You are on page 1of 399
STRUCTURAL MODELING AND EXPERIMENTAL TECHNIQUES. SECOND EDITION ee G ants Philadelphia, Pennsyvania Agoring Edo Tim Peter Prog Foe Corl Woche “Mafting Mangers Rr Cin, ae Lewis Ae May Jn St Cove dein Seton Ponel Pres vn Laon, Cary Bons Mamta Cat Ser Library of Congres Caan Pubcon Data aris, Hay 6 Sects mlg a eines etigns Man Haris sn Gann Satis, = 2nd Fae pb. me: Seta olin al expert ‘etic Inco biog reac nine, ISBN iva) 089524096 (ale pape 1 Sct fee-MeleTesing 2. Seta sen, 1 Sai, Gaon Me Tae, TASIET 1989 aur aat ows his book conti instincts rm ate ad ighy ep soe Reprinted mis qed with emis and ours a Indeed A wie vary of eferes ned Reno fers ave ch ae pe ‘ele dita enfant the aura te pbs ios nepal rte aly of ah mis (forthe omega of i te eke is ook nor ny at my be repro or writen ny am yay mean ceo or chai, Incaing ete ing, mieining, and recring. BY ny fafrmon stage veil sym Wien pe ermine pobier “The eoment of RC Pres LLC dos a xen 1 copying forges Sto, fo omc, freeing et wets ofr eal, Spe pernon mus te sed ewig fom CRC Pes LE Th oping, Dialing 0 CRC Press LLC, 2000 Corte vd, NW. Roca ie 333), ‘Teaeark Notice: Prodat r cores aes my be ademas ope nde ny we fe ican and elation wou tt ge © 199 CRC ren LE ‘No chim to cgi US. German wes Inman Stndrd Book Naber O03 2606 inary of Congress Cad Naber 9846557 Prine ia he United Stes Ameria 1 2.39.45 6.7°8.60 Preface Since the fist edition ofthis book, many developments in the wse of tural modeling have taken place. Significant among these isthe extension of the moving technique io masonry structures and the popular use of models on shaking tables to study the earthquake resance of structures. This secon eton attempts o document these and other developments and present to the student an pe to-date text suitable fr use in advanced undergraduate or graduate level courses in stractural modeling ‘orn the behavior of civil and architectural engineering structures using experimental techniques Structural models have always played a significant role in srctuel engineering research, ‘design, and education. More recently, the modeling technique has been wsed in structural product and structural concept development and i illustrated in the second edition with several detailed examples. Spured by the extensive developments in microminiatarizaton of instrumentation and clectonics and renewed interest in using experimentation inthe new engineering curiculum, the use of structural modeling techniques has gained new importance in civil and architectural engi- neering education. The purpose ofthis textbook is to present curren up-todte treatment of stevtural modeling for applications in design, research, education, and product development, ‘This extensively ilustated textbook teats equally the techniques of modeling of reinforced and prestressed concrete, masonry, stel, and wooden structures under satie and dynamic foads [New chapters on model fabrication methods and use of structural models in civil and architectural engineering edvation have been added to the second edition. Many worked out examples and case studies are used Io guide the student or practicing enginoer through the necessary steps of using {he structural modeling technique for themselves in understanding structial behavioe, Numerous problems of varying degrees of difficulty after each chaptr est he understanding of the reader. he authors view the siracural model as a complement tothe mathematical model and not as 8 competitor nora replacement for analysis. A rule of thumb is that if an appropriate, well-tested analytical approach exists fora given situation, it will usually be less expensive and quicker than an expesimental approach. If analysis isnot feasible, or if boundary conditions are poorly defined ‘or highly variable, the mode! test may be the only solution tothe problem, Models must be used selectively, and their range of application i constandy changing a8 analytical methods get increas ‘ingly more powerful. This book should prepare the reader to form the proper perspective a8 10 ‘when and where models should be lized, ‘Models for determining the elastic response of structures have been used for many years, and considerable information i available in the form of research reporis end books. Many elastic nodeling techniques have been replaced by computer-based analysis methods, and therefore they are treated herein a relatively brief manner, The major emphasis is onthe modeling of the tre inelastic behavior of structures. Compared with ¢lastic models, the problems in inelastic modeling are considerably greater — starting from the selection ‘of materials tobe used inthe models, the technigues of favication, instrumentation, and testing to the imerpreation of mode! ests to predict the behavior of the simulated protype structure Applications ofthe modeing techniques to real structures help one fo ter understand the atuat roces of mode! analysis. They also asist in forming a perspective on the types of structures for which ‘Physical modeling is important. There are some types of special sitar where the models approoch las an essential roe in design. These topics are given detailed treatment in he form of case histories. Chapter I discusses the historical ackground of model analysis and similtade principles that 20vem the desigo testing, and interpretation of models. Fight well chosen case stues illustrate the ide use of model studies covering application to a vaity of strctral forms such as buildings, bridges, and special strctres and a variety of construction materials incluing reinforced concrete, ‘masonry, and wood. Chapter 2 presents the theory of structural modeling under static and dynamic Joading. Chapter 3 deals with the various aspects of elastic models with aditional material added 10 ‘cover the properties of balsa and other wooden models, Chapter 4 teats materials for concrete and concrete masonry models, Extensive material as been added in ths chapter to cover the properties of tock mmomry smtures Chapter 5 presets the material modeling requirements for structural tel rt reinforcement for enforced and prestressed concrete sacar. Accurate modeling tthe Prop tie of both conccte and stel is absoktelyesseitial and is one ofthe more dificult pars of the modsing process. Chapter 6 is anew chapter that covers the various model fabrication techniques. stumncntation techniques are treated in Chapter 7 with emphasis on stain measurement and imerptation, Chapter 8 presets selected laboratry techniques and Woaing methods. Developing f sound faniiarity with these technigues isan imegral part of model analysis. A ew Chapter 9 ‘eats with size effects, accuracy, and reibiity of materials and modes. An understanding of the rte av souee of scale effects is ercial in understanding the capabilites and fmstations ofthe tmodeting tochnique. Chaper 10 covers acta apyliations of structural modeling under static ‘quas-stae Toding. Case stxes of number of important uses of modeling in desig, esearch, lind peaduct development are presented. These are dravn from the experience of Ue authors and their celleagues and cover a wide spectrum of applications. ‘Dynamic load effects suchas wind, blast, impact, and earthquakes are tested in Chapter 11 Case studies of important model structures under dynamic loading are also presented. The problems ‘Gscussed in ear chapters dealing with static loading become more dificult in dynamic studies. ‘dina simltode requirements and new experimental techniques are necessary when dealing ‘vith dynamic stvctural models "A new Chapter 12 dealing with educational models for civil and architectural engineers has been added tothe second edition, A large muaber of applications, mainly from the teaching tapesience of the senior author ot Drexel Univesity, Mustates the wsfulness of the sista Sodeling techie in teaching stectural behavior to undergraduate and graduate students alike ‘We believe that this book will be of substartil assistance not only to students of civil and architectural engncering curricula but also to ter engineering students with intrest in experi ental tecbnigues It should be helpful to practicing engines, designers, and architects who are ‘ealing with structures. It should also be useful to thse engaged in testing large or full-scale Structures since the instrumentation techniques and overall approaches used in testing large stuc- tures are very similar t those utilized in small-scale modeling work Tnformation on typical sacral models courses and how the material presented in this second edition can be introduced into the classroom ean be found onthe senior author's web page: "wovvepages.drexeledufaculyfharrshe! ‘We express many thanks to Dr. Richard N. White, Comell University, advisor, mentor, and colleague who contributed greatly tothe first edition and without whose collaboration this book ‘vould not have been possible, Also, thanks go t» Dr. M, Saced Mirza, McGill University fr his Contributions to the fist edition. The authors are indebted to Dr. Start Swartz, Kansas State University, and Dr Philip Pedikaris, Case Wesem Reserve University, for reviewing the entire manuscript and to Drs. Mohamed Elgaaly, Asa Zerva, Mahmoud El-Sherif, and Prof. James Mitchel, Drexel University, for reviving Chasters 5,9, and 12, respectively. Former grads students Dr. Vincent Cacese, University of Maine, Dr. Bechara Abboud, Temple University, Ivan J, Beviea, Dr. Ahcene Larbi, Dr. Win Somboonsong, and Basem Dow reviewed pars of the hansscript for which we are very grateful, Thanks are also duc to our many colleagues who provided photographs and other material for this book. Among those, we would like to thank Drs. Helmut Krawinkler, Daniel Abrams, Philip Pedikars, Miba Tomazevic, Nabil Grace, Ahmad Hamid, Theodor Krauthammer, Emin Aktan, Vincent Coccece, Bechara Abbowd, Abcene Larbi, Salah Sagur, Keng-Ho Oh, and Win Somboonsong and Mr. Noman Hoffman. heartfelt thanks to George Papayianis, who labored trough the equations and tables of he draft, Finally, we would Tike to thonk our many students over te years who persevered shrough our sacral moves courses and whose needs were the inspiration for this took Harry G. Harris Gajanan M. Sabis Philadelphia, Pennsplvania Foreword “This revised edition of an earir edition entitled Structural Modeling and Experimental Anabyss ty Sabis, Haris, Wht, sad Mia is avery welcome ation to the seta engines laser provides an op to-date exposition ofthe bases, rinciples, and ptics of experimen} yf he tal physical behave ofa broad clas of suet ranging trough al is onnct sy evo constuction with parol trong treatment of conerete and masonry structures I etn from the premise that carefully done experimental studies are estntial im developing x at understanding of bow astracture behaves at high Toad levels (up thrcugh the faire sa). ‘The reoulting insight also is required in the formulation of meaningful mathematical models for further study through analytical means. | ‘Wine the banc thrust of de book is ircted to reduced scale models these same principles ace just a pliable to fll cal progr. One f the stongest fests ofthe book iis ory it engincring approach fo all phases of experimental stody of structures, including iii ‘Sanning inate selection, instentation, ang design and appeain, err analysis and Fine effets, and condocting the est nd reporting on resus “Te cxanded coverage inthis new edition includes detailed information on Fading systems — stats nd dynam, including wid, earksuake, last and impact. Dynamicexperimenta stuctres sath are now pechaps the most important of al, given the fac that nonlinear static response can to increasingly well treated analytically Perhaps the eat way to wu appreciate this book isto study the may excellent applications and hee dosed casestudies cover in Chapter 10. Each case study provides a wealth of infor: Thaton on experimental investigations of complex structural behavior. ae ow festure of the book is a comprehensive chapter (Chapter 12) on educational models for vil and uchitetural engineering stents. The use of physical models in teaching structural Chuincering provides students with aleve of understanding not sehivatleUwough theory) alone ‘Tare ake imalable in motivating and encouraging siden to appreciate the completes and seeks of suctual behavior. Thearly encourage my colleagues in academia wo incorporate theve kinds of experiments in their courses at both the undergraduate and graduate levels Richard N. White James A. Friend Family Professor of Engineering ‘Comell University Inhaea, New York | The Authors Harry G. Haris tas been a professor of Civil and Architectural Engineering a Drexel Universiy since joining the faculty in 1974, Prior wothis, he worked in several postions for Gremman, ‘Aerospace Corporation (Nowhrop-Grumman). He eared his ‘BEE. (cs) degree from the American Univesity of Beirut, S.M. and CEE. degwes from MIT, and his PhD. from Corel] Uni versity where he was an Instructor, He hae authored oF co- authored over 130 publications ad several books dealing with ‘structural engineering problems — many of them related to structural modeling and experimental techniques of concrete, masonry, ste, and wooden structures. He has conducted research forte Air Fore, NASA, DOD, HUD, NRC. NSF, and ‘many indostat concerns. Some of his areas of investigation include progressive collapse and earthquake resitance of precast concrete buildings asks under the enforced concrete and masonry portions of he U.S.Japan Cooperative Program on Earthquake [Engineering Research; earthquake qualification of masonry wall in nuclear powerplants, and the development of ductile hybrid FRP (D-ILFRP) reinforcement for concrete structures He is active in several professional societis inchding ASCE. ACI. TMS, and ERI Gajanan M. Subnis has been Professor of Civil Engineering at Howard Univesity since 1973 Prior to this, he worked for Bech tel Corporationas well a the American Cement Corporation. He {ook his undergraduate studies in Mumbai, India and obtained bis PhD. from Comell University in 1967. He has authored over 100 poblications and several books related to conerete, including rehabilitation o conerete structures. Dr. Sabmis is active in var fous technical committees of professional societies, which inclade ASCE, ACI, ASE, and PCL He has traveled and ected extensively interationlly. In adition to his academic achieve- rents, Dr. Sabnis heads two design and research related firs and isaregisterd professional engineerin many sates, including California, Maryland, and New York. Table of Contents Chapter 1 Introduction to Physical Modeling in Structural Engineering 11 Invoduction sn : 1 1.2. Strutural Models — Definitions an Clasifcations. 2 13 A Bret Historical Perspective on MOdelNg one | 14 Structural Models and Codes of Practice. ~ ced 15 Physical Modeling and the New Engineering Curriculum 8 1G Choice of Geomettie Seale nn = 9 17 The Modeling Process a o 10 8 Advantages and Limitations of Model Analysis a 19 Accuracy of Structural Models... =a) 110. Model Laboratories : sl LIL Modeling Case States 1B 132. Summary a = co Problems, P36 References. aa 37 Chapter 2. The Theory of Structural Models 24 Introduction. a - o 2 22. Dimensions and Dimensional Homogeneity : nn 23. Dimensional Analysis. 45 24 Stmctural Models a 56 25 Simalitude Requirements on 26 Summary . ~ %6 Problems. ~ . oT References st Chapter 3. Elastic Models ~ Materials and Techniques, 3.1 Introduction. se 86 3.2. Materials for Elastic Models . 87 33. Plastics. : : 38 34 Time Effects in Plastics — Evaluation and Compensation . 96 3.3 Bfecs of Londing Rate, Temperature, and the Environment. 100 36 Special Problems Related to Plastic Model. 103 3.7 Wood and Paper Products. —— 108 38 Elastic Models — Design and Reseach Applications. von 3.9. Determination of Influence Lines and Tauence Surfaces Using Indirect “Models — Muller-Breslau Pinciplennunns send 3.40 Sumamary 123 Problems en a 124 References. S — v2 Chapter 4 Inelastie Models: Materials for Concrete and Concrete Masonry Structures 40 General 42 Prototype and Mode! Coneretes 43 Engincering Properties of Coneret 44 Unconfined Compressive Strength and Stess Stain Reto 45 Tensile Suength of Concrete 46 Floxural Behavior of Prototype and Model Concrete. 47 Behavior in Indirect Tension and Shear. 48 Design Mixes for Model Concrete. 49 Summary of Model Concrete Mixes Used by Various lavesigators. 4.10 Gypsum Mortars. 4.11 Modeling of Concrete Masonry Siucures 41.12 Strength of Mode! Block Masonry Assembhges. 413. Summary Problens.. References Chapter 5 inelastic Models: Structural Steol and Reinforcing Bars 5.1 Introdvtion 52 Selon 53. Structural Steel Models... [54 Reinforcement for Small Scale Concrete Models 5.3 Model Prestressing Reinforcement and Techniques. 56 _ FRP Reinforcement for Concrete Models 5.7 Bond Characteristics of Model Ste. 58 Bond Similitde 59° Cracking Siniltade and General Deformation CGonerete Elements 5.410. Summary... Problems. References... Chapter 6 Model Fabrication Techniques 6.1 Introduction. 62 Basie Cutng, Shaping. and Machining Operation. 63. Basic Fastening and Gluing Techniques. 64 Constraction of Structural Stel Models 65 Construction of Plastic Models {6 Constraction of Wood and Paper Models Fabrication of Concrete Models 130 130 BL 133 182 146 8 153 159 165 170 188 2m 202 205 210 210 mal 230 252 255 259 266 267 am 2m a4 280 Chapter 7 Instrumentation — Principles and Applications TA General 72. Quantities t be Measared - . 73. Stuin Measurements. . . : 74 Displacement Measurements alld Stan Measurements and Crack Detain Methods 716 Siest and Force Measurement. ~ 7.1 Temperature Measurements. 8 Creep and Shrinkage Characters and Moisture Measremen 719 Data Acgpistion nd Redhct asm 1.10 Fiber Opies and Smart Stroctures. oe TAN Summary Problems, References. Chapter 8 Loading Systems and Laboratory Techniques SA Invoduction a 82 Types of Loads and Loading Systems. - : 83 Discrete vs. Distributed Loads. : 84 Loadings for Shell and Other Modes 85 Loading Techniques for Buckling Studies and For Structures Subjet to Sway 8.6 Miscellancous Loading Devices : 87 Summary nnn ~ Problems References Chapter 9 Size Effects, Accuracy, and Refiabiti 9.1 General 9:2 Whats Sie Ee? 93. Factors Influencing Size Effects 2 94 Theoretical Sues of Sie Esfees 9.5 Size fects in Pani Concrete—Exprimental Wotk mnsnnn 9.6 Size Effects in Reinforced and Prestressed Concrete effects in Metals and Reinforcements 918 Size Effees in Masonry Mortars. 9.9 Size Effects and Design Codes. 9.10 Errors in Stractaral Model Studies. - SAY Types oF E00. 9.12. Statistics of Measurements. : 9.13. Propagation of Random Errors... = 9.18 Accuracy in (Concrete) Models. a 9.15 Overall Reliability of Mode! Results. a ‘9.16. nfluence of Cost and Time on Accuracy of Modets 917 Sammary em - Problems. : = . References 5 vo 320 32 351 353 361 362 364 367 377 “377 381 in Materials Systems and Models Az au “ald ais 420 431 433 34 “435 437 439 aah aaa 450 4st 458 458 359 4 ‘Chapter 10 Model Applications and Case Studies 10.1 Introduction 7 466 102 Modeling Applications . 466, 103 Case S568 ven nn = — 529 104. Summary... 7 ve 32 Problems nun a 373 References. eee — 379 Chapter 11. Structural Models for Dynamic Loads TT Introd. 00 . —— 586 112 Simitiade Requirements. cso a 387 113 Materials for Dynamic Models. - 388 114 Loading Systems for Dynamic Mods! Testing, 593 11. amps of Dyanic Mods z ~ 08 116 Case Stuies : : 9 117. Summary : nn) Problems. - vom ors References - 2 a 615 Chapter 12 Educational Models for Civil and Architectural Enginecing 121 Inredction ve nn 680 12.2 Historical Perspective. —— 681 123 Linearly Elastic Sirctral Behavion : . 681 124 Nonlinear and Inelastic Stvcturl Behavior = 4 12.5 Stroctrsl Dynamics Concepts vo m2 126 Experimentation and the New Engineering Cunicaln = “ns 127 Case Sui nd Sten Po _ m9 128 Summary S . — “751 Problems . = : 2 752 Reference nnn : _ IS Append A Dimenstona Dependence and Independence Al The Form of Dimensions 131 ‘AZ Method I: The Numeric Method = 159 3 Method T: The Functional Method _ 761 AA Ulustrative Examples. ae 963 ReReMCE8 enero 765 ‘Appendix BA Note on the Use of SI Units in Structural Engoneing Bl Geometry. ss : 68 B2 Densities, Gravity Loads, Weis 8 B.3 Force, Moment, Sires, and Other Stress Resultant . 68 BA Miscellaneous (Angles, Temperature, Enerey, Powes)... 168 BS __S1System Standard Practice... cea 769 Index To Helen and Sharda CHAPTER 1 | Introduction to Physical Modeling in Structural Engineering CONTENTS | LAL fetrodvetion — : 12 Smuetural Models — Definitions and Classifications 121 Models Classification unum 122 Physical Models in Ober Engineering Disciplines 1.3. A Brief Historical Perspective on Modeling... 114 Structural Models and Codes of Prater 15 hlcal Modeling and the New Engineering Camila s.r 16 Choice of Geometric Seale. : 2 17 The Modeling Process a: = 18 Advantages and Limitations of Mode Analysis. 19 Accuracy of Structural Models, 110. Model Laboratories. TIT Modeling Case Staies. LAVAL Case Study A, TWA Hangar Suvetares L112 Case Study B, R/C Frame-Wall Stuctares L113. Case Study C, Reinforced Concrete Bridge Decks. LILA Case Study D, Lightly Reinforced Concrete Buikines. LI15 Case Stady E, Presuessed Wooden Bridges. = LANG Case Study F, Interlocking Mortaress Block Masoary 111.7 Case Study G, Pile Foundations... ae TALS Case Stody H, Exteroalyinteally Prestressed Concrete Composite Bridge System. a at 112. Summary 4 Problems... - 36 References 37 1.1 INTRODUCTION ‘A perspective on physical modeing of statues presented inthis chapter, incaing ss sate rote various ypes of physical model tat hae evolved over the years and some comments ‘STRUCTURAL MODELING AND EXPERIMENTAL TECHNIQUES cn he pr le of hese eters of odin teation, desig, ech, an pdt concept development ‘in North America and elsewhere. “ oceeen Seta models and ede cle strict onetime alld pea mils) fave ala layed a Sica ron stocaraengineig cation, escath sal ge, egress tec sctus anpcints evs ten imper ixhing ocean td sive engintng concept A wie ng of poten et nee, cache ad iterating mci ty osc tr hte sem hh ag on theories oth ate eee dpi of xperinnel se willbe teaed inden sso ht. Tas capers the ale ms eee wa 'shopedan appresation osteoid ya oneal othe eee iting wed dyin th engcrng posi This me expended sca eno took hs aed ne materi onthe tengo sacra cnet co ea aed Ceoineing tens anon theca sacral seg cis soe nt da con ne nw ping rea, Pacly seta se ipa n ling new onto nd creating pce, bt npc md a Conserserebcaseisaeicomplel feet om he Seu, Stal lee wsedinarccctoal egicring efcaton wile canted een, een ev chs fhe tsk se thse come te vehicle by shih te sade sb nance we ope stroctural behavior concept in both analysis and design = 1.2. STRUCTURAL MODELS — DEFINITIONS AND CLASSIFICATIONS Asal mode deed ay phys representation of sce oa prion of ¢ scr: Mos commen te model wll be comet ederd ses ie Seon a ‘ence by ACT Cate 14, Experientia rConce Ses phe eu wel o mol fst ie fay mae eau: fens een aes Janney et al. (1970) is ” met seat 2 sm ys at sey of tl cn i eel ‘ee Gn compron vt lassie) wh ta be te whens soe ‘mast be employed to interpret test results. eis Bath ttn encompass od hs of waking sues on rtp ili) ste sx sing i dam ter ec ves sl tye an win ca etn ours, nds Loins ne sateen es and wind effects. aaacen * Sima a Mary edcesiz str element ae ctl din sac stasis {alors also class these structures as models even though simiitude conditions are not normally {np tele see sec ml nica dia mead canton a ed ay ton he bred thao of th resch med adasngh el ee the sgn fein. sl bent tha me bre ss rane en eae te. An ipo dso is way mae wen protype nal cnet be wd sides. nt ee, apropinte model isnt sated nde Tectia ihen propery deg «mo vipins cnc malo mde of fos prsese ops coc wes witha eon saat of ht a bck sey seat Si scenes sale fcr of 3 Mae is ao provsed ‘sel ad tne ers vhs ones ‘ae ar of 55. Mach them fof snes apie ma oot pe of sts ao iteret see ates INTROOUGTION TO PHYSICAL MODELING IN STRUCTURAL ENGINEERING 3 1.2.1 Models Classification Structural models can be defined and classified in a variety of ways. The definitions adopted dere cela wo the intended faction of the model. That is, what do we expect to achieve from the tests? Do we want only elastic response, of do we expect fo load the model up ofall to observe its complete behavior, including the failure mode and capacity? Ate ve content to work with influence lines determine from the models, or Jo we need actual stan measurements for prescribed Joadings? The models needed in each of these applications have been given well-aceepled names, Elastic Model — This type of mode! has a direct geomesic resemblance to the prototype but is ‘made ofa homogeneous, elastic material that doesnot necessarily resemble the prototype material The elastic mode is resticted tothe clastic range of behavior of the prootype and cannot predict postercking behavior of concrete of masonry, postyeld behavior of steel, nor the many other inelastic behavior modes that develop in actual stractoes when they are loaded. Caper 3 teats clastic models in detail, inclading selection of materials, Plastics such as methyl methacrylate (Plexiglas, Lacite, Perspex) and polyvinyleloride (PVC) are most widely used in constructing clastic models, even though their time-dependent properties present dificulies. Wood of many ‘commercial species and balsa wood are also covered in Chapter 3 since these find many applications ‘in modeling timber structures and aze easy to work with. In many applications of elastic models, specially in the demonstration of structural behavior to students, materials with low modulus of elasticity soch as rubber, paper products, and plastics ae used fo accentuate deformations making it easy to observe the behavior. These materials are also discussed in Chapter 3. Inlrect Model — An indirect model is a special form of the elastic medel that s used to obtain Influence diagrams fo reactions and for internal sites resultants such as shearing forces, bending ‘moments, snd axial frees, Th loading applied to indirect models has a0 correspondence to the actual loads expected oa the prototype structure since load effects are obianed from superposition ofthe influence values. An indirect model often does not havea direct physical resemblance tothe prototype; for example, a frame whose behavior is controlled by its flexural stifines properties (EN) cam be modeled with an indiect model that correctly teproduces the relative stiffness valves, ‘The later can be done without precise sealing of the eoss-sectonal shape (circular shapes inthe indirect model can represen prototype wide lange sections), and the clement areas may be grossly distorted without affecting the reals Inthe pas, most applications of the indirect model have been for nonuniform members in indeterminate frames, but now this type of model find very litle use because these purely elastic calculations are beter done by compte. Direct Model — A diect model is geometrically similar tothe prototype in all respects, and the loads are applied to it in the same manner as wo the prototype. Strains, deformations, and stresses nthe mod! foreach Toadng condition are representative of similar quantities in the prototype for ‘he corresponding loading condition. Thus, an elastic model can also bea direct model. Examples of indirect and direct models are shown in Figure 1.1 ‘Strongth Model — This type of model sslso calle ultimate stength or realistic model or replica ‘model and i a direct model that is made of materials that are similar 1 the prototype materials such thatthe mode! wil predict prototype behavior forall oads upto failure. A strength model of ‘reinforced concrete clement or structure mast be made from model concrete and model-einforcing Clements where each ofthe materials satisfies the smiltude conditions fr the prototype materials ‘The later represents the most dificult problem met in strength models for concrete structures; (Chapers 4 and 5 cover these topics in depth, Strength midels can also be made fr stel structures, STRUCTURAL MODELING AND EXPERIMENTAL TECHNIQUES ENDVIEW DEFLECTION INFLUENCE LINE - MAX NEG MOMENT INTERIOR SUPPORT (@moIRECT MODEL 5.20 UNIF STRAIN GAGES m7 eNDVIEW DIAL GAGES DEFLECTED SHAPE (e)OIRECT MODEL, Figure 14. Exarole of (9) ect and) direct med ner sre, and in each es the mor ote en ding the rope mate a fabrication techniques for the models. ers fines Prone “a "\aengh mel muse a ect mel, by defnton, To wee esi of nist mols, ne musty sgerpeson of rest, ane serpin panes Pt valid fr te poster esse eum alsoength nds isnt conomeal Bad kength models nd {ee them only ine ast ange of bebo Wind Effects Model —There ace various ways of classifying wind effects modeling. We can tlie shape or rigid modes, where citer total ces or dhe wind pressures om he structure may ‘be measured, and aeroelstc models, where both the shape and stfiness properties ofthe prototype structure are modeled in order to measure the vind-indsced stresses and deformations and the ‘dynamic interaction of the structure with the win Dynamic Models — These models are used to study vibration or dynamic loading effects on structures They may be tested on a shaking table for stodying earthquake loading effects or in a ‘wind tunnel for studying aroeastic effects. Dynamic models can also be used to study internal or extemal bast effects or impact effects on structs. These form a very important group of models land are extensively covered in Chapter 11 Instructional, Research, and Design Models — One often sees the classification of instruc tional models, research models, and design models. While the use of each sobviow, tis worthwhile \nTRODUGTION TO PHYSICAL MODELING IN STRUCTURAL ENGINEERING 5 to pint out thatthe degree of sophistication needed in each may be markedly ifferent, Instrctional Tred should be made as simple a possible to demonstrate the concepts und>r study, and simiitode ‘Doteton tat doce not markedly ifuence the desired behavior is permis le. Reseach models from which theories may be substantiated and generalizations made fora cls of structures, sally st be made with as much accuracy asthe lbortory technicians can mast. Design models may tunge m acowacy requirements rom te isirutional mode! tothe ressarch model, depending upon the desired reels. Some design models may be used only a8 a conceptual ool to geta eter idea Df how a proposed structure deforms under Hand; olbers may be expected 0 predict the true Toad Capacity of the structure, The use of srctural models as dzect aids in design is one of the most ‘omerfal applications of strctral models. An enginer is often called upon to design strvctres uch asthe Trans World Airlines maintenance hangar facility located in Kansas City, MO and ‘escibein Chapter 10 n this instance a series of plastic models and a stength model were sed asthe main approach in the final design. “Anothce major physical modeling application in design isto help verify calculations for very targe and monumental stractures where failure consequences could be extemely serious (Such 28 heavy los of life or capital investment, or disruption of essential lifeline services), nuclear reactor strctre is a good example of this application, ‘Perhups the major disadvantages of using a model for design purposes from the sianpoin of the consulting engineer, are the time and money involved in the modeling sudy. This topic is given farther attention in Section 1.8 ‘Other Model Classifications — Okher classifications of models include thermal models, where cifecs of temperature gradients are sted. Thermally loaded models ae usually elastic, direst ‘odes although some attempts have been made (0 combine mechanical ods and thermal loads for steagth models There is also a group of photomechanical models that wsilze opsical effects, sch asthe photoelastic effect for stress intensies and directions and interference effects from trds to measure plate displacements, inleral strain fields, and deflecticns of framed structures ‘Construction procedure models are used to belp plan the building of very complex structures, such sin reinforcement placement in niclear reactor containment, and in cantilever bridges 1.2.2. Physical Models in Other Engineering Disciplines ‘Hydraulic engineering, models have been studied as early asthe late 1800s with considerable succes. Studies of id notion in pipes, pumps. and open channels, wave setion, beach erosion, ‘Shing due to tides an tial caren, and the extent of contaminants de to polation in impertant ‘estuaries and river basins hae all een suceesfuly carried out by means of physical models. ‘Today, all important hydraulic stuctores are designed and built after estin preliminary model studies have been completed ‘Naval architects have fr a Tong ime relied on the use of piysical models in the design of shipa, Al the important features of ship design, such a ship maneweraility in smooth and rough feat, ahip ending and vibration, frictional and wave-smaking characteristics, t., have been studied by means of scale models in model testing tanks or water basins. “The aulomotive and gerospace industries hae relied on the use of physical models for staying ‘a number of phenomena ranging fom the wind flow ever aerodynamic bodies to vibration testing fin thermal Suess elects, Important contributions to the advancement of automotive and aero teal sciences and space exploration would not have been made posible were it aot for model testing. Important examples in these areas ae covered by Schuring (1977), "The geological, peomechanial nd soil mechanics sciences have also used physical models to study sll and rock phenomena. More receny, mel studies dealing witt gesyntheics and geome ‘hati have also been cared oat. Modeling of soil-sirecture interaction is bricfly treated i. the tmodkling of srucral systems using a cenifge in Chapter 2and in an example in Chapter 11 6 ‘STRUCTURAL MODELING AND EXPERIMENTAL TECHNIQUES 1.3 A BRIEF HISTORICAL PERSPECTIVE ON MODELING “The use of small-scale models by engineers and builders dates back many hundreds and even thonsands of yeas. However, these early models were primarily sis for planing and constructing structures and were not useful for predicting deformations and strengths of pottypes, They more nearly resemble the modem architectural model ind should not be thought of in the same context 4s stuctual models ‘Most models uss to predict structural behavior require measurement of srains, displacements, and frees. Thus, the development of modeling practical ool has been sharply inflnced by bilities in experimental sess anlysis. The most-used technigaes in experimental stress analysis have been established ony since the tum of the 2entary. They include: Photoclasticity for elastic stress analysis of complex geometries: 2, Deformeers developed by Beggs, Eney, Gottschalk and oles for introducing deforma- tions into indtect modes and thea determining iaflacnce lines by use of the Muller-Bre- slau principle; Mechanical and pial strain guges for measurement of surface stain; Flecrical resistance strain gages; 5. Linear variable differential transformers (LV), linear potentiometer, and similar devices for electrical recondin of displacements; 6, Brite coatings, moiré and interference fringe metbods, and photoelastic coatings for “full-field” strain measurements on the surface of a structure of model; 7. Automated data acquisition systems that uses minicomputr to contol and process many ‘channels of data Hem 2, deformeters in indirect models, has been used to study iftuence lines and to determine deflections in skeletal stuetures since the early 1920s. Beggs (1932) and Eney (1938) were the firs to use specially designed deformetes to obuin influence Hines for skeletal structures made of plastic and other materials. The simplest appiction isthe use of along, flexible strip of woos, ‘brass, or stel and to measure the model deformations directly. The Gotschlk (1926) continostat is an improvement on this technique. However, lrge defomations must be imposed onthe model, ‘which causes other kinds of errors (Kinney, 1957) Bull (1930) used the brass spring mode! for indirect analysis of aniculated structures. Development of other deformeter i reported by Ruge ‘and Schmidt (1939) at MIT (moment deformetr, and by Moakler and Hated (1953) at RPL (RPL ‘eformete) Ttem 4, the electrical resistance stain gage, is pehaps the single most important development in tems of providing an easily used method for determining either static or dynamic strains in a structure. The same gage forms the sensing devize in commonly used load cells and trandcers ‘Thus its introduction inthe 1940s canbe considered to be the bass for moder experimental tess analysis and structural model analysis (Ruge, 1943/4; erry and Lissne, 1955). Relatively litle model analysis other than photoelastic studies and indirect models was done prior to 1940. The Ysescale Hoover Dam model built in 1930 and models of ether great dams of ‘that era built by the Bureau of Reclamation, Denver, CO were a aoable exception (Savage, 1934). Another notable exception of the exly 1930s wis the work of Goorge Beggs and his co-workers at Princeton on suspension bridge models (Begas etal, 1932). Since that time, the technology ‘needed for rapid constuction, instrumentation, cad testing of stuctural models has continued to develop, The current use of structural modeling is introduced inthis chapter, andthe full ange of| applicability will become apparent as the reader progresses through the book, INTRODUCTION To PHYSICAL MODELING IN STRUCTURAL ENGINEERING 7 14 STRUCTURAL MODELS AND CODES OF PRACTICE Modeling has received relatively ite altetion in most North American building codes and specifications. However, mos codes do contain special provisions that permit the engineer to make rather substantial use of models in the design process. For example, the 1969 City of New York Building Code contains the following paragraph: (5) MODEL TESTS. Teton model es thn fl ie maybe used to determine the elaive intensity, ction, and ditittion of stesses and applied loads, tat shall woe comsred as proper mod. {orevalsting tress in nore suength of individual members unless apoved by the commissioner fortis purpose, Where model anal is proposed a a means ofetaahing the situa design, ‘he flowing coon sal be met. 2 Analysis shal be made bya firm oeosprton satisfactory tthe cmnissone. |b The simiitoe scaling and valiy ofthe analysis sal be axe by an cero pencil ‘ofthe im or cnpration making thesis © A report om the analysis shall be submited showing test sup, equpment, and resting ACI 316-95, Building Code Requirements for Structural Concrete, exits model analysis for shel stractures in Section 19.2.4 192.4 — Expeimentl or numerical snaysis procedures sll be pet where it canbe shown that such procedures provide a safe basis for desion This is further elaborated on inthe Code Commentary R192 Experimental analysis of estic meld" has bon se a. substi foran analytic soation of complex shell rate. Experimental analysis of reinforced miorcomrete model tough the last, cocking, inelastic, and wlimate range should be considered fr important sels of unas sie, shape andor comple, For model analysis, oly thse potion ofthe sate which fet sigicaly the tems wnder stoly net be simulated. Every atte shouldbe made tense tat he experiments reveal he ‘quanta ber of the pottype strstr Some countries, such as Australia, permit complete designs of cern (pes of structures by ‘mode! analysis alone. Thus, there isa relatively healthy potential usaze of modeling in design ‘codes. Many engineers would be even more receptive of modeling for Jesiga if they would only realize that many ofthe code provisions that they apply analytically every day ae in fat derived ‘mainly from tess on reduced-scale models. The engineer who wishes to use models should not Desiate to contact the responsible building official to seek approval and should also seck proper ‘assistance from an expert in model analysis There are numerous situations in which these code provisions might be appiod in practice; in most cases itis where the analytical approach is not fully adequate. Basic doubis may arise in applying existing analytical techniques to new and complex structural forms. Analytical methods are not yet developed to handle te extremely complicated behavior of enforced concrete structures Toaded to nearfllure or certain othe Fimitstate conditions. This is why modeling is often used by engineers studying the failure of structures, ‘Types of sractures suitable fr possible structural model studies during the design phase include: 1. Shell roof forms of complex configuration and boundary conditions 2. Tall structures and odher wind-semstve structures for which wind tunnel modeling is indicated; 8 ‘STRUCTURA. MODELING AND EXPERIMENTAL TECHNBQUES. [New bung stractral ystems involving the interaction of many components; Complex bridge configurations such a moll prestressed concrete box gir highway bes: [Noclear reactor vessels and other reinforced and prestessed concrete pressure vessels; 6 Ordinary framed stractures subjected to complex Toads and load histories, such as wind and earthquake forces, +. Stroctual slabs with unas boundary or losing conditions, oc wit iegular peometry produced by cutouts and thickness changes; 8. Dams; 9. Undersea and offshore stuctures; 10, Detailing, Item 10 points out an important use of modes: fr studying problems that rie in only aimited region of a structure, stch as an involved connection detail or localized stresses de 10 Tange pestressing forces. Carefully designed and tested, he partial model can be extremely important in Clarifying these stations and in leading tan improved design. The major difficulty with the partial model is in providing proper boundary comlitions: an adequate boundary condion in a physical model may produce even worse results ban a poor boundary conlition in an analytical solution, 1.5 PHYSICAL MODELING AND THE NEW ENGINEERING CURRICULUM “The now engineering curriculum, under development and implementation at Drexs! and ater universities with strong udergraduste engineering programs, relies on teaching basic engineering ‘concepts by giving the student hands-on experiences in the laboratory (Quinn, 199). Mastery of ‘eaperimental techniques is expected in the fist aud second years. The students become filiar ‘wth how deta are acquired, processed, and analyzed as well as how basic experimental techniques, devices, and methods are tsed in a wide variety of engineering disciplines. The experiment feperience is not lost in upperfevel engineering courses but is ineprated with the theory. This reans that experimental techniques must become nore widespread fo reach a much wider andience ofthe engineering community. For undergraduate students studying the fundamentals of tuctres, St means thet laboratory feciliies must exist where the design, fabrication, testing, and observation fof the structural behavior canbe carried out. This kn of experience can be invaluable tothe new ‘Student who cannot afford t0 walt to be a practicing engineer with several years of experience 10 tain such information. Stidens” understanding i prealy enbanced by observation and corlation ‘with the theory that they are earning. 1 shoal be pointed ovt that in the new enginceyng cuticular experimentation is an integral pert ofthe engineering stents profesional development The incoming freshmen stat engineer ing courses from day one. Emphasis is given to experimental methods in engineering because of their wide use in analysis, design, development, and manufacturing. Special attention is given to the interpretation and effective presentation of experimental results in writen and ora forms. The computer is used effectively aa rescarch and design tool The students engage in professional ‘design projects solving real world problems. They lear interactively, trough teamwork, gaining life-long leaning skills. ‘Reganlss of discipline, practicing engineers perform a wide variety of experiments throughout their carers. Mastering the techniques of experimentation is therefore very esental. During the freshman year, stents conduct a thre-hovrTabratory each weck in the Engineering Tet, Design, ‘and Simulation Laboratory. Tis state-of the-art facility provides students with opportunites to INTRODUCTION TO FAFSICAL MODELING IN STRUCTURAL ENGINEERING 8 exercise their imagination, satisfy their curiosity, and experience the joy of engineering. Throwgh fhe mar af computers and Computer controlled laboratory instments, stents are able to explore tow experimentation i used in engineering applications. They become famiiar with how data are secured, processed, and analyzed as well as how basic expeimental techniques, devices, and methods ae used in a wide vity of engineering disciplines. ‘During the sophomore and junior years, the laboratory component is shared among several ‘courses rather than being an integrated course at was the case in the freshman yeas, These courses teach the fundamentals of structoral enginecring and form the backbone of the young strvetural ‘engineer's academic taining The laboratory work s competed by the stodent majoring in stractures jn courcs such as Construction Materials, States, Mechanics of Materias, and Introduction 10 ‘Structural Analysis. These courses contin the stents Iboratory training in a variety of hands ‘on experiments and demonstrations wsing the techniques developed inthe fist year ‘During the senior year, the structral engineering major takes courses sich as Project Design, Structural Design of Stee, Concrete, Masonry and Timber Siructures, and Senior Seminar. 18 audition, advanced seniors are allowed to take introductory praduatetevel courses including Model “Analysts of Structures and Experimental Analysis of Nonlinear Sractures 1 an H- Examples of Saractre, drawn from these courses and demonstting diferent levels of behavior, are described jn Chapter 12 to illustrate how the laboratory component of the new engineering curiculum is implemented 1.6 CHOICE OF GEOMETRIC SCALE ‘Any given model being builtin a given laboratory has an optimum geometric scale factor. Very small noes requie Fight loads but can present great difficulties in fabrication and instrumentation. ‘Large modets a easier to build but require much heavier loading equipme. The later requirement is not seus ina laboratory that is flly equipped to conduct tests on large structures, but itis @ Severe handicap ina smaller laboratory. Typical scale factors for several classes of structures are a Follows: ‘ype of Siructure Elastic Models Strength Models stetcot wes yes igh bee BS rs Wind fee dw Neeamtiaie ‘The rationale behind this table should become more apparent as we progress through the chapters of tis book. Strength models of concrete structures have many practical dimensional Timations, such as minimum feasible thickness bar spacing, cover, ct. Maintaining materials similtde requirements a crucial problem inthis category of models. As will become apparent Tater in the tex, tis desirable alays to try tose the Targest physical size model fora particular ‘pplication, An example ofthe various factors that influence the choice of model size is given in the following example 10 ‘STFUCTJRAL MODELING AND EXPERIMENTAL TECHMIGUES Example 1.1 tis requised to test two reinforced concrete strength models to demonstrate the principle of “over” and “under” reinforced concrete beam belavior. The available testing machine ha a capacity (of 5 KN and the work area to accommodate th model i approximately 0.5% 2m, Based on these Physical limitation ly supported eam of up to 50x 100 mm eross section made of sand and fine aggrezate (“pea gravel) canbe fabricaed Using wp toa No.2 bar (6-35 mm) forthe steel reinforcement. A model of the hypoibetical prototype beam experiencing this behavior would be at 4 to % scale. If only wire reinforcement of lest van 6.35 mm in diameter is availble, then a ‘beam of appropriately smaller size will have tobe used with seal as smll as ocr Ya. Obviously, ‘strength model ofa larger portion ofa reinforced concrete structre ta be stiied, sing the sme ‘equipment, wil have to be of a smaller crss section, hence scale 1.7 THE MODELING PROCESS ‘The succesful modeling study i one thats characterized by careful planing of the many diverse steps inthe physical modeling process. An experimental study ofan engineering structure is a small engineering project in itself, and as in any engineering venture, a logical and carefil sequencing of evens isan absolute necessity Detailed planning of an experiment is even more essential than planning of an analytical approach because refinement ofa structural mod! halfway through the modeling process usally imposible. A major aspect of planning is deciding what i expected ror the del. Da we need ‘only clastic stresses and displacements, or do we want to see how the stracture behaves at ovefcads leading upto failure? Is instability a possible fare mode? Are thermal stresses involved? Do we have to simulate dynamic effects? The time quired to complete the model study ean range from pethaps a week or two fora very limited elastic model ofa portion of a structure to 6 months or ‘more fora detailed, ultimate-strength reinforced mortar model for predicting fale behavior of ‘complete stricture. We obviously must guard aginst “overdcing” dhe model stdy just s We Have tw avid excessive analysis of astrctare. The engineer who bears inal responsibility Fr the project rst be the key person in prescribing precisely what the mel i supposed to accomplish, ‘A typical modeling study can be broken inte the follwing maltistep process 1, Deine the scope ofthe problem, deciding what is nceded from the mov and what is not needed. 2, Specify similinade requirements for geometry, materials, loading, and interpretation of resuls. Pay particular attention to those seilitude requirements that cant be met, such the desired equality of Poisson's rato for concrete and plastics when doing clastic ‘modeling of shell and slab structures (Chapter 2), 3. Decide onthe size of model and required kvel of reliability or accuraey, What size mode should be used consistent wih the accuracy? If 230% is adequate for design porposes, ‘hen an temp to achieve 105 accuracy is wasted efort and time (see Chapter 9), 4, Selectmodel materials with properttentoeo steps 1,2, and3 above (Chapters 3 through 5). 5. Pan the fabrication phase in consultation with the technicians who will be constructing ‘he mode, and follow the fabrication activities closely. This can bea ftrating pat of ‘modeling because iis often quite time-consuming (Chapter 6) 6 Sclect instrumentation an recording equipment for stains, displacements, forces, and ‘other quate. This step mast be closely coordinated with steps Sand 6, particularly if ‘embedded strain gages are tobe used in concrete models Special stain gages and other equipment must be erdered well in advanee ofthe actu time of usage (Chapter 7) INTROOUGTION TO PHYSICAL MODELING IN STRUCTURAL ENGINEERING ” 7. Design and prepare the loading equipment; new systems should te thoroughly checked ‘out and calibrated before wse om a mods (Chapter 8). 1 Observe the response of the model ding loading, taking complete notes and photo- _rophic records ofthe behavior. Do not rush though a test, and never leave anything to ‘memory. Some investigators use tape or video recorders inthis pase to record detailed ‘comments onload history cracking development, instability modes, and other information that may be dificult to describe numerically. Approximate calculations should be done before the experiment o estimate expected levels of response. Equilibrium checks should 'be done on results obtained carly in the test. Because of the great importance in properly recording data vhen it must be done manually, afew specific comments ae in order here a. Prepare a ruled sheet with columns; pt the date, names of test personnel, and the model designation on the fist shet. '. Record the readings directly, and do not attempt to reduce the tain one's head «Record the 20 readings, allowing atleast wo lines of space Since zero readings often rust be taken more than once, 4. Allow adjacent columns for reduction of results. “Take readings at lower load increments as failure i approached. “Take readings a8 yielding or fuilure actually occurs, even if the level of accuracy schieved isnot high — approximate readings can give a better idea of behavior. Take final readings when the load is emoved. 98. Anaiyze the daua and write the report as soon as possible, while the ent testi sil fresh in the mind. In addition to reporting the resuls, suggestions for improvement in techniques shouldbe recorded to facilitate better modeling results in subsequent experi- ‘Most of these steps are merely statements of common sens, but i is surprising how often common sense is ignored o¢ left out of a erucial step in an experimental stady. Several “laws” should be kept in mind when thinking bout the difficulties of experimextal work: ‘Murphy's Law If something can go wrong, it will ‘O'Tooke’s Law: Murphy is wildly opimisic. Harris-Sabnis Lave: Things are never as bad as they tur out to be ‘Theres a moral in these tongue-in-check laws, and that s the fact hat experiments most be curcfelly planned, controlled, and interpreted if they are to succeed. 1.8 ADVANTAGES AND LIMITATIONS OF MODEL ANALYSIS ‘The main advantage of «physical model over an analytical mode! is that it portrays behavior ‘of complete structure loaded to the collapse stage. Although substantial progress is continually ‘made in computer-based procedures for analysis of structures, we still cannot predict analytically the failure capacity of many three-dimensionl structural systems, especially under complex loadings. ‘The prime motivation to conduct experiments on structures at reduced scales isto reduce the cost Cost redetions come about from two areas: reduction of loading squipment and associated restraint frames, etc, and a reduction in cost of test-strucre fabrication, preparation, and disposal ‘after testing. The load-reduction factor is most dramatic since the concerrate load ona provtype {reduced in proportion to the square ofthe geometric scale factor ofthe model (a 100-KN protxype 2 ‘STRUCTURAL MODELING AND EXPERIMENTAL TECHNNQUES. Ioad is 0.25 KN on @Y-sale model) This raction is even more dramatic when & low-modulus mates uct ay plat Is used in the model “The major limitation of using structural modelsin a design envionment are those of time and expense. In comparing pysical modes with nati! models, one finds that the later are normally Jess expensive and faster, ane one cannot expect physical models to supplant or eplae analytical ‘modeling of structures wien the later procedure las to acceptable definition of behavior ofthe jrototype structure. Thus, physical models are almost always confined 10 situations where the mathematical analysis is mot wequate or not feasible. Another liiting facto is that changes inthe proatype design resulting from the resus ofa model study may require a second model to chock the desig, Practical considerations therefore often dictate thatthe model wil be used 10 verify 2 “nearly finalized design, “The time involved in modeling is often subjected to Father pressures because the decision to 0 toa piysical sty i often made atthe last minute, after more conventional approaches are proved inadequate. An engineer who is accustomed to geting all answers by analytical means is fatraly hesitant to admit that the analysis is incuficient and that a physical mode is nooded. Suitable efforts mast be made to predict earlier im the design process that atest is needed This ‘would enable earlier planning anda smoother, less hectic approach to the model stody. ‘Design applications of structural models have been oullined eatir inthis chapter. Stuctrat models are also widely employed in research progam in such applications as the following 1 Dele fxpinat da fois f ie ade of ee anti 2. Study of basic behavior of complex structural forms such a sells. 5. Parameitic staies on member behavior. Macs of our basic research on reinforced concrete Aexoral members has been done on lrge-sale models. 4, Behavior of complete strctoral systems sutected to complex loading histories, such as coupled shear walls and connecting beams 5. Development of new stricturl systoms. The“Yry stack” interlocking block msonry wns described in Chapter 10 are an example Many of these areas of rescarch modeling wil be explored through examples in subsequent chapters. tis well ecognized that research models play an invaluable role in improving knowledge fof structural behavior and thercby pave the way for new and improved design methods. This role will always be important in structural engineering because it isa discipline Founded strongly on physical behavior of real systems made of ordinary materials of construction. 1.9 ACCURACY OF STRUCTURAL MODELS “The reliability ofthe results from a given physical modeling study is perhaps the single mast {important facto othe user ofthe modeling approach. This topic is explored in depth in Chapter 9, find only a few general comments will be given bere to stimulate te reader into thinking aboot this important topic. Adequate definitions of reliability and accuracy ar iffielt to formate. One ‘obvious measure isthe degree to which a model can duplicate te response of a prototype sructare. “The problem met in such a comparison isthe inberent variability inthe prototype itself particularly if it isa reinforced concrete structure. Two supposedly identical reinforced concrete elements oF ‘nvctures will normally show differences, sometimes a high as 20% or mor, and when one must ‘compare « model to a single protoype, the difficulty in making a firm conclusion on accuracy ‘becomes rather apparent. Multiple prototypes and multiple models are needed to treat the results Statistically, but the expense of eve a single test structure i usually high, and the availability of sulicient data for application of statistical tests of signiieance i severely limited INTRODUCTION TO PHYSIOAL MODELING IN STRUCTURAL ENGINEERING. 3 The factors affecting model accuracy include model material propenies,iabrication accuracy, loading technsqoes, measurement methods nd interpretation of results. Elst models ean be ail to give cxtemcly high eorclation with detailed computer-based ests, The ony Tintin i in the cost of properly fabricating and foding the model. Elastic models of reinforeed concrete Stroctares can only predict elastic response and thus will have high accuracy (errors on the order of fess than 5 10 10% for sructones with minimal cracking, such as shell. Carcfully Jesigned and tested strength mel of reinforced concrete beams, frames, shells and other strotores normally have maximum errors on the ordcr of less than 13% for prediction Of poseracing displacements and uitmte foad capacity ofthe structure, provided that bond between ste! reinforcement and model concrete not the governing fctor 11 Dehavion "A etter perspective on the degree of reliably tobe expected in any panicular mode! testing ‘program can be achieved only by caefl study of large nuber of individual cases. The material Presented in Chapter as well asin ber chapess and in the ited references will provide the feader with mach of the material needed in staying model reliability. 4.10 MODEL LABORATORIES “There area number of eutstanding laboratories in Europe that have developed excellent ep tations in physical modeling. There ae n similar conymercia laboratories in North America, but there are exellent structural modeling fii at a number of private and initutional laboratories, incloding Wiss Janney Elsiner and Associates in Northbrook, IL, the Potdand Cement Association in Skokie, IL, and edaationalesearch Iaboratries at Come University, MeGill University the Univerity of Texas, Drexel University, and elsewhere, The Boundary Layer Wind Tunnel atthe University of Westem Ontario an a similar facility at Colorado State University are widely wsed for both vescarch and investigation of wind effects on actual stuctoes. Dysamic tests on shaking tables may be done at laboratories lcated atthe University of California UC), in Berkeley end in Las Angeles, the State University of New Yerk (SUNY) Bula, the University of Mbinois, ‘Stanford University the University of Calgary, Cornell University, Drexel Unversity, and elsewhere "The many diverse problems asocited with stractural modeling make it evident that high quality structral modeling is best done by skilled engineers and technicians at established Iabo- Moles This statement is not made to discourage newcomers instead, t is merely a realistic ‘Comment onthe difficulties of good experimental work. Considerable time aad patience are requited {establishing a diverse stuctoral testing laboratory, whether it be a fll scale ora the greatly ‘clued small move sale. One particulary important point o be made is that considerable amounts Of til and.crror approaches to materials development are now in the iterature, and the carefal {vidos can tke Tull advantage of this material in setting up a aboraory. Thus, advances in “ructural modeling since 1960, coupled with the simplicity and reliability of modern instrumen- ation, havea east partaly eased many ofthe difcultes tobe faced in beginning a new labortory operation, 4.11 MODELING CASE STUDIES ‘Several sructral modeling studies will be sed i this text to help illustrate the many faces of modeling. These have been selecied because they representa varety of structural system nates of consaction, and types of loading. Two design model studies (designated A and E), our research model studies (B,C, D, axl G), and two product development model studies (F and HH) ae atlzed as cave studies, General descriptions are given inthis chapter, and the rest ofthe tnateril given in Chapters 10 and 11. Additional case stues of educational modes are discussed Jn Chaper 12_ Other modeling studies elated to more spcislized topics such as prestressed concrete i “ ‘STRUCTURAL MODELING AND EXPERIMENTAL TECHMIOUES Figure 1.2. Completed ane Won Altings mantarsnee faci at Kansas Cty irnaonal Aapor. (Cour {asy of iss orn. Elsner & Assasin, Mera) pressure vessels, long span bridges, dynamic response, reinforced concrete slabs with penetrations, shell roof stability, concrete masonry structures, and fiber-reinforced polymer (FRP) reinforced ‘composite beams will also be given in Chapter 10, 1.111 Case Study A, TWA Hangar Structures The Trans World Aisines Maintenance Hingar Facility located in Kansas City, MO, was designed withthe aid ofa series of structural mols. The total facility consist of four long span sel structures, each Jeated atthe comer of cruciform shaped framed twikdng. Two completed shes are shown in Figure 1.2. Each shel spans about 96 m and consis of hyperbole paraboloid (abbreviated here as fypar) surfaces spanning tween hollow triangular edge members and an arch formed atthe intersection ofthe (wo hypa: surfaces. Each shell varies in thickness from 75 ‘130mm, Four separate model studies were used forthe TWA hangar to resolve a series of design ‘questions. They are refered to as models Al, A2, A and AA Model Al was a Y-seale shape model of tie entre complex of four shell, Ie was placed in ‘wind tunel and loaded to investigate the general character of wind pressures and wind flow over the sit. Only the shape of the model is needed in this investigation; the model can be built of any convenient solid material such as softwood Model A2 (Figure 13) was another wind tunel shape model ofa single hangar shell strcture, built at scale of Yio and tested to obtain a more-detiled and accurate representation of local ‘wind pressures and flow pattems with doors opened, closed, ete The wind tunnel model results were uiizedin the analytical design phase to help generate the proposed final design. Te third model was thea constructed to determine clastic behavior under ‘many different load conditions. Model A3 (Figure 1.4) was Mosca structure made from pas, 1 was loaded with diserete concentrated loads and with a vacuum loading. The evalistion of bending effects produced by the heavy maintenance equipaent suspended from the shell was of particular importance in this model study. The stress resltats in the shell near the stiff edge members were also given considerable attention. Figure 1.3 Duta wind model of hangar wits doors coset (Courtesy a Ws, anny, ner & Associates, bre, Noches 1). Figure 14 Sasi model enclosed in vacuum kame (Courtesy of Wis, Jenny Eltor & Associates, Ie Nemerook 1) Model A4 (Figure 1.5) was a realistic or strength model. Built ta sale of ie, he mrael as esigned to simulate the tree behavior of reinforced concrete sbellstucture Joaded to fue ‘This strength model requced duplication of both concrete and reinforcement at small scale. The Joad- defection behavior ofthe sell was studied for several loadings before the model was loaded to failure to determine its failure mode and load capacity and factor of ety against collapse. ‘The TWA hangar models were done by the firm of Wiss, Janney Elstner and Associates of Northbrook, IL. 1.11.2. Case Study B, F/C Frame-Wall Structures As past ofthe first joint research effort by the planning group of the U.S.—Japan Cooperative Program on Large Scale Testing, a seven story reinforced conerete frame wall earhquake-resistant bulking was designed, constructed, and tested atthe Large-Size Structures Laboratory, Building Research Institute, Tsukuba, Japan. The plan of atypical oor and the elevation ofthe test building prototype is shown in Figure 1.6, As can be seen Irom this figure, this Wst building consists of seven-story space frame with walls. Inthe drcetion of loading end longdinal direction there are 6 ‘STRUCTURAL MODELING AND EXPERIMENTAL TECHMIQUES Figue 15 Compe microcoetete model and hots. (Couey of Ws, Janey, Eker & Asses, tre, Notrost, IL) three frames, cach one of three ays. A fll-heigtshear walls located atthe midbay of the cemer| fame B. In its transverse direction the test building consists of four frames, each one with 190 bays of 6 m each and canilevering 2 m a each erd. I the extreme transverse frames (lines 1 and ‘bin Figure 1.60, there are full-height shear walls which, as their main purpose, provide torsional resistance forthe bulking. The chosen seven-story reinforced concrete test stricture represents 3 ‘portion ofa building having dimensions common to earthguake-resistant construction in both the US. and Japan, The lateral load-resistant consinction is provided by interacting structural wall and frames. Input from both Japanese and U.S. researchers has gone into the development ofthe ‘esign deus, the planning of the instrumentation, and the conduct ofthe testing of the prototype In buildings ofthis type, itis generally recognized thatthe properties of the shear wall wil really influence the aseismic performance ofthe building. From analysis ofthe results obtained inthe analytical prediction ofthe responses of the prototype, and considering the Tmitations of the University of California, Berkeley shaking tbl, a V-scale model of the prototype structure ‘was chosen A detaed dsctssio ofthis selection andthe final design ofthe model structure are siven by Haris and Bertero (1981), Hass etal. (1981), and Aktan etal (1983). The maximum capacity of the 6.1 x 6.1 m University of Calfonia, Berkeley shaking table is 490 KN. Based on these limitations and the desie o test the mode under very large deformations and thes enable the study ofthe Failure mechanisms of the building, a Y+-scale reinforced microconcrete mode! of the prototype structure was chosen; see Figure 1. ITROOUCTION TO PHYSICAL MODELING IN STRUCTURAL ENGINEERING ” a DIRECTION OF (GROUND SHAKING @ @ sua 472" Fas wages Aas less £ eaauec Loiseo eso} ie G@ PLAN WITH FRAME AND GIRDER LOCATIONS. —cIRDER 1179" £19.69" ~ SLAB, ree 6 O08" COLUMN, 19,.69"x 19.69" - SHEAR WALL t 787" 12.30" ieee bieso-t ise" 4 (©) SECTION SHOWING FRAME B Figure 1.6 Pan ond cacon otal tome il rho concele suc The design of the model was chosen to comply with similtode reqirements, in adtion to shaking table Kinittions (Monkarz and Krawinkler (1981). ‘The -scele model chossn forthe Us_tapan program stiles similitde with regard to geometric and loaling parameters and also omplie with all material requirement except forthe mass density. The detailed mechanical Charncterisis of mol matedals and any discrepancies are documented by Bertero etal. (1983; 1984), The mass density is augmented by means of Fed ballast weights attached (0 the roof and 1° ‘STRUCTURAL MODELING AND EXPERINENTAL TECHNIQUES INTRODUCTION To PHYSICAL MODELING IN STRUCTURAL ENGINEERING Figure 17 View of vscale malo the ULC Rrhy shaking bio all oor sats in sucha manner so aso case lite on nence othe sca tiffs, a ‘wel sno significant increase in the mass amet of nero he las (Figure 1.8) In aon, i ded no major changes in the damping characteris of the sete. a Besides th intemal ore transducers t might of the atte columns, the model was instrumented to record overall espns and lea teavioaf thecal ions of tht severe stained members nema and external gages measure average sin along the main reinforcing ‘ars anda the surface ofthe microcnerete were laced at thecal eons fst members Fave 18 One tse mee wh roma ala) Dn fat on or bs) Displacement ansdacers were lcaed at each foe of the model and connected to ste lees ‘Stacort fad gi sc oben othe mol (0) frame placed outside te shaking table fend overall lateral response ofthe mode. a Eo [STRUCTURAL MODELING AND EXPERIMENTAL TECHNIQUES ‘The t.scale model ofthe seven-story RIC test structure tested on the Unversity of California, Berkeley cathquake simulator is one of the largest and most-dtaed sties undertaken thus fat to stdy the earthquake resistance of structures under dynamic leading conditions. Pthaps one of the most important objectives ofthe esearch conducted onthe Yscale mode ofthe structure was to evaluate the reliabty of experimental analysis at all he seismic response Timit states, izing such a smallscale trac replica mode ofthe tes structure. The most important difficulty inthe tinment ofa tre replica model, with the same strain response history 26 the full-scale stractare ‘when subjected to similar seismic effects, was satisfying siilitude requirements forthe response charscersis of the consitent materials ofthe model treture, Deals are given in Chapter 11 and a discussion of some of the consequences ofthe limitations in attaining perfest similitade in ‘the constntve relationships ofboth ste! and corerete as well as their compote action, i the ‘bond characteristics, 35 reflected on the serviceability, damageability, and collapse limit state ‘responses of te structure “The research desried in Case Stady B was cenducted atthe Farhuake Engineering Research Center, University of Califomia, Berkeley under grant CEE-8009478 from the National Science Foundation, 1.11.3 Case Study C, Reinforced Concrete Bridge Decks ‘A research program to siudy the ultimate and fatigue strength of noncomposite reinforced ‘concrete bridge decks under pulsating and moving loads was conducted at Case Wester Reserve University by Perikars and his associates (Perdikaris nd Beim, 1988; Pedikaris etal 1989, Perdikris and Petru, 1991; Peo etal, 1994, 1996). The direct modeling approach was sed to study atypical 15 2$-m-ong protetype highway beige witha 216mm thick reinforced concrete deck supported on four W36 x 130 steel giles spaced at 2.13 or 3.05 m. A series of 17 fall ridge deck ‘Yo scale direct moses (Figure 19) and sis individual panel at Vs sale were tested in his program. “The objectives ofthe research were te obtain a beter understanding of the fatigue response of noncomposte enforced concrete bridge decks an to presen description and rational explanation ofthe various load transfer mechanisms present inthe deck during its fe expectancy. Smal-seale ‘model concrete deck slabs reinforced according to the current American Assocation of Sate Highsay and Transportaion Officials Code (AASHTO) (1989) (ordwxtopic reinforcement) and the Ontario Highway Bridge Design Code (OHEDC) (1983) provisions (isotropic reinforcement) ‘were subjected to static and moving constant whec-load inthe direction of trafic. The tet setup ‘hoven in Figure 1.10 was developed for these stades. close-up of the moving constant whee!- load testing a W-scale pane is shown in Figure .11. Also, fatigue tests under a pulsating load ‘applied a fixed point were conducted to determine the eect of the load movement on the fatigue strength and level of deterioration ofthe deck sibs. The determination of the endurance limit of concrete deck slab reinforced with ether “isotropic” steel reinforcement according to the OHBDC Code or “orthotropic” stel according to the AASHTO Code and subjected to a moving constant wheeLload would give a better estimate of the actual safety factors aginst Gist cracking, steel yielding, or deck punching failure and the tru level of conservatism for the two bridge design philosopi Several important parameters were stdied in this research The effects of the Mexural steel reinforcement ratio, steel sider spacing, rotational and lateral restennt of the deck, and the type and level of applied load on the cracking patie, stifoess degradation, faigue life, and fallore mode ofa concrete bridge deck were investigate. The critical factors contibuting (othe Fatigue failure of concrete bridge slabs were determined, The experimental load-defection static response, fatigue strength under diferent levels of moving constant whee-load, measurements of strains {stce reinforcement, concrete deck surface, and steel girders), cracking patterns, and the flare mode were all studied. The presence of any size effect was investigated by comparing the est results forthe Yee and scale models. These sults ae sommsized in Chapter 10 INTRODUCTION TO PHYSICAL MODELING IN STRUCTURAL ENGINEERING a Plan View pe tnt Section AvA ee = ee ene opinae a @ Plan View ~—t o Figure 1.9. Model at scales of hs ands of noncomposie decks on ste! gir bridges. 3) Type (se Shine rac Jock pane se ge eck mod pon [tine Stem) 6) sens te sea SSS and Poo prope or pang oF 10 “he pching sean was fo toe intertated wih he inst of wo-diensiona toroe-hinge compressive rigid-trat model. The compressive arching ation mechanist in t ‘concrete eck transferring he Toad tothe stel girders s modeled by the compressible stat members ofthe trss, while the lateral re the deck is depicted by elastic horizontal springs. 1.4 Case Study D, Lightly Reinforced Concrete Buildings ie research effort conducted by the National “The case study presented i pat of a comprehensh ch effor Center for Earhquake Engincering (NCEER) onthe damage assessment an performance evaation 2 ‘STRUCTURAL MODELING AND EXPERIMENTAL TECHMQUES Figure 110 Ruston of he moving constant weblad setup. (1) Elactie mor and contol peck (2) ae onder ump) Pyrat pump) presue edrer wheel sen cao (©) presse cl gap. 7) cng var: ) hye ak (0 pahasetane bonded sel whee. (Go) reseton soe ama (1) computer ata segues stem Figure 1.11, Moving constant wteeioad setup (scale whe siz) of nonssismically detailed buikdings during earhquakes. A. sale mode ofa tightly reinforced three-story office building tested on the shaking table at Cornell University EF Attar eta, 1991) and a Yscale model ofthe same building test on the shaking table at SUNY Buffalo (Bracci etal, 1992). The building was designed and dasiled to reflect the common design an practice feature of the Central and Easter U.S. ducing the period 1950 to 1970. Test results from these studies are evaluated and compared with numerical results to investigate the reliability of availabe analytical ool in predicting the response of thistype of building. general layout ofthe prototype building is shown in Figure 1.12. The relative dimensions adapted forthe frame members ofthe idealized prototype stricture were based on a survey of typical construction practices and the Timitations of the NCEER shaking table. INTRODUCTION TO PHYSICAL MODELING IN STRUCTURAL ENGINEERING 2 (Beaten (0) Sieven (oP, Figure 1.12 General yout o he tree Ighty einorcd coveete tulsa Realistic reinforcement details that reflected construction practices during the 1950 to 1970 petiod were adopted forthe models tested. Special attention was paid to critical deals such asthe tbeam-column joints and column fap splices. Figure |.13 shows an inerior and an exterior now seismically detailed join, What characterizes thes joints aro the Following detail: (1) discontinuous bottom beam reinforcement with a very short embedmentIength(150-mm), (2) no bean or joint confinement except for several stops ata spacing of 7S mm (usually; located 75 mim below the ‘beam bottom fae, (3) no joint reinforcement (column stirrups discontinsous through joint length), By ‘STRUCTURAL MODELING AND EXPERIMENTAL TECHQUES eh (0) Interior Joint (0) Exterior Joint Figure 1.12 Benm-cokimn ont deta of he te sexy prototype bul (4) a compressive lap spice that is very short ard Tocated immediately above beam top face, and (5) no special confinement required forthe Lp splice. AI the above features are showin in Figure 1.3. “The Vescale model quired adkitional balls loads in the form of lead bricks symmetrically placed onthe floor and roo slabs as shown in Figure 1.14. The mounting technique was designed 50 thatthe center of mass ofthe lead bricks coiscided with the slab midibickness and minimized theit eccentricity. Bots 13 mm in diameter passing through 19-mm holes held the lead bricks 10 the slab, A set of ses! washers was installed between the slab and the lea bricks to inimize slab stiffening (inser, Figure 1.143), The "scale model is shown on the Comell shaking table in Figure 14 ‘A scale model of the same critical podion of the Tighly reinforced concrete building (igure 1-12) ws built and tested on the NCEER shaking table at the SUNY Buffalo Earthquake ‘Simulation Laboratory. Thee views ofthe Yé-reale model are shown in Figure 1.15. series of varying intensity simulated ground motion teste were performed on the scale building, model using scaled acelerograms on the shaking table 1o represent minor, moderate, and severe earth- «quakes. More deals ofthe Comell and SUNY Buffalo model studies are given in Chapter 11 1.11.5 Case Study E, Prestressed Wooden Bridges Recent concern regarding the decay ofthe infrastructure ofthe nation has focused attention on the need for replacement of many of the mations bridges. OF particular concer, because of the ‘ast quanlty, ate Bridges with spans of less than 50 ft. These make up a majority of America’s bridges (Oliva snd Dimakis, 1987), Many of thes ridges are made of steel and enforced concrete and were constricted in the carly and mid part ofthis century. They have performed forthe most part satisfactoily during their etme: however, they are approaching old age and are in need of repair or replacement. Repair and replacement canbe quite expensive considering the large number fof bridges, and labor an material cost, It is obvious that a rapidly constructible, inexpensive replacement bridge would be desirable. Wood is becom |TRODUCTION TO PHYSICAL MODELING IN STRUCTURAL ENGINEERING Figure 118 One aighih eae mods fey att {b) eed onthe Cora testy shaking tbe) Balas oad st) STRUCTURAL MODELING AND EXPERIMENTAL TECHNIQUES Figure 1.15 Onetnid szae modo! 0 the SUNY Bue NCEER shang ab tenfrced concrete ste esplcement brid. 5 wis. Some wig adntages of wed are pie, {valailiyofshor- and meun-lengh tinker resistance salsa in sn eva oe of ea fe ft tha eae mons nn sd he ve joven, sch as potential dea) hg longer ecep dlectons ow tenth antics whey Compt thes conte nt eo ine ai vey age en mise tase penning ne arate vn or an teal whe akingatbantge of sale sitd members. Orgad in anal inte ae 190s the tenia esn used sucess in seve st spn es in bth Ca an the US (Ontario Highvay Bridge Design Code, 1983) Most ofthe esach and apo he US and Canad wanes pst terioned wood bridge decks hae focused on sated spose An inesting design concept ons Noth Arian hod species, Speticallysbrcoay INTRODUCTION TO PHYSICAL MODELING It STRUCTURAL ENGINEERING a available Pennsylvania rod oak was used in the replacement design ofthe Samuel Road bridze “This is located in West Whiteland Township. Chester County, PA and involves the replacement of severely deteriorated 67I-n skew spon stcl girder and reinforced eonerete deck bridge. The tlesign was carried out in accordance to the Ontario Highway Bridge Design Code and AASHTO ‘equiemeats. The primary exception is thatthe anchorage bulkhead channe! hes been replaced by individual square stct plates 2 the anchor pots. “Two identical Yescale models ofthe Samuel Road bride design (Figure 1.16) with the skew angle eliminated were tested 3 Dresel Universit inorder to verify the design using loca creosote impregnated red oak (Hoffman, 1980), Model E1, shown in Figure 1.17, was used for dead load testing and for sess relaxation meastrements. Live loa simulation tsing the AASHTO HS-20 truck loading was used to sty the behavior of test Model E2, Model F2, shown in Figure 1.18, ‘vas tested elastically and then loaded to ultimate failure. The elastic behavior was compared with ‘an FEM analysis. Results of these tests, conducted inthe Structural Models Laboratory, Drexel University, are presented in Chapter 10. 1.11.6 Case Study F, Interlocking Mortarloss Block Masonry [An ineresting and novel application of the direct modeling techrique that has far-reaching implications in industrial structural product silostated inthis example. It involves the use of faithfully repduced exemplar models in a prodet development sty of new block masonry. Product development involves a large numberof engineers from a varity of disciplines including structural engineering, materials, and construction. The search fr building procedures that are more rapid and less dependent on workmanship has led © the need to develop “dy stackable” block ‘masonry wits which canbe lsd without morta. Introduction of interlocking or dy stack mortarless tasonry systems in fenforced block masonry consruction requires te development of efficent, sj-to-hanle, and yet versatile blocks. Two promising imtertocking block types were developed at Drexel University (Haris eta, 1992, 1993; Oh etal., 1993; Oh, 1994) for application to feinforced masonry construction including earthquake-esistant structures using ¥-scale direct ‘mores. The main aim inthis development was, Fist 10 equal or exceed te structural performance of conventional masonry systems and, second, to provide a more economics nd atonal solution for the masonry sytem thus leading to more-competiiv designs. The two interlocking block units developed are designated the modified H-Block and the WHD Bleck. Productior of the units (ciated by "seal eduction) was followed by strength and stifnes evaluation under compressive, bending, and 2 ‘STRUCTUFAL MODELING AND EXPERIMENTAL TECHNIQUES S- smutate TavcK LOAD *SPRESTRESSED END surrorr Figure 138 Modal E2 during oe load teeing shearing loads. All comparisons ere baslined fo conventional mortared masonry construction. ‘Analytical models were developed to predict betsvior under load The frst ofthe two interlocking masonry systems investigated was the modified H-Block shown in Figue 1.19, Tis isa simple open-ended block whit with tongve-and-aroawe interlock on both the bed and esd joints. Tt can be reinforced in ths vertical and horizontal dretions (Figure 1.19) {VTROOUCTION TO PHYSICAL MODELING IN STRUCTURAL ENGINEERING ~ ae © » @ Figure 149 Moda Hl mascry system. a) Stor) nomial iencons ulsot: (i rae eesang 7 taney bond (2 top sew showing arrent of ces (wa ry ‘Skoog nang bone and is particularly suited ro earthquake resistant construction because it canbe easily placed around orice bas although vericel threading is required. Figre 1.196 shows the dimensions of the fll ‘fae protorype. The main advantages ofthe modified 1-Block wit are case of vertical and horizontal reinforcement, case of unit cell verical alignment, minim tapering of cells, main area of block ‘Concentrated inthe desirable face shell, optimum cll shape for ease of grouting, and high resistanes {water penetration duc to raised tongue and-groove bed joins. These ators are ilustated in Figure 1-19, The problem of controlling the height of units to achieve accurate running bond can bbeachicved by means of specially designed block molds. The grooves can be strengthened by Droper mix design and vibration during production. Dering transportation and handling, simple forking tecinigues ean be developed in packaging 10 minimize breakage, To solve the problem ‘St atabity during construction prior to fll grotig. either preventive patil grouting or extended bracing canbe wed. ‘Te stacking of units in runing bond patter and fe reinforcement placing is shown in hollow wall Figure 119e) and when te wall is fully grouted in Figure 1.19, The hnning bord deystacking technique i lutrated in Figure 119 and f “The WHD Block was developed independent Drexel University and wasstied using ¥-sale ict modeling technigue. Units of the WHD Block system are shown in F gure 1.200, The prototype ‘dimensions are shown in Figure 1206. Reinforcement in th vertical and horizontal directions and ‘Sacking of te bollow nits re illustrated in Far 1.20. Significant horizontal and vertical bending Sita s obtained inthe staggered joints ofthe dey-stacked system (Figure 120e), Fall grouting of a vali lasted in Figure 1-204. The freshly finshed model blocks are shown in Figure 1204. “pocly designed block making machine Figur 1.208 sdeseribd ia Chapter 6. Details ofthe model block strength and tfiness characteristics are given in Chapter 10. “The research described in this case study was conducted inthe Structural Models Laboratory. Drexel University under grant No. MSM-9102769 from the National Science Foundation. Ad3i- tional cases of dzeet models wed in product development are given in Chapters Sand 10. 0 ‘STRUCTURAL MODELING AND EXPERIMENTAL TECHNIQUES. (0 Top view showing alignment of els (0 Wall" ary staked” in running bond Figure 139 (continue 1.11.7 Case Study G, Pile Foundations Pile caps form an important clement between the column and the piles hroveh Which the load {is transmitted to the soil Often, it i overlooked both by the structural engineer and the soils engineer and is designed somewhat empirically, since the code novisions in many cases eannot be used diretly. The problem is one of a desp slab (¢wo deep beams) and has to be handled ‘lifleretly than the coe provisions. "he pile ea study presented here is based on research cared ‘ut at Howard University (Idowu, 1979; Ndukve, 1982; Dagher, 1988) and atthe University of Poerto Rico Vimenea-Perer etal. 1986) and reponted subsequently by Sabnis and Gogate (1980), INTRODUCTION To PHYSICAL MODELING IN STRUCTURAL ENGINEEFING ” » Figur 1.20. YO block masonry stem. (a) Typ of unt) nominal mensions at fl cls snore ‘mont: ()sasing () th mal exracingcxmeed moe! Bork) completed mosel Becks. and Sabnis and Dagher (1989). The overall problem of structural sirengh assessment was tackled through various topics; these include the present practice of pile cap design and the experimental testing program using strctural models designed to cover a wide range of parameters, with particular emphasis onthe variation of reinforcing steel. Two scales for testing were used at tW0 ‘miverties and comelated to demonstrate the universal approach of model studies, The behavior of ple caps was studied by varying the amount of rinforcement, while other parameters, such as concrete strength apd steel depth, were considered secondary based on the cafier research work. Furhermore, de to the difficulty of making larg specimens and handling thom withthe limited available resources, it was decided to use sciled models (Figure 1.21) Avoughout his test program at Howard University. scale was used to suit the facity at Howard Universi and the availabe reinforcement. The scale was large enough er ene person to handle. Larger Specimens at Yas scale were tested a the University of Pvrto Rico Jimenez Perez et al, 1986) ‘The results of the testing program and their implications on cole revisions are discussed in Chapter 10. 1.11.8 Case Study H, Externallyfinternally Prestressed Concrete Composite Bridge System ‘This case study imolves the development of a new stctoal system for highway bridges by eliminating corrosion and the need of shoring and formwork during construction, thus lengthening the bridge life and cuting down on expenses, The system of the main feature is the use of precast ‘modified double-Tee (DT) panels which are prestressed with internal carbon fiber-reinforced polymer (CFRP) strands. A cast in-place deck slab reinforced with CFRP bars is connected tothe DT girders though shear connectors and an epoxy bonding agen. Externally draped post-tensioned CCFRP strands are used forthe longitudinal post tensioning. The varioxs components of the new big system are shown in Figure 122. The new bridge system consss of @ SSIRUCTUFAL MODELING ANO EXPERIMENTAL TECHNIQUES (©) The mold extracting completed model back. (9 Completed block. Figure 1.20. (contin) 1, Precast modified DT girders, prestressed mterally with CFRP strands. 2: Post tensioned cross beams cast monoiticaly into the DT pind. 5. Castin place CFRP reinforced deck, comected to the DT girder with CFRP shear con- 4, Extremal post tensioned draped CFRP prestessing strands “The standard DT girder is modified with thcross beams to accommodate the draped external strands ad to enable transverse presiressing ofthe DT girders. This allows several DT girders to INTRODUCTION TO PHYSICAL MODELING IN STRUCTURAL ENGINEERING = (eye) aot la [ex ese [ee Cal = Lal Figure 1.21, Daresions oc pl cap mode be placed adjacent o each other and te together with CFRP prestressing strands, The unbonded posttensioning inthe vansverse Gietion facilitates fture widening of the bridge system. Tous ty scale experimental moves of the modified DT concrete girder bridge system utilizing ‘carbon fiber reinforeing bars and prestessing strands were constructed and tested at Lawrence Technological University, Southfield, MI to determine their performance under varios loading Conditions (Grace and Sayed, 19964, b; 19973, b; 1998). The large size of the models chosen Stlowed the use of protorype materials throughout, thos greatly simplifying dhe modeling process ‘The test loadings for cach model consisted of saicldynamic fatigue (7 million cycles) an white Srength, Two ofthe est structures were igh-ange o stight bridges and two were skewed aes, “The dimensions ofthe four model bridges are shown in Figure 1.23. ridge model DT-30 (with 30" skew ends 8 shown inthe laboratory undergoing fatigue testing (Figure 1.24). The widest of the foor models, Model DTE-IS (1° skew) is shown in Figure 125 endergoing external post- tensioning Results ofthe testing ofall four moses are discussed in Chapter 16 o STRUCTURAL MODELING AND EXPEFINENTAL TECHNIQUES ata Cen @ Dovater ~ ore Canine eta —/ rs Beam @ suppor —- / ) Figue 1.22 NonconosheOT bridge syst (a) Thee dinentona vow of DT brid sys) components st OT bide Sytem 1.12 SUMMARY Physical models of structures are used in edhcaton, in research, in design, and in product and ‘concept development. A model can be built and ested at a small fraction of the cost ofa prototype (full-scale) stretue because of the great reductions in loading magnitudes and in construction ots. Many diferent types of models are used, and the cost and time requiements vary widely for each type. ‘The applicability of models in design applications changes almost continuously a8 improved anata capabilities permit the engineer o model mathematically inereasngly complex strvctinea, but the development of new combinations of stuctural forms and materials Will mos likely alwayy be one step abeal of analysis. The need of experimental verification wil, therfore, alwaye be necessary to help ws understand the behavior of complen structural stems Experimental sess analysis andthe general principles of experimentation form itera pats ‘of the modeling process. The execution ofthe many steps volved in a stturl mal study fs fan enginering project in itself; hence a considerable degre of “at is imlved slong with the rather well-developed technology given in this book. INTRODUCTION TO PHYSICAL MODELING IN STRUCTURAL ENGINEERING 36 20 — ener once EE / ties Pie | Feed mms 9 at Figure 123. Delis of ested brige made, STRUCTURAL MODELING AND EXPERIMENTAL TECHNIQUES. Ft te oa OT sew. (Catsy Po Nt Ge, Lamers Figure 125- Exual potaneing dg mel 715 (5 kon Ca T1515 how. (Comeay Pa Hab Grace, Lemons Teehnotpea! Ore Sa, i) . “ PROBLEMS 111 Refer Case Stay A. he shops of hi ot ely at 1 oof was completely ft, hos doy think the wind Toads would fer te wind wode ns tested in 3 wind nel? 1.2 What are the main factor for using smal scale modesto cond esearch on sroctral systems ata smal university with ery tinted resources? INTRODUCTION TO PHYSICAL MODELING IN STRUCTURAL ENGINEERING aw 12 imagine tht you ad at your disposal several straight piozes of ste, aluminum. wood, nk Pleiglas rectangular sections of small size and various fengths, Buckling of pin- seal onal column ean be demonstrated by being the cola verically on @ smooth Fonzontal surface and pushing atthe top with your and, increasing the force gradually tnt the column bows ‘Suppose you did not know the Euler buckling equation tat govems the behavior of the imually straight, pin-ended axially leaded column but you had at your disposal @ flee of appropriate curacy that you can ntsdce Between your ind and the boxtom ti he colar to measore the load. How woukl you setup an experiment to dennonstrate ‘hat he buckling Toad varies versely a the eng ofthe coluinn squared and dicey tits Youngs modal and its moment of inertia? {14 Whats the most dificult problem that nus be overcome in executing an accurate small scale model sty ona shaking table? 1.5 The old City Hall Building in your town needs to be upgrade to comply with newer, Inore stringent requirements concemsing earthquake loadings. Being a very impertant fstrca building constructel of unreinforced mason, discuss how a posible program, ising smal zcale models, could belp the engineer in charge of the project to strengthen the Building and make it more earhake resistant 1.6 he new construction materials are developed for stronger and Viger structures, their Tabortory evaluation must precede any fed applications. Think of situations where the testing of new constuction maferals on small-scale sirctaral modes can stv time and money and speed their practical introduction ito construction pracice 117 One ofthe current urgent strtural needs i the U.S, and other snustilized counties fetbe repair of the deteriorated urban infrasracture highways, bridacs, buildings tonnels, (en Te accomplish this in aneconomicel manner, de use of many neve reps techniques (Gurctual concepts and/or materials, et) will be necessary. Think of ways in which the (irl scale modeling technique canbe used wo determine the structural safer, reliability, fd durability of such tepaiemeshods ina timely and economical manner. REFERENCES AASHTO (1989, Spctfcvons of Highway Bridges, 4h ed, Amica Associnion of State Highway ard Tanpertaton Ofcil, Washington, DC. ‘Attn, AEs Bonero, VV. Chowan, A.A a Nagashima, T, (1985). Experimental and Anaytcal cities of te Mechanical Charaterigis of 7-Siery Scale Mode! RIC Frame Wall Busing Free rake Eaginering Resewch Center, Report No, UCBIEERC-8¥13, Univerity of Cah Fai, Bekele Dann 6 (197), Desig of ike Caps, Cena Biking Reser lsat, Roorkee nia, March rane Gt G9%2) An sccrtepechancl soln of tay dtr srctres by we of paper uals an special gages, Proc. Am. Cone st, 18 58-82 eae 6B. Ticy, KE and Bill B. (1932). Suspension bridge ses determined by mel Ent News Record ane 9 Bence VAL Aftan, A E., Haris, H.G, and Chowdhary, A.A, (1983). Mesonical Chareceisties of tural Uued in Scale Mode of Story Reinforced Conercte Test Stace, Eartguaks Ena waren Research Centr Report No, UCTEERC-821, University of Caomio, Berkeley, Oster Bence Ws aktan, AE Charney FA and Sas, ,(1984)-US.lopan Cooerasve Earthquake Reseach rts arhake Simulation Tess and Asia Stes of te Scale Noel ofa 7-Stry Reinforced Coecte Test Suture, Earhguke Engineering Resech Cente, Report No, UCB/EERC 8415, Uni verity of Califor, Berkely, une Blew ad Ficmy, R967), Sur pcos Ann. Tck Bat. Trax Public Pars), 20, bri 8 “STRUCTURAL MODELING AND EXPERIMENTAL TECHIQUES Brac, JM, Reior, A. Mand Mand, JB. (1382). Seine Resisance of Reif Concrete Fane Swetures Designed Only or Gravity Load: Pat I— Design and Proper of Oni Sale Move) Swetre, Technical Reprt NCTR 92-0027, National Cemer for Eagake Engineering Rescarch, SUNY Mutao, Decenber ‘Bull AH. (1930) new method for dhe mechani nays of tases, Ci ng. (2) 181-183, (Cla 1-L. (1973) Behavior and Design of Pile Caps with Four Pes, Techil Report No 2.49, Comeat nd Concrete Association, Wexham Spins, Slogh {RSI (1986, ile caps fer indvida! columns, in CRST Hondbook, Concrete Reinforcing Ste lsat, Chicago, TL, Chap 13 ‘oar, R. (1988). Ivesiion of Thick Pile Caps ME. thesis, Howard Usiversiy, Washington, D.C. rete University (1998). The Drese Engineering Curculun. College of Engineering, Low Engineering Buiking, Pidgin. Lat, AG, Whit, RN, and Gergely. P (991), Shaking Tobe Tet of 2 4Scal Thee Stony Ligh Reinfored Concete Building, Technical Rept NCEER-PLODIS, National Center for Easke Research, SUNY Bull, February 28, EL-Atu,A.G., White, R.N, and Gegey (1999) Behavior of grvty lad designed reine concrete boikings subjected wo eiibguakes ACT Set J, 942), 133-18. ney, WJ. (1939). New deformeer appara, Eng. News-Recon 1227) 16 221. Gotweaik, 0. (1926). Mecha allan of ate syste, J Fullin Ps, Jy. ‘Gres, N-F. and Sayed, GA. (19963). Doale ee aad CFRPIGER bridge system, Cone Int, Ameican Concrete Insite, 182), 38-14, (Grace, N-F and Sayed, GA. (1986). Feast of CFRP(GFRP peste concrctedemonstaton bridge in the USA, paper presemd as Founh Nations! Workshop on Bridge Research in Progress, Nation Cents for Eathguske Enpincering Research, Batale, NY. (Grace, N-F and Sayed, G. A. (19973) Duct of pestestd conc bids esng intemallestemal CFRP sans, paper present at Seven Intemational Confrence & Esti, Sra aul + Reps 9, singh, Scand (race. Fad Sayed . A (1997). shaver of eemallyincrnly pressed coer composite idee system, pope presented at Thi Itenatoal Symposia on Now Meta (FRP) Reinforce for Concrete Structures, Sapper, Jpn, Oot. 14-16 ‘Gres, N-F. and Sayed. G.A. (1998), Dciiy oF peswessel concrete bilge singintemalexteal CFRP stands, Con It, ACL Je Has H.G. an enero, ¥.¥ (1981). One-Fith Sole Mode of 37 Stry UC Fre Wall Bilng under Earthquake Loafing, iouse rept, University of California, Berkely, August ais, HG. and Sabi, G.M. (1996). Physical models fr concrete svctes — thet role inthe nex ‘erancering cium, per presented atthe ACT Fall Convention, New Oran, Nov. 3-8 Manis, HG. Bertero i, and Clough, R.W. (HSI). Oneith sale mal of a seven sory reinforced ‘oncreeframe-wal bing under earthquake loadns, Proceedings afte Joint Sract. E/B.RE- Imernaional Seminar on Dynami Modeling of Sructres, Bung Resextch Sesion, Garson, Watford, England, Nox 19-20 artis H.G.,Ob, Ko and ami, A A. 1982), Deecpment of nm nerocking and morules block masonry units for ecient bldg systems, in Proceedings ofthe Sh Cnaian Masonry Symposia, Saskatoon, Canada, June 15-17, Harris, H.G, Oh, K. and Hamid, AA. (199), Development of New Imerocking and Montes Block Masonry Units Improve the Earthquake Revistas of Masonry Constucton, Final Report othe National Science Foundation under Grant No. MSM-9102769, Departs of Chil and Arteta Engineering, Dreve Unive, Pitelphin ‘Hobbs, NB and Sti, (1957). An imestgation fo the stress ditibton in pile caps with some notes ‘on design, in Proceedings ofthe Istintion of Cl Engineers V7 a ‘ona, N.S: (190), Betav of Transversly Pesessed Wood Bie Decks unc Simulated Short Term ‘snd Long-Term Tac Loan, Me. hess, parent of Civ amt Archiectral Engineering, Drexel Univesity, Philadelphia, Mech ow. (197) Imei of Thiek Pie Caps Reseach Post Rept Howard Universi, Washington, DC. anes JR. Breen, JE and Geymayer, (1970), Use of msi srctralengneting in odes for Conrete Stace, MC SP-2, American Coast Insite, Det, M19. 18. |NTRODUCTION TO PHYSICAL MODEUNG IN STRUCTURAL ENGINEERING. 2 Simener-Pee,R., Sans, G.M. and Gogte, A.B (1986). Experimental bevoe of hick pile caps, paper ‘resented at the Strvtral Assessment Sema at Londo, ‘Kinney. (195) Indverminare Structural Analysis, Nason: Weskey, Reading, MA, 655 pp ‘The Mins, (1983), Ontario Ministry o Transporation and Commonicaice, Ontario Highway Bridge Desen Cove, 2 ed, Downsview, Ontario, Canada Moar, M.W. afd Hafield LP (1983). The design and constncton of defommeter fr se in motel ‘analysis, M.Sc. thesis, RenlarPolyecae Insti, Toy. NY. Monkarz PD. and Keavialder, H. (1981). Theay and Application of Expimental Model Analysis in Eathquake Ensioering, Rept No.5 the Joh A. Blume Earhqanke Engiacring Center, Department ‘of Civil Engaceig. Stanford Universi, CA ‘Nauk, A (1982) Comparison of Method of Analyst an Design of hick Reinforced Concrete Pile Caps, Research Projet Report, Howard University, Washington, DC. (Ob, K-HL (1991). Developmen and Investigation of Failre Mechann of leeching Moxtaess Block Masonry Systems, PhD. hess, Deparment of Civil and Architect Engineering, Doel University, Pile, June (Ob, K, Haris, HG, and Hamid, A.A. (1993). New interlocking a morass bck masonry units fo anuake resistant consti, in Procedings of th Sth Nort American Masonry Confrence, une 6-8, Piadetpia (Oto DOT (1947). Imesigion ofthe Stength ofthe Connection betneen a Cee Cp and the Embedded End of a Size H-Pile Research Rept No. I, Ohio State Departs of Highvays tiv, M.G. and Dimas, A. G, (1987), Bebavor of Pos-tensoned Wood Dig Decks Pull Sale Testing, “Analytical Corton, Design Gidlines,Suctrs and Materials Test Laboratory, 87-1, University of ‘Wisconsin, My 150 pp, Pena, PC and im, 8. R (1988). RC trig decks user puting and moving load ASCE. J. Strat Eng. 11403) March, 391-47 Pela, .C. an Petr, M, (1991), Coe prictions vs. smalscale ridge ceck molest measurements, ‘Transporation Research Record No. 1230, TRD, Vl 1, Bridger & Srutres March 10-13, 179-187 Pena, PC, Bei, S-R. and Bousas,S (1989). Sab contin effect of uhimate ant Tague stent Of RIC be deck model, ACT Since J, 844, Joy-Avgest, 483-191 ery, CC. an Lise, H.R (1958). The Sian Gage Primer, Mera il, New York Peru, (199). Behavior of Concrete Bridge Deck Models Subject o Concenrated Load — Ontario ‘i AASHTO, Mse_ thesis, Case Westem Reserve Univer, Cleveland OF, Ms Perou, MF, Perdis, Pan Wang A. (1994) Faigoe behavior of non omposite reinforced concrete ‘ridge deck models, Transporation Reseach Recor No, 460 Bridger aud Siractures, December. Paro, M. E, Penta P.C, and Duo, M, (1956), Sate behaior of on compost coneete ridge docks ‘odor concentsted ha Bridge xg, ASCE. (Quinn, R199), The mathematical and sckeniie foundaion for an ieee enginecingcurculum, Eig. Edu, Avpst Rage, A. C-(1943/). Th Bonde Wire Gage Tague Meter in SESA Proceedings, VI, No.2 Rue, A.C. an Schnih F-O. (1939), Mechanica sirctral atalysis bythe ment inden, Trans. ASCE, TOA: abo Pre: ASCE, 65, No 1, Janay, 16-170 and No. 6, Dune, 103"-1040. Satis, G.M. and Daghe,K (1989), Investigation of reinforced concrete ile caps, in Proceedings of One- ‘Day Conference on ifeSirucures, Brion, Enea Sains, GM. and Gorse, A. B. (1980) Ivesigation of thick ple caps ACT J, 77), Jan-Feb, 18-24. Sovage, JL. (1934), Dam stesss and strains std by sce models Eng, News-Record, Dee: 6 720-72. Schuring DJ. (1972 Seal Models in Expinceing Fundamentals and Aplications, Pergamon Press, New ‘Yer ‘Whine RT and Beatie, D.(1972), Stadadpleeap and Concrete (0), onary 4-36; 6(2), bray, 28-30. ‘Yon HT. (195). The design of pile caps, Chl Eng, Public Works Re, May and Jue CHAPTER 2 The Theory of Structural Models CONTENTS 2 nteodaction a2 22 Bimeson a Deon Honey 42 23° mensional Analysis... . 45 1 23.1 Buckigham’s Pi Theorem 49 232 Dimes Inepenec nd Foraton fF Te. “$0 23.3 Uses of Dimensional Analysis. 32 F34_paonel Considerations In Using Dimersional Analysis co 8 : 24 Structural Modes. . - 36 SE Modes with Compt Similarity 36 342 Tecnologie! Dieaies Assocsed with Complete Sinan 38 242.1 Other Types of Distortion 39 22 Relaxation of Design Requitement vm 0 24.3 Mode ih ist Oder Simian : . a) 244 Distorted Models. —— 6h 25° Similitude Requirements oe e 251 Reinforced Coneree Models ae B i 25.2 Models of Masonry Strctores 61 35:3 Sutures Sbjeced 1 Termal Landings onsen 6 25:4 Structures Subjected to Dynamic Loaings 70 234.1 Inroduction ~ =) 2842. Vibrations of Elastic Stactues Ss 0 2543 Fideasic Models 7 uB 2544 Blast ant Impact, Load Modeling - 7” 1. External Blast om Stretes cn Oy 2 oteral Blast Eset. a 15 254.5 Earthuake Modeling of Sitar... 16 215:46 Modeling for Centrifuge Testing. 76 26 Summary sn = 16 Problems = n References. — st “ | | | | 2 ‘STRUCTURAL MODELING AND EXPERIMENTAL TECHNIQUES 21 INTRODUCTION ‘Any structural model must be designed, loaded, and interpreted acconting toa st of smilitude requirements that relate the model tothe prtcype structure. These similitde requirements are based upon the theory of modeling, which can be derived from a dimensional analysis of the physical phenomena involved inthe behavior ofthe strveture. Accordingly, this chapter examines ‘two distinct topes: 1. Dimensional analysis and similitude thecry. 2 Actual similitude requirements for diferent types of structural model, aimed at studying their response under elastic and ultimate load conditions as well as vk dynamic and ‘thermal loadings. Paticularemphasisis given to models of structures undergoing dynamic loading effects because this isan area where physical seale model testing can be of significant help to te structural engineer, must be emphasized that a strictly formal application of modeling theary 10a stuctural problem, without atleast some understanding of the expected stuctral behavior, can lead to an inadequate and even incorret modeling program. Similtude theory must be viewed simply as one aspect ofthe total modeling problem, 2.2 DIMENSIONS AND DIVENSIONAL HOMOGENEITY ‘The use of dimensions dates from carly bisry when human beings first stempted to define ‘snd measure physical oanites. It was essential for these descriptions to have two general chor actersies: qualitative and quantitative ‘The qualitative characteristic enables physical phenomena to be expressed in certain fda ‘mental measures of natre. The three general asses of physical problems, namely. mechanical (static and dynamic), thermodynamic, and eleical, are conveienly described qualitatively in ‘enms of the fllowing fundamental measures: 1. Length 2, Force (or mas) 3. Time 4 Temperature 5. Bleeric charge ‘These fundamental measures are commonly referred to as dimensions Pll chapters of books (such as those by Ipsen, 1960, and Bridgman, 1922) are devoted to establishing and categorizing the fondamental measures ‘Most structural modeling problems are mechanical; thus, the measures of length, force, and ‘The quanttarive characteristics made up of eth a number and a standard of comparison. The standard of comparison, also called the standard uni, was often established rather arta from trational usage (such 35 the inch). Each ofthe fundamental measures, or dimensions, hus has its associated standard units inthe several different unit systems in use today (U.S, Customary, SL etic, et.) Dimensions and units are such logizal quantities tha they are naw taken completely for grated It is ificnlt o realize thatthe present state of physical description of oecurrences did not always exist, ‘THE THEORY OF STRUCTURAL MODELS i s Keeping the above definitions of dimensions and units in ming, the sheory of dimensions can bo summarized ia two ested fate: 1. Any mathematical description (i.e, equation) that describes some aspect of nature must, be ina dimensionally homogeneous form, That is, the governing equation must be valid regardless ofthe choice of dimensional units in which the physical variables ae measured, ‘Asamexample, the equation for bending sess, = Mor, i comeci regardless of whether force and length are measured in Newtons and meters, pounds and inches, of other 2. Asa consequence ofthe fact that all governing equations must be dimensionally homo- ‘ean be shown that any equation of the form FU XoonX,) =0 en cam be expressed in the form Olmos) =0 22) where the m (pi) terms are dimensionless products of the m physical variables KX) and m= wr, where ris the numberof fundamenial dimensions that are involved in the physical vanables. ‘This second fact, that any equation of the form FOX, XX) G0, y-) = 0, has two very important implications: is expressible as 1. The form of a physical occurence may be partially deduced by proper consideration of the dimensions of the m physical quanities X,iavolved. The deductions are made by dimensional analysis, which is discussed a length in Section 23 2, Physical systems that difer only in the magnitudes ofthe units used to measure the ‘quantities X, such as the quantities for a prototype tueture andi reduced-scale mel, willhave identical functionals G. Similinde requirements for modeling result fom forcing the pi terms (1, ,) 10 be equal in model and prototype, which is a necessary ‘condition for the full functional relationships to be equal. Sector 2.4 expands upon this coneept. Example 2.4 ooke’s lw furnishes a good example f dimensionally homogencas relation. Fora stretched bar ofa perfectly elastic material the equation which describes the stnss-stan relation is o-ke. Wis important to notice thatthe fundamental measures tha describe the physical quantities in this ‘equation combine in a certain fashion, Thus: forceunit [foe wit. tengthunit "Genin wih? [Gene ani? Teng wit “a ‘STRUCTURAL MODELING AND EXPERIMENTAL TECHMIQUES force nit force unit * Gength unit? (length unity” and the equation holds for any system of force and Jength units. Such an equation i sid (0 be dimensionally homogeneous. ‘Consider the special case ofa bar made of steel with modulus of E = 200 Nim Thea the ‘equation is 6-206 =0 “The equation is po longer dimensionally homogescous. Is obviously corect only when & and E are measured in terms of newtons and milimeter “The following examples further verify the fact that physical phenomena ae always expressible in dimensionally homogenous form. Example 2.2 ‘An algebraic equation describing the deflecticn of simple prismatic elastic beam witha span {at is subjected to the ssiangularlydissbuted total oad W (Figure 21) is Ws ime 19 (0-10 +71) <0 All quantities may be expressed inthe fundamental measures of feng (L) and force (F¥: thas the slimensions of the let side of the equation are on Free UE +0) Example 23 |A general nonlinear ordinary differential equaion describing the undamped motion ofthe simple ‘pendulum (Figure 2. 1b) is eo 128 sino ae Dimensional, the left side of the equation is Example 2.4 ‘A partial differential equation describing srall deflections of @ flat plate subjected to Tatra and edge loads N,N, and, (Figure 2.16) “THE THEORY OF STRUCTURAL MODELS 6 Figure 21. Examples of phys! eytems lasing fo omogenems equations. te gy aw de IAYI(, gD gy PH yay, Be 2aeaF af (oom SRN SF he aa and the dimensions of the left side become da heed agtlna (Get beth EL) Bree (Freee ee ce ei, ‘Note in each example that the right-hand side of the governing equation is zero, Te fat that 2610 js. constant and thus has no dimensions would seem to indicate a contradiction. Of cours, there 4% no contafiction. Zero is a very special constant as far as dimensional considerations ate ‘concemed, an one could iter transpose one quantity fom the lett the right or nondimension- lize the equation, 2.3 DIMENSIONAL ANALYSIS Dimensional analysis sof substantial benefit in any investigation of physical behavior because it permits the experimenter 1 combine the variables into convenient groupings (perms) with 9 subsequent reduction of unknown quantities. “ ‘STRUCTURAL MODELING AND EXPERIMENTAL TECHNIQUES Example 25 AAs an introduction, consider the problem of experimentally determining the maximum stess ata section of a muispan girder subjected to known uniformly distributed loading 4 per ont length. The analytical equation for stesso has he form of Equation 2.1. Assuming that one had 8 good insight into the nature ofthis problem, it would be appre that the sess cis a Futon ‘of loading q and a represeniative eth J, oF Hato) ° @ ‘Buckingham (1914) proves that any equation ofthis form can be represented as prevhit of powers {nthe form below: a= ky! en ‘where K's a dimensionless parameter that may islf be afinction of dimensionless grouping of| the pertinent physical qoantes, but is more ten simply constant. In dimensional terms, Equation 2.4 takes the Form® elie ASPs es) ‘The dimensional homogeneity requirement forces equal dimensions on each side of the equation, ‘or the exponents on each of the fundamental meastres must be equal forthe wa sides of the equation. Thus, one writes exponential equalities for both F and, ot Bote 26 be teas tom wis snd on(? . (’) em Jn which K can be determined experime to be a linear fanetion of (@/. iy. Note that dimensional analysis alone has shown 6 = Reza wis who ssn on mnt ed m= wl be fr i pie ‘enorme as : Spin sc botnet ef pin 2 3) : sn wot pte as em ge es is vnknown functional G. Recourse toa mathematical solu em) ofthe problem would lead to Me aa la!) a0, a TT af af here a, ay anda, are constants that depend wpon the geometry of the girder and ayaa, = K in Equation 27a. Ofcourse, the dimensional analysis ofthe problem could not have determined the magnitude ofthe constant aaa, Example 26 [Now suppose thatthe load in the previs example of a dynamic mitre and, therefore, has «variation with time, fone is interested in determining the maximum e'astic displacement, then logical set of pertinent physical parameters would include modulus of elasticity F, geometric Tenth time nading g, specific weight of steely, and the acceleration of gravity g, The last 180 are needed account for inertia forces ofthe girder. The functional relationship implied FUE legrg)=0 e9) ‘or writing the displacement 12s a function ofthe other variables, P(E Lta.8) ‘which can aluo be expressed inthe continued product fmm we keregit e1 “The dimensional equation for his expression is ca(aey er (rey (rey (erry Forcing this expression to be dimensionally homogeneous, one then has three equations for the ‘three fundamen measures of force, length, and time, an 6 “STRUCTURAL MODELING AND EXPERIMENTAL TECHNKQUES. “The three equations (Equations 2.11) ave six uuknowns and thus allow a vreefold infinity of soltions. Selecting de and J afbitaniy; the ayations can be solved foro, band ein tere of de, and f. The solutions are bendte-s4l ‘Then eK EMO GLY) fairey] on wa wee) diget le am “The important fact that can be drawn from this analysis is thatthe dimensionless ratio involving the desired displacement u can be expressed as function of a set of dimensionless ratios, or em “Tins, we have been able to use dimensionel considerations alone to deduce a substantial insight Jno the problem and to eliminate a number ofthe exponents involved inthe original formulation. “The for of the functional relationship would have to be determined by experiments in which he several dimensionless ratios were systematically varied. Murphy (1950) presribes tests tha can be made to detrine ifthe parameter K in Equation 2.12 is merely a constant or if tsa fanction ofthe dimensionless parameters. ’ fully equivalent formlation ofthis problem that involves one fewer variable can be obtained by realizing thatthe acceleration of gravity g emers the problem only indirectly in converting the specific weight to a mass density, and that if mass density p had been used instad of specific ‘weight then g would not have been needed. This altemate formulation leads 10 es) 216 “Later in te chapter i Beco even ht i cess ebay: tee hoe ponents st embrace {alte enna meses “THE THEORY OF STRUCTURAL MODELS ny 2.3.1 Buckingham's Pi Theorem “The disceming reader will have noticed that an analysis of Uincnssons fs tod nthe Fst ‘example from 0 ein and inthe second example from )=0 0 . a) +o )n0 Tae “These two examples ae illustrations of general theorem stated by Buckingham (1914). This theorem states that any dimensionally homogencous equation involving certain physical quantities can be reduced toon equivalent equation involving a complete st of donensionless products. For the stvctural models engineer, dhs theorem sales thatthe sTution equation for some physical quantity of iterest, ie FX XpeoX JO a1 can equivlemly be expressed in the form (pRB, )=9 2.20) “The pi terms are dimensionless products ofthe physical quantitis X, Xy--» Xy A complete set of dimensionless products ae the m =n independent products that canbe formed from the physical (umes Xp Xo X, Inthe previous two illsvatons it tard out that three and six physical ‘Tarables reduced to one and three dimensionless produts, respectively. Generally, team be stated thatthe numberof dimensionless product () is equal tothe difference between the number of physieal variables (y) and the number of fundamental measures () that are involved, The first ‘ample was a static mechanical problem: the fndamental dimensions were force and length, oF tema red-2/ 1 The secont problem was a dynamic mechanical problem: the fundamental dimensions were force, length, and time, and m = 6—3= 3. ‘Buckingham’s pi theorem occupies a very important place in the theory of dimensional analysis ‘Before proceeding to the applications of dimensional analysis, the net sections consider in some etal he procedures used in obtaining the dimensionless prodvcts that go into Buckingham’ pi ‘equations. 0 STRUCTURAL MODELING AND EXPERIMENTAL TECHNIQUES Table 21 Typleat Lit >t Physical uanties ‘Guaniiy ‘a 7 ese 7 @ Force F u Mus nore ° sre a e Sian = : Acetic ae 5 Dipcoret i ¥ Pan's 0 = E Modis oestcty FL 2.3.2. Dimensional independence and Formation of Pi Terms. ‘Table 2.1 presents a sample of physical quant that might be involved in structural problems, and the dimensional measures required to describe them. Some strctral problems will involve {emperature and heat considerations; the extenscn of Tle 2.1 to cover these quantities is done in Section 2.53. ‘An examination of the dimensional measures i Table 2.1 makes it obvious thatthe dimensions ‘hat are reqited to describe any physical quantity occur inthe form of single product. The fact ‘hat all such dimensional deseriptions occur in the produc form isa dict result of the nate of ‘dimensions and the basi foundations upon whic eiemific messorements were fist established. Now, from any set of physical quantities, suchas that sted in Table 2.1, iis possible dmen- sionally to form certain quantities in the set by combining other in the form of products. Thos, the dimeasions of length () divided bythe dimensions of time (0) yield the dimensions of velocity (©). Also, the dimensions of stress (6) divided by the product of the dimensions of acceleration (a) and of time (squared yield the dimensions of specific weight (. Consequently, its een that the products v1 and yar are dimensionless. In any set of physical quantities there i @ mite ‘number of quantities that cannot themselves be combined with other quetiies inthe set 0 yeld tmensioness product. The quantes involved inthe linked set are said to be dimensionally indepen- dent, while the otber quantities are dimensionally dependent upon the special linited set, The main “question met in applying the Buckingham pi theoren peti othe formation of appropriate pi terms. ‘The following simple points are the only guidelines needed in the frmation proces 1. All variables must be included, 2 The m terms most be independent 5. In general, dere is no unique set of pi tens for a given problem: ltemate formulations are possible cither by forming the pi terms in several differen ways or by suitable transformations of one set of pi terms. Thust isnt possible state that st of complete, independent pi terms i either “right” or “vrong” for a given problem. ‘The best method for ariving atthe groupings af pi terms i open to personal preference there are a number of rather formal methods which involve seting up the appropriate dimensional ‘equations. One less formal approach involves the fllowing steps 1. Chooser variables that embrace ther dnrensions (fundamental measures) required in ‘cxpressing all variables of the problem, aad that are dimensionally independent, This ‘means that if problem involves the dimensions of force F length L, and time T, then the three variables chosen must collectively have dimensions which include FL, and T, but no two variables can have the sume cimensions. Variables that afe in themselves ‘dimensionless (strain, Poisson's ratio, angks) cannot be chosen in the set ofr variables, {THE THEORY OF STRUCTURAL MODELS 5s soa ee| 1 ~ Figur 22 Boam stress example, 2. Form the m pi terms by tking the remaining (n— 1) variables and grouping them with the r variables in sucha fashion that all groups are dimensionles. This procedure will guarantee aset of independent, dimensionless terns. Tt shouldbe note thatthe r variables osen instep 1 above will in general sppear more than once in the toal set of pi terms While the remaining (n~°) variables will each appear only once. Example 2.7 How does he clastie stiffness ofa rectangular cantilever beam dependon its properties? Assume ‘the beam is loaded at its end, as shown in Figure 2 2, and that stifiness defined as the force per unit displacement measured atthe locaton ofthe load. With no knowledge of the actual relationship ‘between beam properties and stifnes, one might choose the stiffness Sto be a function of beam length J depth h, width w, elastic modulus , and Poissons ato v. “The variables and their dimensions are conveniently represented in tabular form: All six variables are expressible in two dimensions: force and engi. Ths, there willbe 6~ 2= 4p terms. Since the desired relationship i forthe stifines Sit is Bast otto include it inthe two ‘variables that may appear more than once. A convenient choice forthe two multiple variables is span length and modulus of elasticity E, These variables embrace both fimensions (F and Z) and do not have identical dimensions. ‘The pi terms may be formed by inspection by appropriate grouping of I and E with the other variables: ' 21) ‘he tv thea th be ni Set s)oo oxi, om Ey SSTRUCTLAAL MODELING AND EXPERIMENTAL TECHNIGUES Example 2.8 FForm a complet, independent set of dimessionless pi terms from the quantities listed in Table 2 “The quanstes and ther dimensions are Hist in aray Form as In selecting the thee independent quantities that cvident that either mass M of acceleration a must be inluded, as these ae the only 180) ‘quantities which possess the dimension of time. The tee quantities chosen here are length J, ‘odilus E, and acceleration a, The six pi terms are then formed by inspection: 229 A more rigorous treatment of dimensional independence is given in Appendix A. 23.3 Uses of Dimensional Analysis, Dimensional analysis can be used by the engineer in two separate ways. Firs, it can be aseful in deducing, from experimental observations, catan theoretical reslts regarding the behavior of ‘ physical phenomenon. Such a situation coulé arise if one knew the relevant physical variables that affected the state of some oter physical variable but didnot know the mathematical reatioship that connected these variables. For example, if only thee or four physical quantities are involved, ‘dimensional analysis may reveal the solution -o within some constant value or some unkngwn function of one or wo variables. Rater simple experiments ean then be performed to determine the constant vale othe functional relationship. Ofcourse, if here ae 10 or 20 physical variables to begin with, here wil sill beso many dimensionless products remaining as to make experimental analysis diticult or impractical. _imsecond eine se of sacral sin wrk een eed ery sity Biman 1922): “There are in engineering practice a lage number of pubes So compliatd hatte exact soon {snot caine, Under hee conitons dimensional analysis males ws to oblin etn information ev the format the resol hich coud be cei in practic onl by experiments wih an impossibly ‘ride varaion ofthe arguments ofthe unknown function. In ode 0 apply emesional anys we merely ave to know what hn of a physical tem iis hat we are deaing ith nd what he ‘orale are whieh enter into the equator we do not even have to write he equations dawn xpi, rch Test ele them. ‘oppose thatthe variables ofthe problem ate denoted by XX ean thatthe dimensionless proves re fun, and that he resis how in the fem | N= AB LEAP PAP) (24) {THE THEORY OF STRUCTURAL MODELS = “nee the aguments of the Fncton alte far ease embrace ll he dimesioness prods Torta the rl a sho is general Now in pusing fom one physical syvem to ath, the “Setary fonetion 6 wil in nel chage m am ekonn way v0 dt i f ny wef information al he obtained by insrimiate mac expesnents. But the mods ar chosen in such @ ‘att way tht ll the anges of he aka fanetion have the same ae for he roel a8 Tarte Fltcsle example then the only vrs in psig from rol to fll scale sin he factors tide th anton sig, ad the mano of varton ofthese factor ison Fm he drensiona aly, ‘Stated in this way it would appear thatthe stractural design engincers model problems are solved, and, infact they are if the dimensionless products that are te arguments ofthe unknown Fonetions have the sme valve forthe model s fr the prototype. As will be sce later, technological problems may make it impossible to sity satis this conition,particolnly For modes that sre Intended to reproduce the inelasti response of reinforced coneree stuctars, 2.34 Additional Considerations In Using Dimensional Analysis tis essential thatthe experimenter who wishes 10 use dimensional an Insight int the problem to choose the proper physical quantities in the forma beam problem of Example 25 is a convenient example for discussion her. To the person who doesnot have a Keen insight into beam bending, i might not be immediately apparent that the ‘modulus ofelasiity E doesnot enter ino the stress problem. If E was aed tothe problem, then the governing relationship would be Fe.giE)=0 225) whic is expressible in the two fundamental measures F and L. By folowing the procedures set Gown in Section 232, itis scen that g and J, q and 0, q and E, 1 and 0, and {and £ are all ‘Simensionally independent. Taking 4 ard asthe dimensionally independent physieal quantities, the complete st of dimensionless products is 226) Accor e2n “There was nothing unique about the selection ofg and fas the independent physi js son that 9 and E could just as easily have bee chosen, and a procedure identical .o the above ‘would have led to o-n(2) a : STRUCTURAL MODELING ANO EXPERIMENTAL TECHNIQUES Equation 227 should be compared with Eyuation 2.7a, The unnecessary inclusion of E as relevant physical variable does not make Equaion 2.27 incoret. However, it is nels com plicated, and it would require addtional exprinental work to find out that he beam stress vas in fact independent of F. The experimental design engineer thus is faced withthe following dilrama: 1. Ifa relevant physical quant is omittc frm a dimensional analysis ofa problem, the investigation cither mects an impasse or lead to eroneous esis, This statement is not rigorously proved here, but reason alone indicates its validity. f not only the relevant physical quantities but also some irrelevant ones are included, the ‘imensional analysis wil lead to a res that will make dhe experimental investigation ‘much more dificult than need be. In fa, it may eliminate @ model study a a practical ‘means of obtaining the desired information, In this tight, Langhaar (1951) has stated, ‘requenly the question arises: How do me knew that a cersin vale afta phenomenon? > ‘nswer this question, one must understand cough shoot the problem 10 explain why and how the ‘rile infuencs the phenomenon. Refer one undertakes the dimensional analysis of a proble. he should to forma theory of the mechanism of te phenomenon, Even arate thoy stall discos the actions ofthe more important variables. Wf ke fleet equations hat gover the phenomenon are avaiable they show diet which variables ar igi, they are not available, then the engineer must have some other insight nto the phenomenon, ‘because itis clear tht dimensional analysis ca be of no use unless one can identify the relevant physical variables, Example 2.9 Free transverse vibrations ofa at ela plate are known to be govemed by the paid iterenal equation aw yy tw dw, pha. a aca 9 ETAT) ar ae subject to certain prescribed boundary and inital conditions, withthe pertinent variables being “%y = evordinates of points on the surface ofthe plate plate thickness = modulus of elasticity Poissons ratio = mass density = out-of plane displacement of pate mide surface h E P Its desired to make a model study of a lage inregulaly shaped plate inorder to determine the lowest natural frequency of vibration. What ae te relevant physical variables? Inthe ight of 4 dimensional analysis of the problem, how sheuld a model stidy be conducted, and is it likely ‘hat useful results wil be obtained from a mod investigation? ‘Assume thatthe solution equation forthe netural frequency Js ofthe form FUP Lp. ean {HE THEORY OF STRUCTURAL MODELS ‘where isa characteristic plate dimension for either plan or thickness dimensions. Taking fp, and EE tobe dimensionally independent, Equation 231 can be reduced 9 {i »)-0 aa er ‘orn the solved form 233) ‘Byuton 2.2, whieh the mot hat an be etic fom amensonl alysis canbe compared ‘with the mathematical solution of the simply supported rectangular plats. Such a sofution yl forthe funda * aa p(t") where a and b ae the side lengths and m and are integers indicating the number of half sine waves inthe deflected shape ofthe plate. For a square pte with j and # equal to 1, Returning tothe model problem, Equation 2.33 can be written once forthe prototype and once for the model. Dividing the prototype equation by the model equation, one obtains Losin (4) (E,/e,) 0), “Sousa (U1, {(E-/ Pa) We 1 Poisson's ratio in the model and the prototype are equal, then the megnitude ofthe function is fan identical constant foreach, In that ease 4b FEPa 03H Sra = Sno) VF 9 : iy interes py phenom: tn fait igh be sone ht Poisons i does ot elise his py Chon nh expen ig en a te el marl ohne a Poin i et fomint of Se ope These mst bear oo at he ral te ubje to tirenor anid wih he departs rm estates the mensional mals “he pounce soot be male. Clea i tees of the prottpe ate ae sy aero snore hn cigs of emo poet lobe ey Bd ll suena, cay eamed egse rae Be Me 2 Inte of eres it Sl be expe the ecological polens associated wi providing “ir bord otis nh polo wos be of ser more concen tan the eet ci Pasion a. ‘STRUCTURAL MODELING AND EXPEFIMENTAL TECHNIQUES 2.4 STRUCTURAL MODELS eis a relatively simple matter apply dimeasional analysis principles to the structural model {As the discussion is developed, toe types of sructural models will be described. These are 1. The true model which maintains comple siilarty. Any mode! that stisies each and evry siplton se forth bya proper dimensional analysis would be sid to have complete Similar. 2. The adequate mode, which msintsins “fi-order” similarity. Ian engincer has a special insight ito a problem, thea it may he possible to reason that some ofthe stpaations set Toth by proper dimensional analysis are of “Second-order” imporance. For example, in rigid-frame problems itis known that sxial and shearing foross are of second-order importance relative to bending mements:nsofir as deformations are concered. Tus, it ray be adequate to model the moment of inertia but not the eros-scetional areas of ‘members. Thus, any model which satisfies each and every first-order stipulation which is set forth by ® proper dimensional analysis but which may not satisfy certain second- ‘order stipulations would be sai to have store similarity “The distorted model, which fails to satis one or more ofthe first-order stipulations 2s set forth by proper dimensional analysis: ‘Of course, complete similarity is desirable mall structural models, but usually the economic and technological conditions preclude a model study that maintains complete similarity withthe prototype, By neglecting certain second-order etfects, ti usually possible co make an adequate ‘model sty to obtain results to predict dhe behavior Of a prototype structure accurately 2.4.1 Models with Complete Similarity Iuhas been sen from Buckingham's theorem that the mathematical formulation of any physical Phenomenon can be reduced fo an equation insalving a complete set of dimensionless products, BOR Rook 235) If Equation 2.35 is writen once forthe prototype and once forthe model, the following quotient can be formed (ry 7 036) a a here refers 1 min the movel and mis rofxs fo, in the prototype, ete. Complete similarity is defined to be that condition in which al of the dimensionless products ae the same in both ‘model and prototype. When complete similarity is maintained, 2 {he THEORY OF STRUCTURAL MODELS so that Equation 2.26 may be written By Arteta) Bag WE a Maan) 238) eee 23 rslaton btwoen model and ations 237 and 239 ae the bass fr the mol tho The a Sgeaton 237 and 22 ip Egan 237 a cal the dein ev opera comin: Fa et Spine predtn eon othe dependet aise of th problem. has insane 29. =D = uation 2.39 took he form (ee py fom Hate fe ect Sain pis thr ae heat ie p tc vtonataentge tan ay sa Sew {eh cnave Reema rine 20) he ideal forte odes Stresses shuld me hat adel ao prot nn mest be swe oa eas ety ea oma hen he se materials wed a ode an on of cing aon or any uve modeling problem canbe exis aie 2 eee amano ems eed ee factors The ede emo 19 aa tact of quanti fr some more complied probe, rive SIS ents oe and examin tn cry Py ‘STRUCTURAL MODELING AND EXPERIMENTAL TECHNIQUES. 24.2 Technological Difficulties Associated with Complete Similarity Several types of departures from complete similarity can occur, eluding 1. Accidental overlooking of a pertinent variable; Deliterate violation of a siilitude requirement that s considered tobe not critical, sch ‘8 usinga model material with Poisson's aio differen from that ofthe prototype materi, "Necessary deviations fom true modeling, such as using a dserete load s)stem to replace 8 continaoas bad Lack of complete similarity in the model and prototype dimensionless suios means that the ‘atio 6/04 in Equation 2.38 is no longer unity. While i is not feasible in mos instances to make ‘an evaluation of the te value ofthe rai, itis essential to realize that this departure from tue similarity affects the model result the degre fat the rato departs rom unity. Lack of similarity, ther known or unknown, offen las to diferences that ae misleadingly calle “size effets Size effets as such do not exist if complete similarity is maintained in geomeiry, materials Propeties, nd loading (Chapters 3 to 5 and Chapter 8 dal withthe later two aspect in Jee, Example 2.10 Gravity Load Simulation ‘This example ivtrates « departare from similitade of the second type mentioned above. Corresponding to the fact that the maximum numberof dimensionally independent quantities equals the umber of fondamental dimensions involved in the quantities, the model engineer can select only resticted number of mode! quantities without regard for the poterype. Ths, static and ‘dynamic problems only two and three model goattcs, respectively, can be atitaiy selectod, Practica considerations generally demand thatthe model geomet scale and certain model material Properties be selected to be compatible with the available equipment and moter. 1 the problem involves only static response and the deadweight of the srocture exerts an ‘important influence, then one of the dimensionless products inthe problem willbe WE where ys the specific weight ofthe material, isthe modoks of elasticity ofthe materia or some equivalent ‘quantity srhich represents the stress-strain characteristics ofthe matvial, and li a epresentative length. Now only two model quantities can be selected arbitrarily, and thereafter the preceding imeasionies product must have the same maghitude in model and in prototype, This Ow ae 1 7 ay Youssef ean AAs typical illustration, considera reinforced concrete prototype structure simulated witha o-sale polyvinylclorde plastic model, Then ous 9010/1140 kg/m?) but the dead-oad smite requirement sas hat the deasiy ofthe model materi should be smo $3000) 22516 (3600 kg/m") °° “THE THEORY OF STRUCTURAL MODELS Figre23. Model br conteuous brio under emai loading (Coutesy cf ISMES, Berger, ol) ; te sntasy 6 necessary 1 he a te {Eure ne ete te erg ned oe ones ns ely motte my et ste si tt On Be he ‘hand, a “massive” structure may require that the addtional mass be dispersed irougoat be vlame ae ney a ml Sis Fg conte de have ie Tee fact should not ree ‘model quantities can be arbitrarily selected in dynamic problems. This ea se ald La sea ie at ep eb on ey Leb ey wcrc wa dnc nde uni Te swe ses ens SK pcaccnn theefore alice mens must rina Be sed to pie the sim iso Ofc scl sph of meet pen say om ce SS ats nat apopen prcte my be singly © eles te Sep) Sonne 2.3 shows bow atonal artical mass was attached to abridge vibration model to elencap env a 24.2.1 Other Types of Distortion There sre a numberof distortions that are met frequently in modeling, including Discrete loading in place of distributed loading; stein ’ : A Pantie size a 4 1 5 wit ae : 6 Sian as ' 7 Liga density % es 1 Surface tension 6 oo 1 a 8 cute . Mea f . oi a ' ° 1 ida 2 2 emt » os . Fa Paitin : ee & clon oo ' compen of ES : 14 Inertia " at 1 Nt F 5 mint at 7 16 _ cop oi Nae Wo USin ere ‘Sie: Fhang Do Kis Ores, Nin Coma oS Mech, WH Cig ‘als, AA Halle Pals, Booka VE T96 ih pra, 23 For the case of , #1, S,# 1, and, # Ey derive expressions for a reinforced concrete beam for a. Required area of model reinforcing 4: 2, Yield strength of model reinforcing a6 function of yield suength of proorype stl. (in: This can be accomplished by considering the basic exural mechanics of a simple reinforced conerete beam.) 24 a, Using a dimensional analysis, develop 2 general expression forthe distance that a freely falling object will drop in time 1, noplecting resistance affrded bythe ax, ‘we entneously assume thatthe weight of the object Assume the dropped object stars with an intial velocity vj Develop an expression for distance tavcled in time. ‘THE THEORY OF STRUCTURAL MODELS 25 Determine the x terms forthe beam-stiffness problem discussed inthis chapter, expressing ‘te parameters in terms ofthe fundamental quankites of mass length and te. What 1s the basi difiulty met inthis formulation of the prem? 26 ‘The following equations are tobe checked for dimensional homogeneity. Give values for the dimensions attached to any constants in the nonhomogencous equations. a. Column strength formula raaloast'esn) 1b Bending of a wransversely loaded plate: «© AISC column stess form quiet : 112 cot ey 4. Manning's formula for open-channe! flow: 16, v, «Buckling load of pin-ended coh P= EL & at end of cantilever of length L, ith tp load P. PLIBEL 2, Rankine or total active pressure pon each foot length of wall of height H, caused by cobesive soi Ua — 2clt 2.7 You are working in'a design office and have just completed the calculation to determine ‘he maximum bending monnent i a uhree-span continuous girder carrying uniform Toad ‘of 3.0 hipsf. Your supervisor then informs you thatthe spans have been increased from 460, 80, and 60 ft 1072, 96, and 72. Can you, without going though the cau again, quickly find the new maximam bending momen? 28 Given ses, Lo 5hog 0 “STRUCTURAL MODELING AND EXPERIMENTAL TECHNIQUES ston et = terme pi by ping sven i Staab ) Moasla nd forming the ratio of M, 10 Me HE A \ 29 Prove o disprove the validity of Galileo's statement (given below) by applying smilie theory to the problem of deadweight steses ina structure of constantly increasing size {i mou be impossible to bail up the bony structures of men, bores, oF eher animals 30 a5 to ‘hold tothe and pfonm their moral fui these animals were wo be nreased enormously Jn heights cease in eight cab be accomplished only by employing material which is Tarver an song than usta o By enlarig the ie f the Bones, thus changing tr shape tit the form and appearence of the animal suggest a monstrosity. the sizeof a bay be ‘imiised the sreagh of tat bay not dished inthe sane proportion; indeed the smaller the body the greater is relive stengh, Th, 4 seal dog could probebly cary on his back to ‘rire dogs of his ov sae, ut elo a ore could not ary even one of Ms own ie 210 A rectangular tubular steel section i subject othe loading shown below Use the pi theorem to develop a general expression forthe deflection of pont a with respect to point b. 2.11 Use the pi theorem o develop an expresson for the natural frequency ofa freely vibrating fixed-end beam of prismatic erss sectcn Sone = Sally JE: Pal EP “THE THEORY OF STRUCTURAL MODELS ot 217 tis desired to build and festa Plexiglas model of a Tage cast tet flywheel having fesvy rim and radial spokes. Establish the simiftude conditions. Fora Yx-scale model, ere the sees and velocity seas fora protoype that has an angular veloviy of $0 dis? 2.13 The natural Fequoncy of a steel tuning fork is 200 vibrations per second. What is the rata fegueney of aV-scale aluminum mode ofthe tang fork? The unit weight of ‘lominom i 36% that of steel, and its Evalue is % that of stel. (A tuning fork is a Srely vibrating elastic system) 12.14 The ideale dynamic loading on a bidge is given by a forcing function that varies periodically with Gime The exciting forces are fully specie by the frequency Fand the Taximum value of force Fy= 10,000 1 (445 kN), Determine simiitude and scaling ‘elations fora model brie with a geomeric cake fctrS, = 30. Bath bridges ae made “Of stet. Is itncceasary to ad ational mass (weights) tthe bridge satisfy smiitude? Explain carefully 2215 For the beam of Problem 2.14, develop an expression forthe maximum siess due 19 ‘concentrated load P st midspan, the beam self-weigh, and a uniform temperature rise of AAT Assume elastic action only. 22.16 The torgue onan airplane propller depends only on the diameter d ofthe propel, ts angular velocity the velocity of advance ¥ the mass density and viscosity Hof the tie Using dimensional analysis, fin an expression forthe torque, and show that i the “fect of viscosity canbe neglected then the torques proportional to ir density. (Courtesy Of W. Godden, University of California, Berkely) 2.47 Tass the performance of a high-speed tran omits track, a Yn seale model is made of the complete system, The track is both curved and banked (ranverse slope) in places. “The model train is tue scale, and its effetive density is found ic be twice that of the rototype. The model sto be wed to measoe the forces exerted bythe tain on the tack in toca the tendency forthe train o overturn on corners. Specify the following model ratios required wo simolate prototype behavior ‘Track: a, Horizontal radius of curvature, + 1 Transverse slope, @ Coefficient of friction, Train: d. Velocity, » fe Acceleration, a Forces: Centrifugal frce onthe track, F, at points of eurvatre ‘Arial forces on tack due to accleration, F ‘Wil this model comet simulate the tendency for the train to overturn? (Courtesy of W. Godden, University of California, Berkeley) REFERENCES ‘Anton, Bf (1987). The we of small scale drect modes for consete block masonry assemblages and “lender rvforced walls wader out oF pane lade, PRD. thesis, Departmen of Civil and Architectural Engineering, Dress! Universi. Piadlphia. [Alesanier SI and Harby, © F 1970), Design of strates to withstand gxsows explosion, Farts 126d 2, Cone (2), Febuary; 413), March Ane 1, Sts H. Dan Hansen). (1960) toy of te Apiaility cf Model for Investigation of “hr let Eves on Saces Technical Report to the Defense Atomic Suppor Agency, Department ‘rch Engineering, Massachusetts Insite of Technology, Cambridge, Ostet we ‘STRUCTURAL MODELING AN EXPERIMENTAL TECHNIQUES, Ave, Smith, HD, Sharma, H.D, apd Haris, 4G. (1962). Elston of Techies for Consisting Model Stncural Elemens, Research Report RS2-15, Dparment of isl Engvcring, Masachsets Insite of Technology, Cambridge, May. Briley, A. ad Vincent, N.D.G. (1943). Wind pressure on bldings ncn he effet adjacent uiings, I Inst. Cis ng London, 20, 243-273. Baker, W.B (1973. Explosions ia Ais Univesity of Texas Ps, Astin Beayjint,N. (1960), Simitde and theory of mod, RILEM Bul, Pai, No, 7 one Socio, 3. and las H.G (1977. Erbation of Tchniges athe Diet Modeling of Concrete Stace, Structural Mogels Laboratory Report No, M771, Deparment of iil Enginesng, Drexel University, syste, Phys Res, London, 4345) {Cstld A and Casa, M. (1976), Experimeta Techniques forthe Dyaanie ali of Complex Stac- ‘ures, Repom No.7, [SMES-Ialo Sperimenle Modell c Site Beran lly, Feeney (Cermak, JF (1975). Appiations of fié mechan 1 wind engineering. A Freeman Scboer Ler. lads Eng. ASME, 971, March, 9-38 Cermak, 16. (977), Windunel sing of stucites, J. Brg, Meck, Dix, ASCE, YONEMO), December, 1125-110, Dragosai, lM, (1973), Stertal measures aint astral yt explosions in highs ck of Mats. Heron, 1944), Departmen of Ciil Engineering, Technlogial Univers, Delt the Neha 5-1, Dryden, HL. and Til, GC. (1933). Wind presse on mel ofthe empire tate baling, Rex Nol. Bur ‘Stand, 10, 493-535 ‘Pohang. D: and Krchs Ovesea.N-(1988.The apaicton of the theory of modeling to cet ste, in Cente in Sot Mechanics, W.H. Crip RG, ames, ant A'N.Scbofici Eds A.A Bateme Poblshrs, Brooke, VT, Maris, H.G., Bd, (1982), Dynamic Modeling of Conroe Stace, Publication SP-73, Asercan Concrete Insite, Det, 282 pp Marrs, H.G. and Basics, 1-1. (1977). Dict stl seae models of concrete masoay statues, paper sen atthe 2d Anmoal ASCE Enineting Mechanics Division Spay Confereace, Noh Carina State University, Raligh, May 23-28, 197, published in Advances in iil Engineering tough Eng reering Mechanics, ASCE, New York, 11-108 ‘ais, HG. and Bic, 1.3 (1978), The behavior of conrete misony strictures and jim deals sing sinallscle diet modes in Proceedings of the North American Masonry Confronce, Univers ef Colorado, Boule, August 14-16, Hanis, HG, Pal, PJ. and Sharma, S.D. (1962. Dynamic Sunes of Stas by Meas of Models, ‘Report 63.23, Deparment of Chil Engineering, Masachrots Isat of Techlog, Cambri. anti, H.G., Schvind R Taber, I, and Were, S (1963). Techies and Matis inthe Modeling of ‘Reinforced Concrete Snetues under Dynamic Loads, Rept R63-54,Departnet of Ci ginceeng. ‘MassacsetsIsint of Tecnology, Cambridge, December, ao NCEL-NBY3228, 0, Naval Ci Engineering Laboratory, Pot Hosneme, CA Hudson, D.. (1967). Scale model pincpes, in heck and Weraton Handbook Hai, CM. and Cred, CB, Fas, MeGrawHil, Now York, Chap. 77 Isen, D.C. (1960), Unis, Dimension, ad Dimensionless Mumbers MeGrav-Hil, New York Innnger, JO. Vand Noklanved (1930). Wind presore on bldg Ingenoeridenst Ske Issn, L (1966. eae modes in fe esearch on cone srores.. PCA Res. Dev. La, 83), Sete, 10-26 Isyumov,N, (1979). Modeling of Wind Efects on Stntues and Buildings, lecture pots. Jensen, M, (1958). The model’ for phenomena in nat wind, genoven inetationl eon, 4, 121-128. Jensen, M. and Franck, M1965). Movel eae Tests Turblot Wind Fans 1 and The Danish Technical ress, Copenhagen ‘Kato 8.1963) Sinitiude in Thermal Models of Spaceraft, NASA Technical Note D-1631, Api Kroner, H. Mills, R.S, Monerz, PD. etal (1978. Seale Moding a Testing of Siretures for Reproducing Response wo Karhgtake Excatiot, the Jobn A. Blane Earthquake Engincring Cee Deparment of Cl Engineering, Stanford Univesity, Sted, May ‘THE THEORY OF STRUCTURAL MODELS. 8 Langhas HL (1951, Dinensia Anat and Teo of Models Jha Wey & Sons, New Yak [Une A an ey 98) Macht honda deformed one A Cone is tl November 172 stir, MLS Has HG an Sabi, GM. (197), Sra mols sathqukeengacrg, in Procetings fhe Third Canton Conerce of ErhqueEngneering, ee «6, Mote, Qube, Seu ses ‘up, (950) Sime Engng, the Ronld Press Copan, Now Yk Nite, (197 lel aw oe nse an ows with pctv wo mal Arch Me, “ei 29, War ath, (190), Exons in dei src pr: the eet of gape explosions in ome tacts, Su Eg (Landon, #10), Oct. tM OH6) Dei of al oss Inch dns by meas of eel, RILEM Bll ah Ci. Roce, KH (1969 Si an ml tJ Sin A,X, 7-4 Scan, RH (1973), Dynami tude in ed inPocedign EDF Conference sur Ac Hhdo- laste CEA-EDR, Chat Fane, Sptenb 9 1972, Ei, Pa 17328 Sesln H(197) Sele Mi an Modeling Lave nF: sy, Pre No. 2247, ASCE National Sica Egineeng Meng, Cae OF, Api 22-25. Scouing, DD. (570, Sele Moe In Exgicering~ Pandan nd Apion, Pergo rs, To York. Slo, 1.197, Exlsosn biigs the bebnvir of einfred cone frames, Concrete 5, Apa, iia Sah ILD. Ch RW, and Mayor, RP. (1960, vation of Model ehnius fr he vestigation of Sancta Respone io Blt Lan, Rept R63 16, Deporte of Chil Engineering, Masschsets ine of Teche. Cg Tey. Saye, WHC (1972) Soy eri fo the application of id dl he say of a polation meter, Beandary Laer Metal 3), Sepenbe Sct, 1.0), Expos in mete arte: Pat: the taloshperwen cote chars Wists an gc eto, Sic. Brg, Lamon, 70, Other Sony, Dod uo, N (178). Mos ke of wind elec Fence on te robles of ‘peincnalechgue sd isunenato, paper presaed tI. Cons on Instrumentation i ‘Secor Situlton Faller, nave, a, Whi RN ad Van Hor, DA. (970 Pains of mods nis in Mod for Cone Scares, NSP, Arca Conte nie Det, Ml 19-2, CHAPTER 3 Elastic Models — Materials and Techniques CONTENTS 3.4 Introduction a so 86 32 Matra for lstic Modes. 87 33. Plastics. 88 TEL Thermoplsis and Thermoscting PSUS enon 88 332 Tension. Compressia, and Flexural Characteristics of Plastics 89 33.3. Viscoelastic Behavior of Pastis cD 333 Mechancal Popes af Polyester Rein Combined wit Cate Filler. 34 Time Effects in Plsties — Evaluation and Compensation 96 Tear Deteinaon of te Time Dependent Modus of stich amd Poisson's Ratio... 91 342 Tring Testis Accu ine Dependent HE. 98 35. Effects of Loading Rate, Temperature, and the Environment 00 5.1 Influence of Sita Rate on Mechanical Properties of Pasi “100 332 _ Fifects of Temperature and Related Thermal Problems 101 35.21 Temperature Etfets on Elastic Constants, 101 352.2. Temperature Effects on Stength = 101 35.3. Coeliciens of Thermal Expansion 101 435A Thermal Conductivity. 353 Softening and Demolling Temperate nn : 3136 Influence of Relative Humidity on Flastc Properties on. 103 3.6 Special Problems Related to Plastic Modes. 108 3.6.1 Modeling of Creepin Prototype Systems... 103 362 Poisson's Ratio Considerations. 104 1363 Thickness Variations in Commercial Shapes 108 264 none of he Caing Pres on he Mesh of Stasi 108 3.7 Wood and Paper Products, : 108 37.1 Balsa Wood 7 : cals S711. Balsa Wood Shapes Available xno lO 37.12 Strength Properties. 107 3.7.1.3 Applicaton of Balsa Wood to Medel Stas... 110 1. Two and Three Dimensional Trusses 110 2. Blase Buckling Std ens 12 28 ‘STRUCTURAL MODELING AND EXPERIMENTALTECHUQUES 37.2. Modeling of Structural Lamber.. na 3.7.2.1 Strength Propetes of Wood Used for Stuctiral Models. ta 372.2 Examples of Wooden Modes as 37.3 Sinal-Scale Modeling of Gtue-Laminated Stuctres TT) 3173.1 Paper Products Used for Structural Models. fen 3732 Examples of Paper Models. no 38. Flastic Models — Design and Research Applications wenn 39° Determination of Inuence Lines and Induence Surfaces Using Indinect Models — Muller Brest Prinipe . sev 3.10. Summary vs = Se ES Problems, aes . 124 References aa : 7 3.1 INTRODUCTION An clastic medel can be sed to study the bebavior ofthe prototype only inthe Hncaly elastic range. It cannot be used to predict any inelastic behavior of a loaded siuctare resulting from ‘material nonlinearities such a the postracking bsavior of concrete, postyielding behavior of tcl, ‘oF the postbuckling behavior of a plate or column. Historically, clastic models were wed study thetinearly elastic behavior of complex redundart structures prior to the development ofthe digit computer. Elastic moves have been used extensively to sind the response of maltistory bildines, bridges, nuclear reactor pressure vessels, dams, and other types of structures subjected to state, «dynamic (including earthquake, bast, and wind lds), and thermal loadings. Elastic models of structural components such as columns, frames, sh, ad shells have been used in elas stability studies. Many such applications can now be performed by a myriad of computer codes that are readily available to the engineer. Usually the computer solution is more economical and efficent. However, thee exist many structural situations where the use of elastic models can be very beneficial Inthe ese of educational models and architectural engineering models the use of elastic ‘materials is sil essential. A greater emphasis or laboratory and structural concept demonstration inthe new engineering curriculum also requites physical model testing. A majority ofthese moet applications can be salisted with linearly elastic model materials. Is with these applications in mind that this chaper is retained in the second ition and is somewhat expanded to satisfy the ‘neds ofthe civil and architectural engineering sent. Similitade requirements for static elastic madeling were presented in Table 22. Independent scaling factors were chosen for modulus of elasiity and length and all remaining sale factors \were established as functions ofS, and 5, It follws from Table 22 thatthe material for an elastic ‘mode! of prototype structure need only sitisty th requirements that it must nt be Toaded beyond the tinea elestic range and that it must have the same Poisson's ratio as the prototype tera, althoogh this later requirement can be waived in one-dimensional structures. Howeves, even structural situations with more complex sates of stress, elastic model materials having & Poisson's ratio else to that ofthe prototype can be found because the range in this material constant is small (0.15 to 0.35 for most materials of interest). Ehstc models can be used as deet models (ihe pattern of mode! loading i similar to that of the protatype) o as indirect models (io derive infhuence lines an intluenee sarfaces, where the pattern of prototype lading doesnot have to be rece), In indirect models, iis nt necessary to satisfy the condition of equality of stains at comesponding (omologous)poins inthe prototype andthe mate! (€, = ¢, It follows that the intensity of applied Toaingcan be vated at wil; the resulting deformations will be propononal to the aplied loads provided the Finear elastic range of the materia i not «exceeded. Conceptually, such experimental mode's are very similar tothe availabe procedures for ELASTIC MODELS - MATERIALS AND TECHNVQUES a clastic analysis of strictures and account forthe stilfesses of members and joints; in some ways, spealy wid espect wo boundary cuits, de piysival uodelscpenc a eter idealization fof the siuctue than the mathematical meds. The reason for ths is that analytical modeling of ‘actual support conditions i over simplified in most, if ot all, computer codes of structural analysis. 3.2 MATERIALS FOR ELASTIC MODELS Pec ari if heirs ony cl a ayn reser nt fir proeiypesctres. The alasiages, adage, nd Linas com Imonly ved model mae have een examined by ssvrl iver. The sled materi ins at ews of snl resend in Chap, snd bees ang eprobe mchan ial properties and geomet ably, shld e readily aval, ex ablated, and aoe A imspensne. A dad dsciscon of he physal ad chemical poperies of matrls Subir fe consacion of models is beyond the scope of is et ad only fe relevant ‘ropes ofthe more commonly wd model mater represented to asst ith he selection roves The mos sigicat ol matrilopeis ae Proportional (limit) stress Siitfness Failure mechanism Taflcnce of temperatre aed humidity on materatpropestis Creep characteristics Load rate and stain rate effects Effect of size and shape on material properties All material properties should be confined by appropriate tess, as data given inthe various handbooks or manufacturers eatlogs represent average or “obtainable” values and are often unreliable. The properties of some matrils such as plastics, wood, and concrete not only show large variability from one sample to another, bu are also significantly dependent on the type, shape, and sizeof the specimen and the rate of loading, Three major types of materials suitable for the constuction of elastic models are presented in this text: plastis, wood, and paper (Chapter 3), ‘cementitious materials (Chapter 4), and metals (Chapter 5) ‘The only condition required to be sisted for indirect models is thatthe material exhibit 2 linear elastic stess-srain relationship; in general, plastics or metals are used for constructing indirect models, Plastics have low elastic moduli, and this leads to smal load requirements and measurable large deformations in sal scale models. By using strucual mechanics, it can be shown thatthe principles of superposition and reciprocity are valid within the linear elastic range of behavior for the prototype or model structure (this also implies that there are no stability oF ctenary effects, the deformations ae small, and the limit of proportionality of the material isnot exceeded). These two principles are useful for determining deformations and stresses in direct models asa result of combinations of a variety of foading conditions such as dead, lve, and wind Toads. Fora general cas, to simulate a given sate of deformation inthe prototype, the following Smiltude relationship fer Poisson's ratio mst be satisfied (see Table 22 en In planar skeletal structures, suchas frames, russe, arches, and cables, torsion is nonexistent and shearing effects are not as predominant as the Hlexural and axis effects therefore, the similitude equiement (Equation 3.1) can be relaxed for these structures, However, if the response of the cy ‘STRUCTURAL MODELING AND EXPERIMENTAL TECHNIQUES structure not independent of Poisson's rato, as n grids, dhnee-imensional frames, plates, an Sls, then Bquation 3.1 shouldbe sauisied. 3.3 PLASTICS “The term plastic is normally used for & material derived by chemical synthesis and containing carbon compounds (Preece and Davies, 1964). 8 wide variety of plastic materials with varying ‘chemnical compositions and mechanical properties is available under specific trade names from ‘manufacturers and local supplier. Only those progertes of eommonly used plastics that ae relevant to the construction and testing of elastic models ze presented inthis section. ‘eggs (1932) is perhaps the fist known invexigator to use plastic models fo solve statically Indeteminate structural problems. Since the exly 1960s, plastics have been used effectively for the constuction of direct and indirect models cesigned fo simulate the linea clastic or linear viscoelastic response of the prototype. Comprehensive studies of structural models coastacted from plastics have heen reported by Fialbo (1962), Lite and Hansen (1963; 1965), and Carpenter etal. (964), Mechanical properties of plastics directly relevant to structural modeling, were reviewed by Rotve (1960), Preece and Davies (1964), and Roll (1968). 3.3.1. Thermoplastics and Thermosetting Plastics hastics can be classified into two categories: thermoplastics and thermosetting plastics. Ther- -moplastes become progressively softer a lempemtaes between 93 and 149°C and can be formed 0 complex shapes such as shells of complex geometry with litle or no pressure and yet resin their shape upon cooling. Ia thermoplastic material is reheated, it ean be remolded into another shape, although some types have a “memory.” This characteristic of thermoplastics is extremely ‘sel for the vacuun-forming process, which will be described Tate. The commonly used ther- maplastis (namely, seri plastics and polyvinylchloride, PVC) are generally avilable inthe foc ‘of sheets, ods and tubes I must be note tht several types and grades are commercially available ‘under a singe ae name; fr example, there are many different grades of Plexiglas. Plexiglas G, commonly used for many applications, wil hick about 2.2% during te heating process, with 2 corresponding incesse i thickness. By contrast, Pexiglas II UVA, used extensively for research work {2 at Comell Universi, is preshrunk and is manofactred to mone exacting standard of optical “quality, surface quai, and thickness tolerances. ther preshrunk types include A UVA, 55, ax 5009. “The acrylic plastics Pexigas, Perspex, Lucite) can be easily machined and cemented, and sccurate models can be rapidly aisembled. Shets can be softened by heating and formed shells of single or double curvature using the vaeuum-forming proces. Comprehensive studies of acrylic plastics used for models have been repaid by Fialbo (1962), by Carpenter etal. (0964), ‘and later summarized by Roll (1968) Fomagalli (1973) has reported om the use of Lucite for several clastic model stdies atthe Istituto Sperimentale Modellie Struture (ISMES) in Hay PVC is normally available in thinner sheets nd sheets of more uniform thickness than acrylic plastics and are specially sited for model studies of various kinds of shells. series of vacuum: formed shells of PVC have been tested at Cornell University and the Massachusets Insitute of ‘Technology (MIT). Lil (1964) has reviewed the use of PVC for model testing and has reported considerable amount of strength and stiness ats. Thermosetting plastics differ from thermoplastics in that they cannot be remolded by beating ‘once they have been cast into their orginal shapes. Thermosetting plastis such as the epoxies {Aalite or polyesters (Marco, Palatal) can be sed foreasting models at room temperature with upper and Hower melds without the use of pressure or ovens (Rol, 1968) Since these casting resins ‘atin Tiga form prior to polymerization, they are equently used for casting very intrieate models; [ELASTIC MODELS ~ MATERIALS AND TECHNIOUES ® however consirable sill and experience are prerequisites to achieve succesful results, Tero- wet slssis tr poferred to thermoplastics inthe manufacture of sell models with varying ‘Teak Any complex curved surface with ny desired thickness variation cin be cast conveniently tang hemasting pestis. One oust remember ta the thickness of suck shells fected by we Ravone in thickness due to the forming proces. Tis variation is governed by dissipation of ‘he heat of polymerization, and therefore care must be exercised in using thermosetting pasties ‘The advantage of using epoxy resins Compared with thermoplastics is hat the Hmite devel- pment ofthe eat of polymerination assares a more homogeneous hardening process, which rests poiMesant elastic modulus throughout the mass (Fumagal, 1973). Also, the elatively lower “Fvinkge that occurs in epoxy resins after casing rests ina significant decrease inthe internal san These are particularly wsefl in models of varying thickness where the intemal stresses sitsrren lead to fracture ofthe mode. Properis of some plastics commenly used for strvctral vaoukts are deserbed in Table 3.1. Ie must be noted that ony typical values ofthe properties of Tastes sed hy some investigators are Histed. Thee ae oter avaiable pastes that may e suitable or model work pony resins also offer the possiblity of modifying tsi physical properties by adjusting the quantty of hardener or by adding an inert mister such a6 a ile dispersed homegeneovsly ‘Afavghowt the mass andor reinforcements consisting of inorganic or organic fibers. Silica sand, yromdeed met (luminom orion), cr, lea shot, polystyrene granules and ote ingredients Pane boon sucessflly used a fillers. The addition of llrs alters the material density and modifies the modulus of elasticity within wide Tinits. It also decreases the temperature rise duc to the {femted eat of polymerization and reduces shrinkage athe associate itera sess, Pow: ered cork, sand, and polystyrene hep reduce the values of the Poisson's ratio, while the wse of eecluminum powder increases the eral conductivity. which helps disperse the heat generated ftom tbe cleteal resistance strain gages later applied tothe surface, Properties of Some epoxy resin mixes with varying anvouns of selected fillers are shown in Table 32 3.32. Tension, Compression, and Flexural Characteristics of Plastics “The strength and stress-strain characteristics of plastics ae dependent on @ numberof factors, ‘sch asthe type of test (ension, compression, of Hexore), the specimen size te rate of Ying, and te previous tess history in tems of creep and relaxation. The mechanical properties of ‘lass are alo signiieany inuenced by temperature and relative humidity, which ae discussed Jhrmore dealin Section 35. The measured properties vary not ony frm batch to bath but also ftom one abect thickness Io another within the sare batch, Reasonable care must therefore be tretised to determine the properties in the laboratory under conditions of temperature and relative humidity sir to sein which the model will e both cast snd tested Also ifs important that Ff the model is subjected principally to dicot stresses, the modus of elasticity indirect tension ‘or compression should be determined fom tension oF compression tests on suitable specimens, ‘Simla, ifthe model is subjected principally to bending, the moduhs of elasticity sn flexure ‘should be determined using a cantilever beam test or smlar flexure tt. Tis recommended tht the tension and flexural specimens should be at fast 8 in. (200 mi) Tong, randomly seleced from the material for model construction. A brief description of the Specimens recommended by the American Society for Testing and Materials (ASTM) Standards tnd of specimens used by some investigators i presented in Table 33. Deas ofthe tension specimen Type 1 recommended by ASTM Standard D638 Standard Test Method for Tensile Properties of Pasties are shown in Figure 3.1. The specimen has a uniform ‘ross action (12.5 mim seet thickness) over ength of 57 mm and » gradu transition to the two enlarged ends to prevent failure atthe grips. I is recommended tha! this specimen be used in ‘model material evaluations 0 STRUCTURAL MODELING AND EXPERIMENTAL TECHNIQUES. Table 2.1 Properties of Some Plastics Suitable for Stucturl Models Tense Compresone rexural Woguus of ‘Thecmst Avaliable Strangih, Strength, Strong, ‘Elastic, Plastic (Characteristic “Shapes pal, pal pal Pal ELASTIC MODELS —1ATERIALS ANO TECHNIQUES ey ‘able 1 (command) Properties of Some Plastics Suitable for Structural Models "Elongation “Coetcient Poisson's at pure. Specie Softening of Expansion, sointng alo Pereont Gravity Temperatre-C —infinrC” Machinabity Characteristics {Cao wiaesTeopsie Shas, roa 300-7080 360-30 HND-F7 ADO ACD Guo 50 135. TIGHT Toe Cap comme ih ‘cali ‘nd ier ‘ten cent ey see ne) Cet sectates Themepaic Spt ros, 2250-1160 2209-19900 200-1500 65-2005 10" oo Hae as050 set ante cemented wih “oli ber ‘thee nd ehh Mato ‘Temeplstic Shes ok, 7000.00 12000-29009 30-1790 420-500 1° ose0s8 51D La) SDF xn Cane eee wt ethene td ee Sa ese (resis, ‘Slr te pla het Fagen) Ertom| PVC oka) Temeptcis Stet re, 350-1350 350-400 «10 02.040 $5.10 136140 BIOS S810 ale Can be welded wit PVC “odes ‘aro cemered ih foxy cements obetptencs —Themphsie Sheet, ros, 1000-5000 390-7000 17-80 1 045.050 nora saz toe Cin welded bt it ait) ‘as, lng ‘coment pin autor ements Shes am 10004000 = 290 050 Moan 095129 7S Oe con creed See br eroded pe ‘ats cements vse sins" Memesetin, Cagrenne Smee00 pm20009 so et “Miro Fs) ‘else po sine” Themoseing Casing esse SH0-12600 1500-3000) — 430400108 ones 51m = 2.910% Gaal Cam Be erent with pon Ale) cro amen ser Figure 81 Tension test specimen dimensions in mitre (or shes, pi, and mode plastics) (Courtesy orAStM Pilon) ‘The ASTM Standard 695 Standard for Compressive Properties of Rigid Plastics recommends that a right cylinder or prism whose length i wice is diameter or is principal with be used ‘determine the compressive properties of plastics; the preferred specimen size is 12% 12 25 mm for prisms snd 12 mm diameter x 25 mm high fr cylinders. For detenmination of elastic mods and the yield stress, the test specimen should be of such dimensions to avoid the slendemess or instability problems. The sendemess ratio geverlly used is between 11 and 15, The prefered specimen size fo prisms is 12 x 12 x 50 mm and for cylinders the recommended dimensions are 12mm diameter and SO mm height tis not necessary 0 machine the test specimen cross sections to the preferred sizes; compression specimens of suitable height can be cut from the material used forthe model consiracton according to the ASTM Standaed 695. Several investigators have used ASTM standard specimen to determine the flexural properties of different plastics. According to the ASTM Standard D790 Standard Tes Methods for Flexural Properties of Plastics and Electrical Insulating Moterals fr plastic matrils 1.5 non ot greater in thickness, the depth of the specimen for tise teste shall be the material thickness and for txlgewide tests the depth shall not exceed the wii; fr al tess the support span shall be 16 times the depth with sufficient overhangs to prevent the specimen from slipping though the support. ‘Nowe: Al plastics with specie gravity of 36 than 1.0 wil oot in wae ene eis. ‘able 322 Composion and Properties of Some Epory asin Mixes with Adi Fiera ‘er Fanaa ‘Table 3.3_TestSpuckmene Used by Various Researchers for Tete on Pastles Fecearcher ‘Specimen et. Test ‘size Commente ‘Sinminny (971) Ten ASTI Sadr Sesion Tea ropis of eas enoeed owes Cremer ta (196) Tewion 1.3025 16a “late popes of Plexi: pesimen (GF*6x 1 nm) ended in dam Fi Pee 7 ih veyingaouns of cate ‘6m Sam 200mm) "Br wed ee Asta Sere models: aed scien em ames pope of he above ie ai at Show (1965) Temon Bain zi Shun (1975) —Comprion 2. am (0 mada, 10) ASTHE Std Spsinen Flexwal provers of Serge eineced pele ain sal oes Peis Sian W971) eae ‘ob (ITD, Fam TD) Fee Hn sheet heb Sin. mm = set kee 20m) ‘STRUCTURAL MODELING AND EXPERIMENTAL TECHNIQUES, 3s a2 i @ i © Figure 22. tase tan-tine characte ola toear"Vacoeeaste ats Dimensions ofthe roller supports and the rounded eating pose, along with ther details, are also specie. 3.33 Viscoelastic Behavior of Plastics. ‘Most of the commonly used plastics exhibit Tinear viscoelastic properties; that is, the stress-stain relationship at any particular tine after leading is linear although the modulus of Clastcty varies with time. Inothervords, the stess-strai relationship doesnot conform to Hooke’s Jw, and the stain sa function of time ater losing, the Foading history and the stess level. Thos, if a plastic specimen is subjected to stress 6 tat is maintained constant for duration of time ¢ (Figure 32a), the strain response ofthe specinen will be as shown in Figore 3.20. There will be ‘an instantaneous stain inthe plastic immedinely upon the application of stress. Under constant ‘tres the strain increases frther with ime, and this ate of strain increase (del) is dependent on the stress intensity. According to Preece and Davies (1964), in most model studies the maximum stress intensity must be controlled so that (ede) is negative. If the stress intensity is above the rep strength, the vale of (eld) wil be postive nd the stain will proceed to increase at an increasing rte until he specimen fractures. This problem can normally be avoided by limiting the value of the maximum sites in the plastic well below the creep strength ofthe plastic. ‘Unde a constant stress, the total stain & at any time interval (Figure 3.28) consists of two ots; the instantaneous clastic strain, ¢ and ths time-dependent creep strain €, When the stress Femoved, there will bean instantaneous recovery of stan approximately equal to & followed by f slower stan recovery that gradually disappears with time. 1 the stess intensity ts doubled 0 2, the eovresponng teal strain atthe same me ¢ will be 2c, With he total instantaneous elastic ELASTIC MODELS ~ WATERIALS AND TECHNFOUES 93 oT 1 en erin 20 be ic Figure 33. Stoss-snin cons for PVC and Peas (ler Pall and Soo, 1964) and the creep stains being 2 and 2, respectively, ifthe plastic exhibits nea viscoelasticity Tt Fellows that atanyeiven fim interval the ratios (Gi, (26/2e... willbe constant, and thas the pfectve elastic modulus @2 is instantaneously constant wit respect to time. "The typical mechanical properties of commonly wed plastics are summarized in Table 3.1 The moduli can vary from about 42 10 58 Nimm? for plastic foams to values betwesn 6900 and T3800 MPa for some plas-einforced plastics. For common acrylics, polyesters, celloloids, poxies, and other plastics, the modal valves range between 2070 and 4140 MPa. These relatively eye model values (compared with sel nd concrete) result in measurable strains and defections inplate meds without requiring lage loads, Typical tension and compression stress-strain curves for Plexiglas (grade G and grade II UVA used at MIT) and atypical stess-sirain curve for PVC (oonmalimpact grade Type Fused at MIT}; set thickness of 0:8 men ond 1.6 mum are shown in Figure 33. A series of Comell University ess for tension on instrurented Plexiglas coupons “Showed tht afer about 10min of foading, the modulus of elasticity decreases to about 907% ofthe Jnstantancous value, The specimens ceased 10 remain Tinear vscoelaste after a strain of about 1700 pinJin, which isin agreement withthe range of 1300 {0 2000 yinJin. observed by other jnwestiguors, Alhough these Plexigls specimens underwent lage foal deformations, they were pot duct and exhibited litle waming of failure. However, some of the softer PVCs are more ‘uctile. Unlike other plastics, ethy! cellulose and polycarbonate (tirmoplasic Kind) have Stress stain curves in tension similar in shape 10 that of steel (Figue 3.4) but without strain hardening aod have been used by Haris etal. (1962) for model studies of welded steel frame Thructures subjected to blast-ype dynamic loadings. Mt should be notec that at small strains, the ftresestain curves for most plastics in tension and compression and the corresponding moduli of ‘asticiyaralmost identical Thus, the moduls of elastic in Hexare wil be equal othe modulus ‘of elasticity in tension or compression, However, here is a variation between the compressive and the tensile yield suength, the former being greater o ‘STRUCTURAL MODELING AND EXPERIMENTAL TECHIMQUES, VP _ 1 1] Figure Siess-stain curv of ey cadose, (ter Hae otal, 1962) 3.3.6 Mechanical Properties of Polyester Resin Combined with Calcite Filer Polyester casting resin with the trade name of Plastrene 97 (Balint and Shaw, 1968) is another ‘mate used fr elastic modes. The propenics of the resin are varied by using different amounts of calcite filler; thus itis possible to match the modal of elasticity of the dense and Highs concrete of the prototype sreture with that ofthe model in order to fall the similtud require ments (See Figure 35). The influence of varsion of the calcite filler on material shrinkage, ‘oeficent of Tinear expansion, strength, ceep characteristics, and gluabiiy have also been investigated. This matral was used in the malel studies of the circular, [83-m-high Australia Square. Syney, whichis constructed wit lightweight concrete in the upper stories (Gero and Cowan, 1970), Plastene 97 is a moldable material and was wed fortis model because of the Inricate shapes involved, such as recesses around doors, tapered columns and beams, and vated slab thicknesses; the change nthe modulus of eksticiy of the concrete through te building height; and the need for 3 low Poissons ratio. At room temperatare, complete polymerization of Plasttene 97 normally took place 90 days after the aiion ofthe catalyst andthe accelerator. However, his process was accelerated at elevated temperatures, and curing for about 3h at 65°C resulted in fully polymerized produit (Balint and Shaw, 1955), Best results were obtained by slow and even cooling ofthe heat-cred components, guarding them aguinst warping. This slow cooling process resulted in a practically stress-free material. Reheating the cured prodact to 81°C softened the atrial and easily eliminated any unwanted warps u indicated Inear viscoelastic behavior and equal moduli of elasticity in tension and compression. A S-min test vs, a 3-h test showed the presence of ereep strains and a varying modulus of elastty. Typical ereep curves for Mix 23 23% calcite filler by ‘eight showing variation of strain with time an temperature st two stress levels 1.7 and § MPa ae shown in Figure 36. The tensile strength of Mix 23 and Mix 130 were 36 and 23 MPa, respectively. Variation of the modulus of casicty of Mix 23 with ime and temperate is shown, in Figure 37. Testing these flly cured specimens at clevited temperatures reslted in a lose of tensile strength LASTIC MODELS - MATERIALS ANO TECHNIOUES. 2 a “tC i O Sag" onan 7 Figure 335 Varaon of modus of eatty of Plastene 97 wih cate tx. (Aor Ban and Shaw, 1965) 2 | 1D eet et 4 Figure 36 Tiel crep cures fo Plastene 97 mix 20% cae ery weight) (Aor Balt nd Shaw, 1965) 98 STRUCTURAL MODELING ANO EXPERIMENTAL TECHNIQUES cla of ay, hal MP Figure 37 vaiaten of moclue of saxty of Prana 97 mix (20% caste by weit) with re and ompertre (Mor Balt snd Stam, 1963) 3.4 TIME EFFECTS IN PLASTICS — EVALUATION AND COMPENSATION ‘The time-dependent stain or creep should be accoanted for in interpreting the experimental data, especialy i several sian gages ae wo be read and there is a ime delay in readings. Because ‘of reap, strain readings unde a given applied load are changing with time: therefore, no unique value of stress is associated with 2 measured strain value, and the time-dependent modulus of elasticity E, must be used wo evaluate the stess values fromthe coresponding stain values a a specified time 4 In an indirect mode tet, one derives the infvence lines or surfaces from the distorted shape ofthe move, and time dependent properties do not nfence the results. In these «eases, the change inte sess with ime at any’ pein, that i, elaxation, i inconsequential. However, in a direct model one cannot ignore the creer. of the plastic, and proper care must be used in iespreting the test data, Ihe sess values ae low (about 20% ofthe material yield strength), then creep and creep secovery are linear functions ofsress and all creep is eventually recovered upon the removal of stress. Similarly, if the materials subjected to strain, there is a stress corresponding tothe inital siran, and this stress decreases with time under a sustained strain. Tis relaxation is als a linear function of strain if the imposed stain is smal. Basi creep and relaxation behavior of plastics are normally determined by simple uniaxial or xara tests, and it hs been observed that in general the creep modulus of elasticity ad the relaxation modulus are about equal fora specified time 1 Hence, only one time-dependent modulus, called the apparent modulus, is necessary ELASTIC MODELS ~ MATERIALS AND TECHNIQUES te) Semen (ec a Figure 38. Time-dependent behav f last. (Ate Rh, 1968) 3.4.1 Determination of the Time-Dependent Modulus of Elasticity ‘and Poisson's Ratio Roll (1968) suggests that the tine-dependent sess curve or the effective modus of clastcity can be easily determined by conducting crep tests in tension or flexure Tor as few as two vale of ses Ts recommended that for ter acuracy these ets be conducted for thee (ifferent stress values 6, Oy and 0, As shown in Figure 3.8 atthe tne fg when the streses are fist applied the corresponding inal strains ae and, espestivey or stress values Below the proportional limit, these points (6 £)(By &)- ad (GE) lie ona straight ine conresponding to tani Figure 386), the slop of which isthe nti tangent modulus of elasticity ; (Figure 3.8e) "At sometime 1, the corresponding tal stain vles ar Gy Ey and yg respectively Figure 3.84), ‘phich result inthe stress-strain curve (Figure 38) andthe modulus of elasticity Ey (Figure 3.8), Ina similar manner, we can determine the stess-srain curves fortes f, and f; and the corte Sponding moduli of clasicity Ey and Ey respectively. Ths, tbe varaton of, with ime can be etermined (Figure 3.8). “The curve in Figore 3 8e may be represented by the equation: 63 oe STRUCTURAL MODELING AND EXPERIMENTAL TECHNIQUES where (isthe unit creep function (Le, the eeep due to unit stress. It must be remembered ‘hat Equation 3.3 is based on the assumption tat the material is sranffee when the sresey ane ‘upplied. However, like all viscoelastic materials, plastis exhibit a "memory" elec thas Fnction Of the stress history, including the currently applied tess. As fr elastic mseials, the principle of superposition is applicable for viscoelastic hehavior provided that the governing differential ‘equations are linear and thatthe resulting Jefermations do not significantly change the boundary ontitions, In soch a case, i can be shown that fora linea vscoelstic material, the total stain at any time is piven by eet Ffo Me as Where: 0, = the curently applied stress = the stress previously applied or removed t time (=the unit creep fonction EE, = the inital modulus of elasticity 4 can be noted that if dhe unit creep function cis independent of stress at low stress levels and ifthe material is strain-fee whea stress is applied, Equation 3.4 reduces wo Equation 3.3. Iti clearly advantageous o eliminate the complications ruling from the memory effects by loading the ‘model in asain free condition. This can be easly achieved by removing the load afer the sain readings have been taken and then allowing enough time for creep recovery before reloading the ‘model, Because of the principle of superposition, oe can expec the model to recover completly itis left unloaded for as long a duration as fer which it was loaded - The Poisson's ratio v, can be easly determined by using ational stain gages installed perpendicular othe longitudinal gages in a uniaxial tension of flexural test. These gages ae used {to measure the lateral strains. The value ofthe Poisson's ratio at ime ts given by tera srnin Tongindinal sain 6s In genera, the Poisson's rato is not as sensitive to time effects as is E, “Another indirect method to determine the Pcisson's ati is to determine the value ofthe time

You might also like