute 1.87e-02 1.00e+00 Yes
wind_example_5 16.1 v_ideal 5.00e+00 v_mean 4.97e+00 Absolute 2.75e-02 1.00e+00 Yes
wind_example_10 16.1 u_ideal 1.00e-12 u_mean 7.62e-03 Absolute 7.62e-03 1.00e+00 Yes
wind_example_10 16.1 v_ideal 5.00e+00 v_mean 4.97e+00 Absolute 2.54e-02 1.00e+00 Yes
wind_example_32 16.2.2 u_ideal 7.64e+00 u 7.63e+00 Absolute 1.20e-02 1.00e+00 Yes
wind_example_32 16.2.2 T_ideal 1.85e+01 T 1.85e+01 Absolute 2.89e-02 1.00e+00 Yes
zone_break_slow 10.3.1 Pres1 1.97e+03 pres_1 1.97e+03 Relative 1.36e-03 2.00e-02 Yes
zone_break_slow 10.3.1 Pres2 1.00e+03 pres_2 1.02e+03 Relative 1.31e-02 2.00e-02 Yes

300
zone_break_slow 10.3.1 Pres3 1.00e+03 pres_3 1.01e+03 Relative 6.30e-03 2.00e-02 Yes
zone_break_fast 10.3.1 Pres1 1.97e+03 pres_1 1.97e+03 Relative 1.36e-03 2.00e-02 Yes
zone_break_fast 10.3.1 Pres2 1.00e+03 pres_2 1.01e+03 Relative 1.18e-02 2.00e-02 Yes
zone_break_fast 10.3.1 Pres3 1.00e+03 pres_3 1.01e+03 Relative 1.18e-02 2.00e-02 Yes
zone_break_slow_ulmat 10.3.1 Pres1 1.97e+03 pres_1 1.97e+03 Relative 1.36e-03 2.00e-02 Yes
zone_break_slow_ulmat 10.3.1 Pres2 1.00e+03 pres_2 1.02e+03 Relative 1.31e-02 2.00e-02 Yes
zone_break_slow_ulmat 10.3.1 Pres3 1.00e+03 pres_3 1.01e+03 Relative 6.30e-03 2.00e-02 Yes
zone_break_fast_ulmat 10.3.1 Pres1 1.97e+03 pres_1 1.97e+03 Relative 1.36e-03 2.00e-02 Yes
zone_break_fast_ulmat 10.3.1 Pres2 1.00e+03 pres_2 1.01e+03 Relative 1.18e-02 2.00e-02 Yes
zone_break_fast_ulmat 10.3.1 Pres3 1.00e+03 pres_3 1.01e+03 Relative 1.18e-02 2.00e-02 Yes
zone_shape 10.3.1 Pres 2.52e+03 pres_1 2.50e+03 Relative 8.99e-03 1.50e-02 Yes
zone_shape_2 14.10.6 Delta p 1.00e+02 Delta p 9.95e+01 Relative 4.64e-03 1.00e-02 Yes
zone_shape_2 14.10.6 Mass 3.13e+02 Total 3.13e+02 Relative 4.83e-04 1.00e-02 Yes
humidity 10.2 H2Ov 6.03e-02 H2Ov 6.03e-02 Relative 2.43e-05 1.00e-03 Yes
matl_e_cons 12.2 M1 -8.65e+02 H_M1 -8.65e+02 Relative 1.39e-07 5.00e-03 Yes
matl_e_cons 12.2 M1 -1.86e+03 H_M1 -1.87e+03 Relative 2.68e-07 5.00e-03 Yes
matl_e_cons 12.2 M1 -8.65e+02 H_M1 -8.65e+02 Relative 1.98e-04 5.00e-03 Yes
matl_e_cons 12.2 M2 -1.86e+03 H_M2 -1.87e+03 Relative 2.68e-07 5.00e-03 Yes
matl_e_cons 12.2 M1 -1.14e+03 H_M1 -1.14e+03 Relative 8.77e-08 5.00e-03 Yes
matl_e_cons 12.2 M2 -6.95e+02 H_M2 -6.95e+02 Relative 1.15e-07 5.00e-03 Yes
matl_e_cons 12.2 M1 -1.14e+03 H_M1 -1.14e+03 Relative 8.77e-08 5.00e-03 Yes
matl_e_cons 12.2 M1 PC -6.40e+02 H_M1a -6.40e+02 Relative 1.56e-07 5.00e-03 Yes
matl_e_cons 12.2 M2 -6.95e+02 H_M2 -6.95e+02 Relative 1.15e-07 5.00e-03 Yes
matl_e_cons 12.2 M1 -1.31e+03 H_M1 -1.31e+03 Relative 2.45e-06 5.00e-03 Yes
 Case Name Section Expected Expected Predicted Predicted Type of Error Error Within
Quantity Value Quantity Value Error Tolerance Tol.
matl_e_cons 12.2 M1 PC -9.73e+02 H_M1a -9.73e+02 Relative 3.52e-06 5.00e-03 Yes
matl_e_cons 12.2 M2 -6.95e+02 H_M2 -6.95e+02 Relative 1.15e-07 5.00e-03 Yes
matl_e_cons 12.2 M1 -4.29e+02 H_M1 -4.29e+02 Relative 1.17e-07 5.00e-03 Yes
matl_e_cons 12.2 M2 7.27e+02 H_M2 7.27e+02 Relative 2.75e-08 5.00e-03 Yes
matl_e_cons 12.2 M1 -3.53e+02 H_M1 -3.53e+02 Relative 1.20e-03 5.00e-03 Yes
matl_e_cons 12.2 M1 -3.53e+02 H_M1 -3.53e+02 Relative 1.20e-03 5.00e-03 Yes

301
Appendix B

Installation Testing

The use of computer software for safety purposes is sometimes regulated under a software quality assur-
ance (SQA) program by the Authority Having Jurisdiction. For example, in the United States, both the
Department of Energy (DOE) via DOE G 414.1-4 [111] and the Nuclear Regulatory Commission (NRC)
via NUREG-1737 [112] impose specific requirements on software used for certain safety purposes. In ad-
dition to imposing general requirements on configuration management and verification and validation, SQA
programs often impose an acceptance testing requirement. For simulation software, this involves a test pro-
gram in which the software is used to execute a number of sample cases and the results are compared to
accepted values. The purpose of this testing process is to ensure that the software installation was successful
and that no corruption of the software executable or libraries has occurred. The acceptance test procedure is
as follows:

1. Obtain and install the version of FDS you will be running. Note that this procedure will need to be
repeated for each version of the FDS executable being used.

2. Find the files shown in Table B.1 in the Examples folder of the FDS-SMV installation package. The
Folder listed in the table is a subfolder of the Examples folder.

3. Run each of the FDS input files listed in Table B.1.

4. Compare the results of each FDS Output File listed in Table B.1 with the exact/expected results in the
corresponding Expected Results File. The Expected Results Files can be found at
https://github.com/firemodels/fds/tree/master/Verification
This is a link to the FDS repository hosted by GitHub.

5. Verify that the expected results match the results from Step 3 using the criteria shown in Appendix A.

303
 Table B.1: FDS Installation Test Suite. Details of each case may be found in the listed section or reference. Note that the light blue colored links
refer to the FDS User’s Guide.
Folder Reference Input File FDS Output File Expected Results File
Controls 18.5.11 activate_vents.fds activate_vents_ctrl.csv activate_vents.csv
Pressure_Effects 7.2.3 isentropic2.fds insentropic2_devc.csv isentropic2.csv
Energy_Budget 7.2.5 energy_budget_tmix.fds energy_budget_tmix_devc.csv energy_budget_tmix.csv
Energy_Budget 7.2.1 energy_budget_cold_walls.fds energy_budget_cold_walls_hrr.csv energy_budget_cold_walls.csv
Fires 15.3.1 spray_burner.fds spray_burner_hrr.csv spray_burner.csv
Heat_Transfer 11.2 heat_conduction_kc.fds heat_conduction_kc_devc.csv heat_conduction_kc.csv
HVAC 14.11 ashrae7_table.fds ashrae7_table_devc.csv ashrae7_exp.csv
Pressure_Solver 21.1.2 duct_flow.fds duct_flow_devc.csv duct_flow.csv
Pressure_Solver 21.1.2 duct_flow_uglmat.fds duct_flow_uglmat_devc.csv duct_flow.csv
Pyrolysis 12.7 enthalpy.fds enthalpy_devc.csv enthalpy.csv
Pyrolysis 12.6 shrink_swell.fds shrink_swell_devc.csv shrink_swell.csv
Radiation 9.14 radiation_shield.fds radiation_shield_devc.csv radiation_shield.csv
Species 10.4.1 reactionrate_EDC_1step_CH4_nonmix.fds reactionrate_EDC_1step_CH4_nonmix_devc.csv reactionrate_EDC_1step_CH4_nonmix_spec.csv
Species 10.4.2 reactionrate_arrhenius_2order_1step.fds reactionrate_arrhenius_2order_1step_devc.csv reactionrate_arrhenius_2order_1step_soln.csv
Sprinklers_and_Sprays 13.3.1 water_evaporation_1.fds water_evaporation_1_devc.csv water_evaporation_1.csv

304
Sprinklers_and_Sprays 13.1.3 particle_drag_U50_N16.fds particle_drag_U50_N16_devc.csv particle_drag_U50_N16.csv
Appendix C

Unstructured Geometry integral

conservation tests

Consider an immersed body B, with surface defined by Nτ triangles, completely immersed in a domain Ω
divided in M cartesian mesh blocks. Three geometry conservation tests can be defined for the polyhedra
definition around this body:
1. Surface area: The total body B surface area can be computed from the triangulation AB,τ = ∑Niττ=1 Aiτ ,
where Nτ is the number of triangles in the surface and Aiτ their areas. It can also be calculated from the
N f
boundary cut-face areas computed by the cut-cell definition algorithm, AB,CF = ∑icbcf =1 Aic f , where Nbc f
is the number of cut-faces that have been computed on the body surface. These two areas should match.

2. Gas (G) region volume: This is the volume of region Ω − ΩB . It can be computed from the solid volume
VB and the domain volume VΩ as VGB = VΩ −VB . Here:
M
VΩ = ∑ Vnm ; Vnm = Lx,nm Ly,nm Lz,nm (C.1)
nm=1

where Lx,nm , Ly,nm and Lz,nm are the nm mesh sizes in each cartesian direction (M = 1, Lx,1 = 3DB ,
Ly,1 = 3DB and Lz,1 = 8DB in the cone problem of Sec. 16.3.1). The discrete volume of the solid can be
computed from the triangulation applying the divergence theorem to a field Fv = xî, such that ∇ · F̂v = 1,
to the surface triangles:
Z Z Nτ
VB = ∇ · Fv dV = Fv · n̂ dA = ∑ (î · n̂i )xci Ai (C.2)
ΩB ∂ ΩB i=1

where xci is the ith triangle centroid position in the x direction, and Ai its area. The G region volume can
also be computed from the sum of cut-cells and regular G cells volumes over all M cartesian meshes:
Ncc
VGRC = ∑ Vicc + Nrg ∆x∆y∆z (C.3)
icc=1

where Nrg the number of regular cells in the G region, and ∆x, ∆y, ∆z the uniform grid cell sizes. The
Ncc cut-cell volumes Vicc are computed in similar manner to equation (C.2):
Z nc f Z
Vicc = Fv · n̂ dA = ∑ (î · n̂ic f ) x dA (C.4)
∂ Ωicc ic f =1 Ωic f

305
 where nc f is the number of faces (cut-face or regular) which are boundary of cut-cell icc.
In this test VGB and VGRC should agree.

3. G region centroid location: The centroid of the G region can be computed from cut-cells and regular G
cells as: " #
Ncc Nrg
1
xGRC = ∑ xc(icc)Vicc + ∑ xc(irg) ∆x∆y∆z (C.5)
VGRC icc=1 irg=1

Where xc(irg) is the cell center location for regular cell irg, and the scalar components in each coordinate
direction {î, ĵ, k̂} of cut-cell centroids xc(icc) are obtained from:

1 1 1 nc f
Z Z Z
xci(icc) = ∇ · Fc(i) dV = Fc(i) · n̂ dA = (êi · n̂ic f ) xi2 dA (C.6)
Vicc Ωicc Vicc ∂ Ωicc 2Vicc ic∑
f =1 Ωic f

where Fc(i) = 1/2xi2 êi . Here, xi = {x, y, z} and êi = {î, ĵ, k̂}.
The centroid of the domain Ω is:
1 M
xΩ = ∑ xnmVnm (C.7)
VΩ nm=1

where xnm = xlow,nm + 12 (Lx,nm , Ly,nm , Lz,nm ), where xlow,nm is the vector of low boundary coordinates for
mesh nm. In the single mesh cone problem (Sec. 16.3.1) xlow,1 = (−1.5DB , −1.5DB , −2DB ).
Also, the centroid xB of the discretized solid can be computed applying equation (C.6) to the surface
triangulation τ, in same manner to what was done for cut-faces. From domain and solid the centroid of
the G region can be obtained as:
1
xGB = [xΩVΩ − xBVB ] (C.8)
VGB
Consistency of the computed centroids requires that xGRC and xGB should correspond.

306