KEMBAR78
Dynamics of Structures | PDF
Scribd
Upload
10K views654 pages

Dynamics of Structures

Uploaded by

givir
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
10K views654 pages

Dynamics of Structures

Uploaded by

givir
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
You are on page 1/ 654
Dynamics of Structures RAY W. CLOUGH Professor of Cell Eginering University of California, Berkeley JOSEPH PENZIEN Profesor of Struturl Engineering Unbersity of California, Bereiey “This book was set in Times New Roman, ‘Tae editors were B. J. Clade and Michael Gardner; the cover was dsigned by Pens Portfolio, Ines {be production soperizor was Chaces Hes ‘The drawings were dane by ANCO Technical Services. “Kingsport Pres, Ta. us printer and binder. TAGS 4 cé tle cg a5 : : Prmeics gr, etre age one. ten 270189 ‘Library of Congress Cataloging in Publication Data Clough, Ray Wate Dyoamis of structures Joclude indo. 1 Stuctural €yaamics. 1. Peazen, J, joint author. Tie Tassh.cs ee. 7420868 ISBN 0.07-0113920 DYNAMICS OF STRUCTURES Copyright © 1975 by MoGraw-HIl, Inc. AU rch reserved, Prisiod i the United State of America. No pat of this publication may be reproduced, ‘sored in eretoal sytem, or ansite, ia ay form or by any means, estronic, mechanisal, photocopying, recording, or otherwise, ‘without the prior mation permission of te publisher. 1819202122 KPKP 9987654321 16 CONTENTS 007 KOT ‘> 8ST MET. LBRARy Preface x List of Symbols xix Overview of Structural Dynamics 1 Fundamental Objective of Stuctural-Dynamics Analysis 1 ‘Types of Prescribed Loadings 2 Exseatial Characteristics of a Dysamie Problem 4 Methods of Diseretization 3 Lumped-Mass Procedure 5 Generalized Displacements 6 ‘The Finite-Flement Concept 7 Fonmulation ofthe Equations of Motion 9 Direct Equilibration Using @’Alemberts Principle 9 Principle of Virtual Displacements 10 Hamilton's Principle u Summary n Organization of the Text R 23 24 25 26 27 3 32 33 41 42 43 45 46 ‘SINGLE-DEGREE-OF-FREEDOM SYSTEMS, Formulation of the Equation of Motion ‘Components of the Basic Dynamic System Methods of Formalation Direct Equilibration ‘Vietual-Work Analysis Aoplication of Hamilton's Principle Infuence of Gravitational Foress Infuence of Support Excitation Gcaeralzed SDOF Systems: Rigid-Body Assemblage Generalized SDOF Systems: Distibated Flexiility Expressions for Generalized System Propertis, Free-Vibration Response Solution of the Equation of Motion, ‘Undamped Free Vibrations Damped Free Vibrations Csitical Damping ‘Underdamped Systems Overdamped Systems Response to Harmonic Loading ‘Undamped System ‘Complementary Solution Particular Solution General Solution Response Ratio Damped System Resonant Response ‘Acoslerometers and Displacement Meters Vibration Isolation Evaluation of Damping in SDOF Systems Free-Vibration Decay Resonant Amplification HalE-Power (Bandwidth) Method Energy Loss per Cycle (Resonance Testing) Hysteretie Damping 62 1 82 83 bt OT 92 93 Part Two 10 Tot 10-2 103 Response to Periodic Loadings Fourier Series Expression of the Loading Response to the Fourier Series Loading Exponential Form of Fourier Series Solution Response to Impulsive Loads General Nature of Impulsive Loads Sine-Wave Impulse ‘Rectangular Impulse “Triangalar Impulse Shock or Response Spectra Approximate Analysis of Impulsive-Load Response Response to General Dynamic Loading Duhamel Intepral for an Undamped System ‘Numerical Evaluation ofthe Duhamel Integral for an Undamped System Responte of Damped Systems Response Analysis through the Frequency Domain ‘Numerical Analysis in the Frequeney Domain Discrete Fourier Transforms Fast Fourier Transform Apalysis Analysis of Nonlinear Structural Response Analysis Procedure Incremental Equation of Equilibrium Step-by-Step Integration Summary of Procedure ‘Vibration Analysis by Rayleigh’s Method Basis of the Method Approximate Analysis of a General System ‘Seleotion of the Vibration Shape Improved Rayleigh Method -MULTL-DEGREE-OF-FREEDOM SYSTEMS, ‘Formulation of the MDOF Equations of Motion Selection of the Degrees of Freedom Dycamic-Equilibrium Condition ‘Axial-Force Bifects 80 at 3 7 88 31 92 100 100 102 105 108 43 a m4 118 us 9 m1 123 129 129 131 133, 137 145 14s M6 19 u Ul 2 m3 mt Is us 2 ara 2 23 D4 Bs 3 IB 132 133 Be ‘Evaluation of Structural-Property: Matrices ‘Hlastic Properties Flexibility ‘Stiffness Basic Structural Concepts Finite-Element Stifiness Mass Properties [Lumped-Mass Matrix Consistent-Mass Matrix Damping Properties External Loading Static Resultamts Consisteat Nodal Loads Geometric Stifizess Linear Approximation Consistent Geometric Stifness Choise of Property Formulation ‘Undamped Free Vibrations Analysis of Vibration Frequencies Analysis of Vibration Mode Shapes Flexibility Formulation of Vibration Analysis Influence of Axial Forces Free Vibrations Buckling Load ‘Buckling with Harmonic Excitation Orthogonality Conditions Basic Conditions Additional Relationships Normalizing Analysis of Dynamic Response Normal Coordinates ‘Unconpled Equations of Motion: Undamped ‘Uncoupled Equations of Motion: Damped Derivation of the Bquations Conditions for Damping Osthogonality Damping Coupling Summary of the Mode-Superposition Procedure 15h 151 151 182 153 155 160 160 161 165 165 166 166 167 167 168 in 176 175 1p 182 182 182 183 184 185 185 186 187 191 191 193 194 194 195 198 199 “4 ua Wad M3 4 165 M6 Wer Ma 1s 161 15-2 153 154 ISS 16 I61 16-2 163 16-4 Practical Vibration Analysis Preliminary Comments Stodola Method Fundaniental-Mode Analysis Proof of Convergence Analysis of Second Mode ‘Analysis of Third and Higher Modes Analysis of Highest Mode Buckling Analysis by Matrix Iteration Holzer Method Basio Procedure ‘Transfer-Matrix Procedure Holzer-Myklestad Method Reduction of Degrees of Freedom Background Discrete-Mass Allocations Rayleigh Method in Discrete-Coordinate Systems Rayleigh-Ritz Method Basie Concepts in Matrix Iteration Eigenproperty Expansion of the Dyoamio Matrix Iterative Solution of the Eigenproblema Iteration with Shifts Subspace Iteration Symmetric Form of Dynamic Matrix Diagonal Mass Matrix Consistent-Mass Matrix ‘Analysis of Unconstrained Structures Analysis of Nonlinear Systems Introduction Incremental Equilibrium Equations ‘Step-by-Step Integration: Linear-Acceleration Method ‘Unconditionally Stable Linear-Acceleration Method Peeformance of the Wilson 8 Method ‘Variational Formulation of the Equations of Motion Generalized Coordinates Lagrange’s Equations of Motion Derivation of the General Equations of Motion Constraints and Lagrange Multipicrs 208 213 21s 219 22 226 230 232 235 2s 236 237 239 243 283 245 m7 232 233 254 255 260 262 263 265 268 an an 213 219 Part Three cg Ind ID 73 Ir4 Irs 176 7 18 181 18:2 18:3 184 185 195 9 19-1 192 193 194 D2 203 205 205 2d 22 23 2nd 21S ‘DISTRIBUTED-PARAMETER SYSTEMS Partial Differential Equations of Motion Introduction ‘Beam Flecure: Elementary Case Beam Flexure: Including Axial-Foroe Effcots ‘Beam Flexure: Including Shear Deformation and Rotatory Inertia ‘Beam Flexure: Including Viscous Damping ‘Beam Flexure: Generalization Support Excitations Axial Deformations Analysis of Undamped Free Vibrations Beam Flexure: Elementary Case Beam Flexare:Incinding Axial-Force Effects Beam Flexure: Including Shear Deformation and Rotalory Inertia Beam Flexure: Orthogonality of Vibration Mode Shapes Free Vibrations in Axial Deformation ‘Orthogonality of Axial Vibration Modes Analysis of Dynamic Response Noma! Coordinates Uncoupled Flexural Equations of Motion: Undamped ‘Uncoupled Flexural Equations of Motion: Damped ‘Uncoupled Axial Equations of Motion: Undamped ‘The Dynamic Direct Stiffness Method Introduction Dynamic Flexural tiffiaess Matrix Dynamic Stifiaess for Flexare and Rigid Axial Displacements ‘Dynamic Axial-Deformation Stiffness Matsix ‘Combined Flexural- and Axial-Deformation Stifiness ‘Axial-Fores Effects on Transverse-Bending Stiffness Wave-Propagation Analysis Basic Axial-Wave-Propagation Equation Consideration of Boundary Conditions Discontinuity in Bar Properties Stress Waves Developed During Pile Driving Shear-Wave Propagation in Buildings 293 293 294 296 298 301 302 305 308 317 318 321 323 325 328. 328 331 336 338 345 345 353 357 358 360 364 364 369 am 317 382 Part Four 2d 222 223 224 25 226 RI 228 229 23 21 232 23 234 235 236 237 238 BO 28-10 23-11 BID 28-13 21 24-2 m3 24 RANDOM VIBRATIONS Probability Theory ‘Single Random Variable Important Averages of a Single Random Variable One-dimensional Random Walk ‘Two Random Variables Important Averages of Two Random Variables ‘Scatter Diagram and Correlation of Two Random Variables ‘Two-dimensional Random Walk ‘m Random Variables Linear Transformations of Normally Distributed Random Variables Random Processes Definition Stationary and Ergodic Processes ‘Autocorrelation Function for Stationary Processes Power Spectral Density Function for Stationary Proossses Relationship between Power Spectral Density and Autocorrela- tion Functions Power Spectral Density and Autocorrelation Functions for Derivatives of Processes Superposition of Stationary Processes Stationary Gaussian Processes: One Independent Variable Stationary Gaussian White Noise Probability Disteibution for Maxima Probability Distribution for Extreme-Values Noastationary Gaussian Processes Stationary Gaussian Process: Two or More Independent Variables Stochastic Response of Linear SDOF Systems ‘Transfer Functions Relationship between Unit-Impulse- and Complex-Frequeney- Response Functions ‘Relationship between Input and Output Autocorrelation | Fanctions [Relationship between Input and Output Power Spectral Density Fanctions 389 389 394 396 42 416 419 430 432 436 436 438 uy 451 454 455 437 4 416 416 482 485 gz 245 246 mr Bel 2.2 253 254 Part Five 26-1 26-2 EERE 27 272 3 74 5 26 Response Characteristics for Narrowband Systems ‘Nonstationary Mean Square Response Resulting from Zero Initial Conditions Fatigue Predictions for Narrowband Systems Stochastic Response of Linear MDOF Systems ‘Time-Domain Response for Linear Systems Frequeney-Domain Response for Linear Systems Response to Discrete Loadings Response to Distributed Loadings ANALYSIS OF STRUCTURAL RESPONSE TO EARTHQUAKES Seismological Background Introductory Note pear Elastic-Rebound Theory of Earthquakes, ‘Earthquake Waves Measures of Ground-Motion Characteristics Selection of Design Earthquakes Deterministic Analysis of Earthquake Response Earthquake Input Mechanisms Excitation by Rigid-Base Translation Lumped SDOF Systems Generalized SDOF Systems Lumped MDOF Systems Comparison with Uniform Building Code Requirements Distributed-Parameter Systems Excitation by Rigid-Base Rotation Maliple Support Excitation Infuence of Foundation Medium in Earthquake Response “Modeling of the Foundation Medium Soil Modification of Earthquake Motions Soi-Structure Interaction: Equations of Motion Soil-Structure Interaction: Response Analysis Nonlinear Response to Earthquakes Need for Noalinear Analysis ‘Method of Nonlinear Analysis 495 498. 508 510 su 313 sn 2k 25 331 282 283 Bt 285 ‘Typical InelastioRespoose Behavior 598 Influence of Strength Variations 599 Duetilty-Factor Method 602 ‘Nondeterministic Analysis of Earthquake Response on Stochastic Modeling of Strong Ground Motions eu Stationary White Noise ou Stationary Filtered White Noise 613 ‘Nonstationary Filtered White Noise 613 Analysis of Linear Systems 616 SDOF Systems 616 ‘MDOF Systems 67 Analysis of Nonlinear Systems 6i7 Extreme-Value Response of SDOF Systems 618 Stationary White Noise Excitation 618 Stationary Filtered White Noise Excitation 619 Extreme-Value Response of MDOF Systems 626 Index 29 PREFACE ‘This book is the result of more than 25 years spent in developing the program of instruction in structural dynamics at the University of California, Berkeley. Under standably, the material has undergone considerable revision during that time. Three different sets of notes ave been prepared ang distributed to the classes at widely separated intervals, and local versions of these have been used as class notes in the curricula presented at such diverse locations as Santiago, Chile, Trondheim, Norway, and Tokyo, Japan. ‘In the initial organization of this material, Professor Clough was strongly influenced by the lectures presented by Professor R. L. Bisplinghoff at the Massa- chusetts Institute of Technology, and he acknowledges his indebtedness to that superbly prepared course on the dynamics of airplane structures. Subsequent orientation of the text toward civil engineering problems in structural dynamics reflected the work of Hohenemwer and Prager in their pioneering treatise “Dynamik er Stabwerke.”! Similarly, Professor ‘Penzien acknowledges the great benefit he received from the lectures on random vibrations given by Professor S. H. Crandall at the Massachusetts Institute of Technology. The continuing development of this 1. Hobenemie and W. Pree, “Dyzamle do Stabwesia Juus Springer, Bers, Germany, 1933, subject matter, however, has ben essentially the work of both Clouzh and Peazien. Contributions to the literature by numerous writers have been incorporated appro- Dratly into the sequence of lectures; most such coatributions are so wel established in the field of structural dynamics that itis dficult to assign credit for them. Coo- sequently, few credit references are given, and the authors offer their apologies to those who may feel slighted. Although this text materi has undergone contiouous revision during its evelopment, the general orgunization bas remained unchanged. ‘The logical transi- tioa from steuotures with a single degree of freedom, to generalized single-depree systems, to the mode-superpostion analysis of multidegree discrte-coordisate structure bas provided a dimple route for the “statics trained structural engineer to follow in being introduced to the special problems arising from éynamic loadings. Moreover, it has always been considered essential to emphasize traasent dynamic response anclysis rather than to be concersed only with vibration analysis. As a prerequisite to the study of structural dynamics, i has proven elecive to require @ solid background of static structures theory, including matrix methods, and it is assumed that the readers ofthis text have bad such preparation: Probably the most obvious and farreaching change that has occurred during the development of this ext material has boon the acceptance of high-speed digital computers asa standard too! for structural analysis. Before computers were widely tied in structural ofces, major emphasis in structural dynamics was placed on efficient methods for slide-rule and desk-caleulator analyses. Such methods still have 4 prominent place inthis text because the avtbors ae convince of their value ia the study of the subject. Ifthe details of a hand-solution procedure are thoroughly ‘understood, itis not dificult to write or use a coresponding computer code, bat it ray be impossible to use a “black box” computer program effectively without ‘knowing the computational details. Nevertheless, itis zeognized that any significant, practical dynamic response analysis roqures so much numerical efort as to be ‘economically feasible only by computer. Therefore, the solution techniques emphas- ined herein are generally those that may be employed effectively with a bih-spoed computer, 28 well as by hod, The porpose ofthe presentation is to explain the basis fof che methods; coding technigues and efclnt computer usage are not discussed in detail. “The subject matter of this tent is the bass of a sequence of graduate-level quarteiength courses given atthe University of California; however, ros oft could be covered just as well a the advanced undergraduate (fourth-year) level. The basic course on dynamics of structures treats the material in Pat 1 and in Part 2 through the ist secions of Chap. 14. Itis expected that nearly all Masters-Degree studzats in structural engincsring should ave at lees this much contact with the dyeamics field, “The subsequent “advanced” stroctural dynamics course teats the reminder of Part2 smersce sil ‘and all of Part 3. In addition, much of the earthquake engineering material in ‘Chaps. 26 and 21 is presented in these first two courses to provide practical applica tions of the theory. The basic material of Pact 4is covered in asingle-quarter course om random vibrations, together with some of the earthquake-enginecring spplications from Chap. 28, Finally, the entire book serves 2s background and reference material for a course on structural design for dynamic loads; the prerequisite for this design ‘course isthe basic course on dynamics of structures. Although the frame of reference for most of this material is the feld of civil-engineering applications, the same basic techniques of structural dynamics are applicable in aerospace engineering, naval architecture, automotive engineering, and any field in which structural systems are subjected to dynamic loads, ‘A large number of examples have been incorporated into the text because the ‘authors have found that much of this material may be taught most effectively by that means, Moreover, many homework problems have been provided with most of the ‘chapters because it i essential for ths student to actually make use of these analytical techniques to master them fully. However, such problems must be assigned sparingly because dynamic-response analyses are notoriously time consuming. The authors have found that from one to four problems may constitute an adequate weekly assignment, depending on the subject matter and type of solution procedure required, Consequently, the book includes many more problems than can be assigned during a one-year sequence of courses on structural dynamics. RAY W. CLOUGH JOSEPH PENZIEN LIST OF SYMBOLS distance Fourier coeflicients, constants area, constant constants distance, integer Fourier coefiicients, constants constant damping coeficient generalized damping coefficient critical damping coefficient Fourier coefficients damping influence coefficients normal mode generalized damping coefficients dynamic magnification factor, plate stiffness dynamic matrix = k7'm constants axial displacement ‘Young's modulus ynamic matrix = D“* 4 Ska se Se So Is a % é GG, G & HO H(@), Ho) Hz r I 1 i expected value, ensemble average flexural stifness natural lic frequency flexibility influence coeficients inertia, damping, and spring forces, respectively acedleration of gravity general displacement coordinates, stress wave functions shear modulus constants plate thickness, story height unit impulse response fonction complex frequency response function Hertz (measure of frequency, eyles per second) integer Jimpulse, moment of inertia identity matrix integer spring coostants generalized spring constants efeativestifiness geometric stiffness stiffness influence coeficients combined stiffecssinuence coefcints eometicsifness influence coeficients zeneralzed stiffness of nth normal mode length earthquake excitation factor mass, integer sass infiuence coeficients uniform massunit length rotatory mass moment of inertia ‘generalized mass of nth normal mode generalized mass internal moment at @ section constant sual Joad, number of time increments, number of degrees of| freedom xtc axial load time varying axial force Pp Pe Pate Pe) PO,3) pel) PCy ap ++ Hq) PD PX), PY), Pr % a RO RQ) Re) load ‘uniform loading/unit length, generalized loading effective loading probability density function Joint probability density function ‘conditional probability density function ‘multivariate probability density fumetion forcing function of mth normal mode probability distribution functions probability ith generalized coordinate ith generalized forcing function radius of gyration response ratio. autocorrelation function eross-correlation function, constant power-spectral density function cross-spectral density fimetion spectral acceleration response spectral displacement response spectral pseudovelocity response response-spectrum intensity time impulse duration ‘transfor influence coefficients ppetiod of vibration, kinetic energy period of nth normal mode period of loading transmissibility ) 00 5 () G-z) Denoting the uniformly distributed mass percent length by 1% the uniform sifoss by El, and the uniformly distibated load per unit length by A), evaluate the generalized ‘hice properties mand e* andthe geseralized loading 2%). 20, fd gee) 20) 1 sions eas £1 Sena ity FIGURE P24 25 (@) Wa downward load Nis applied at the top of the column of Prob. 2-4, evaluat its ‘combined generalized stiffness &* using the seme shape function (3). ® Repest pact (a) assuming that the axial force in the coluran varies linearly along its Jength as N(@) = NO ~ 3/2). . 246 Assume thet the uniform elab of Fig. 2-7 is square, with side length a, and is simply supported on all four edges. (© Ii te mass per unit aea is y and its flexural rigidity is D, determine its generalized properties m* and A* in terms of the central displacement coordinate Z(t). Assume the displacement function is in WG) = sin (The uniformly distributed extemal loading per unit of area is p(t). Determine the generalized loading p*(t) besed on the displacement function of part (2). 27 The outer diameters, height, and material properties of a conical concrete smokestack are shown ia Fig. P2S. Assuming a uniform wall thickness of 8 in and thatthe deflected, shape is given by compute the generalized mass mi aad stifaess h" ofthe structure. Use Simpson's rule 40 Drewocs oF emRUCTURES to evaluate the integrals, inckuding ia the summations the integrand values for the [bottom, middle and top sections. For example +») me Z oo +4, ere 7, = my? evaluated at level “L" += aoe! “hie vouRe is 3 FREE-VIBRATION RESPONSE. 3:1 SOLUTION OF THE EQUATION OF MOTION Ht was shown in Chap. 2 that the equations of motion of any SDOF system can be reduced to the form. m2) + ctZ(0) + BOZO) = plo) This is entirely equivalent to the equation of motion of a simple spring-mass system with damping, as shown in Fig, 341, which may be written mile) + oft) + ke) = pt) Gt) ‘Thus in the present discussion it wil be convenient to use Eq, (31) and to visualize the response of this simple system, However, it should always be remembered that these results apply equally to the generalized-coordinate responso of any complex system which has been represented as a SDOF syster, ‘The solution of Eq. (3-1) will be obtained by considering first the homogencous ' ‘equation with the right side set equal to zero: mile) + cife) + ket) = 0 2) Motions taking place with the applied force set equal to zero are called free vibrations, and it is the freewibration response of the system which we now wish to examine, 42. macs or emveronss : : £: no eee FIGURE $41 Basie SDOF system, ‘The solution of Eq. (3-2) is ofthe form W=6r Ba Substituting this into Eq, (3-2) leads to (mt test GE =O Bs) After dividing by mGe* and introducing the notation G5) Eq, (3-4) becomes 6) ‘The value of s which can be derived from this expression depends on the value of e: ‘thus the type of motion represented by Eq. (3-3) will depend on the damping in the system. 3:2 UNDAMPED FREE VIBRATIONS Af the system is undamped, Le. if ¢ = 0, itis evident that the value of s given by Eq, CO) is =i 67 "Thus the response given by Ea, (3-2) is H@) = Gel + Ge 38) in which the two terms result from the two values of s and the constants G, and G3 represent the (as yet) arbitrary amplitudes of the motion. Equation (3-8) can be put in & more convenient form by introducing Euler's equations 8m coset t isnot — B) The result may be written *@) = Asinor + Boosor —G-10) in which the constants and B may be expressed in torms of the initial conditions, ie, the displacement 2(0) and velocity €(0) at time ¢ = 0, which initiated the free ‘amvmexnion niseoxse 43 FIGURE 32 ‘Vodsanped fre-vibration response vibrations of the system. It is casily seen that o(0) = B and &(0) = Ao; thus Ea, G0) becomes 2) 63. = Qeinor+ aera — G.t0 x = ©) em ‘This solution represents a simple harmonic motion (SHM) acd is portrayed graphically in Fig. 32. The quantity cis the ciculaefequency or angular velocity ofthe motion; it s measured in radians por unit of time. The eyeio frequency J shih is uovally refered to ierely asthe frequency ofthe mation, i given by 2 a f=z, @2) an its reciprocal is ale the period 7, G-13) ‘The motion represented by Eq, (211) also can be expressed in the form 2) = poos (ot 8) (B14) ‘as may be noted from the Argand diagram or vector representation of Fig. 33. The response is given by the real part, or horizontal projection, of the two rotating vectors. ‘Thus the amplitude of motion is given by the resultant tor=[S] ems and the phase ange by o=tatO Gg x0) twill be noted in Fig, 3-3 thatthe phase angle @ represents the angular distance by which the resultant motion lags behind the cosine tem in the response FIGURE 33 Romaing-vector representation of five vibrations. 33 DAMPED FREE VIBRATIONS If damping is present in the system, the solution of Ea, (3-6) which defines the response is ‘Thee types of motion are represented by this expression, according as the quantity under the square-root sign is postive, negative, or zero. It is convenient to discuss fist the limiting case, when the radical vanishes; this is called the critcaldamping condition. Critical Damping I the radical in Eq. (3-17) is sot equal to zero, itis evident that o/2m = a; thus the critical damping value cis (= 2m 18) Then the value of sin Eq, (3-17) is G19) and the response given by Eq. (3) is of) = (G, + Gye" G20) in which the second term is mltpled by r because only a single value of sis available in the solution, Eq. @-19) Introducing the initial conditions in Eq. (3-20) leads to the final form of the critically damped response equation a 6) = [ONL + 8) + HOYJer* G21) SRIEATERATON nEEON 45 “0 wo) FIGURE 34 Froo-vibration response with critical damping. hich is portrayed graphically in Fig. 34, Tt will be noted that the free response of a critically damped system does not include oscillation about the zero-delestion position; instead the displacement returns to zero in accordance with the exponential decay term of Eq. (3-21). One useful defiition of the critically damped condition is that itis the smallest amount of damping for sehich no oscillation ocevrs in the free response. ‘Underdamped Systems ~ the damping is les than critica, it ie evident from Eq, (318) that © < 2mo and thus thatthe radical in Eq. (3-17) anst be negative, To evaluate the free-vibration response in this case, is convenient to express the damping as a ratio ¢ tothe critical darmping vals thus fm Gm) imo C9 in Which £is called the damping ratio, Introducing Ea. (-22) nto Eq, (17) leds to s=-tos Gayo or changing the sign of the radial and introducing a aew symbol gives se -totim 623) where anavi-F Em ‘The quantity top is called the damped vibration frequency; for damping ratios to be ‘expected in typical structural systems (£ < 20 percent it differs very little from the ‘undamped frequency, as may be noted in Eq, (3-24). To estimate the influence of ‘damping on frequency, i is convenient to remember that a plot of the ratio of dazaped to undamped frequency ep/o vs. the damping ratio ¢ is a circle of unit radius, 28 shown in Fig. 3. ‘FIGURE 35 Relationship botweon damped frequency sod damping ratio. ‘The freo-vibration response of an underdamped system can be evaluated by substituting Eg, (3-23) into Eq. (3); ths Hi) = Gyertetlent 4 Gyertetntons = gte(G, cont 4. Gaeta ‘The term in parentheses represents simple harmonic motion [compare with Eq. (3-8)]; thus this expression ean be writen more conveniently as XO = eM Sin opt + Beosost) — 6-29 Finally, when the initial conditions v(0) and (0) are introduced, the constants of Eq, (3-25) can be evaluated, giving fs ) + wO}o ae % ‘Aterativey, this eipons expression can be writen in rotating-vector form H) = per 608 (oot = 8) 27) 9 = fA eHOHET 5 cxowy} a ann? XO) + WOR @px(0) A plot ofthe response of an underdamped system to an intial displacement o(0) ‘but starting with 2ero velocity [H(0) = O] (in other words releasing the mass from a stationary displaced position) is shown in Fig. 3-6. Itis of interest 1o note thet the underdamped system oscillates about the neutral postion, with a constant circular frequency és, The rotating-vector represeatation of Ea. (3-27) is equivalent to Fi 3.3 except that the length of the vector diminishes exponentially as the response damps out, The true damping characteristics of typical structural systems are very complex and difficult to deine. However, its common practice to express the damping of such sia ot + enon] 626) in which, RSE-VURATION ESPON: 47 FIGURE 36 Free-vibration response of underdampad system, ‘eal systems in terms of equivalent viscous-damping ratios £ which show similar Gecay ries under freesibraoa conditions. Therefore, let us now reate more fly the viscous-damping ratio ¢ to the free-vibration response shown in Fig. 3-6. Consider any two successive positive peaks shown in Figure 3-6, that is, v, and tunx- From Eq, G-7?), te ratio ofthese to suoesive values ie given by a (= 2) 29) “Ting the natural logarithm (n) of both sides of Eq, (9-29) gives the loadin decrement 8 seta 39 on, wits Eq, 028), ean For low damping, Es. (31) caa be approncnated by e where the symbol s represents “epproximatsly equal” 1 soeh cses, Eq, (225) can be writen a6 a seis expansion 3 eta e i By. Gay For low values of { sufficient accuracy can be obtained by retaining only the first ‘wo terms in the series, in which case Ss 3.34) FIGURE 37 Damping-atio correction factor {to be ‘applied to resale obtained from Eq, ape00) cor oso To illustrate the accuracy of Eq. (334), the ratio of the exact valve of € as given by Fg, (3-30) tothe approximate value as given by Eq. (3-34)is potted vs, the approximate value in Fig. +7. This graph permits one to correct the damping ratio obtained by the approximate method. For lightly damped systems, greater accuracy can be obtaived in evaluating the damping ratio by considering response peaks which are several cyeles apart, say m ceyeles; thea rring 2.35) hich can be simplified for very low damping to the approximate relation Inn OO When one is observing damped free vibrations experimentally, a convenient method for estimating the damping ratio isto count the number of cycles required to give a 50 percent reduction in amplitude. The relationship to be used in this case is presented graphically in Fig. 8. As a quick rule of thumb itis convenient to remem- ber that for 10 percent critical damping, the amplitude is reduced by 50 perceat in 1 oyele, Overdamped Systems ‘Although structure ystems having areater than critical damping are aot encountered fn normal conditions, it is sefel to carry out te response analysis of an overdamped system to make this discussion complete. Th this case £ > 1, and Eq, (+17) may be written fa-feteVB=1=-t4d 63) in which asm-vRAmON rasPONs? 49 ' ag OO se [\T tT tf $a 3 4 FIGURE 35 a” i Daring row, number of eee rx quel ty redwe amplieds by 50 OOF OIF -EIs Bao fone {ance e Substituting Eq. (3-37) into Eq, (3-3) and simplifying leads eventually to 6) 01d sinh @t + Boosh Ot) 3-38) jn which the constants A and B could be evaluated by consideration of the initial conditions. It will be noted from the form of Eq. (3-38) that the response of an over damped system is not oscillatory; it is similar to the motion of the critically damped system of Fig. 3-3, but the recur toward the neutral position is slowed as the damping ratio is increased. EXAMPLE E31 A one-story building is idealized as a rigid girder supported by weightless columns, as shown in Fig. E31. In order to evaluate the dynamic properties ofthis structure, afcce-vibration test is made, ia which the roof system (rigid girder) is displaced laterally by a hydraulic jack and then released. During ‘the jacking operation, tis observed that 2 force of 20 kips is required to displace the girder 0.20 in. After the instantaneous release of this initial displacement, ‘the maximum displacement on the retum swing is only 0.16 ia and the period of this displacement cycle is T = 1.40 s. From these data, the following dynamic behavior properties are to be determined: 1 Effective weight ofthe girder: Qn [Ww 28 og [H 723°" G 4, e 0 20 a = 1920 Kips i w @& ig = 0.0496 29 386 = 1920 kip 40s 50 DyNAncs oF smRUCTURES wage Wome opp . [Loa sta ‘| FIGURE EM Fe Vibration test ofa simple boling. 2. Frequency of vibration: 1 A noms 140 O74 tie r @ = 2sf = 448 rad/s 3 Damping properties: Logarithmic decrement: Damping ratio: Danpiagonfceats —€= fr = Be = 00355222) aap = 1,584 kips/in-s Damped frequency: = oY 1 ~ F = 00.999)" 2 @ 4. Asap after tx cytes : = @) 6 = (9020) = PROBLEMS Sl For the uniform caatilever tower of Example 2, the following expressions for the ‘generalized ass and stiffness were determined: pt = 02280 roeaeee Be ‘Based on these expressions, compute the period of vibration for a concrete tower 200 ft 32 33 3S risrymaron nssosee SL ‘high, with an outside diameter of 12 and wall thickness of 8 in, for which the following properties may be assumed: = 0B se = EL = 165 x 10" Test? Assuming that the tower of Prob. 3-1 supports an additional point weight of 400 kipe at the top, desermine the period of vioration (neglecting the geometric etifiaess effect). ‘The weight 1 of the building of Fig. E3-1 is 200 kips and the building is set into free vibration by releasing it (at time « = 0) from a displacement of 120 in, If the maximum displacement on the retum swing is 0.86 in at time ¢ = 0.64, determine: (@) the lateral spring stilfzess ©) the damping ratio & (© the damping coeticieat © Assume that the mass and stifaess of the structure of Fig. 3-1 are as follows: m 2k-sFjin, km AOkjin. Ifthe system is set into foe vibration sith the inital conltions 2(0) = 0.7 in and (0) = 5.6 in(s, determine the displacement and velorty at t= 1.05, assuming: (@) © = 0 (undamped system) @c= 28es/in Assume that the mast and stifaess of the system of Fig, 3-1 are m = Sk-s%/in and ‘= 20¢/in, and that it is undamped. Ifthe inital displacement is 0) — 1.8 in, and the displacement at ¢ = 1.2 ie also 1.8 ia, determine: (@) the displacement at = 2.48 (@ the amplinade offre vibration p ODTU KOTUPHANES! M.6.T.U. LIBRARY 4 RESPONSE TO HARMONIC LOADING 4-1 UNDAMPED SYSTEM Complementary Solution 1 will now be assumed that the system of Fig. 31 is subjected to a harmonically varying load p(?) of amplitude pp and circular frequency &. In this case the differential equation of motion becomes mi(e) + ci(t) + ke(@) = po sin (41) Before considering the general case with damping, itis instructive to examine the response of an undamped system to harmonic loading, for which the equation of motion becomes mife) + ket) = po sinar (4-2) ‘The complementary solution of this equation is the free-vibration response of Eq, Gao, aff) =Asinot + Boosor (4.3) RESPONSE To MAMMON LOADING 53 Particular Solution ‘The general solution includes also the particular solution, i.., the specific behavior ‘generated by the form of the dynamic loading. The response to the harmonic loading ‘can be assumed to be harmonic and in phase with the loading; thus ef = Gsinae — 44) in which the amplitude G isto be evaluated. Substituting Eq. (4) into By. (42) leads to mi sin Ge + KG sin 1 = pysinw (48) Dividing through by sin Gt (which is nonzero in general) and by k and noting that ‘jm =o results (after some rearrangement) in ‘The amplitade of the response therefore becomes a in which f represents the ratio of the applied load frequency to the natural free- vibration frequency; Le, as) ele General Solution ‘The general solution to the harmonic excitation of the undamped system is then siven by the combination of the complementary solution and the particular solution, in which the value of G is given by Bq. (47); thus =, 0,(0) = Asin of + Beos ot + 20+, 6) = oft) + oft) temo Bote In this equation, the values of 4 and B depend still on the conditions with which the response Was initiated. For the system starting fcom res, ie, for “at rest” initial conditions (0) = (0) = 0, ican easiy-be shown that the constani take the follow ing values sinwt (49) a-- te ano em ‘Then the response given by Eq, (4-9) becomes = Pe On TIF Gino Bsino) 11) ‘54 prunes oF sraveTERES where po/k = 2 = static displacement, ie., displacement which would be produced by the load p applied statically 1/0 — 6?) = magnification factor (ME), representing dynamic amplification effect of harmonically applied load sin Gt = response component at frequency of the applied load = steady- state response, directly elated to the load B sin cot = response component at natural vibration frequency = free vibration effect induced by the initial conditions Since in a practical case, damping will cause the last term to vanish eventually, itis termed the transient response. (For this hypothetical, undamped system, of course, this term would not vanish but would continue indefinitely.) Response Ratio ‘A convenient measure of the influence of the dynamic character of the loading is provided by the ratio R(t) of the dyzamic response to the displacement that Would bbe produced by the static application of the same load: RO ay % Polk From Eq, (+11) itis evident that the response ratio resulting from harmonic loading of an undamped system (starting from res) is R= gles pine — G19 It is informative to examine this response behavior in more detal by reference 10 Fig, 41, Figure 41a represents the steady-state component of the response. Figure 416 is the so-called transient response, the free-vibration motion initiated by the copulitions atthe start of the response. ‘In this example it is assumed that f= */, thet is, thatthe applied load frequency is two-thirds ofthe free-vibration frequency. “The total tesponse RI), the sum ofthese two tex, i shown in Fig. 4-le. Two points are of interest: (1) the tendeney for the two components to get in phase and then out ‘again, which causes a Beating effect in the response, and (2) the zaro slope at time 1 = 0, showing that the velocity of the transient-esponse term is just sufficient to cancel the steady-state velocity and thus to satisfy the speciied initial condition. 42 DAMPED SYSTEM Returning to the equation of motion including damping, Eq. (41), dividing by m, and noting that cfm = 20 leads to 8) + 2eoHe) + ool @1) RESIONSE TO HARMONIC OADNNG 55 ® oS Frequent, FIGURE 441 Response to harmonic load from atest initial conditions: (a) steady.stae ay () wansient; (9 total RC). ‘The complementary solution of this equation is the damped free-vioration response sven by Eq, (3-25) (assuming that the structure is less than critically damped, as is ‘tho cose for all practical structures): alt) = eM(A sin gt + Boos oof) (HIS) ‘The pasticular solution to this harmonic loading is of the form 20) = G, sin Bt + Geos at" (416) jn which the second term is required because, in general, the response of a damped system is not in phase with the loading. Ree "Substituting Eq, (+16) into Eq. (4-14) and separating the multiples of sin at from the multiples of cos Gt lends to [-G,57 — 6,0Q5o) + G,u7] sin Bt = PO sin awe (4-170) [-6,0? + 6,a(2§a) + 6,0] cos Gt = 0 (4178) 56 prauacs oF smueToRES FIGURE 42 ‘Steady-state displacement response ‘These two relationships must be sutiied individually because the sine and cosine terms vanish at diferent times, Dividing both by 0%, reprouping terms, and canceling ‘out the trigonometric expressions yields ~ 6) - Gat = 60-6) G2) =2B Gig G1 ~ 6°) + 6,26) = 0 Solving these equations simultaneously then results in expiessions for the response- amplitade factors: a= BY + @epP ~ Gee ak — BY + OP Introducing these expressions foto the parculr solution [Eq 4-16] and combining with the complementary solution finally yields the general slution a(t) = eM(A sin @gt + B cos @pt) Poy — py sin oe — 288 cos wr 2 + Parr al ~ Pana it] 4-20) ‘The first term in Eq, (4-20) represents the transient response to the applied loading, The constants A and B could be evaluated for any given initial conditions, ‘ut this term damps out quickly and generally is of ite interes; therefore its evalua- tion will not be pursued here, The second term in Eq. (#20) is the steady-state response, at the frequency of the applied loading but out of phase with it. This steadyestate displacement behavior can be interpreted most easly by plotting its two vectors inthe Argand diagram showa in Fig. 42. The resultant p ofthe two vectors RESPONSE TO HARVONIC 10ADING_ ST a} FIGURE 43 siteas Force equim in steadyatate 5 ty Ayre represents the amplitude of the steady-state response e= Bla ~pyr sary 2p ‘and the phase angle 6 by which the response lags behind the applied load is given by 2p es tant a2) e here it s understood tat the phase angle is limited tothe range 0 < @ < 180" "Tus the steadystat response can also be writen ot) ‘sin @t — 6) 423) “Te rato of the reskan response amplitude tothe ae dplacement which would be produced by the force will be called the dynamic magnification factor Ds thus Paps epoye — a2 Pen Mt ~ PF + O80T 28) It abo is of interest to consider the balaace of forces acting on the mass inthis steady-state vibration condition. ‘The force components are conveniently expressed in tes of the dynamic magnifcaton factor and plotted in an Argand diagram ig, 3), Note that the elastic force sets in the dicetion opponite to the resultant Asplacement vector of Fig. 42 Similarly, the damping and inertia foees actin directions opposing the velocity end acceleration vectors, respectively. Finally, iti evident thatthe resultant ofthese reiting forces exactly balances the applied Toad ‘Do 8 it must to maintain dynamic equilibrium, FIGURE 44 ° Y 7 3 Vasinton of dynanse magnifeaton 6 factor with damping nd frequency Te may be soon from Eq, (4-24) that the dynamic magnification factor D varies ‘ith the frequency ratio fi and the damping ratio ¢; plots of these relationships are showa in Fig. 44, The phase angle @ also varies with these same quantitics, a is evident in Eq. (4-22) and as shown in the plots of Fig. 4-5 EXAMPLE E41 A portable harmonicoading machine provides an effective means for evaluating the dynamic properties of structures in the fied. By operating the machine at two different frequencies and measuring the resulting structural-respoase amplitude and phase relationship in each cate, itis possible to determine the mass, damping, and sifioes of a SDOF structure. In 2 test of ths type on a single-story building, the shaking machine was operated at feoquencies of , = 16 radjs and , = 25 rad[s, with a force amplitude of 500 bin each case. The response amplitudes and phase relationships measured Jn the bvo cases were Ppa 72% 10%in = 15" 608, = P= 145 x 10%in, 0, = 55” cos Bs 1966 sin 8, = 0.259 1574 sin 8, = 0.819 To evaluate the dynamic properties from these data, rewrite Eq, (421) as is convenient to cos 6 @ t a + [EBL ar i ‘where the trigonometric function kas been derived from Ea. ($22), With farther algebra simplifeation this becomes Ka By 0.605 8 ° ESPONSE TO RARMONE LANG 59 Poa anle? ° r T 3 Frequeney ro. 8 FIGURE 45 ‘Variton of phate angle wth damping and frequency, ‘Then introducing the two sets of test data leads to the matrix equation 0.966 1-16] Te Tax = 50015 (ll) os 43 x 10 which can be solved to give = 100 x 10° Ibfin m= 128.5 tb s*fin that i, W = mg = 49.6 x 10" Ib ‘From these the natural frequency is o- 27.9 rads To determine the damping cociicieat, two expressions for cos @ can be derived from Eqs. (@) and (4-22). Equating these and solving for the damping ratio Teads to ‘Thus with the data of the fist test ‘and the same result (within slide-rule accuracy) is given by the data of the second ‘test. The damping ratio therefore is e_ _ 12se79) pated Tijo ~ 200 x 10° ~ 7% ut 60 orwoncs oF STRUCTURES 43 RESONANT RESPONSE, ‘From Fig. 44it may be noted that the peak steady-state response occurs ata frequency ratio near unity for lightly damped systems. The condition when the frequency ratio is unity, ie., when the frequency of the applied load equals the natural vibration frequency, is called resonance. From Bq, 4-13) itis apparent that the steady-state response of an undamped system tends toward infinity at resonance, A more general result may be obtained from Eq. (424), which shows that for resonance (f = 1) the dynamic magnification factor is inversely proportional to the damping ratio: 1 Dyes = 5 425) “However, slough itis close to the maximum, this does not represent the maximum response for any damped system; the feequency ratio for maximum response may be found by difereatiating Eq. (424), with respect to P and equating to zero, For practical structures having damping ratios ¢ < 1/-/2, the peak-response frequency is found to be Boa = N12 (4-26a) and the corresponding peak response is 1 Dass = ——2—= (4-268) mice For reasonable amounts of damping, the difference between Eq. (4-268) and the simpler Eq. (425) is negligible. Fora more compte understanding of the nature of the resonant response of a structure to harmonic loading, it is mecessary to consider the general esponse equation (4-20), which includes the transient terms as well as the steady-state term. At the resonant exciting frequency (B = 1) this equation becomes () = eA sin wot + Boos gt) — 22 8-2 Ht) = CHA sia ost + Boos ost) — BSA 427) ‘Asraming that the system starts from rest [1(0) = #(0) =O], the constants are os KE (428) ‘Thus Eq, (4-27) becomes Bel ee won neem For the amounts of damping to be expected in a structural system, the sine term in this equation will contribute little to the response amplitude; moreover, the damped sin opt + 00s nt) = eos « 29) Uesamged ‘ost amped FIGURE 46 ‘Response to resonant loading f= 1 for atest intl conor, frequency is neacly equal tothe undamped frequency. Thus the response rato in this ase approximately wt Polk 2 For zero damping, Eq. (425) becomes indeterminate, but when L’Hfospita’s srule is applied, the resonant response of an undamped system is, RQ) = (]— Yeoser 30) RG) = fin ot = or cosa) — 31) Plots of Eqs. (4-30) and (4-31), shown in Fig. 4-6, show how the response builds up in cases of resonant excitation, with and without damping; in both cases, it is clear that the respoase builds up gradually. In the undamped system, the response continues to grow by the amount x for each cycle; thus it will eventually produce distress in the system unless the frequency is changed. On the other hand, the manner in which the damping limits the resonant-response amplitude is clear in the lower sketch. The number of cycles required for this damped resonant response to reach 6 8 a Dass of ein FIGURE 47 Rate of buildup of reeonantresponte from rer. essentially its peak amplitude depends on the amount of damping. Curves showing the buildup of the response envelope (dashed lines in Fig. 46), plotted for several values of damping as fimetions of the nomber of response eycles, are given in Fig. 47 Note how few cycles of excitation are required to reach nearly the full response amplitude 44 ACCELEROMETERS AND DISPLACEMENT METERS At this point it is convenient to discuss the fundamental principles on which the operation of an important class of dynamic measurement devices is based, These are seismic instruments, which consist essentially of a damped oscillator as shown ia Fig. 48. The system is mounted in a housing which may be attached to the surface ‘where the movement is to be measured; the response is measured in terms of the ‘motion o(f) of the mass relative to the housing. ‘The equation of motion of this system is given by mi + ch + ho = — m0) peel?) (2-21) If the base on which the instrument is mounted is moving harmonically with an acceleration amplitude 3(f) = Oe sin Gt, the efective loading of the mass is Pan?) = mij, sin Gt. ‘The dyoamic steady-state response of this system then has the RESPONSE TO MAMONTC LONDON 6B ute retire apiece [us scree FIGURE $8 Schematic diagram ofa typical sesmometer, amplitude given by Eq. (421), ee) in which Dis given by Eq. (+24) and is presented graphically in Fig, 44, Examination of this figure shows that fora damping ratio of € = 0.7 the value of Dis nearly constant forthe frequency range 0 < < 0.6. Thusit is clear from Eq, (4-32) that the response indicated by this instrument will be directly proportional to the base-acceleration amplitude for applied frequencies up to about six-tenths the natural frequency of the instrument. Hence this type of instrument when properly damped will serve effectively as an accelerometer for relatively low frequencies; its range of applicability will be increased by increasing its natural frequency relative to the exciting frequency, ie., by increasing the stiffness of the spring. This isthe basio principle of seismie acoelero- meters Consider now the response of this same instrument to @ harmonic base dis- placement 1, = 249 sin Gi. In this case the corresponding acceleration is ~@7ay9 sin Gr, and the effective loading is puy(t) = miB*eyp sin we. By Eq, (4-21) the response amplitude then is - me = %6'D (4-33) ‘A plot of the response function f*D is presented in Fig. 49, Tn this case tis evident ‘that 6D is essentially constant at frequency ratios § > I fora damping ratio ¢ = 0.5. ‘Thus the response of the properly damped instrument is esentially proportional 10 the base-displacement amplitude for high-frequency base motions; ie, i will exve asa displacement meter in measuring such motions. Tis range of applicability for this purpose willbe increased by reducing the nstural frequency, i, by reducing the spring stffaess or increasing the mass. | | Z Respons amptude a= 92D Frequency to? FIGURE 49 Response of seimometer to harmonic base displacement 45 VIBRATION ISOLATION Although the subject of vibration isolation is too broad to be discusted extensively here, itis important to note that the basic principle is closely related to that of the seismic instrument. Two different classes of problem may be identified in which vibration isolation may be necessary: (1) operating equipment may generate oscillatory forces which could produce harmful vibrations in the supporting structure, or (2) seasitive instruments may be supported by a structure which is vibrating appreciably. ‘The first situation is illustrated in Fig. 4-10. A rotating machine produces an oscillatory vertical force pp sin ®t due to unbalance in its rotating parts. If the ‘machine is mounted on a SDOF spring-damper support system, as shown, its steady state displacement response is given by Ht) ~ BD sin@—o (434) where D is defined by Eq. (4-24). Thus the force exerted against the base by the spring sopporsf Ss = Kult) = poD sin (Gt — 6) At the same tine, the velociy ofthe motion relative to the base is prea 1) = 22 Da cos (ar ~ 8 40) = 22 D0 cos (at ~ 6) ‘hich ead to a damping fore en the base fo = et) = $2208 cos ot — 6) = 2ApoD cos (Wt — 6) RESON TO RARNONIC LONDEN. 65 AO py sniit FIGURE 410 ag SDOF vibrationitoltion sytem (p- 1 ed loading). Is ‘Since this force is 90° out of phase with the spring force, i is evident thatthe amplitude of the base force fis Saas = nas + FB. naa)*!* = poD[t + EAR}? ‘The ratio of the maximum base force to the applied-force amplinude, which i known 4s the transmissibility (TR) of the support system, thus is given by TR . DvI+ OF 433) A plot ofthe transmissibility as a function of the frequency ratio and damping ratio is shown in Fig. 411. Although itis similar to Fig. 4-4, all curves passthrough the same point at a frequency ratio of f = V2. This difference from Fig. 44 is due, of cours, to the influence of the damping force, Because ofthis characteristic, itis o 1 Frequengmtio,2 FIGURE 411 ‘Vibration tracsissibilty rato (applied load or dsplacement). 66 pevivacs oF smvcroms a, GURE2 _[romiesesr Hortons ase Gor pos etn. evident that damping tends to reduce the effectiveness of a vibration isolation system for frequencies greater than this critical ratio. ‘The second type of situation in which vibration isolation is important is ius trated in Fig. 4-12. The mass to be isolated is supported by aspring-damper system con foundation lab which is subjected to harmonic vertical motions. Te displace- rent of the mass rlatice o the base then is given by [see Eq. (#-33)] 2(0) = tyef*D sin (Gt ~ 8) 14-36) However, when the motion of the base is added vectorial it can be shown that the total motion of the mass is given by v{@) = ov + GPF D sin Ge ~B) 437) in which the phase angle 8 is of no particular interest. Thus, if the transmissibility in this situation is defined as the ratio of the amplitude of motion of the mass to the ‘tase-motion amplitude, it can be seen that the expression for transmissibility is the same as that given by Eq. (4-35). This can be expressed mathematically as mat =D/iF OF (438) Peo and Fig. 411 serves to define the effectvenes of vibration isolation systems for both tasie SDOF iolstion stations f For the design of a vibratonislaton system, itis convenient to express the behavior ofthe sytem in terms of is iolaton effectiveness ratber then the tens nissiiliy, where the eflctivenss fs Gefned as 1 — TR. Also, when i noted in Fig 4-11 that an ication sytem is effective oaly for frequency ratios > 2 and that damping is undesrebc in this range, itis evident tbat the isolation mounting should have very litle damping, Thus itis acoptable to use the transmisbity expression for zero damping Ba-2 1-TR= Pai (39) RISPONSE TO BARONE LOADING. 67 Input eens ZH i 8s Aig OS OD OTT OS OA OS OT OSS Ts Ste defection dein FIGURE 413, ‘Vibration izlation design cart. in which it is understood that f > (2. Finally it may be noted that f* = */o* = Blk) = BWW a) = GUA, where g is the acceleration of gravity and A, — Wikis the defection that tae weight of the system tobe isolated will produce onthe vibration-mounting devices. Thus itis evident thatthe effectiveness of the mounting system can be expressed ia terms of the frequency of the input motion & and this static-deffection value A,,. Solving Eq. (4-39) for the frequency ratio in terms of the isolation effectiveness leads to §* = [2 ~ (1 — TR)J[I — (A — TR)]. This can now be expressed in terms of the input frequency (J = 2nt) and the static defection F=33 Vet =(.— TR) rh vere Fis in Het ad Ay is in inces, A plot of a, (M40) presented in Fg 413. ‘Knowing the frequency of the impresedexcation, one ean determine diel from this graph the supportpad defection ,, required to achieve any desired level of ‘iertonclaon, asain hat the lator ave ile damping also is apparent in is raph tate lation stem yl haves dsleteous eet if i oo sf EXAMPLE E42 Deflections sometimes develop in concrete bridge girders due to creep, and if the bridge consists of a long series of identical spans, these deformations will cause a harmonic excitation in a vehicle traveling over the FIGURE Bet [ealized vebicle traveling over an uneven bridge dock. bridge at constant speed. Of course, the springs and shock absorbers of the car are intended to provide a vibration-isolation system which will limit the vertical motions transmitted from the road to the occupants. Figure Eé-1 shows a highly idealized model ofthis type of system, in which the vehicle weight is 4,000 Tb and its spring stifiess is defined by a test which showed that adding 100 Ib caused a deflection of 0.08 in, The bridge profile is represented by a sine curve having a wavelength (girder span) of 40 ft and a (Gingle) amplitude of 12 in, From these data itis desired to predict the steady- state vertical motions in the car when itis traveling at a speed of 45 mph, assuming thet the damnping is 40 percent of exitical. ‘The transmissibility for this case is given by Eq. (4-38); hence the amplitude of vertical motion is [ 1+ ey |" La BY + Gee ‘When the car is traveling at AS mph = 66 fy, the excitation period is tt 1, = BX ~ 0.6065 66 ft/s i tetra peso ea wei i 2 fw eX am Zeom 2 an [Rose age 9 = 17, = OS790606 = O94, and with {= 04 the sepone since ae = 120.642) = 1.97 in It also is of interest to note that if there were no damping in the vehicle (& = 0), the amplitude would be 12 thay = 290 = 2 = 109 in “TF Ont MUSHONGE TO EARIONE LOADING 6B ‘This is beyond the spring range, of course, and thus has litle meaning, but it does demonstrate the important function of the shock absorbers in limiting the ‘motions resulting from waviness of the road surface. Mi EXAMPLE B43 A reciprocating machine weighing 20,000 Ib is known to develop vertically oriented harmonic forces having an amplitude of 500 Ib at its operating speed of 40 Hz, In order to limit the vibrations excited in the building in which this machine is to be installed, it is to be supported by a spring at each corner ofits rectangular base, The designer wants to know what support spring stifiness will be required to limit to 80 Ib the total harmonic {force transmitted from the machine to the building. “The transmissibility i this case is 80/500 = 0.165 lence by Ea. (4-39) aR 7 P= 625 from which =125 = 2% kg ‘Solving for the total spring stiffness gives ow k= OY 6 451 x 10% tyfin 7259 “Thos the stiffness of each ofthe four support springs is Hs = if, = 113 kipsin It is of interest to note thatthe static deflection caused by the weight of the rmachine on these spring supports is ae = fess = 0.0446 in Hh 46 EVALUATION OF DAMPING IN SDOF SYSTEMS Ii the foregoing discussion of response analysis in SDOF structures, it has been assumed that the physical properties of the system (mass stifizess, and damping) are known, In most cases, the structural mass and stiffness can be evaluated rather easily, either by simple physical considerations or by generalized expressions such as Eqs. (2-37) and (2-39). On the other hand, the basic energy-loss mechanisms in practical structures are seldom fully understood; consequently it usually is not feasible to determine the damping coefficient by means of the corresponding generalized damping ‘expression. For this reason, the damping in most structural systems must be evaluated isectly by experimental methods. A brief survey of the principal procedures for evaluating darnping from experimental measurements follows. 70 pyswacs oF srRocTURES Free-Vibration Decay Probably the simplest and most frequently used experimental method is measure- reat of the decay of free vibrations, as mentioned in Chap. 3. When a system has been set into free vibration by any means, the damping ratio can be determined from the ratio of two displacement amplitudes measured at an interval of m cycles. Thus isis the amplitude of vibration at any time and 9, isthe amplitude m eyels later, the damping ratio is given by é fn (at) ~ Fran(alag) ~ 2m where 5,, = In (2y/tyeq) Fepresents the logarithmic decrement and and op are the undamped and damped frequencies, respectively. In most practical structures, the damping ratio is less than 0.2, so that the approximate form of Eq, (4-4), based on neglecting the change of frequency due to damping, is suficiently accurate (the error in €is less than 2 perceat). A major advantage of this free-vibration method is that equipment and instrumentation requirements are minimal; the vibrations can be initiated by aay convenient method, and only the relative displacement amplitudes need be measured. Resonant Amplification “The other priocipel techniques for evaluating damping are based on observations of steady-state harmonic response behavior and thos require a means of applying bar- monic excitations to the structure at prescribed frequencies and amplitudes. With such equipment the fregueneyzesponse curve for the structure can be constructed by applying a harmonic load pg sin Gat a closely spaced sequence of frequencies ‘which span the resonance frequency and plotting the resulting displacement amplimdes a5 a function ofthe applied frequencies, A typical frequency-esponse curve for & moderately damped structure i shown in Fig. 414. ‘The dynamic magnification factor for any given frequency isthe ratio of the response amplitude at that frequency to the zerosfrequency (stati) response. Tt was shown earlier, by Eq. (4-25), that the damping ratio is closely related to the dynamic ragnieation factor at resonance, When the statioresponse and the resonant- response amplitude are denoted by fy and pus, respectively, the damping rato is elven by tfo ay 2 bres In practice, however, itis dificult to apply the exact retonence Frequency, and it is tore convenient to determine the maximum response amplitude Pau Which occurs a | 3 at t | f g i g z Br By OE i Lu | al Li i | | | ¢ ae 7 Fron nioure 414 ata slightly lower frequency. In this cas, itis evident from Eq. (4-266) that the damp- ing ratio may be evaluated as follows: paleo 2 1 po "2 peu Op 2 Paar (443) ‘The error involved in the simpler expression again results from neglecting the dif- {ference between the damped and undsmped frequencies and is unimportant in ordinary structures. ‘This method of daraping analysis requires only simple instrumentation, capable of measuring relative displacement amplitudes; however, the evaluation of the static displacement may present a problem because many types of loading systems cannot ‘be operated at zero frequency. 72 mawancs oF sreverums Half-Power (Bandwidth) Method It is evident from the general harmonic-respoase expression [Eq. (4-21)] that the shape ofthe entire frequeney-response wave is controlled by the amount of damping in the system; therefore, its possible to derive the damping ratio from many different ‘properties of the curve. One of the most convenient of these is the bandwidth, or half-power, method, in which the damping ratio is determined from the frequencies At which the response is reduced to (1/~/2)ppux, that is, atthe froquencies for which the power input is half the input at resonance, ‘The values of these hal-power frequencies can be determined by setting the response amplitude in Eq, (4-21) equal to 1/-/2 times the resonant amplitude derived from Bq, (4-42), that i, a8--lonol i- _,, ae Ot Ge Por? swiee from which (neglecting € in the square-root term) the two halftpower frequencies are or, squaring both sides, Solving for the frequency ratio then gives Poz1-%—-2 Bat-g-e Prat +2%- 2 felei—e Hence, the damping ratio is given by half the difference between these half power frequencies o= 1B: — Bi) 4A) ‘This method of evaluating the damping ratio also is illustrated with the typical frequency-response curve of Fig. 4-14, A horizontal line has been drawn across the curve at 1//2 times the resonant-response value; the difference between the fre- ‘quencies at which this line intersects the response curve is equal to twice the damping ratio. It is evident that this technique avoids the need for the static response; how- ‘ever, it does require that the respouse curve be plotted accurately in the half-power range and at resonance, EXAMPLE E44 Data from a frequency-response test of a SDOF system have been plotted in Fig. E42. The pertinent data for evalnating the damping ratio are also shown, The sequence of steps in the analysis after the curve was plotted were as follows: 1 Determine peak response = 5.67 x 10°? in RESPONSE TO MARVONIC LOADING TB Bek comes 267 1 in 2x12 | > E 401X107 zg fama I \ iat i aia, 4 g | er Ea xen fegeney fe FIGURE E42 Frequency-response experinent to determine damping rato. 2 Construct tine at peak//2 = 401 x 10°? in 3 Determine the two frequencies at which this line cuts the response curve: fy = 19.55, fy = 204 4 The damping ratio then is given by po Boh 2 th 218% Mt Energy Loss per Cycle (Resonance Testing) TFimironeotatin is vale mere he pate eticnpbevee he at force and the resulting displacements, the damping can be evaluated from tests run fy oes oan ae a Sd © ees wpe ae Sm potsdur vas exsbhing unas byaosag opt foroery aa Se Ripon 50 out pase mh taped leing Then the apt ed trot lance by he dap oc ht i he eons th ppc wa end eating apace pled for or lating ele he Be Ds te reultcan te nerpoed se he dangngfoce-ciptcenen dpa, ‘If the structure has linear viscous damping, the curve will be an ellipse, as shown ty tr data nen ig 15m th ee he damplag concent ent dere Mad det fom th ta of ie mazinum damping foe tothe manu weely: cxfarnB cs 74 Dewees oF smucrenas fo (9 atx) hip 0s damping) 1 eet) FIGURE 415 ‘Actual and equivalent damping eneray per ejte, Where it is noted that the maximum velocity is given by the product of frequency and displacement amplitude. Ifthe damping is not linear viscous, the shape of the foree- displacement diagram will not be elliptical; a curve like the solid line in Fig. 4-15 might have been obtained, for example. In this case, an equivalent viscous damping ‘coefficient can be defined which would cause the same energy Toss per cycle asin the observed force-displacement diagram, In other words, the equivalent viscous damper is associated with the elliptical force-displacement diagram having the same area and maximum displacements as the actual force-displacement diagram. In this sense, the

You might also like