DETECTION AND ASSESSMENT OF SEISMIC STRUCTURAL

2m ago
13 Views
0 Downloads
2.47 MB
98 Pages
Transcription

(', PB88163712i, 1111111 /I /111111111 II 1/1111/111'-----,--II'\NATIONAL CENTER FOR EARTHQUAKE.ENGINEERING RESEARCHII IIState University of New York at BuffaloDETECTION AND ASSESSMENT OFSEISMIC STRUCTURAL DAMAGEbyEdmonda DiPasquale and Ahmet S. ,CakmakPrinceton路 UniversitySchool of Engineering and Applied ScienceDepartment of Civil Engineering. Princeton, NJ 08544NCEER-87-0015 ."'August 25, 1987This research was conducted at Princeton University and was partially supported by theNational Science Foundation under Grant No. EeE 86-07591.REPRODUCED BYU.S. DEPARTMENT OF COMMERCENATIONAL TECHNICALINFORMATION SERVICESPRINGFIELD, VA. 22161I;;;;: --.- ;;o- )

NOTICEThis report was prepared by Princeton ,University as a resultof research sponsored by the National Center for EngineeringResearch (NCEER). Neither NCEER, associates of NCEER, itssponsors, Princeton University or any person acting on theirbehalf:a. makes any warranty, express or implied, with respect to theuse of any information, apparatus, method or processdisclosed in this report or that such use may not infringe. upon privately owned rights; orb. assumes any liabilities of whatsoever kind with respect tothe use of, or for damage resulting from the use of, any information, apparatus, method or process disclosed in thisreport.

50272 -10111. REPORT NO.REPORT DOCUMENTATIONPAGE.,4. Title and Subtitle3. Recipient's Accession No.2."88NCEER-87-00l'S,1 6 3 7 1 2/1C5. Report DateDetection and Assessment of Seismic Structural DamageAugust 25, 19876.7. Author(s)8. Performing Organization Rept. No:Edmondo DiPasquale and Ahmet S. Cakmak9. Performing Organization Name and Address10. Project/Task/Work Unit No.National Center for Earthquake Engineering ResearchState University of New York at BuffaloRed Jacket QuadrangleBuffalo, NY 1426112. Sponsoring Organization Name and Address 11. Contract(C) or Grant(G) No. ECE-86-07CC)L-' 59186-501313. Type of Report' & Period CoveredTechnical ReportSame as. box 9.14.15. Supplementary Ne:rtes .This research was conducted at Princeton University and was partially supportedthe National Science Foundation yThe problem of earthquake damage assessment is defined. The role ofthe analysis of strong motion records in damage assessment codes based on expert systemis pointed out. A literature sur ey on damage assessment and structural system identification is presented.16. Abstract (Limit: 200 words)A program for the identification of linear ,structures, based on strong motion records,has been written and implemented by the authors. The theoretical and computationalaspects of the identification algorithm are discussed.Strong motion records from six buildings that experienced different levels of damageduring the San Fernando earthquake (1971) are analyzed. The analysis of strong motionrecords of buildings that have been damaged during the San Fernando earthquake shows agood agreement between the numerical values of the damage indices and the levels ofdamage observed in the actual structures. If this result is confirmed by furtheranalysis of both strong motion records from actual structures and small scale experimentsthese damage indices can be used as a measure ,of structural damage. Areas of futureresearch are outlined.17. Document Analysisa. Deseriptorsb. Ide'ntlfiers/Open路Ended TermsDAMAGE INDICESSTRUCTURAL ANALYSISSTRONG MOTIONEARTHQUAKE ENGINEERINGSAN FERNANDO EARTHQUAKEc. COSATI Field/Group18. Availability StatementRelease unlimited19. Security Class (This Report)unclassified21. No. of Pages9 I-------------i---"'- ----20. Security Class (This Page)22. Priceunclassified(See ANSI-z39.18)See InstructIons On ReverseOPTIONAL FORM 272 (4-77)(Formerly NTIS-35)Department of CommerceJ

ABSTRACTThe problem of earthquake damage assessment is defined. The role of the analysis ofstrong motion records in damage assessment ,codes based on expert system is pointedout. A literature survey on damage assessment and structural system identification ispresented.In order to quantify the damage to which a simple structural element is subjectedunder earthquake or earthquake-like excitation, .several indicators (damage indices)have been proposed. Numerical sImulations have been used to study the dependenceof different damage indices on the paramenters of the structure and of the groundmotion. Furthermore, it can be shown that many of the proposed indices are statistically equivalent, and therefore carry the same informatiqn about the damage state ofthe structure.Both peak deformation and fatigue load contribute to the damage of a structural element. A global damage index for a complex structure should as well consider the twocomponents of damage. Damage indices can be computed from the optimal time variant linear model, that is fitted to recorded strong motion accelerograms. Two of theseindices are proposed, and their performance in estimating other response based indicesis discussed.A program for the identification of linear structures, based on strong motion records,has been written and implemented by the authors. The theoretical and computationalaspects of the identification algorithm are discussed.Strong motion records from six buildings that experienced different levels of damageduring the San Fernando earthquake (1971) are .analyzed using the techniquesdescribed in chapter 5. The damage indices proposed in chapter 4 are estimated. Thecomputed values are found to be consistent with the level of damage.The analysis of strong motion records of buildings that have been damaged during theSan Fernando earthquake shows a good agreement between the numerical values ofthe damage indices and the levels of damage observed in the actual structures. If thisresult is confirmed by funher analysis of both strong motion records from actual structures and small scale experiments, these damageir .dicesstructural damage. Areas of future research are outlined.can be used as a measure of

TABLE OF CONTENTSTITLEPAGE11.11.2INTRODUCTIONStatement of the ProblemOrganization of the Work1-11-11-32DAMAGE ASSESSMENT AND STRUCTURAL SYSTEMIDENTIFICATION: A LITERATURE SURVEYIntroduction;;Damage ofSimple Structural ElementsDamage Indices For Complex StructuresIdentification of Structural SystemsEffect of Structural Damage on the Vibrational ParametersDatabase on Seismic Structural DamageSECTION2.12 稤AMAGEINDICES AS A MEASURE OF DAMAGE DURINGEARTHQUAKES: A STUDY BASED ON NUMERICALSIMULATION :Introduction:Description of the Experiment.Correlation Between IndicesInfluence ofthe Parameters of the Ground Motion and of theStructure on the Damage E INDICES BASED ON SYSTEM IDENTIFICATIONUSING LINEAR MODELS;IntroductionDamage Indices Based on Equivalent Modal Parameters.4-14-1.4-3MUMOID: A PROGRAM FOR THE IDENTIFICATION OFLINEAR STRUCTURAL SYSTEMSIntroduction;:Model of the StructureEstimation of the ParametersImplementation of the Procedure5-15-15-15-4.5-6APPLICATION TO STRONG MOTION RECORDS FROMTHE SAN FERNANDO EARTHQUAKEii 6-.1

TABLE OF CONTENTS :7-18-1APPENDIX A: DERIVATION OF EQS. 5.2.5a AND 5.2.5.b .;A-1APPENDIX B: DERIVATION OF THE DIFFERENCE EQUATION 5.3.1 FROMTHE MODAL EQUATIONS OF MOTION 5.2.3a AND 5.2.6 .B-1APPENDIX C: UNCONDITIONAL LlKELlliOOD FUNCTION FOR ATRANSFER FUNCTION MODELC-1APPENDIX D: DESCRIPTION OF MUMOlDD-1111

LIST OF ILLUSTRATIONSFIGURETITLEPAGE1.1"Conceptual Flow Chart for Pamage Assessment Using ExpertSystems (from Yao (1982禄1-23.13.23.33.43.5Simulated Earthquake: Magnitude 8.1, RockSimulated Earthquake: Magnitude 8.1, Hard Soil,Simulated Earthquake: Magnitude 8.1, Soft Soil.Simulated Earthquake: Magnitude 8.1, Very Soft SoiLComparison Between Energy Index and Normalized CumulativeDisplacement.路Comparison Between Stephens and Yao's Index and its SimplifiedVersion':Comparison Between Energy Index and Stephens and Yao (1)Comparison Between Energy Index and Stephens and Yao (2) . Comparison Between Ductility Index and Energy Index:Relation BetweenDuctility Index and Normalized MaximumAccelerationRelation Between Energy Index and Normalized MaximumAcceleration:Normalized Frequency and Cumulative Energy for the EarthquakesSimulatedIndex Ratio vs. Normalized Frequency: Low Cumulative EnergyIndex Ratio vs. Normalized Frequency: Intermediate CumulativeEnergyIndex Ratio vs. Normalized Frequency: High Ctimu1ativeEnergyComparison of the Resonance Curves.3-43-53-63-74.1.4.24.3Comparison Between Ductility Index and Maximum SofteningSymmetric and Nonsymmetric Hysteresis Cycles;Comparison Between Energy Index and Cumulative Softening".4-8.4-9.4-116.1 .6.2 .6.36.46.56.6Accelerograms from Millikan Library (EW)Accelerograms from 611 West 6th St. (N38E) :Accelerograms from Sheraton Hotel (S90W)Accelerograms from Holiday Inn Orion (S90W)Accelerograms from Bank of California (N79E)Evolution ofthe Equivalent Fundamental 183-193-20.3-213-22

LIST OF ILLUSTRATIONS (Continued)FIGURETITLEPAGE6.76.8Maximum Softening vs. Damage LevelCumulative Softening vs. Damage Level6-96-10LIST OF TABLESTABLETITLE1.1Damage Indices 路PAGE 1-4v

. SECTION l:INTRODUCTION1.1.Statement of the Problem.When a major earthquake strikes an urban area, one of the most compelling problems. that engineers face is to evaluate the safety of existing structures. This task is usuallyby inspecting the buildingsin question. Visual inspectioncan identify. \accomplished,., .cracks and permanent deformation, and field testing can measure the degradation ofthe structure. Eventually, expert engiheershave to decide what,' if any, action shouldbe taken, ranging from cosmetic repairs to the demolition of the building.The analysis of strong motion accelerograms will play a major role in an automaticdamage assessment scheme. In: fig. 1.1 (from Yao,1982), a flow chart for a damage.assessment procedure based on expert systems is illustrated. From the accelerationresponse records, informations about both the three damage classifiers (global,cumulative and local damage) can be obtained).In this study, an attempt is made to define procedures and algorithms, so thatinformations about global and cumulative damage can be extracted from strongmotion records.When laboratory tests are performed, model structures are extensively instrumented,and local damage can be analyzed in .great detail. However, the number ofobservation points is, in real structures, very limited. One accelerogram array isplaced in the basement of the structure, and this record can be used as input to the .structural system considered, if soil-structure interactions are neglected (the absenceof soil structure interactions is a working assumption throughout this report). Another路array would usually be placed at some upper level, so that the response of the buildingcan be observed. These data can still give some information about the globalbehavior of the structure, when system identification based on modal decomposition isused.1-1

MeasurementSpace6 PatternSpace.I Decision Function or ClassifierI ClassificationO\JtputDamage dueto Driftf---------路 --- I Data fromI Load iestI"'.StiffnessDegra da tionI.J------tion-路,--1.!. . AeceleraDamping Rela tedDamageI: Response.L RecordsFatigue &.Creep Damage.- -- -- ---,;-.: Npeetion : Results and': Observations .' '1Damage路 toFrames(Global &. LocI I)I----------J""'- ---------Ft!arure selection and extractor with intermediate diagnostic staresFig. 1.1 : Conceptual Flow Chartfor Damage Assessment UsingExpert Systems (from Yao (1982))1-2Damage State

1.2. Organization of the workThis report stans with a literature review (chapter 2) of the assessment of damage forsimple elements and complex structures, and of system identification techniquesapplied to structural dynamics. 'chapter 3 presents the results of a series of numerical simulations, that have beencarIjed out to test the performances of several damage indices in the damage analysis,.of SDOF (Single Degree of Freedom). non linear systems to pseudo-earthquakeexcitation.Damage indices based oIl system identification are introduced in chapter 4. Thesystem identification algorithm that has been developed" and implemented by theauthors is described in chapter 5.Strong motion records from the San Fernando earthquake (February 9, 1971) havebeen analyzed. The results are presented in chapter 6.Several damage indices, both proposed in" the "literature during the past years orintroduced by the authors, are described. A summary is presented in table 1. L1-3

TABLE 1.1: DAMAGE INDICESIindexIformulanotesChapter 2 (literature survey)XMductility ratioNewmark and Rosemblueth (1974)xy-;t1 . OJi BerteroBertero and Bresler (1971)W(T'liSi1 Yiri1 OBanonf(i ' leBanon and Veneziano (1982)lei I)eyXM ,XuFyxOParkPark and Ang (1985)u1t(ll PtOStephens1 II pf Stephens and Yao (1985))i aChapter 3 (numerical simulations)ductility index ( Il:c )physical component of OParkXM'-Xuphysical component of OPa k' energy index ( IlE )FyxuChapter 4 (system identification)ultimate stiffness degradationmaxim msoftening ( oM)cumulative softening ( OE )(T O)final - (T O)inirial(T 0) initialglobal degradation(llT O)ii l,nwind (T O)initialglobal damage due to peak deformationmaxnr1 (llT O)i(T O)initials'1(TO)i1-4global damage due to fatigue

SECTION 2:. DAMAGE ASSESSMENT AND STRUCTURAL SYSTEMIDENTIFICATION: A LITERATURE SURVEY2.1.IntroductionStructural damage is a complex phenomenon that is very difficult to modelanalytically or to reproduce in laboratory experiments.A certainu,nderstanding has been achieved when two limit cases for the failure. mechanism are considered: static load, when the strain is increased monotonicallyuntil the element that is being tested breaks, and low cycle fatigue, when repeatedstrain- cycles, below the