Chapter 1

Process Simulation in Chemical Engineering

1.1 Introduction

Chemical process simulation aims to represent a process of chemical or physical

transformation through a mathematic model that involves the calculation of mass
and energy balances coupled with phase equilibrium and with transport and chem-
ical kinetics equations. All this is made looking for the establishment (prediction) of
the behavior of a process of known structure, in which some preliminary data of the
equipment that constitute the process are known.
The mathematical model employed in process simulation contains linear,
nonlinear, and differential algebraic equations, which represent equipment or
process operations, physical–chemical properties, connections between the equip-
ment and operations and their specifications. These connections are summarized in
the process flow diagram.
Process flow diagrams are the language of chemical processes. Between them
the state of art of an existing or hypothetical process are revealed. Thereby, the
process simulators are employed for the interpretation and analysis of information
contained in the process flow diagrams in order to foresee failures and evaluate the
process performance. The analysis of the process is based on a mathematic model
integrated by a group of equations that associate process variables such as temper-
ature, pressure, flows, and compositions, with surface areas, geometrical configu-
ration, set points of valves, etc.
In most of the simulators the solution of the equations system is made linearly,
solving each unit separately and moving forward in the system once the variables
required for the calculation of the next unit are known. However, that process is
useless when there are stream recycles in the system since some of the variables to
calculate are required for the process initialization.
An alternative solution for that type of problems consists in taking one stream as
tear stream. That means assuming the initial values of that stream to start the
calculations; later on, based on the assumed information, each of the following units

© Springer International Publishing Switzerland 2016 1

I.D.G. Chaves et al., Process Analysis and Simulation in Chemical Engineering,
DOI 10.1007/978-3-319-14812-0_1
2 1 Process Simulation in Chemical Engineering

is solved obtaining new values for the parameters of the tear stream. Subsequently,
the new values help to repeat these calculations again and again, until the difference
between the initial and the calculated values fulfill a given tolerance; that point is
known as convergence.

1.2 Chemical Process Simulators

A process simulator is software used for the modeling of the behavior of a chemical
process in steady state, by means of the determination of pressures, temperatures,
and flows. Nowadays, the computer programs employed in process simulation have
broad in the study of the dynamic behavior of processes, as well as to the control
systems and their response to perturbations inherent to the operation. In the same
way, software to perform equipment sizing, cost estimation, properties estimation
and analysis, operability analysis and process optimization, are now available in the
market; all those characteristics can be observed in the Aspen Engineering Suite,
and some of them are presented in last chapters.
Process simulators allow:
• Predict the behavior of a process
• Analyze in a simultaneous way different cases, changing the values of main
operating variables
• Optimize the operating conditions of new or existing plants
• Track a chemical plant during its whole useful life, in order to foresee extensions
o process improvements
The appearance and development of digital informatics determined the advance
of different areas of the human knowledge. Chemical engineering was a part of that
development, particularly in the application to the process simulation. The first
attempts of mathematical modeling refer back to the 1950s with the debut of
FORTRAN language (FORmula TRANslating). Afterwards, in the 1970s, appears
the first process simulator, known as FLOWTRAN, which would frame the begin of
an ceaseless research work principally performed by the academy, in some cases
financed by the industry, and routed to make the process operation more profitable
and to access to the evaluation of different alternatives in a relative short time.
Process simulators are built in libraries of subroutines or models, generally
developed in FORTRAN, Cþþ, or Visual Basic, that conform algorithms for the
solution of equations. The subroutines or models are known as “procedures,”
“models,” or “blocks.” To make an effective use of the simulators, process engi-
neers must know the guidelines and assumptions of the models provided by each
simulator. Those assumptions are described in the user manuals. Furthermore, it is
always important to take into account the criteria used in the specification of phase
equilibrium and the models used for that purpose, as the accuracy of the obtained
results of a simulation is affected by them.
1.3 Types of Process Simulators 3

The commercial and academic process simulators more divulged actually are,
among others: SPEED UP , ASPEN PLUS , DESIGN II , HYSYM , ASPEN
HYSYS , CHEMCAD , and PRO II .

1.3 Types of Process Simulators

Process simulators are classified according to the simulation strategy that they use
to set the mathematical model that represents the process to simulate. The simula-
tion strategy refers to the way in which the problem of the model solution is
boarded. Generally the strategy depends on the complexity of the model and the
calculation mode. The first one, understood as the different existing possibilities,
since linear to sophisticated models with equations of mass, energy and momentum
transfer rates. The second, referred to the information (input variables) that is
necessary to specify to solve the model in terms of the remaining information
(output variables).
The subroutines of a process simulator are computer programs supplied initially
with vectors containing the information corresponding to the feed streams of the
process and some of its parameters. The subroutine takes the vectors, interprets the
information, and looks for the appropriate model to solve the problem. The results
are, basically, the product streams of the process. Thus, the subroutines permit
working with two calculation modes in a process simulator:
• Design mode: according to the required process conditions, a desired perfor-
mance is used as starting point to find the process or equipment specifications
that allow the accomplishment of those conditions.
• Rating mode: according to some design specifications provided to the simulator,
the performance of the process or equipment is evaluated to meet some specific
conditions of the process.
The two fundamental strategies used in the solution of simulation problems are
the sequential strategy and the simultaneous strategy.
The concept of sequential simulation comes from the necessity of calculating
different process units, which are part of a flow diagram, in a rating mode. In this
mode, some of the values of the feed streams and the parameter specifications of
each one of the units must be known. As it is impossible to specify all the streams
entering to all the units in a simultaneous way, it becomes necessary to use the
outcoming values obtained from the calculation in one unit as input information in
the next unit. In that way a sequential order of calculation is established to solve
one-by-one all the units of the process. The calculation order is fixed automatically
by the simulator, ensuring that it would be consistent with the information flow. In
almost all the cases the calculations begin in the equipment for which the feed
streams values and equipment parameters are known. In Fig. 1.1 the typical
calculation order in a flow diagram is shown with arrows. The calculation order
normally matches with the direction of mass flow and is modified or interrupted
4 1 Process Simulation in Chemical Engineering

Fig. 1.1 Calculation scheme of a flow diagram. Source: Authors

with the appearance of recycles, that make mandatory the implementation of an

iterative process.
A simulation becomes more complex as the number of recycles increases, thus it
becomes more difficult to implement in a successful way. For that reason, a second
option for the solution of the model consists on taking all the equations at a time and
builds a unique simulation model by the simultaneous solution of all the equations.
In this way it is not indispensable to evaluate all the equipment units from the input
values and their parameters, allowing the manipulation of design or evaluation
specifications without any distinction.

1.3.1 Sequential Modular Simulators

These are simulators in which each process unit is represented with a module in
which the operation model and the numeric algorithm used for the calculation of the
outgoing streams are included. The modules are totally independent from each
other and the information flow for the calculation in the simulation coincides with
the “physical flow” in the plant. The advantage of using modules is that each
equations system is solved using its own methodology.
The module of each process unit must contain routines with the models and
procedures for the solution from a set of predetermined variables. Additionally,
when the process to simulate has several recycles caused by recirculation streams
1.3 Types of Process Simulators 5

and counter-current operations, the executer module follows a methodology com-

posed by three stages:
• Partitioning: recycle detection
• Tearing: selection of tear streams, over which the iterative process is made
• Ordering: establishment of a logical sequence for the operation units
The recycles owe their existence to all those processes in which reversible
reactions and competitive phenomena take place. In the simulators there are
specific subroutines for the calculation of recycles and the pursuit of convergence.
Between them an initialization value for the tear stream is established, the calcu-
lation is executed and, finally, the values are compared and new initialization values
are given until the convergence is achieved. The better-known methods of conver-
gence in the simulators are Wegstein, successive substitutions (or direct iteration),
Newton–Raphson and Broyden Quasi-Newton. The Wegstein method is employed
in those situations in which the “successive substitutions” fail or a high number of
iterations is required, as will be explained in Sect. 1.5.
The recycles constitute an extra unit in the simulation flow diagrams. It should
be noticed that it must exist a convergence algorithm that allow the adjustment of
the recycles and, therefore, their computing must be performed separately from the
normal calculation units associated with unit operations. Generally, the recycles are
not installed, that means, they must not be specified as calculation blocks in a
simulation. Nonetheless, as will be seen further in Sect. 1.6.2, in the specific case of
Aspen Hysys some blocks that denote the recycle streams over which the iteration
is executed, that is, the tear streams selected for the problem solution must be
defined.
Summarizing, the principal features of a sequential modular simulator are
following:
• Contains calculation libraries and routines
• The iteration variables are in the recycle streams
• The individual models are solved in an efficient way
• It is easily understandable for engineers that are not “simulation specialists” due
to the existing correspondence between the mass flow and the calculation
sequence
• It involves complex convergence methods (direct substitution, Wegstein, etc.)
• The information entered by the user (related with streams and equipment) is
easily verifiable
• The design problems are easier to solve (parameter selection)

1.3.2 Simultaneous or Equation Oriented Simulators

In this type of simulators the mathematic model that represents the process is set
building a large algebraic equations system. Here modules representing process
6 1 Process Simulation in Chemical Engineering

units or subsystems may also exist, but these do not contain numerical methods for
the calculation of output values; contrary, they possess the required information to
provide the equations that represent the mathematic model.
The process model is made of the sum of the models of all the units that comprise
the process or plant in the simulation. Due to the compilation and clustering
philosophy of all the equations that build the process, this type of simulators are
known as “Equation oriented” or “Equation based.” The concept of modules is
maintained here in order to facilitate the interaction with the user and to allow the
specification of the required information for the problem.
The main problem associated to the concept of simultaneous or equation ori-
ented solution is the convergence of the system and the consistency of the solutions
found. Thus the highly nonideal systems that correspond to the chemical plants
models could, for instance, produce multiple solutions. Additionally, the numeric
solution of problems consisting of large equations systems requires proper initial-
ization; this is close to a solution surrounding (status).
Summarizing, the principal features of the equation oriented simulators are:
• Each unit is represented by the equations that model it. The model is the
integration of all the subsystems
• The distinction between process variable and operative parameters disappears.
In consequence the design problems are simplified
• Simultaneous solution of the system of (nonlinear) algebraic equations
• Higher convergence velocity
• They need a better initialization (the higher the complexity of the problem to
solve the better the initialization to be provided)
• The higher the complexity, the less the reliability in the results and the more the
convergence problems (solutions without physical meaning)
• Easier to use for “nonspecialists”

1.3.3 Hybrid Simulators

This type of simulators uses a strategy mixture of the sequential and simultaneous
ones. Each iteration consists of two steps: a first model solution employing a
sequential strategy, and the upgrading of the lineal coefficients to find a solution
with a simultaneous strategy.
The simulation begins with a sequential step in which the output variables are
determined from input information and parameters in an initial scanning of the flow
diagram. Then a simultaneous step starts for the solution of linearized models in a
second scanning of the process.
Some simulators of high industrial application appeal to a hybrid strategy to
ensure convergence, even in the worst cases.
8 1 Process Simulation in Chemical Engineering

1.4 Applications of Process Simulation

Process simulation is a tool for process and chemical engineers that can be used in
the execution of repetitive tasks or in activities of high complexity that must be
solved in relatively short times.
The various applications that process simulation has found are result of the
necessity of:
• Making a better use of the energy resources
• Minimizing the operating costs and the emission of waste streams that may be
contaminant
• Increasing the yield and process efficiency
• Improving the process controllability
• Propelling the teaching of process design
Some of the principal applications of process simulation are discussed as
follows.

1.4.1 Computer-Aided Design

The steady-state simulation of mass and energy balances constitutes itself the center
piece of the computer-aided design. The principal reasons are: (a) the results of the
calculations in the design stage are necessary for further stages; (b) during the
design, in order to meet the economical and operation restrictions, the information
changes dynamically, in such a way that it is necessary to adjust and actualize the
result of the balances in a continuous way; (c) such a large quantity of information
is generated that the only way to administrate it is if it is consolidated through the
process simulations and the cases of study developed with the simulations (Yee Foo
et al. 2005).
Process simulation allows making a study of different process alternatives (flow
diagrams) in order to determine in a reasonable time which of them are not feasible.
Besides, a flexible simulator develops different cases of study in the search of an
optimal configuration of the process, making possible that the design moves to
more advanced stages rapidly. In the same way, a more flexible simulator can be
used to conclude about to rival technologies, to design or evaluate the more
adequate operating configuration for a process or to plan in the most economical
way the laboratory and plant experiments required for the design.
1.4 Applications of Process Simulation 9

1.4.2 Process Optimization

The optimization of chemical processes has its origin in the linear programming at
the beginning of the 1960s. This task has as fundamental goal the comparison of
different alternatives to select the best according to some process response criteria.
In an optimization process it is important to identify the independent characteristics
that lead to different results (independent variables) and the variables that make
possible the measurement of the relative excellence of a solution (dependent
variable). The set of interactions between the dependent variables that conduce to
a response is known as objective function, in which the parameters of the system
relate with each other.
In a general way, the optimization of a process brings on the minimization of the
operating costs, the energy consumption and the contaminant emissions, or to the
maximization of the yields and operation productivity.

1.4.3 Solution of Operating Problems

A process plant never operates under the design operating conditions, either
because the composition of the raw materials is different from the one considered
initially, because the environmental conditions are not contemplated in the design
or because the configuration of the plant or some equipment is modified as
consequence of the materials availability or costs. It is often used an overdesign
in engineering stages in order to have spare capacity for future plant modifications.
Process simulation permits to predict the effect of changes in the operating
conditions over other process variables and, in that way, to establish control points
more favorable, through dynamic simulation. Likewise, it simplifies the supervision
of the changing conditions in large periods of time; for example, the deterioration of
the random packing in a distillation column, the fouling in heat exchangers or the
catalyst deactivation.

1.4.4 Other Applications

The commercialization of turnkey processes demands the demonstration—by the

concerns—of the capabilities of the technologies that they are buying. Such dem-
onstrations are made using process simulations, since it is impossible to find pro-
totypes or pilot plants similar to the design to be commercialized.
The teaching of process design is an action that may be more enriching when a
simulator is used, as it allows to evaluate different alternatives and to solve various
cases of study without recurring to a large quantity of calculations that make the
task tedious and impractical. Additionally, the integration of creativity and the