Text

Recommends That NUREG/CR-0345 Be Sent to All Boards Before Which Applications Are Pending.W/Evaluations of Rept & Marked Up Rj Mattson 800303 Memo Commenting on Research Info Ltr 79 Re Equipment Seismic Testing
	ML19309G506
Person / Time
Site:	Trojan File:Portland General Electric icon.png
Issue date:	03/03/1980
From:	Knight J; Office of Nuclear Reactor Regulation
To:	Varga S; Office of Nuclear Reactor Regulation
Shared Package
ML19309G497	List: ML19309G496; ML19309G500; ML19309G506; NUREG/CR-0345, Advises That NUREG/CR-0345 Has Been Reviewed & Provides Useful & Independent Confirmation of long-time Concerns;
References
	RTR-NUREG-CR-0345, RTR-NUREG-CR-345, TASK-AS, TASK-BN-80-09, TASK-BN-80-9 BN--80-09, BN-80-9, NUDOCS 8005070063
	Download: ML19309G506 (8)
	v • d • e

{{#Wiki_filter:, ^ p* "8 49 ,,(7) e rj'(C, UNITED STATES NUCLEAR REGULATORY COMMISSION Ob 3 WASHINGTO N. D. C. 20555 8005070 5.lL -

,I i

,./ MAR 3 1980 MEMORANDUM FOR: S. A. Varga, Acting Assistant Director IN,..., for Light Water Peactors Division of Project Management FROM: J. P. Knight, Assistant Director for Engineering Division of Systems Safety

SUBJECT:

BOARD NOTIFICATION REGARDING t;UREG/CR-03 5 ENTITLED "AN EVALUATION OF SEISMIC QUALIFICATION TESTS FOR NUCLEAR POWER PLANT EQUIPMENT" We have reviewed the attached subject report. The staff evaluation of this-report is attached. Because the information in the report and the staff evaluation are relevant to all plants in licensing, we recommend that this information be provided to all Boards before which there are pending applications. Since equipment cualification was a matter before the Diablo Canyon Board and l since the Appeal Board is now considering the Intervencr's brief on the Diablo Canyon appeal and the Staff's and Applicant's responses, we request that Oe Diablo Canyon Appeal Board and Licensing Board be provided with this information as soon as possible. J. P. Knight, Assistant Director for Engineering Divisicn of Systems Safety cc: R. Mattson D. Ross D. Eisenhut L. Shao i V. Noonan R. Bosnak i

i;RR Staff Evaluation of i;UREG/CR-0345' I I. PURPOSE OF THIS RESEARCH: To meet seismic requirements safety related equipment is generally qualified by testing. The test input (shake table motion) is expected to adequately simulate the specific seismic environment and to consider.its sensitivity to equipment response, which may vary greatly from case to case. Since the regulatory position must be general in nature, the selection of a test input for a specific application needs experience and e7;ineering judgement. Be-cause simpler test inputs have been used in many cases.for equipment qualifi-cation before the existence of the current criteria, this research program was requested and intended to provide a basis for comparing tne effectiveness r of various test inputs. 4 II. ACHIEVEMEllTS & CO.5tEriTS: Studies were conducted by subjecting one typical electrical cabinet to shake table tests using different wave forms. tio internal electrical equipment of any kind was tested with the cabinet, and the effects of the test input on equipment operating function was not included in the investigation. Primary findings and staff comments are as follows: 1. A numerically defined " Damage Severity Factor" (DSF) was developed and introduced as a way for comparing severity of various types of seiscic qualification test inputs. When the DSF is fully developed, it may have the potential to Nsess relative damage that can be-inflicted by earthquake transients or test inputs to structural components. However, no immediate application of the DSF to equipment seismic qualification is recommended in its.present fo rm. The relationship between DSF and equipment operability requires further investigation. 2. The research results concluded that the single frequency sine dwell'and sine beat tests are far more severe in general than the biaxial random . tests for verifyin. structural integrity of passive equipment and supports. We are aware that the single frequency sinusoidal test input at resonance is generally a very severe structural test, but this is not the case from-the standpoint of verifying the operability of active equipment. Single frequency sinusoidal testing also does not necessarily.' yield valid results when testing.to determine resonance. This fact was known to the staff through licensing reviews since 1972 when a revision of IEEE Standard 344, 1971 was initiated at the request of I'RC. tievertheless, these research resalts provide a useful independent conformation of the above facts. 3. It was found that there were somt differences between modal data obtained from the same cabinet when mountid to a concrete floor and when counted on the :, hake table. i ~ c e r

.g. = ~. _2-It is a well known fact to dynamicists that modal data will be effected a whenever the boundary conditions are changed and whenever. dynamic coupling e.ists between the fixture and the test item. An MEB Eranch Position f.eveloped in 1973 emphasized that items being tested should simulate service mounting and should avoid dynamic coupling with the fixture. This position was later adopted in the Standard Review Plan, Section 3.10 issued in 1974 and also incorporated in the revised IEEE Standard 344 in 1975. Equipment mounting has remained a concern of the SQRT audit program since its initiation in 1974 and continues to receive special attenticn in our review process. The research results provided further justification that our concerns are valid. 6 4. It was revealed that some deficiencies may exist in the use of response spectrum for seismic qualification testing.

t was stated that the criterion requiring the response spectrum of testing input- (TRS) to envelope the specific response spectrum required for the equipment qualification (RRS) may not ensure proper energy distribution through the range of frequencies tested and may actually indu:e an excessive zero period acceleration (ZPA), which, in turn, may cause an over test.

Although the development of explicit and generic guidance to achieve proper energy distribution and proper 2PA has not yet been completed, ~ these shortccmings can be avoided if the wave form of the test input is carefully reviewed. The complex wave forms used by. Westinghouse in their 1974 and 1975 generic testing programs were typical examples of carefully reviewed test input wave-forms. The staff has constantly addressed such concern 3 in-licensing reviews since 1974, when SQRT started systematic plant seismic audits, especially on those' items of equipment tested at an earlier date. The research results have provided further evidence of our concerns. In order to improve the regulatory process, further efforts in this area have been recommended and have been incorporated into a request for contract proposal to be issued by the Office of Nuclear Regulatory Research in the near future. The IEEE Standards Committee responsible for developin; equipment' seismic qualifi-cation guidance was also informed of the research results for possible refinerent of their current criteria. III. CONCLUSIONS: In summary, the research results provide a useful and independent confirmation ~ of certain staff concerns which have existed for several years. These con-cerns are_either already explicity stated in the existing regulatory position,- or have-been addressed in past licensing reviews. These research results do not. impact the regulatory process at-the present. ime'but future. efforts by RES or the IEEE Standards Committee ' refining their current criteria may have impact'. i 4 -

r---- f gb y7 pa ar c, o [{$ e fP j c UNITED STATES o NUCLEAR REGULATORY COMMISSION $? N, E W ASHINGTON. D. C. 20555 1 ..... of L,,. ,, 1333 ,j l ~ MEMDRANDUM FOR: William T. Russell, Acting Chief i Technical Support Branch, NRR d [ FROM: Roger J. Mattson, Director Division of Systems Safety, NRR NRRCOMMENTS0,'RIL79)-EQUIPMENTSEISMICQUALIFICATIONTESTS SUEJECT:

~ d

REFERENCES:

(1) DDR memo on RIL 79, 2/12/80 g/ ~ g[ (2) TSB memo to DSS & DDR, DSS and DDR have reviewed the subject RIL and the associate EG/CR-03h entitled "An Evaluation of Seismic Qualification Tests for ff Equip,ent". This memorandum incorporates the DDR comments (Reference 1) on /% RIL 79, and in addition, incorporates the DSS experience perform d in licensing revien's by the Seisr.ic Qualification Review Team (SQRT), and should be considered the coordinated NRR response requested by Reference 2. zS, 4, I. PURPOSE OF THIS RESEARCH: &hMr 4"//7 To rieet seismic requirements safety relatec equipment is generally queiified by testing. The test input (shake table motion) is expected to adequately simulate the specific seismic environment and to consider h its sensitivity to equipment response, which may vary greatly from case to case. Since the regulatory position must be general in nature, the selection of a test irput for a specific application needs experience and en ;ineering judgement. Because simpler test inputs have been used in many cases for equipment qualification before the existence of the current criteria, this research program was requested and intended to provide a basis Tor comparing the effectiveness of various test inputs. II. ACHIEVEMENTS & reg",Eigjs: Studies were conducted by subjecting one typical electrical cabinet to shake table tests using different wave forms. No internal electrical equipment of any kind was tested with the cabinet, and the effects of the test input on equipment operating function was not included in the investigation. Frimary findincs and our ccments are as follows:

Contact:

S. N. Hou, DSS:MEB X27538/72 ~ ,,e .. w., see a _ s

W.T.Musse11 1. A numerically defined " Damage Severity Factor" (DSF) was developed and introduced as a way for comparing severity of various types of seismic qualification test inputs. When the DSF is fully developed, it may have the potential to assess relative damage that can be inflicted by earthquake transients or test inputs to structural components. However, no immediate appli-cation of the DSF to equipment seismic qualification is recommended in its present form. The relationship between DSF and equipment operability requires further investigation. 2. The research results concluded that the single frequency sine dwell and sine beat tests are far more severe in general than the biaxial random tests for verifying structural integrity of passive equipment and supports. We are aware that the single frequency sinusoidal test input at resonance is generally a very severe structural test, but this is not the case from the standpoint of verifying the operability of active equipment. Single frequency sinusoidal testing also does not necessarily yield valid results when testing to detennine resonance. This fact was known to the staff through licensing reviews since 1972 when a revision of IEEE Standard 344, 1971 was initiated at the request of NRC. Nevertheless, these research results provide a useful inde-pendent confimation of the above facts. 3. It was found that tnere were some differerces between modal data obtained from the same cabinet when mounted to a concrete floor and when mounted on the shake table. It is a well known fact to dynamicists that modal data will be affected whenever the boundary conditions are changed and whenever dynamic coupling exists between the fixture and the test item. An MEB Branch Position developed in 1973 emphasized that items being tested should simulate service mounting and should avoid dynamic coupling with the fixture. This position was later adopted in the Standard Review Plan, Section 3.10 issued in 1974 and also incorporated in the revised IEEE Standard 344 in 1975. Equipment mounting has remained a concern of the SQRT audit program since its initiation in 1974 and continues to receive special attention in our review process. The research results provided further justification ti,at our concerns are valid. 4. It was revealed that some deficiencies may exist in the use of response spectrum for seismic quali fic3 tion testing. It was stated that the criterion requiring the response spectrum of testing input (TRS) to envelope the specific response spectrum required for the equipment qualification (RRS) may not ensure proper energy distribution through the range of frecuencics tested and may actually induce an excessive zero period acceleration (IPA), which, in turn, may cause an over test.

N 2 1333 W.T.E,$ sell '3-The staff has constantly addressed such concerns in licensing reviews since 1974, when SQRT started systematic plant seismic audits, especially on those items of equipment tested at an earlier date. Although the develo:r.ent of explicit and generic guidance to achieve proper energy distribution and proper ZPA has not yet been completed, .these shortcomings can be avoided if the wave form of the test input is carefully reviewed. The complex wave forms used by Westinghouse in their 1974 and 1975 generic testing programs were typical examples of carefully reviewed test input wave foms. The research results have provided further evidence of our concerns. In order to improve che regulatory process, further efforts in this area have been recommended ar.d have been incorporated into a request for contract proposal to be issued by the Office of Nuclear Regulatory Research in the near future. The IEEE Standards Committee responsible for developing equipment seismic qualification guidance was also infomed of the research results for possible refinement of their current criteria. III. CONCLUSIONS: In su=ary, the research results provide a useful and independent confirmation of certain staff concerns which have existed for several years. These ccncerns are either already explicitly stated in the existing regulatory position, or have been addresseo in past licensing reviews. These research results do not impact the regulatory rrocess at the present ti e but future efforts by RES or the IEEE Stanc 'rds Corittee refining their current criteria may have impact. Enclosed for your consideration is a copy of suggested input to " Official Progra-Office Comments on Utilization or Value of RIL to the Regulatory Process". i p i // fi / / j' b4Nf % Roger J. Mattson, Director Division of Systems Safety cc: H. Denton, NRR D. Eisenhut, DDR J. Knight, DSS L. Shao, RES R. Bosnak, DSS F. Schauer, DSS V. Noonan, DDR D. Crutchfield, DOR 1 W. Anderson, SD J. Richardson, RES G. Bagchi, RES W. Paulson, NRR P. Kuo, DDR P. Chen, 00R J. Lankins, RES S. Hou, DSS

u m ~- c.gy; ~ ..s n-y..:- ..> xQ&O.W-fi $*N.4 ~

n..

MRR I W T TO ~ ^ OFFICIAL PRO 3PM OFFICE C0%4ENTS .. ' S ^. L.. - ~

,gqply.g.e g.g3qf-9j..

_e:>

w.- .n w 'x,n .-*.;:,i:,..

- c-ON UTILIZATION OR VALUE OF RIL TO tie REEPJLATORY PROCESS (RIL!79issueden 12/28/79)

+ . --. fw.; 3.&.W.' v', ;:c:y - NRR CO WENTS, FEBRUARY 15, 1920, R. J. MTTSON & D. G. EISENHUT DESCRIBE APPLICATION TO REGL"_ATORY FROCESS: The research results have s+ A. provided an independent confiration of the following licensing positions: .y (1) The single frequeny test in?u: r.ay be severe for verifying the structural integrity of passive equipment and supports, but may be inadequate for verifying the operability cf active equipment. ~~ -(2) Test items should simulate che actual service mounting during the test, and dynamic couplin; with tne fixture should be avoided. B. identified areas of futu e resea c3 for potential assistance in the licensing process: (1) The Da. age Severity Factor, if further developed, may be useful to assess the relative damage that can be inflicted by earthquake transients. :- or test inputs to structurai c:rponents. Currently the DSF is not useful in assessing functicnability but additional work is warranted in this area. (2) Explicit guidance sh:uld be developed to handle generic testing boncerns-relative to: (a) energy content versus frecuency distribution and 2_.. (b) pmper value of the ZPA level in the test input. .mo* o o Ju o MU U Alfta i [ "M

e. ~ . DESCRIEE IM ACT 0: RESL"_TS : The research results are esser.tially confirnative in nature. Areas identified have already received staff attenti:n even prior to the beginning of this research prograr, consequently the research results have no impact on regulatory requirer.ents. However, future efforts by RES or the IEEE Standard Committee on refining curren criteria maj nave ic;act. COMMENTS /REKA.r:: This research program has had relatively limited scope and resources and ccnsecuently has not or;vided dire :ly usable new procedures or methods of qualificaticn of eq;iorent. Fo-ever, it did serve a useful purpose in identifyir.; the poter.:ial a eas nee:e: for further investigation which sh uld be continued. Eased on the research results, NRR has taken the follwing actions: 1. The IEEE Standards Co r'ttee reso:nsible for developing equipment seismic qualificati:n cuida,ce nas been ir. formed of the research results. Invest-ica:icn by the Conr.ittes to deten-ine if any possible refinement of current i criteria is necessa y is underway. 2. Further resea-ch on the Danace Severity Factor and the criteria to impro've the frequen:y versus ener;y distribution in the test input specificaticn continues to be a re:c rerdatier cf NRR for consideration by the Office of Nuclear Regulatory Research. t 3. Reviesers have beer ir.stru:ted cf the continued need to maintain a careful l review cf all tes; in:u: functi ns used or proposed for equipment qualification. P l0 [}3[} X}l@rb ..kD ~ D llMU20 i

NUREG/CR-0345 SwRI PROJ. 02-4675 R0 AN EVALUATION OF SEISMIC QUALIFICATION TESTS FOR NUCLEAR POWER PLANT EQUIPMENT FINAL REPORT September 1,1976 - August 31,1978 Daniel D. Kana Robert W. LeBlanc Manuscript Completed; February 15,1978 Date Published: September 1979 Southwest Research Institute 6220 Culebra Road San Antonio, Texas 78284 Prepared for Office of Nuclear Regulatory Research l U.S. Nuclear Regulatory Commission Under Contract NRC-04-76-372 NRC FIN NO. B6000 I 79/0/00 70 7

ABSTRACT A series of different seismic qualification tests has been conducted on a typical nuclear power plant electrical cabinet in order to provide comparative data. Acceleration and strain responses were measured for four different ground level and six different floor level specifications. The test types include resonance ;earch, biaxial independent random, biaxial de-pendent random, uniaxial random, sine beat, and sine dwell excitations. Tests involving random motion were derived both from a random generator and earthquake signal source. Response data are initially presented in terms of transfer functions, time histories and response spectra. Then, analytical parameters are developed for correlation of the data in terms of peak re-sponses, tbne-average RMS responses, and a new parameter defined as a damage severity factor. Several important conclusions result from the data correlation for the various tests. Typical sine dwell and sine beat tests are found to be far more severe in general, than biaxial random simulations. The developed damage severity factorn indicate this result vividly, and also provide a useful design tool for comparison of test severities before the tests are conducted, so'that a choice can be made. It 1:s found that the choice of random generated or earthquake sources is immaterial for test development. Modification to test procedures are recommended for cases where differ-ences may be anticipated in floor mounted and simulator mounted resonance tests. Furthermore, a significant discrepancy is discovered in the simple specification that a TRS match or overlap an RRS. In certain cases this requirement is found to be inadequate for assuring a valid test in which all structural modes respond properly. This result is particularly impor-tant for those cases where subsequent component qualification tests are to be based on response spectra genernted from response measurements at compo-nent attach points on the basic cabinet. iii

1 1 TABLE OF CONTENTS Page i ABSTRACT 111 LIST OF FIGURES Vii

1.0 INTRODUCTION

1

2.0 DESCRIPTION

OF TEST SPECIMEN 4 1 2.1 Physical Design 4-2.2 Instrumentation 4 3.0 TEST PROCEDURE 9 4 I ~ 3.1 Floor Mounted Tests 9 3.2 Earthquake Simulator Mounted Tests 9 3.3 Earthquake Simulation Test Matrix 9 3.4 Allowsnce for Cabinet Failure 15 j 4.0 DEVELOPMENT OF EARTHQUAKE TIME HISTORIES 17 l 4.1 Ground Level Tests 17 4.2 Floor Level Tests 19 4 5.0 RESULTS FOR RESONANCE SEARCH TESTS 21 l 5.l_ Floor Mounting-21 l 5.2 Earthquake Simulator Mounting. 21 6.0 TYPICAL RESULTS FOR SIMULATED EARTHQUAKE TESTS 30 6.1 Ground Level Tests 30 6.2 Floor Level Tests 38 7.0 DERIVATION OF DATA CORRELATION PARAMETERS 52 7.1 Response Spectra Relationships 52 Development of Test Severity Factors 55 L 7.2-8.0 ANALYSIS OF' CORRELATED RESULTS 59 i 8.1 Mechanical Behavior of Cabinet-59 f 8.2 Comparison'of-Severity for'Various Tests '69 8.' 3 Floor Versus Simulator Natural Modes 72 V - ___=

TABLE OF CONTENTS (Cont'd) Page 8.4 Component Excitation 73 3.5 Effect of Excessive ZPA 75 9.0

SUMMARY

OF CONCLUSIONS AND RECOMMENDATIONS 79 10.0 ACKNOWLEDGEMENTS 82

11.0 REFERENCES

83 4 vi

4 1 LIST OF FIGURES 1 Figure Page i 2.1 Sketch of Electrical Panel Specimen 5 2.2 Positions for Instrumentation 6 2.3 Detail of Cabinet Base, Left Front Corner 8 3.1 Floor-Mounted Arrangement for Cabinet and Apparatus 10 3.2 Simulator-Mounted Arrangement for Cabinet 11 and Apparatus I 3.3 control and Analysis Diagram for Biaxial Seismic 12 Simulator 4.1 IIorizontal Ground Level Response Spectrum 18 4.2 Vertical Ground Level Response Spectrum 18' 4.3 llorizontal Floor Response Spectrum 20 4.4 Vertical Floor Response Spectrum 20 5.1 Cabinet Natural Modes Below 3511z 21 j 5.2 Top Acceleration Responses for Floor-Mounted 23 j Sweep Test 3 5.3 Interior Panel Acceleration Responses for Floor-24 Mounted Sweep Test - Y-Axis Fxcitation 5.4 Strain Responses for Floor-Mounted Sweep Test 24 I 5.5 Top Acceleration Responses for Simulator-Mounted 25 Sweep Test 5.6 Interior ' Panel Responses for Simulator-Mounted - 26 Sweep Test i 5.7 Strain Responses'for Simulator-Mounted Sweep Test 27 5.8- . Transfer Function Characteristics 29 ~ 6.1 -Responses for Biaxial Independent Random Ground 31 i Level Test, YZ-Excitation ~ - 6.2 Response Spectra for Biaxial Independent Random

32 Ground Level Test, Y-Z Excitation vii.

4 6 c. .. ~. e

4 LIST OF FIGURES (Cont'd) Figure Page 6.3 Responses for Biaxial Independent Random Ground 33 Level Test, XZ-Excitation 6.4 Response Spectra for Biaxial Independent Random 34 Ground Level Test, X-Z Excitation 6.5 Responses for Biaxial Dependent Random Ground 36 Level Test, YZ-Excitation, Phase 1 6.6 Response Spectra for Biaxial Dependent Random 37 Ground Level Test, Y-Z Excitation, Phase 1 6.7 Responses for Biaxial Independent Earthqrake Ground 39 Level Test - XZ Excitation 6.8 Response Spectra for Binxial Independent Earthquake 40 Ground Level Test, X-Z Excitation 6.9 Responses for Biaxial Independent Earthquake Floor 41 Level Test - XZ Excitation 6.10 Response Spectra for Biaxial Independent Earthquat o 42 Floor Level Test, X-Z Excitation 6.11 .Respones for Biaxial Independent Random Floor 43 Level Test - XZ Excitation 6.12 Response Spectra for Biaxial Independent Random 44 Floor Level Test, X-Z Excitation 6.13 Responses for. Uniaxial Random Floor Level Test - 46 X-Excitation 6.14 Response Spectrum for Uniaxial Random Floor Level 47 Test, X-Excitation (ay) 6.15 Responses for Biaxial Independent Sine Beat Floor 48 Level Test - XZ Excitation i 6.16 Response Spectra for Biaxial Independent Sine Beat 49 Floor Level Test, X-Z Excitation 4 6.17 Responses for Uniaxial Sine Dwell Floor Level 50 l Test - X-Excitation 6.18 Response Spectra for' Uniaxial. Sine Dwell Floor 51 Level Test, X-Excitation l viii.

l LIST OF FIGURES (Cont'd) Figure Page l 8.1 Peak Acceleration Responses at Cabinet Top 60 l 8.2 Peak Acceleration Responses at Cabinet Interior 63 Panel 8.3 Peak Strain Responses at Cabinet Base 64 8.4 Time-Average Acceleration Responses at Cabinet Top 65 8.5 Time-Average Acceleration Responses at Cabinet 66 Interior Panel 8.6 Time-Average Strain Responses at Cabinet Base 67 8.7 Response Spectra at Interior Panel Biaxial Independent Random Ground Level Test, X-Z Excitation 74 8.8 Parameters Which Describe Ground Acceleration 76 (a ), Y-Z Excitation 8.9 Parameters Which Describe Interior Panel Accelera-77 tion (a2y), Y-Z Excitation 4 I l l t i i 4 l 1 r ix l

1.0 INTRODUCTION

Seismic qualification of Class I equipment for use in nuclear power plants is a rather compicx process which is influenced by a variety of factors. Qualification can be demonstrated by analysis or test, or both I in combination, depending on the exact nature of the equipment and its function. Applicable procedures are affected by the location of the equip-ment within the plant as well as the geographical location of the plant, along with the particular characteristics and function of a given item of hardware. In view of the many combinatione of parameters that are possibic, it is obvious that some standardization of qualification procedures is neces-sary, and for several years the Nuclear Regulatory Commission along with other organizations have developed guidelines for this purpose. The NRC 1.100(I) and other regulatcry guides, as well as IEEE standards ( ' } R.G. specifically govern seismic qualification tests of Class I equipment. For years, these guidelines generally have increased in complexity, as safety requirements have become inercasingly more rigid. In view of the vast variety of equipment and parameters that must be considered, useful guidelines must of necessity be general in nature, and application of them to specific cases must be accomplished with considerable experience and engineering judgement. This is certainly true of the present NRC and other guidelines mentioned above. Furthermore, the use of simpler // procedures fer qualification of earlier items poses the question of a possible requirement for requalification to newer sr.andards for some equipment already in use. As a result, it is possible that several significantly different de-4 tailed qualification procedures, all of which fall within che general guide-lines, could be prescribed for a current equipment item, and these procedures may or may not be more conservative than earlier ones. On the other hand, very little quantitative data is available to date as to_which procedures a best represent the design environment, or indeed, which detailed practicos might even cause significant differences in the final results of the quali-fication procedure. It is obvious that the existence of any comparative data would be extremely useful in the decision to use a specific'qualifica-tion procedure, or to determine whether requalification of operating equip-ment is appropriate < Further, such data would be indispensable for con-1

..~. 4 i i vincing an equipment vendor that a more complex (and expensive) procedure is more appropriate than a simpler, less expensive one. The purpose of this report is to present the results of a research program having the objective of providing some answers to the above-described questions, as they affect qualification by testing only. It was considered reasonable to concentrate on the test phase of qualifica-i tion, for testing is generally recognized as the preferred method of equipment enalification, since functional operability is usually diffi-cult, if not impossible to demonstrate by analysis for many items. Further-3 i more, we emphasize that this has been a research program, so that all in-formation required for a given qaalification test will not be given in every case, but in fact, a much more detailed analysis of data will be per-formed. Thus, the objective was to conduct a series of tests which provide data with which to compare the results of several tests that can be pre-5. scribed for a typical Class IE electrical equipment item under the general guidelines, or may.have been prescribed under earlier versions of-the guide-lines. The series was divided into two distinct groups, ground level and floor level tests. At the same time, particular attention was given to de-tailed procedures which experience has indicated may cause significant dif-ferences in the final results. A technical paper which summarizes some of these problem areas was also developed under this program, and has already-been published elsewhere.( } The findings presented herein are intendi.d l to provide a quantitative basis to aid in the decision'to use a p' articular type of test for a given item, to help determine whether requalification of existing items is appropriate, and to provide a basis for possible future refinements of the currently accepted standard guidelines. Although atten-tion has been focused on an electrical equipment specimen, virtually all of the conclusions can be applied to seismic qualification' tests of mechanical .and other types of equipment, in general. We"begin with a description ofIthe test specimen chosen for the test series, the apparatus and instrumentation, and an. outline of the test' matrix and associated procedures. Thereafter, results of resonance search tests are presented,:followed.by: typical samples of preliminary data acquired' from the'various earthquake simulated tests. ~This information leads to'an '2.

analytical development of correlation parameters which are designed to pro-vide a comparative basis for the effects of the various tests. Subsequently, a thorough analysis and comparison of test results are presented.

Finally, a summary of conclusions and recommendations of further work are included.

l 3

2.0 DESCRIPTION

OF TEST SPECIMEN 2.1 Physical Design The test unit is a Bailey Meter Company Control and Instrumentation Cabinet as described in Figure 2.1 and Bailey Meter Company drawing No. D 3052169. The cabinet contains two interior panels. Each panel consists of four panel sections. Mounted in the top panel section are sixteen male, 18-pin connectors. Heavy duty instrumentation cables with female connectors were installed in the male cennectors and routed out through the top of the cabinet. The cables were then routed through the top into the cabinet on the back side and installed in the second panel. On the three lower sections of each interior panel are mounted cable termination strips. Each panel sec-tion contains eight 12-point terminal strips; see Figure 2.1. The four panel sections are mounted in a common mounting frame which, in turn, is bolted to the cabinet by a series of 8 bolts, four per side. During the initial setup runs, these bolts were found to be vibrating loose. To eliminate the possibility of the interior panels becoming loose during a test, these bolts were replaced. The replacement bolts were installed using a second nut as a lock nut as well as using " lock-tight" on the bolt threads to insure that the nuts would not vibrate loose. The electrical cabinet was welded directly to a 1-inch thick, 4 foot by 4 foot steel plate. The cabinet's base was welded along its front and back edges as shown in Figure 2.2. The terminal connections on the interior panels were connected to a series of wires to simulate a possibic control panel wiring configuration. Since this arrangement was not considered a' typical electrical system, it was concluded that the most reasonable approach to evaluating the cabinet was to measure' mechanical responses which could readily be related to the operation of specific electrical components. It is our opinion that the cabinet can be considered mechanically typical, and for this reason we have concentrated on mechanical failure criteria. 2.2 Instrumentation The tape recorder. channel assignment and transducer locations are shown in Figures 2.1 and 2.2. Nine accelerometers were required'to measure 4

I l 33 [9 @ i i e l l UUb$0 "N 9 [ 1. $1 1 aDDGDC N s 6,A,, 7 l EDEUU 56.000 ~ f -rr:~ ~rc a3 s, ~ 8 4" E \\ ] I _y_ _l j_ cq\\r-] i I l /E E E E r y / G E E : t_ _ J J l ,, / _E Ea _E.L l 2no 8" , g /., 3 dawn fren 30.000 24.000 r top of block Front Face Top View Te rrninal Panels 46 Total Points. ,s" ll 3.000 Front View with Door Removed FIGURE 2.1. SKETCH 0F ELECTRICAL PANEL SPECIMEN 5

a, 3 g %D Mb ix=a a)l TAPE RECORDER l 3 CHANNEL ASSIGNMENT l l l FRONT l CHANNEL OSCILLOGRAPH TAPE l PANEL l 1 a a 3x 3x l 2 a a 3y lH Aa 3 a a I 2z 3z 3y g 8 a l - -- a 2x lT l )2Y l 3z 5 a a 6 a a g i l lNTERIOR l E a g 2X l PANEL l 8 E E 4y lX E lz 2y 10 a E l l lX lz 11 a a l l yy 2z l ( + tension ) l 12 a E lz I E g lz 13 XCH E h C2 4y +- l=E ( + tension ) w lX ( + tensI0n I a;x a CONTROL ACCELERATION, ag,ay,alz C01.mMND DISPLACEfMNT, XcH' *cz Figure 2.2. Positions for Instrumentation 6

the cabinet response at the three measurement locations. The accelerometers were calibrated in accordance with SwRI NucIcar Projects Operating Procedure X11-EE-101-0. The arrow for each accelerometer indicates the direction of positive acceleration. The coordinate system differs from the normal right-4 hand system because the original data analysis was performed utilizing a standard table output format which later disagreed with the planned cabinet coordinate system. In an attempt to keep the results consistent, the format indicated was utilized throughout the testing. In the production of this report, acceptance of this coordinate system was found to be expedient. The location of the three strain gage installations is also included in Figure 2.2. A detailed sketch of the c4y strain gage installation can be found in Figure 2.3. Two gage installations were used to record data for this location. During the floor mounted resonance searches, the c4y data was recorded f rom location 1. Figure 2.3 shows the final configuration of the welding in the area of the c4y strain gages. The initial weld did not extend past the gusset as shown, but stopped 0.4 inch before the gusset. The floor mounted and initial simulator mounted resonance searches were the only tests performed with this configuration. All the earthquake time his-4 tory runs were performed using the gage at location 2, with the welding as f shown. It should be noted that the resonance searches were repeated after the weld change. The modification is covered more fully in Section 3.4. 1 l 7 L

i i SIG Location 2 CABINET / i SIG Location 1 0.20 " l 0.75 " 0.65 " [ l l 0.30 " L / g +y i Figure 2.3. Detail of Cabinet Base, left Front Corner 4 4 8

3.0 TEST PROCEDURE 3.1 Floor Mounted Tests The test unit was welded to a mounting plate which, in turn, was bolted either directly to the floor or to the seismic shaker facility. The photographs in Figure 3.1 illustrate the mounting arrangement for the floor-mounted resonar.t frequency searches. The resonance searches were performed utilizing a swept sinusoidal input over a 2 to 50 Ha frequency range. A magnetic shaker was mounted to input a constant peak force excitation level $nto the top of the cabinet. The frequency sweep rate throughout the reso-nance searches was one octave per minute or less, to assure maximum response at resonance. The floor mounted tests were run for one axis at a time in the two major horizontal axes as defined in the X and Y directions, previ-ously identified in Figure 2.2. Response curves for all data channels were recorded individually to obtain an accurate three-dimensional transfer function for the responses to each axis of excitation. 3.2 Earthquake Simulator Mounted Tests Figure 3.2a shows the test unit mounting as it was bolted to the seismic shaker facility. Although not shown in the photograph, it should be noted that the cables at the top of the cabinet are supported by an overhead crane. Figure 3.2b illustrates the instrumenta ion equipment utilized to control the seismic shaker table as well as record and analyze the test data. A control a d t,alysis diagram of the equipment utilized for the earthquake time hist.ory tests is shown in Figure 3.3; for,more information, see Reference *

series of resonance frequency searches were performed on the simulator mounted test unit individually along the three major axes.

The hydraulic simulator table was controlled to input a constant 0.2g acceleration into the base of the test unit. The controller was used to sweep from 2 to 50 Hz at a sweep rate of one octave per minute, or less. The outputs of the data channels were individually recorded on an X-Y plotter during the resonance sweeps. An elaborate series of earth-quake tests were also performed and are described below. 3.3 Earthquake Simulation Test Matrix i A planned series of seismic tests was conducted according to the Test Matrix included as Table III-1. The order of the tests as listed in S O. n, y, 3,'; g N'], lU Q:J,7,a 9 Y h y&

l l AfE f I 5l + ( c755 r[..f,+' 1; f l? W p I O. [ 1 \\ il q / g., v. ., + d,' W jf l i L .t ? ll A

o.

2 to e _7_.; a,; ~f. I n . ; p.; - - p-e, M I d hEH :,,; '.'.'.T 4.. _r ~ - {% a. {' O 3 91. fe'

J SA;Tp.'?: xte,m.-

-i t

{!;

1 # O#y, m y:Ec. n h h.*.* 58 )e

- - 3 4 giq,i exe l.

n \\. '

O Y 3,.

h

. >) Q k, ~. "i y:d'l 4, a [_ 'ji 7 4 [L t ~ %]

k i"

il ,l' _ w..; " c ?, : n[ 3:%3;,

,i.

.) a.aw.u~. ,u-_c w__ _a. a.. :a. -.- -N C (a) Cabinet Internal View (b) Overall View I c i w{ FIGURE 3.1. FLOOR-MOUNTED ARRANGDIENT FOR CABINET AND /1PARATUS wr w

1 i I _ q:f j' q .~ . _ p_., : ;,34 ~ 7, i l "y A ~ t :...- 4 p .d j .. yy * '. - [; -

g -

p } [.., - ~ N t. "e A c ?. 'th ' g_ l, 4.c ., / c, .. '. y - . i%s 1_ .g, y - ) t -

g. g

_ g. c., -.- ( w;

4. ; e

[{3' ~_' - - ' . 4 7. _. 4% _ er.. qit] 2 % ; a. Cabinet on Simulator l 5: .e, 1 Y[ / ,]h ~'iW e ttti N $ g,q p!t t '{ 'I w' W ~

- [;

b. Instrumentation View FIGURE 3.2. SDIUI.ATOR-MOUNTED ARRANGDfENT FOR CABINET AND APPARATUS , h l Q %N WJus&&R(. 1

MUlit-CHANNEL ANALOG RtmRO RtsPON51 X-Y yg3r uoogL C ON TAPE PLOTTER RECORDER @RIZMTAL HORIZONTAL ACTUATOR i / TABLE 3 ' i" Q 8 gtcuitt i M' ' ' ] \\ _yg,7,c,L INDIVIDUAL GAIN g T '1 iT:{ MOTION SHOCK CONTROLS N, , Asstusty SPECTRUM ANALYZER 9 SUAVAER g7eg i at g CLIPPER I i TRANSIENT MEMORY L UNIT DOUb'I INTEGRATOR HORIZONTAL VERTICAL c VARI ABLE ~ SERVO CONTROLLERS t AND OSClLLOGRAPH TRANSDUCER AMPLIFIERS l Figure 3.3. Control and Analysis Diagram for Biaxial Seismic Simulator hy).kik hih bY i

TJALE III-1. TEST tiATRIX FOR ELECTRICAL CABINET ~ l Test Runl Test ' Ncitation Signal Remarks l No. No., Type t Directions Source Ground Level Tests l 1 1 Biax. Ind. Y-Z Random (2H.2V) 1 2 Biax. Ind. X-Z Random (3H 3V) l 2 1 Biax. Ind. Y-Z Random (1H,1V) 2 2 Blax. Ind. X-Z Random (4H.4V) 3 1 Biax. Dep. Y~ Random (2V,2V) Horiz.+Z; Vert.+Z I 3 2 Biax. Dep. Y-Z Random (2V,2V) Horiz.-Z; Vert.+Z 3 3 Biax. Dep. X-Z bndom (3V 3V) Horiz.+Z; Vert.+2 i 3 4 Biax. Dep. X-Z Random '3V,3V) Horiz.-Z Vert.+2 t 4 1 Biax. Ind. Y-Z Eartnquake ($H,$V) 4 2 Biax. Ind. l X-Z !arthquake (6H.6V) Floor Level Tests 5 1 lBiax. Ind. ! Y-Z Earthquake (7H 7V) 5 2 Biax. Ind. X-Z Etrthquake (8H.PV) i l 6 i Biax. Ind. Y-Z Random (10H 10V) 6 2 Blax. Ind. X-Z Random (118,11V) ? l 7 1 Biax. Ind. Y-Z Random (9H,9V) l 7 2 Biax. Ind. X-Z Random (12H 12V) l 9 1 Uniaxial Y Random (10H) + 8 2 Uniaxial I Z Random (10V)' S J Uniaxial X Random (11H) 9 1 Biax. Ind. Y-Z Sine 2 eat 13.0 Hz (H&V) 9 2 Biax. Ind. Y-Z Sine Beat 27.0 Hz (H&V) f 9 l 3 Biax. Ind. Y-Z Sine Beat 7.0 Hz (H); 16.5 Hz (V) l 9.{ 4 Biax. Ind. X-Z Sine Beat 7.0 Hz (H); 16.5 Hz (V) 9 i $ Biax..Ind. X-Z Sine Beat 9.8 Hz (H&V) l 9 l 6 Biax. Ind. Y-Z Sine Beat 23.0 Hz (H&V) L 10 1 Uniaxial Y Sine Dwell 13.0 D, 0.17g i 10 2 Uniaxial Y Sine Dwell 7.0 Hz, 0.75g 10 l 3 Untaxial Y Sine Dwell 27.0 Hz, 0.75g l 10 a 4 Uniaxial Z Sine Dwell 16.5 Hz, 0.21g i 10 Uniaxial X Sine Dwell 7.0 Hz, 0.65g 10 6 Uniaxial X Sine Dwell 9.8 Hz, 0.20g i 10 7 Uniaxial Y Sine Dwell 23.0 Hz, 0.37g 13

the matrix, is not indicative of the order in which the individual tests were performed. Since each test was considered to be completely independent of the other tests, the Y-Z direction tests generally were performed prior to remounting the cabinet for the X-Z tests. As each test was performed, all data were recorded using an analog tape recorder, for later playback and analysis. The test matrix is divided into two classes of tests, the ground level tests and the flo3r level tests. The ground level tests, Tests 1 through 4, are generally more severe at the lower frequencies than the floor 1cvel tests. The details of the specific requirements for, and the development of these earthquake time histories is covered in Section 4.0. The time history carth-quake signals were used to form biaxial independent, biaxial dependent, and uniaxial command sigt.als. The type of signal configuration used for each test is listed in the column entitled, Test Type. The Excitation Lirections column designates the direction or directions in which the electrical cabinet was excited during any given run. The Signal Source column specifies the type of source from which the drive signal was derived as well as identifying each time history with an individual source number. Referring to Table III-1, it can, therefore, be seen that the same vertical source signal was used for Test 1, Run 1 and Test 3, Run 1. Each of the tests, 1 through 10, are comprised of a group of runs which, when combined, form a typical possibility of a present-day seisr.ic test method. Tests numbers 1 and 2 are both biaxial independent random ground Icvel tests, but they were created independently for the purpose of comparison. The same is true of floor level Tests 6 and 7. The Test Matrix presented in Table III-1 includes, as Test No.:9, six biaxial sine beat tests. :These tests consist of series of sine beats of 30 seconds total duration applied to each cabinet resonance found during the resonance searches below 33 Hz. LThe sine beat signals were designed to provide 10 cycles per beat with the peak acceleration amplitude set equal to the zero period acceleration (ZPA) of the floor level response spectrum. This required a 0.73g-horizontal input and a 0.21g vertical acceleration input. -one additional sine beat test was run'at the frequency of peak re-sponse requirement from the floor levelfRRS, shown in Section 4.0. The horizontal and vertical command signals were developed independently such that no special phasing existed between them. 14

r t Test No. 10, listed in Table III-1, includes seven uniaxial con-tinuous sine dwell tests. A sine dwell of 30 seconds duration was applied at each resonant frequency observed during the resonance searches below 33 IIz. In addition, one sine dwell was applied at the peak which occurred in the floor level RRS, shown in Section 4.0. The peak amplitude of the inputs was adjusted to be aqual to the ZPA of the floor level RRS, when possibic. Reduced amplitudes were required for sevetal of the more severe resonances (see remarks in ',able 111-1 for amplitude utilized), in order to avoid immediate failure in the cabinet support welds. 3.4 Allowance for Cr.oinet Failure At the start of the research testing program, it was clear that the test unit was to be subjected to a much more severe program than would normally be encountered during a standard seismic qualification test. Although there was no way to predict whether to expect test damage to the cabinet, prior to the start of the program, it soon became obvious that the test scope would eventually cause failures. For this reason, a pro-cedure had to be formulated to take failures of the test unit into account. The primary objective of the testing program was directed at record-ing the mechanical response of the cabinet. Since the mechan! cal response of the cabinet was initially considered to be typical, this response, as defined by resonance searches, was required to remain unchanged due to repair of failures. During preliminary runs, required to develop the earthquake time histories, checks of the electrical monitoring circuits revealed damage to the wiring. Since these changes did not affect the mechanical response of the cabitet, the electrical checks were discontinued. At the same time, the interior panele were found to be loose and were repaired as pre-viously described in Section 2.0. Upan completion of the preliminary-setup runs, a comprehensive examination.;f the test unit revealed that several of the velds attaching the cabinet to the mounting plate were cracked. The cracked welds were removed and rewelded. Resonance searches were repeated, and the mechanical response was found to be essentially unchanged. 15 l L

t The earthquake tima history tests were completed with no signu of cabinet damage. During the setup for the sine beat test, the welds to the mounting plate again cracked. The welds were repaired and the resonance searches were again checked. Since the mechanical response was again un-changed, the sine beat testing sequence was restarted. The remainder of the testing was completed with no further failures. t 4 I 1 16

~ 4.0 DEVELOPMENT OF. EAR 111 QUAKE TIME HI' STORIES 4.1 Ground Level Tests Four independent synthetic time histories were generated to produce a 30-second full scale simulated seirmic event at the table level on the simulator. The process was carried out in several steps which are described in tin following paragraphs. The basic full serie specifications for the earthquake tests are given in the form of the horizontal and vertical ground level generic response spectra shown in Figures 4.1 and 4.2, which are taken from Refer-ence 6. Note that a +3dB tolerance is allowed, and the Reg. Guide 1.60 spectra are given for comparison. Since Reference 3 usually requires an envelooe of the RRS by the TRS, the present signals were designed to match the generic RRS, but to envelope the Reg. Guide 1.60 spectrum. A " dummy" specimen of approximately the same weight as the cabinet was attached to the shake table and used to develop initial excitation (command) acceleration time histories. A signal was synthesized by com-l bining six narrow band signals, each of which was filtered from a differ-1 l ent frequency band of a random noise generator. The levels of each band l were adjusted under successive trials antil the TRS, as computed from the input acceleration signal (alH), sufficiently matched the RRS as given in Figure 4.1, for the horizontal axis SSE condition. A separate time history was then similarly generated for the vertical axis. Subsequently, the " dummy' specimen was removed from the table, and j the cabinet was installed S its place. Then, preliminary simulated earth-quake runs were conducted to allow final adjustments And refinements of the command earthquake time histories. Each frequency band of the random i l signals was further adjusted until an optimum matching of the RRS by the TRS was achieved. l The next step in the production of the_ command signals was to sum I all six channels into a single signal, and retape the sum onto a single channel of the analog tape. Accordingly, this signal formed the horizontal command displacement (xCH) placed on channel 13.: A similar, but completely' l 17 1 4 Y f f

-.= l t Damping Ratio 0.050 Zero Period Accel. 1.0g SSE: 0.5 OBE 9 E-W and N-S 5.0, 4.0 .sr e 3 30 s~ e, s s / s 5 N \\ 2'0 U ~ ,/ s N s _g i. .h Tolereice + 3dt 2' 0.0 3dt 1 2 .3 4 5 6 8 9 to IS to 25 30 35 Frequency Mz) l Generic Level - -- Reg. Guide 1.60 Figure 4.1 Horizontal Ground Level Response Spectrum Damping Ratio 0.050 lero Period Acceleration 0.9 SSE: 0.5g 0BE 9 50 4.0 30 ~~ '~ y ~.,,,,,,. 5 2.0 / 2 w '\\ d / N 3 1.0, L f 3 rtolera ice + 3db D - 34t 0.0 l Z 4 3

7 89 to 15 20 25 30 35 Frequency (Hz)

Generic Level - -- Reg. Guide 1.60 Figure 4.2 Vertical Ground Level Response Spectrum 18

(x Z) was synthe-independent signal for vertical command acceleration c sized and put on channel 14. This process was then repeated for the other pairs of command signals until all four sets of time histories required for Tests numbers 1 and 2 were completed; see Table III-1. In order to form the dependent biaxial signals required for Test No. 3, the six vertical narrow band signals used for the two runs in Test No. I were readjusted to envelop the horizontal RRS when used as a horizontal input. In this way, following the same gen-eral proccdure, the blaxial dependent command displacement signals were formed. The development of the two sets of command signals, derived from an analog signal of an actual carthquake event, were performed in a similar manner. The analog signal was used instead of the random noise generator as a signal source for the filtering. The remainder of the procedure was the same. The actual earthquake used was the El Centro Earthquake of 1940 (Illinois Version - Ahmin). Our horizontal component was derived from the F-3 component of the El Centro Earthquake. Our vertical component was de-r rived from the E-W signal which was similar to the actual vertical signal. We were unable to use the actual vertical signal, as a defect was found in 1 our copy of that trace. 4.2 Floor Level Tests Four independent synthetic time histories were generated to produce 30-second full scale simulated seismic events. The signals were generated utilizing the method outlined for the ground level signals. The required response spectrum for which the floor icvel signals were shaped are shown in Figures 4.3 and 4.4. These signals were also utilized for the three uniaxial tests; see Table III-1. Two sets of floor level command signals also were formed, based on the El Centro Earthquake and shaped to meet the RRS specified in Figures 4.3 and 4.4. The signals were generated as outlined for the ground level tests. The various taped signal pairs produced by the processes described above were then reproduced on the same channels at a sufficient number of tape segments on the analog tape to provide control and recording capability for the required number of runs. 19 i

8.0 1.2 $ 0.010 6.4 5.6 a 3 l li j 4.s 4.0 4 ~ 3.2 ?d 2.k k 1.6 0.9 m I O O.1 0.2 0.6 0.6 1.0 2.0 4.0 4.0 10. 20.

60. 60.

100. Frequency (Hs) Figure 4.3. Horizontal Fioor Response Spedrum J 3.0 2.7 0 0.010 2.4 2.5 Y ^ ww g 1.5 1 k., 1.5 l 4 ~ 1.2 t 2 l U N 0.9 1 0.6 I l 0.3 M i i 0 O.1 0.2 0.4 0.6 1.0 2.0 4.0 6.0 10. 20. 40. 60. 100. rreguency m o Figure 4.4. Vertical Floor Response Spectrum 20

5.0 RESULTS FOR RESONANCE SEARCH TESTS The results of resonance search tests were sought as in any typical qualification test, to learn something about the basic harmonic tchavior. Some differences between results of floor mounted and simulator mounted tests were observed. 5.1 Floor Mounting Figure 5.1 shows the general mode shapes of vibration for the first four modes observed during the floor meanted tests. The resonance frequen-cies and dumping are given for the corresponding modes. No damping figure is available for the second mode since we had no accelerometers mounted on the side panels. Note that two modes were identified to occur dominantly along each horizontal axis of excitation. Figurec 5.2a and 5.2b illustrate the cabinet transfer function at the top, due to Y and X axis excitation for the floor mounted sweep tests. The cabinet response for a Y-axis input at the interior panel is shown in Figure 5.3 for the floor mounted condition. Figure 5.4 depicts the strain responses for X and Y axis inputs for the floor mounted resonance sweeps, Recall that for this test the c4y strain gage was in Location 1 (see Figure 2.3). 5.2 Earthquake Simulator Mounting Figures 5.5 through 5.7 contain similar data for the simulator mounted test condition. The resp nse magnitudes and resonance frequencies were found to have changed somewhat. The resonance frequency for the lateral bending mode in the X direction decreased from 11.7 Hz to 9.8 Hz. The side panel flapping riode, however, remained approximately constant at 16.4 Hz versus 16.5 Hz for the simulator' mounted, condition: see Figure 5.5b. For a Y-axis input, the fore / aft bending mode r6sonance frequency was observed to have changed from 19.1 Hz to 13'Hz; see Figure 5.5a. The l Interior panel mode was'also changed from'a' single resonance at 31 Hz, to i two resonances at 23 Hz and 27 Hz for a Y-axis input into the simulator mounted test condition: see Figure 5.6a. Figure 5.6b includes data for the three aces at the interior panel, due to an X-axis resonant sweep input. Figures 5.7a and 5.7b show the strain responses of the cabinet for Y and X inputs, respectively, for the simulator mounted 'sucep tests. Recall that for this test the c4y strain gage was mounted at Location 2 (see Figure 2,3). 21 l

f f z z g g ~X "X I ~l i \\ / / I \\ I I I \\ I I I I I I I I l I I l I l I I I l l \\ I I \\ l 1 ,,i/,<</ / /, d' ///, 4 FIRST MODE SECOND h00E LATERAL BENDING SIDEPANEL FLAPPING 11.7 Hz, E 0.041 16.4 Hz z z d u LY -Y T \\ 1 L \\ \\ l \\ \\ \\ l \\ \\ \\ \\ \\ \\ \\ \\ \\ l ? l l 1 I I \\ \\ \\ \\ \\ l \\ l i f ,,,,<//,, 5 THIRD MODE . FOURTH MODE FORE / AFT BENDING FRONT, BACK, INTERIOR PANELS l 19.1 Hz, ( - 0.042 31.0 Hz, C - 0.025 ( Figure 5.1. Cabinet Natural Abdes Below 35 Hz 4 4 h.. i i

l l 1.2 i i l j a, 3 ~~ 1.0 3Y

a 3z J 0. 8 mz 8

2 cf i~j = 0' 6 l 5 ] = =5.E m a g.E c w.E 3 $ s: 22 O 8E e 38 O 0.4 /\\ me a o< n j \\ /\\ \\ / (fg-

0. 2 f

g t i \\ jy j/ \\ \\ /\\ N N./, Y,/. X 'T- ' - ~ ' - i ev_ ___,,- 0 5 10 15 20 25 30 35 40 FREQUENCY, Hz (a) Y-Axis excitation 1.2 i e i i i i i a, 3 1.0 ~~# 3y a 37 l m g 0.8 z w.E 8 e? W 33 m

0. 6 g

c< b00.4 1E E E 5.

s gs 4

52g88 h c i mm mm 0.2 / n \\ / / _ 4. ' \\.... a_\\ m d... g - - --- - r b --4.. ..x i 0 l 5 10 15 20 25 30 35 40 FREQUENCY, Hz (b) X-Axis excitation Figure 5.2. Top Acceleration Responses for Floor-Mounted Sweep Test 23

s a 2.8 a2x -- a 2y 2.4

a 2z o.

- 2.0 .'52 e wm z2 5% A m 2 9 DE PR EE A ~ 31.2 / \\ l l I / \\ / \\ W 0.8 J \\ <\\/\\ / \\ / \\fN 0.4 j x' / rJ _a_....~~-.. -c- ....a 5 10 15 20 25 30 35 40 FREQUENCY, Hz Figure 5.3. Interior Panel Acceleration Responses for Floor-Mounted Sweep Test - Y-Axis Excitation 120 4 i i i i E FOR CONST. F 4 3x --E FOR CONST. F 100 Iz 3y

E FOR CONST. F g

1x 3 E WE EE E 80 1 E I 3cE< g 60 z2 m E 40 2!L E E E Dm E as z- $ $ m$ m $j 5 m. g 'l 20 p / _/ w-_ /,.. _V j, ~ -_.-1 5 10 15 20 - 25 30 35 40 FREQUENCY, Hz I Figure 5.4. Strain Responses for Floor-Mounted Sweep Test 24

i l ig & a 4 4 5l.E $1,] a3x - _ __. a 1.6 3Y a g 37 z2 Il a = 0.2g I ly m M

1. 2 j

z \\ a Q i \\ .e a u e t b. 0.8 i e* U / '\\ _E E a o< l \\ / \\ A 0.4 / \\ / N N /~ N '/ ,A \\. / NJ N 0 '-~- 0 10 20 30 40 50 FREQUENCY, Hz (a) Y-Axis excitation

2. 0 5h

'3x $5 --_a 3y "l"

a

1. 5 3z N

a 0.2g Z 1X 2 = 2 1.0 9 E mBU E 0.5 N 0 = /% ' " - - ^ O 10 20-30 40 50 FREQUENCY, Hz (b) X-Axis excitation Figure 5.5. Top Acceleration Responses for Simulator-Abunted Sweep Test 25-i e-I

I l 2.4 i i g, 5 a 5E 2x l 2aa a 2.0 a g I g a 0.2g D w-f{ fk I c-1.6 e n \\ I \\ l\\ = z I 91.2 l \\ g \\l\\ 1 I il a \\Jv\\ d E e \\ l \\ u W 0.8 \\ l \\ n \\/\\ 0.4 / / \\ M, _x- - -- - N> n 0 10 20 30 40 50 FREQUENCY, Hz (a) Y-Axis excitation 2.0 a2x ...-- a -a 1.5 2z a = 0.2g lx 2 am m i I e ~ 1. 0 z9 4 md O 0.5-0--^ A-0 10 20 30 40 50 FREQUENCY, Hz (b) X-Axis excitation Figure 5.6. Interior Panet Responses for Simulator-Mounted Sweep Test 26

L I 80 i i i i E lx ___E w 60 gy =' E lz J l a 0.2g yy 40 / m E 'k z 4n i I!h plq aw l q A" h 0 M -- ' ~ " # *- 0 10 20 30 40 50 FREQUENCY, Hz (a) Y-Axis excitation 80 i i i i Elx w 60 dY " ' ~ m-Elz d W a 0.2g 1x 2 40 g t,\\ a< $ 20 l / 'A

j D'

- 2 0 0 10 20 30 40 50 FREQUENCY, Hz (b) X - Axis excitation Figure 5.7. Strain Responses for Simulator-Mounted Sweep Test 27

Results for Z-axis excitation are not presented, as the cabinet was found to be nearly rigid along the vertical axis. Further discussion about the importance of the above changes in responses will be discussed later. Finally, the results of a linearity check of the responses for two dominant modes are shown in Figure 5.3. Nonlinearity is evident for the Y-axis excitation, but is very significant for the X-axis excitatica. Similar nonlinearities were found for the other modes as well. The influ-ence of this behavior on test results will be noted carefully later. 28

l 6 i 60 3 13.0 H: Resonance }g 7 2y E@ - z2 a a g g 3y 5 5 d m m g1 ~< 0-0 O.05 0.1 0.15 0.20 ACCELERATION, a -g yy (a) Y-Axis excitation g_ 3 9.8 Hz Resonance E y 40 g2 a z-E 3x 2 5 d m 20 u1 a 2x y i 0-0 O.05 0.1 0.15 0.20 ACCELERATION, a -g 3 (b) X-Axis excitation t Figure 5.8. Transfer Function Characteristics l l 29 l l

6.0 TYPICAL RESULTS FOR SIMULATED EARTHQUAKE TESTS A selected set of typical preliminary results for simulated earth-quake tests will be presented in this section. The results are displayed in the form of oscillograph time hietaries for responses at various speci-men instrumentation locations, as weU as comparisons between Required Response Spectra (RRS) and Test Response o, RS) for the simulator motion. Such data are generally required for sei.aic qualificaton tests, and are useful for obtaining an overall picture of the effects of the tests on the specimen. A much more detailed analysis of these and other data will be presented in Section 8.0. When reviewing the data of this section. it is useful to refer occasionally to Figure 2.2, which identifies the transducer locations and orientations, and to Table III-1, which provides a more detailed identifi-cation of test type. Furthermore, only ten channels of the time history data are displayed in each case, in order to provide the optimum clarity of the results. One strain (cix) and the transverse horizontal acceler-ation (aix or aty) were, of course, recorded on tape, but were dropped from the display as being negligible. Furthermore, the command displace-(xCH and xCZ), were also omitted as being of no consequence to the ments analysis of the excitation motion or the cabinet responses. 6.1 Ground Level Tests Figures 6.1 through 6.4 show recoits for Test I which is a biaxial independent axis, random source test. Responses for both the Y-Z and the X-Z excitation orientations are given. Several general observations can be made from these data, and are also reflected in the results of-subsequent-tests. A careful scrutiny of the alz, a2z, and a3z traces in Figures 6.1 and 6.3 shows that these accelerations are essentially identical,.with only a slight amplification occurring toward the top of the cabinet. These results indicate that the cabinet was essentially rigid in the vertical direction, and that there was negligible cross-coupling between the herf-zontal and vertical responses for both orientations.- Furthermore, there -30

4- ! ! l-i : i c ^3x-g o ..t =.. .w c- =. :- 4-a3y-g 0 4-i T1MIl'M'9?T-7ISH.7'D!- - - dhtatikbl o. n,A au ni i. u...,_ 1 __ 'iL21.do. Mao a,-g o 3 vg'

r i y Uprir,-'; n y 'ygpr 41 "

Tl ', q vl" mp.,'77 ', ~ ' f'1 i F!,i;l +? I f 4- ~; -f. fh- -h-f f ' 7-t --[- p-f l 2x 0- 'f I ' ' l. ' l ' : s I 1-m -.m 3 '2 pl.__ u'. _ __ u..:_ _'.a L J-4t-f :,. m-l 4-i + + ..t a 3 ___m -.m y ! t '} 4

st.4},

i ' j 4 a... 4 .-J

..f,j { - }0NI.f j j ELb ba1 4 d NA A LU a A11. m'1I 1 h _ _

"2z-8 0 -Ay, i lj .qh^l,j.{p""P9I1,71M'sIU,!_U..,'~~' ] T:".1.l r l '[F 'f) U I o j ~ l*f(k'~I.*icl-j1!. f. { .-hl,,f..f.f..k,h,q.hd8 y l j@y .] y ,,, --m~g-r-- F - i 200 +j.j , - p j _pf - m c W r-~ +- ;' - : y. i r, ~r t m p r-q.;.., 4y-M' C 8

' 'E

'k* 0 I [ ,,--[iv.7* F.C D

7 f p i _ - rd. '.

200- -. i __ - - >. g..p ,4,., c r,, - n.,- I' ' $ I I'

'I I^

- I' ~ 8'Iz-DC -> m aia L L - Zl-A.. d e.' 'j,/i l m m,1._m m,, ' m,,,., j '?

o-i j

N-.% % @:IN'T T-9 -., j 7 p g g,,jp,'4 '-P}i. ' D T' Z T f_T Q.". '- y .a, 4-j dp. j.j .- j

gy I

,; 4 y- "ly-E 0% [ p) l s 4-y

7

.J.! l.j ] 5-l9 i ^1z-E r ad e IJdEd c kEl_b b'd NJ dd ILNMGJiili o ' " 'I R Wyl' T "', r 3'[F f f1n 'M rl"11=' wrviT" ' ~ ' f. _ _ s '__ d._._..l _ l .__C ..:. l_ l l

p. i. l.

E t; m. : _ L j y.__ 0 5 10 15 20 25 30 Time, Seccads lx' "lx - Negligible Test 1 - Run 1 c FIGURE 6.1. RESPONSES FOR BIAXIAL INDEPENDENT RANDOM GROUND LEVEL TEST, YZ-EXCITATION / 31 f

8 .4 4..-4 7.. f T + wt ee ,.4%,. .g t- +-+ i + - ;-e; m 5 t M se. e 1p*5

!' d4

^ '(; i i. .a 4 l j j ) !$ d j lli l l' 1 i , 11s ! IN f.i, k:@4 1 i ~ j.,.!.lJ fl p..('thi N u . f.f} $t i I 5 j if f' I r ,.4 e .L pg i M.;Ijj n11 . l[ d l y :1 ( p 4-P.,: J i , h, u a m O.11.. %,,..h!l9, s;+;::/., },f. l a--+*--+++ t y 18, ,4- .1 p ,a s e T 8 :l ! d.I i, %, ! U!d...;, R t-

- s***.--..

t~ *'it1, H H I, L 1 } 4 o , =, '*"4 .N m.,+. +y,t- +t. ,.t."t: +t- ,D aJ P 'm r i . j~ rn j n m,i i ?' i ""t"*. m, f"'*" 'y -?E o -' * ".-h. 4 i 'o~+,tii'*!*T b" i

."e

.o +}.14_ N " al' i 4 l.. . [. $ $..,h, . 7 --y 1..ft . p+1 .ih.f,... D". Q l!d }- .)" 4

L1 i

e c

.h$

Ij i,. .tv 'wJql,; g! NN l $dl ..{ ,9 = u f 1 ! U iN li l IUl! $ h U "J' [N Tilh*f l d j' .i a-f'* t' (li N"**h T I I k a...i. i Il h! I f $f y + A D' iJ y u n o-M';I.. h s.h i . 0 5 NI,h! ! I WE ,l DIl P :-! l O;.r q. b at m t t - +: w d.e*Q:

r. j:

} ul,M:,1ljd.,:'a...._" N

l!,

l i.u o,h..: n j o a s 4 ~ nt. - j l I' y' p: .o -M 't~ 'T. ^ F' -+t+ltl W N, x-,. ,,f-74 y 1 -H

d Q,.

+- m,:

,~n Tr

...'~ o g :, 4,i.a, .t.

Ji T

' 7 i g g 3 .T3 no m 4; pi ,d.,. . q l e a w~ p -14 a,,. & %m. i i i H o u + i qijP iki! l .7+ + y:I < I d A "u i t m.ti: it ,3.4. 4 1 w e 71

.4.a.,

u....,, h 9 eh[L 4 "d! bE!!h.. o "M9! %r

!-l %mnh n

I h ' f l E I L Yh 4 9 a ! 1 y .p;.u 1 I NI!M [fl-db'} $![!!!Nll b. $A i l% y al.p II : a i d'm: i ;j a';;p !st .tg; l l i A 4 d f:4 M r i 1 ....'5 1.;'Wl N i;b ;i 5 ! l t""; d;} t .4 t$ 4 TM ;;, i i . ~.... '. d E! Cl w o o o. MH o a H .M O O 3 *uop uotaaay g] b dE Ud l , p.- M wg o 7 - '."{.~.;.t...y-1p;pern._.+ I L.. r-g $. J q. ' ~~.I mrM .o a p._,,,

s..

t r ~.,i:. ..,..i. wg . + . y., .v 4 g mo . h (( M . { M, P' h j Hi 4 i1li d'P---Q y - -.. :s.' t -++r [.,. -r<'t j q m s s < w ~ l ;KQ ;y f y 4 J- ( p(' y a y h, --- 4mf ht- : ri u o W...,! .a. -...L, l n.': ry! i q e e . ~..J.. ..J.: ;I;! ,N..l. m m

, i n.

s m + L l.

b 1

.i '.: 7 c. p g ' i W jh

p1 t ! J.f.. ri n1 3

a ...y 1 i..o .xe w g .w g +.. g; p p.. g l_ g g -?'+ r}-t* -t+-q++;ty Q

.r-n i ?+

t-. 1 5.

+

-'l w t p , ~ ~.-.t.* II '. 4, - :C 9 %_ r'- '*r' .;,l ' t a ';._n... c + 1..a! m..-.-, !; t",. t ~.. .e i s' f

9i 9

p n a ]: t" q. g j .. Iy n ,; j i....j h o e , y q. , -yg-;.j.A

..a

.h h .q t_ l j t 4 .Nk ..1..j,.M ... u. g .u i -.m p 'l .q, -.-..~.L h h h@ /,2

k. $.[.hg

[ _y p j y 4 -...e __..n,_.d_db. . w w - i L l a vaa_.,a. g p_ m g .g ~~ 1 I J 7 o q. _ _.**-Q 6_ . v,, s ., m. y . l.. t ? o, m' <l. .l. o I 1 E .I.. .+. ii. -+ e .j A.., .l, a w,_ .i J. j M.,. ] I q;n

4. 4 m'
g 4.

2.gg __ i q:. - m L .i, n!.

~

1:: . i r. '.. l ' F] 'l i......,.. 8 . o.. o-s M .o 2 *uopuagaaay &~ l'

l l I I 4-ll l[ re f 4 l l I I il ^3y-S 0-7- q - _ - y- .m x . _.,_ 1. l L f !,!L 1i 1 1. ( 8 "3z~8 - dy,t

k*

t A 1 Mi i{ [ lj/; "f,.' f' 'l' i j, I l i "2y-R 0 , r :1 - ' .= : - 4 . _.g[

LL;; 3; 1.

'y- ,.4 - i' l ..j; 4r oQ#g4 f 4 .,3,n, _ a2z-g ,. ;, l ' g,Q k i 8 + 200r g ""*1%<Nkki h /h ff jd,f. 'k$ 4~ 0 ~ Y '{ i I jlf f Il, ~ 11 1 l f E ~F I ,1 fil/ ' lz O -- I j 4 l i! Q~ ~ "lx-8 _ ) 9((tht ! I d ig,1) I

4) ip ~,l g (df i

"Iz~8 0 -det,.*4 g f Il d .} [ ! l l. j IlI ![ l I.!. I l ll j O 5 10 15 20 25 30 t Time, Seconds Test 1 - Run 2 cyx, ayy - Negligible FIGURE 6.3. RESPONSES FOR BIAXIAL INDEPENDENT RANDOM GROUND LEVEL TEST, XZ-EXCITATION l

i o .o v i -+t't i , -----e-eo g i r i i j + m. +. 1 97-i n i 9 n Up i ~T i 4.1p 4 i ::.jn i 4 i %t .f4 4 l i M'M M l i d: !f pi l "j " ~* a I[ F 1h 1 di: ' .'j db % I l Midy

t +@!d @ -n i.i il}.

.i T.,; !1.- I d . t .ldii i '4 .,i.) ij: Pi J g: 4 4 7 p. ,jijil: il N., ad.'::O,,J. ; I i;iiid :} h. 9!!.!jd.i !,;!!ih.i, F i l. H H J t% i a o g

m. f

--t W k . :+t ~ 7 ,i, e .y 4 i ~.. ,,. f.. i 4 u 1 p H . m;'!!! P

98 1

i m ,.e i .' jh 1 iM.,h'Ig : j n;# e i:iq : i 4 .1 l l bin! ! i ! ! i kd% lb!! ! l Silh i I :tII!ld ? h U c Sa i ,.-1n a. , i + a, nis wi i T.t - W N. hirj i } ,it p; .) $!!$-M.;p! l .y. t ',p s Q i 3 yt bll i d ,.e lii j Ti 4 '.l j7-g. - fjj;j*t* y; i i, m* -, y. N t 1...:il)@.4.......h ] A, l3; gi.i f. o 1 s u n ji i J a. f + t .-n D, . {...gn .;g ,,a... o ..4.. 4 Fie Qo w

c s.

en n i - mcm n -.a v. iiii;:0 sip 4] M! y gi:P .} Q i'$ - i s i ! i idi*' O klh ! kH 'h!!it l Pi i [h tini !.i WU i 6 _c h..- t.I, N !. i,f [ %i. t ~M. l"} J .a a map w m r $h!b MNL I !1 .t .+ i } N, i.. ni!M..i! $$ l Ni ! N O !' .s, W.l.: f.e'.itu } E...F.. I..,1.!! a

i e.

,1 ,4tt 4 s u .4 ...... ~. g ny o o o. es o o M et o 8a

J

. S 'uopeastasay g a ea M,e o em_ .o m5 i . u., u se ~~ i i i.i i r I i j e f i e i i a 1 a @ p .m . ) p g ** y I m n Op ..'c w I ) I!! 00 i l 1N h j k h I b' j l ; n I d . M. i i M ~ d . m a o s eu 4 Lm :. :. t .u. a i f'. ., ll,$ [f,) H H .i $..l. f.'hh, k.I I .'h, N.!h i ^ 4 i u..t x w m. .i, ,,o.4 o i ~~ r gag q 1 af j - v n mi L,;(rgi.;,;n..!! I - 12 ' W I i ~ m j' _d .la ! H!! m e + 1 d@ h! l. _,k "'l .t}: 4!- jg ;ii e y I].:,{]h ,fl

i, l

l i ,_g s :lil y;l - ~D 8 l! I j dj! i 1l! 4 3 @ i l ,e i t- .- M g t + y'; m m ! {W,ik,~i ; 'i'

i @a7! ~7 Dj!.

. g, i -a i

1. a m

.,i..fnN 1 h Mi Et p. 4 { _-i o m y r N h Ih 3 i .h 1.I l l m.. y!i.:.,M db- - o mu q -.M l fj be e i w e

d. }c, h

..e .. d. t 3 1 lp Ti wi. " "th M u.. l. o.

a..o m

y. i i N^

~' ,i o

W . f'h -iHR l . 9,' m m. ,fv .:HI s. M o! i '*-'t-* ^ m,;i it i 1 1 3.ab.u +

t"::

~ i e w m. p. n.m s '- u f!' Y. w ig i I T o T .,l 8 ff 0 3 ! !f !9i

I

[ -..+ g k U k -Of ~ r *-N('n ? ~ } +p p ci i p h .u.n 1 .l - ..d id i a, , n p u._' ahq Hli g y q., yl-dl di 04 {!! N 9;! if

M t

~ -s i O ,,a O O.e O. e

'4 A

o S 'uopeastasay i ,1 i -t

is a similarity between the horizontal responses at a2 r a3 with the horizontal excitation at ai, although a significant amplification of motion occurs from base excitation to the upper cabinet positions. Both of the latter observations are consistent with the transfer functions pre-viously described in Section 5.2. Note also that the strain response cig displays a strong, one-sided clipped nonlinearity. This type of response apparently results from the fact that the cabinet base is not welded on the X-direction sides. Thus, clipping occurs when the support frame re-laxes contact with the base plate in one direction of the vibratory mo-tion. This appears to occur above a threshold level of about +75 pc ten-sion, and no corresponding limit exists for compression. Finally, note that only negligible responses occur at a2 and a3 in the transverse horizontal directions for each respective test. Figures 6.2 and 6.4 show the respective response spectra for the table motion al. General match!ng of all response _pectra was held to about 3 dB below 10 Hz, but higher levels were allowed at higher frequencies, where the test zero period accelerations (ZPA) generally is much higher than the specified ZPA. This particular type of apparent overtest at high frequency is typical for simulations produced on mechan-ical-hydraulic systems, and further attention to its consequences will be discussed in detail in Section 8. 5. It should be mentioned that the matter of a 3 dB tolerance was typical in this case, since the RRS is a generic (i.e., all encompassing) specification. Many less severe specifications require that the TRS envelop the RRS at all frequencies. (Note by compar-ison to Figures 4.1 and 4.? that the respective TRS's would indeed envelop the Reg. Guiae 1.60 RRS's.) We assert that the conclusions of this rep >rt would not be altered if either set of tolerances were applied consistently. The vertical TRS (Figures 6.2b and 6.4b) indicate the presence of excessive excitation energy at about 12 Hz and 48 Hz. Further comment on the signif-icance of this result will also be covered more in detail in Section S.S. l I Partial results for the biaxial dependent random ground level test (Test 3) are shown in Figures 6.5 and 6.6. Here, it can be seen from the similarity of aiy and alz that the identical drive signal was used for each axis, and it corresponds to the z-axis excitation source for Test 1. l 35 Q%$$ W

4- "3x-8 0 -- e^- t 7 t'a de '".. - - ? # - 4 ^ ^ yet ^^ ^^ r'ce-=c-4- "F' " ^ " - ~ a3y-g 0 -5'pm'igy]TlP '] - t'{f i ' I ll lFil ply j'p ll tif'11 P TF T,4 i. l.l f f} lfl j l i 4-b

N'

'""'"m-- 0-4/p FT{TP] "3z-M '8 n r y r y' ' - i'"' m p-{ {' p1 Wi .__,_y ___...p +...--q_..- 2x-R 0 .- ; e , _- t,-

- Ac:," -

N ., '...L i _l j. _ 4-I .!__.i

j_.;J.{, I :-

[,-[he "2y-R 0-(hl ), i l).t i ! I. ' l-h -{ I, l i. l I ICI ! I 4-4 i... 2z OM h hy ]I( 3h,h.1 [ v. i 200( _ }, _,,, .,_,n f,,,, ' 4 y* dly[ f '[p. ,_t 0-gh M/p>%hg yIYfM#k8IfhN "h v)'ikr I ~=i - -7 'i 1 ~ E -pc 'l 1z T ...-y...- .: >.. a..; - .n ...u... [ + J ~ b.t - ! d,Mf p,9 [.f/, h,8,4hf[pf,h,h/h,A, !A- ((f,3h,% " l y'~ 8 0 1,.. p I!;;i: I!jiiii !i11!!!j 4r i i y (k "1z-E 0--~h//4 3 f 1 l l l l I l !l l l i ! lll ' ' ! ! I I ] Ll. !' ! 'l I f i ! ! I i l, 'l' I I l l i 8 i t i l l 0 5 10 15 20 25 30 Time, Sc,:ond s cix, atx - Negligible Test 3 - Run 1 FIGURE 6.5. RESPONSES FOR BIAXIAL DEPENDENT RANDOM GROUND LEVEL TEST, YZ-EXCITATION, PilASE 1 d

o-i i I i f ( i 1 0 l .o 1 ...i. i Ma i f ..t m

t--

eo o m i J n

t'T v

.o. W i i n m ~1 a. a., q v;;

t--' T._ittM.i. h, 4

..il4 ai i a a 4,, .y: , i s s _.c 1..pl J 1 l l 1

L 1i -l

.-.i.+l ! im i.1 h f 4 a + m, a < h W:t - pg L .4g;y@ . i, M.4 n i q.@ m i m e u'i.n /$. tehu T'. M t. ~

m!$. :

a u ! h - v 4 i m a a.

'g N.!

/x -J Pt* h, .e ]1l.l. +

't r '.

J, t,jb '. @.h.. .4 7 l d

- H

. I.

rli.p

+ i e .i+, ., o.4 ~

.y a

.,-4 ..4 h ,8, ..3 7 j .~ d . F*b .,t" h-

~

?; u j 3 +~- w J l 5 !l "i I i yow ',e;t [{. t 4 '8!1 '"^ i D* M O ! l idl ~ j,j 1# i r. ii4 4-id. h i.dl..l D U l l S L.H,.i}Wi. ,L tt mt#- 4y t b, i.

b..P 9

P3' e e u.. s,j !i r r c w ,,+. n T. u m q 9 1 ~ }6t. t r.i i %;r i i t i o p,. i e c-4 1 m

b

'* *tt+ d, h.%;

?

e 4.

! l,--

!h,@r j i r ..ak..

p i w

.o ~ Mr[ n mra fttit r 6T p k. A i ' t-S.1'i![ r!' y ir. %!H :. 1 q e., r. s.: ,+ ._o, qiq n i o, - gj .tl W '.l J. H.:h %,x ...l $ & -tfl'..y,C i 4t" '4 '-' '..to .a."*....r t.- i* ,,,3 >- ..;;,4- . [ Q .t..q

- - 1 m

i i t.31

o. gtipl

-b tt!:!

i rug.uj.-"4"*4r.c, 7

n 5c a-4. .'i*1 t .a. n Sp 4. i

p O$

J any d.ib i i i IO"'"m"h edj l [ Tt* 7.Pth ht h +syD m Us t C a: . i ,.,+,se w

. m

.. h f f jj f hh d ao 1 1 m 3d e.e. m c 3 s q.gj i ..i. t a .m n in a. d r l ., um J .,*-:-+ m t !, e. - t-i j - : i , mt'-

t- ;! !

..rn: 'p T t^t . +

~. *i d gr i,,1J 1

+ .S.e!nt. ea" : A NqM .[ 5y 2 c:a. n e..n j +, n .....u r o m> o o.4 e. o .o p I 2p 4 o m 2 'uopeastaaay w ge Ud o r e**. *a 3 em 3. qf-+m.-e,g-e}.a. T,..e 4. m. 4 %. ;...y.. v4+i-s n a n o g ma - o i 9 wp . b- --i ~' fE+d% f y j h $ d ^ l a c r.2 9 a 3 g h a, y d. ..,. q.u.,,,a, . w~. t s ; i me m.w., . 4.,, , 4 w l ' I i N .i 9 .~ l hh h m .M.i t.t...in&[ b fl f., n..+1 $.,.n!f !.,d.4..I M...j p g; J. M. 4 i i u..p. 9 ,l h,,! ,, l.,1' L _ ^ I I ^ '* .g g .7, .'*.,_"*) ! c i J N i m r. g'4*4 =.**'I=.Mg+<. - 7 ^ o s w + . P'4 m. OI i.p. A *1-- -4.. -.v 44 ,a.e.t * . j;jj eg ..p.gpi.~

+1 +

v

'aT*7 ep 1.. *;rt- :.

i "t

tyt.

p, 4 a,p g 4 t . #. '"'.'*,r n.q 64 1.*". ' '.?**N';,P". } i..]p.5 .ta.f'" T N N 'N

- 1t

+ f r .l m e. .i i J l ] n..'W.i .p tsE v.4 . n' p 4d %l q 4; t.... yat. a4.l .li 1 9 W .o r ' jr. 9 p .i.%ap +p7-1 h k .+ ibfli 1:.f d d !,,.+I! L c

'I..$

MI h I N lI I g ,.d. 1 a}ll 1 i e.ph. 4 4'.. t G_.. i i c 4.: w 41; a,j t e u .n.ii .J, ;n I o o p l - e'd'h -4.a,,4 p i i [ J} cr -m . a. + 4 a qC ! h.d h.5 [! R@ ..5. f,. ,4}.; ..@ I. j Jh,.. y t 3 t .g a 'l s I {t.m + 7 .1 < j1p .j 6 ..a g. y o m 9 u 3,. *,'%g!~n U :r : lt } _ o. .J.,,.:- d.h U I i.l s .l,1 @.U 1 fil 4 . ~. . +.z*(.~. ..e . L.,..p. a.i. ,p .f'" !h ?* n - a [ 'h k h" i

T in

. g ', ', " '-aii

+

34 o. 4l :'i i

i

-l 1,. 4,-.g a v m. fj 7 ..ju .e .c i - ( o ..p i., .i ~4 op p.. tt y". _'q! M 44.l.. I a q; 1 e-1 i Y J O 7 !'h.!.b i d ld h b .k l 7,k t... N S N..i. .l 1 .'$jli ..i y-t L ,.j. ..., 7 l . f .hi )ip .j s "O i

l.,

'l.IjI l', M 3l d ;"h. i:1p. '; U _*., ';k.$ ".C !!! 191 1 4 O + O O 'e 8g _ -s g O 3_'uopesaisaay D -37 . WW W - [ ",} M"n t-

2 -

3

l These results are typical of Runs 2, 3, and 4 of Test 3, although the upper cabinet responses are different, of course, for the respective horizontal orientations. 5 A final sample of ground level test data is shown in Figures 6.7 and 6.d, which were generated from the El Centro 1940 Earthquake source. Matching of the vertical TRS and RRS were particularly difficult with this source, as can be seen from Figure 6.8b. A large overtest is seen to occur. This apparently results from the large initial downward acceleration re-quired in alz at about two seconds into the test. It is probable that velocity limiting of the vertical drive system occurs, so that this peak is more pronounced than is required, and the excessive buildup at 14 Hz, plus the excessive ZPA occur. A closer matching of this response spectrum probably could hav been achieved with more attention given to the equaliza-tion process during the development of the command-signal time histories for this test. However, this would have required significantly more test set-up time, and was not implemented in order to illustrate more vividly the discussion of excessive ZPA to be presented in Section 8.5 6.2 Floor Level Tests i ~ Similar test results for the floor level simulations are provided in this section. A greater variety of types of floor level tests were con-ducted (see Table III-1). However, all preliminary results in-this section l will be presented in a similar uniform manner, including time histories and response spectra for selected runs. This will.be done for comparison pur-- poses only, and would not, in general, be done for sine beat and/or sine dwell tests when performed individually. Figures 6.9, 6.10, and 6.11, 6.12 show results for X-Z excitation' runs, respectively, using the El Centro 1940 source and a random source'(Test 5 and Test 6). In general, the floor level test'is less'cevere'(on an absolute scale) than the ground level' tests previously described. _However, this results merely because of the particular original choice _of independent required re-sponse spectra (Figures.4.1 through 4.4), and one should not be too hasty in comparing the severities at this point. Much more will be_ developed on the comparison of test severities later. j 38

4-ei ,). N % fih E h v hh*hhMkhG'^^- 3*~* 0 k "3y-g , go;,. o _,d a n g,, 2 ,,,,m g(,g g g ] g, g,, g ,, gg a32-g o u ff rT- %7 fi'p: q P'r 1 r var ns ' '* 'I T T '-r 4- .3 8 a j g ! i

9.. ;

g ig,'l l yg . yh.9, q g g [.p 7,;,/ l} ij A y 7,(17 n s. g ^2x-8 o,1 . tc 4- . j- -+4 q,.~ . i i I 4 f ' $^ : - ^lt :: ] - q...... } ? : r^;^ ' ^ ^t" "2y~8 OM I_ 4 t

.., t q.

i 4- > 4, t . #., i.

p. w.-

-t;.t..L.'. 4 J%' g$f'- }A ddt 11 d.' WtJmh "'rm pr p Lhll] ,_b Lbl.1 J.31 . G A_ a22 -g 0 pqy 7 ,, gg w m, t r- ', " .. ;;. j .. p. ...;. j. [ p. 4. p. q.

a Jt

. : 3..;. - j, 3. p... j.., s,,,. ;.. -h - -- 4 --- l--I;%s I-.- I-' -h-4--j ^ T.. 4.#C4,6o. %, d k--~,- --c ,?-4 1 J'? U.i '. 200r ~- ..1 i 4Y 20 ~ s .,jhh f-h, I l-c -pe 1z 0---f 4- _l t-, p. .1. m. g& &glej!ffrt:E'f.4l$.=W..rf5??hW!lE&NL'$,2OSkw ..3 .,..r. ; _ 7i .t .- a l W"^- I "1x^8 i 0 r,.., Mi Us W,loi de liL1Ill .A s.dilLt. s I '. t. t i. 4.., [s i.'..t.l.a '..'...l. ' I a -g 9 p g

g*g...

grqM l Iz ),,- w p, l l!i I !- l l l 3 l ! ! I

! I i j i

l., . 1 1 ]._.i i

....l i

1 .m a... g 0 5 10 15 20 25 30 Time, Seconds - Negligible Test 4 - Run 2 c1x, ayy FIGURE 6.7. RESPONSES FOR BIAXIAL INDEPENDENT EARTHQUAKE GROUND LEVEL TEST - XZ EXCITATION i +b-

.m a' I 8 = f'+~ L i gr: .y,..,..;.w t s .4+...6 m';F .i

~

,ae 4 } r- +: e ...7 t J 7 i g o r 1! .ig.O j - + - i. :' 'l i m g f f, I j.di.J.w n a pim& ni + - -hAd j, = '} ! b 1 d s w a o 8 4 ~ e ( Y 4 Mh. NV + - h[t L.31} ! $0{sdI . m [' 3-h! ., N .t.;p d f .n ,m -j j /1 9 4 mqP3' W ! M h. 'IY*Ii ..a ,j (aM}ui"" [r i I'" "?i* M d i i H is E 'E k.j.. 1 i o m w,h I t p tj l d p~j r H H

t

.t - h! o i t m b y: t w a e.s o p*., m m ~t $$L g% as, - ++Y 4Wii.% y ~ . L ~ Qp P- . W3-4 4 M i i o + u i %jii Wp 6 f i f f j i,-

j ddsU U

! l J Q1 t t. d."itQia 1. W i .d c s a @ t tj q. +h .) r W W t 4 1 t f '. ks t ] l h,. h h .4 %! b. 3 t l s a 1 m M T 1 Wik i ! I 9ith !!I, 'J e ll p i%Q M i l N! 2: ii.J

+

1 bhk ! Ah t 4.41.- . M <~** 4 .6 a 4?.. M'*4 "b 7. a,. 4..a.4...44,4. ..t.**t* m .i 6 .a

je i

.n7 ,. g,,, +3 4 aiu**.++ t . r er o .A s + + L -m. ,.. i a ~ i 9 g f 'g ti{ M' lip 6'Ptt~' ., '1 dh 't P i t f3 H f ii ~ 'M*i e m i 15-4t n 9 h ], j}} g',h n yj F M w n i 1 05- [ j b bu Ji ! y*l !j O j,' jp :9ll;j,.. p ti j .e

a a m

. ; w f Z.! a2 @d 5 i OM F fd4 f I m l i ey j T' 1 + n - i N 4 N e -'4pb i-t t i m q' 6 i 1 (h ..A . ++ L

il' 'j d, ih.i.}

. 1.nr e w:[g , -{.%

!h M H

K..;i 4 l ':1. L a;{h l i i f.. 4 2 o .p q $ f 1 b E 't wH ......7.'.-- ..... ~. d a g! MH o O t g o o t g.e w S 'uaise2atasay g Ha l 4' g O, k 4 ' W .,ipu j! m. NW .-+ -+- p. ..t+tt t---- .g. -i

f?****-'+h!'-M'-+ ---

no a m :3 74 + H'. m : :* tit'+-* n ~ .s t:~ gp*

'*-~~?*
+*

e O gg -i --'-1Id,.+ n o.++4 , ?Tp't m. y

2. : s;. t-l 4

g mtt . g. n mo f4 +-- . E 4; Q t % Vht Klt y .j i-_;n;E g-l % altj q i I,, l } T ,,g' ppd}; t j gl Q-l l ~. g g

,T%.f-n.;-

-3g.,

hf':

I 4,.. ,.? H H y 7.- 3 e t1 7 I f Ck,Nr! ! h.*.I y, I .. '!I.[Y>'.[ 7 F kl -t N . 4.i i&.. M..}I . I i J m g 1 N .. ># !i , t a %M 4-f -M F 'l $N !H M 4,.. .-a n .+. ,.,...m 4. !.1 ~r t .L jb b .d.h ~~ . q.7.g.. ~.,,[/ b ,..e ,e 4 a a... -o ,s,e, g 1 m 4 .y -9 ..t,p4-~tr7 i(*i:**- " t, es d M-m W D y, i a1 1 r ~ M*T-E., H. e r h t'"1* tit'.'.'*.l.i'.."".'". -]-7,T,'.*.-t'ttti. ...j _. - 3 ,.?. c. @. y Q i e .L _.,.,9 m4{ W},j,- - t %jp

r. ! - -

ti .m .p, .,19 c-.u i {..l.- y N s f ,4. a. o g u.. I { 1 % d.5,Y..M.m.[ Rf jll h O-4 }M (h h $e E.! M l ? .E ytn n 4 ,,..g 4 a a.,d' . n .,6--g4 %% .m i44 - q- .y w a ! ; 3.. ,44 .a.p 34 t i a A a:13 th m E8 O t n-a g. v 2 c p 7 b.. Q l ..a h 4 .Y d d]d Ei' M j" 4,1! { l l 4 M yM q t ---*t;,97 t+ +- m. te 1

P vW i

t. ." 1 # i h,, Nil ?c "1 i >l '.j, .a. t inr -.ih --1 e I -o, y ~ .1L %.t.n. %;3 .. g+... - ~....,.. i.l.%.o o %p tm ..r-w. + or r. 'T ~ T[ M,*? *"M il

t. t*

c o ... i :W' T.., r'4 *tti._ .'J'r.. e .h ;, + ..4 o. a t.s p 1 O QS. >A.l's t --ql ,' t u, ' H' j E .l. i.?, ~"r.it 1 o i- {" } e. .A-p'- i .I,.. t, si __..i. _i .m.,. .a;., s g: m i t ! J y.m.s H,. ;J6. . 4 .b. t. u, A>

4.

J.L, I, 4 m:;. A m I - _I m c; .a._ :,h h 'y r g-m p;: 4 +* ,t 'n m 9; MS +- l { s .j, 4"T t y 3.'.f... F f.,,

4...in.. y, t yI 4T*t y t

1 m: ..1....ip..a.,;;i :l.l.

4 t 4 f.

1 f .o_ o w g-4 d. - S 'uopeaaT*33Y , h l 4 4 e .u---.r

l 4-0-+gM M.-.pr g.th,iieari.er p -pyyp@.W gy(,hwb,h*-ub-k ) "3x-R g t I a3y-g ( e' 1 i i i i ! 4-i 1 l "3z~8 0 --th- : ; g

I : r i ~.i t

I j t e i 4 a: .,.,,,1 t !I j ! I

1

.....t. ._.,_.}_..,.__i.., 4- .._. w.... 0 -+%*t4,h.k! l.J,)ff?. h % # V. a $.!' ;'l,"[.4 $-l. M h. I "2x-8 . i I .,,-i

3

, y ;;,.,., t . -r^ 4-

".),.}_.;

6 i l i i .t 4 e , i ) {' e I 's i -. r 9 2y-E O 'l ' f' I. iII I i ' 3 IJ l~ :l ' I 4- '.. Jt ? ' j : '- !.s.: - 4, ' ;,4..J..'. ; L...,!- d 4 -.. 1 S: _g O r- - .s s. .z 'I I !' d 1. .... '. J.. ! 200- ~; ;. ; ;_. 4y~UC 0 -w spig ≷ je:::';;,,;=2: 6.^;; 4l:.' s .o i i.,. 3 C "'^;^... ,,' 3 .._.h 1-h' t 200-1s p J .t

-[

t b !.t

4.i

.I .i, -l { r ! j t 4. l i cy7-pc - - - ' " - -, ~ ', g' ~'rr,.~. ~, ~ ~~ ^ ~~ ~ ^^ "- ' ~^^^'" " " ^ '* ^ ' - - " ~pn.,"p.'.-. q :r u.

y "u iir..u,-. 'y," 'm y - -

0-4- . be.NM = .. 1

1x'E O w,%y.%4:.";.u.l:e..[:.:'.'lbt..%.; f.},%.;;: ^:... ; : : :::V ;.

?' ! 'l ! l i l-}, -l } l i l-i- l ll l 1 l' i i i I t 4-i .: i i 1 i j ],.i i e ayz-g O' .hw J., I i 1 l i l l' 4 .l. .1...- .J. ..l.. .l. .l 8 [ ) .m 0 5 10 15 20 25 30 Time, Seconds - Negligible Test 5 - Run 2 clx, ayy 1 l FIGURE 6.9. RESPONSES FOR BIAXIAL INDEPENDENT EARTHQUAKE FLOOR LEVEL TEST - XZ EXCITATION h

i g tI 17 . i ,,7m#+, 4 -..g.~ .-3 -4 m --.,1 ,g y i n 41,' idi i i -% n n. .t i t .3 &,. 4 =, C , t L..h j j yg C J @: w,!!

Hli,

! !v.n ; 4, .u. 4, gif ;: 3 ,h:l a s drP, + y 7iy :.,.s.t a 8 < o ;, b l *i ,v-er w . l."l '. so m , J; a. n w g n .e p: ~. e a p i s @l v u u iw i ~~ L.y .Q.,,ane q '"l4' ,p f.' y n

}

't f]

s e e

l;:IH b b 4.l..

t. !

.{ f 'yll lp, f lij![ a,.et 'l 4;t .t d M ? u

' +

++ r-1.- q q ;gm h. _7- +O ? TP in tn? I.f. r- = -7 n t I

,g; g J

y' m %y ~",, .- ]/ . if 1-p"- i' ul! >n h .r

1 q

~ ul s d, i #kil n h O q - 1 ? '~ ~ d $$ { y jH ! M y M g ^ ! i' If-ik h[l.- hd l' f' 0;tik ~' a ~ m-Mis t . i ~w T;,i I 8:a' dhi 1-- : ir + l' ! dh! i $ii'i v'. a.. 7

i W

o A E m 4! .i "efiiM.h. y 0-? s d!$ $lh] . i l[ g Nb i l e,eno Jf a O e D%, IIZ:'I' I!O";*;?QT'~~-. H w i""dW l d$ h' i W;; w;;; ?,.!"t't,h,"'+H-l 1 l

U i

ld ~ 'Q S T g%].. h,l p, titp ..{ ar;.y 1H 3+ i h li y c a ia, h. t- .l ,4, o - i i a' .+ . C ]t a.q 4,*, s $d n. i .w + o 3,

i1

+ 4 o ql:q. r i i.. l t;p :lll e," j gu ,y'q + ", $g e 7' i \\ t+ p}', ~ N sd %g. j;l;g -'l y l l ,1 .. }% %{- l,kh" I sni + .2 1

j. l l

f an ,+ I I Otti #f'if - f M!j UIWij 'I T71 p t AA y lr l { 7; d l:ll@l f f, - w j 8. j q h,h id m. .s., ~. ms m o un 8 o o e kb f El n ~ % 'uoJggjagg33y Ca N

g

&= 2 .u m wmp ,,, k n@@JU+: t? - n o- ~ t QT m E: " a: l 5 e L ^ 1:ty", p: . i.n 1, e'[ 8" m.. o. 5" o 94 a *n uJ. i m.ua.eg# p.. t' Hs Jy A i e o' q, p 4 l u) pyg i 1dp 9 .h' y + 3f a, a l i i i.e fu4,} f' t' i" ~ 1 .r

m. m x

-i, L ve e: a t lg." al*'ff' A. d 4 x 4 4 i

y

-~ g m n p r. m. t d's .? s d uu IM.t m. " b ];jjp. rp s. AI[ i yi..m ;

ni@

,t.I:;j, i t n e a ?"yt ++; J

n; I : %iYp' !s 3 Tftt id.,!:lf _..o - g 7:: 4 g 'I.f l 2 n 1 ..~ u r a -t.. y e -+ t 7. iW ~ m i .u g Al' m'N, e' f-5 f-ma n

pj ;

I$' i- . +

)

i! 3 s. t GN ~ u 01IN ! g.' ~ f).au,.;.;4OJ ! u a Ty-s 3 I a ;h i 'nik 1 H. !t 4 ;; h D f N:#i $ II re ]4 '"t a :1 yQt'if 6 .o qJU 4 i o y ~ j,tjp i.p... y I o-l [ a Ag .4

t b IM [f,,

i + d)N[ flhf h,. a 4 3 .Nf f!I!bff - j

gy'r 3"

o Q:t i ! l 3) + x e -a . 4. 1jj .u....1:.#. 7 m . w +: +-o.7 4. + ~ 2 ) r.. o I tt i ' hf j 1 1 i i 4 EM i 4 ill l H 1 m '.i 4 y' i d l Mhl,q' .;!}

e

.lt-8 , 3.$!:'b L , q;;'m 4 i { p 4,dp',;4. Lj 3,lj [ f k . ] !4 Y

v

'l l l $ Y 2 AOh d h lj' ! jf. ? l.&,I& id [! .a4 g: l l M e k r ^ ^,- s a,-.. o o N 8 o E g 8 *uoitusage33y t 4 JJ gy s m wn

1 4-4r. a3y-8 o l ' l i.. . '--{+ 5- {.i f l,4'_. c i. ^;, j, '. l'H ~ ',', ! Iy q,I 8 i 4-l t s x

v. i;

. {.3.a. g'., 1 i g ~ a ' ' k -,i r,,,- t m m.' ^3z-S 0 ,r' r n-- mt-t t i 4 r s r. y I I .I I i -l 1 '-J,. 7 l-l.. p' 4- - l l' 1 I. l '.j L a' 1 e s s .,_ ? ... ---. L...:.. < d 1 3 ' 2.' / q;, j 4lI. pa i w f, i i . M I .i I l-},./j - &,:.,- i' -?t'!.!.t e 'l .. ~ ! w I -Z I J-i a2y O r. .g.4-: fe.t.t; j,. {. M.,a.fpl, 4 -; } ;4 -- j - jyipj.p{.j,jy;gghj;.%#,{~, ~! e s. - " -) P $ b I O -.L ' d P d'J.'t.J D 1.b T 1-- d-- h--Q', W f.,.. a -- g o 1 ---,r-,-,-I -_;,'>> 2z u-L i.L; , l. j. n.a '- ' ...;- p u,w;.;]d* 4~ ' _: _... , p.l';1. 4 L, .c,..'.i1

i. -C*

i 200[. . <2 i. i..q .;p.p ,. c -j r i- -,,, r. 2 214 -'. j - * " * ' " " ~ " " ", - E E, ' .,r y.*,' c4 -pc 0 T v- " ' ^ ^

7 srv p - yr-r 7 iprrrr=-

-m 7sur r-

p. L ;' l ;.l,3 s.1, j

l.o 4 *

. j. 4 ~., j ' l: '; 1.. M*a t :,i,'

q;.- 3 --,'fi-I:t,f.d.il' g!tti;7tyi;ii 200- '] g1. t l~

!t, i

'1z-PC 0 -4,--i,~! Q.;i.: f i d a!,,7j. l l v.;; 7:l; l J., f,. ] $.y. j j. j j 4 L c 3., 8 ~ v.w,$.h,;f'ggQl1'7i? 1 g;,p g,y_ < n. "1x-8 0 r >r .v u ,r-i i i, ' I ' P l I~ i ' I 4- .I

O l

.l ;_.s_.Y w _ l. l .' <l-8 l rl i 3 -~~-----m a -g 9 i ,i .) !. 'f, f t _.l 7 l .1.. .(. 6 1 l ~. 1 0 5 10 15 20 25 30 Time, Seconds cyx, ayy - Negligible Test 6 - Run 2 FIGURE 6.11. RESPONSES FOR BIAXIAL INDEPENDENT RANDOM FLOOR LEVEL TEST - XZ EXCITAT N nwn%d J0R ogggq

I 8.4 9 E .- p .l4. J.4., .W ,+* T' ,s M tadt e 4,. 7 .h- - 5 .n i e, ' 9',j .+ + [ r k-4 +h m[', ,n,, a +~~ h n .e ~ e d h;- ,p i,LF P ' $.*iti ~ 'tiiff"~ .t i r .jp g' 4,hk h V I f M.i.l,hhI l . -7

p. y a

I d t i 4 e..c .} 4-4 ++. f I Ii j ) b.Y 3 4 8 C.e 7 . ( I l. N.!N.l.. .!l .h_,i.i b.i.I"N f. . f. 5[ +Ib Np.Nl Y. $. I ..{.'.} } 4 AM.d+,. .+ $A1.e4 e.t .. 4.. ' q'i.+ e e W I ! ! M l. E

7W ch

= N .O .M w I 6"**t****t-.- n .17i """*'** E "".*t ?*t*4 1 +

}

V ! I . ya.;b.iF[ 1 9 Ph[-.@+ ; [ 5 s f 6. 4 L t .N,- '3 ) 4., y l 6...1.

-: r fB 4 W" L

5, c 4 ti I 9 -= u l ..i.,.+.4 } Y 7Old m e.;. n: c.L I b; h. Sk --+- >= +4 re .-4 a t! b + i, a .kld-i I .N U. IM i g (u.! p( i~ fi!' $mrf tt t a r rl rh i u n 1 fl. E.

k. T f

. N F i l-l Y

1'i i

.f.,+M. S h;L f fI klf. ! l@h h.

- f*4' E

4 @; dbI I l i I +2 t J. 9 4 ... p-i "f j h i d'. E k 1 b W 4+ i i. - 7 Z.i r; o a a i r m. g c ;> g mx .e.s n. .4 r 9. Lal O .+a.l 4 'ab as ..14aqng p.'N ' ++ f 1 l ' p-MIAM.hQ.- e ! O ui 4' 1 . x.: d a'n' QN .. 1' .. l...g ? c d t! i 4 i A # q la3 ' - o a . I 9d d AH 2 41.'..tP h } aj UA* t y 9 $ Q ..p s li, ? jii M $h" h I ! l d t!, if, U

hlh

$.h.e-9U ..l

h.

1 J xg M q'"d4 {l

i 4

ytwii + E a m ;;.4j. # 4 4

4 41%

~' ? .!;i Si:d2,2! 4 .1 1 dx iW i a.1}i a, # l. n

! a J

$- 4s,, N. a n l $q9,h.. N;a j hlh qf f. h f, h ? b 1 g ilp l' k 4 ti i J !!!. al_ 9 i M gg3 p q p d gy 7 t' F:, i a a .3.. 8 O O M OH H e e Q H H Q S 'UOJ3WJals.13y ga O

t

.e 4.L., .lL.+ 2.. '"%. h .L L ~ TT T j n. .w ...4 8 a m. L-4 4 -4 .g tal m ,_4 a .s I{.1 t .m i g' j . n# d M

h

. g 4 .U 4 t 4 s# A 7@l ~ - re11 ". a3j*ti-v 1 8 i 4 i a; 2g h 9 [ $p ]. --. e f f[,) [ ]{ Q,}T ,j N Mq i I' s n e c' I T- - -$ % &ll, d.1 'S.,

l...@. )

.ty l ( d e. t.o ji y* p 1 H H ll. 9 4

. {. !

6'q " *!.,t, t. "7 m l 4,. N 7:n ,1 4b.i C a e. j it il l

! I I,8 11 i

M t sg II"J% 7%l$f;7 '.~ " {.'",D*8 ' 2l'fa. M'..h ee .. Wr* - e-@F.I".. ~4,.. htH-1 d. } a - -v? "r~MW" .. --t. r m ~t w m

, "+r~n a

a m m&.g~w" ~ ~ t t ',i'i7"fb."

2 z

-kq u b)v ^ d ~iv 9 4] . M..*iti'MIph - ++ -ot-1 } Gd p e

1. Q ]!

I M di M.li@ '. p %,r i 1.h,;t +tet 4 d t .:n - l! l-- e

71. *.

!! @7,h o h,4 h 't I kh Nh f' U N 4+' M} c 4 h . 'h f I ih*h'. 4 fl!'- f.A r*. 7 7-h! h.

d.,

I s. ,. i --1 p; +

34 rd y

si-m : { 4;i: -n. .g m y 3 .a.Jum4; i N t

w

~ q a

,W vir 4c:

t m. 1, 'i; ;f.1il a y i; i , g %' d m. .I q ;. v s ] 1 I;. ' ll

I!,cU yq'r-is-C l

".r.!.l17.'Qj.3[.- f M.. r. o .u , OM, "*p"".;..W;, ..a m pi :- t. ~~ 9: !; *i . _u 1 Yd%dW- -v ip +"* t ~h'tf'I t *7k*.. 'T.l "i.l 9 d Ml' j '~1 jiT*' "I' 17 7' d .-e c U I i ..... i. 'f'ftMi .o T'" i i Tdt'~~~~ 1 ~ i Ti"" b++++t- ! 4.+ ' ir ;il li i i',- h f *: i pli 1. .; C -at. y!.!...; j, m .is. l...i }}i.L.-.% .' t.!.!.t. a u i ....%j .n h 4

U j

...N j'i )L. q L ;q i ol: t. . n p __,1 4 r y 4 i .y.' 2 7 ~-. 4 l .. h.. a t lll1 j

'..%._..a y htLl l. '..i.h, q..?

M i s

4 an i

,g

jn ti J.y-

.u lt t g..v l .'bo t l'. N. i _a s 1i ? g ..... m D b O 4 d d

3.

uop eaa pasy

It can be seen that the required vertical motion is quite small com-pared to the horizontal for these tests. Very little vertical response occurs at a2v and a3y. Furthermore, in Figures 6.13 and 6.14, when inspect-ing the results for Test 8 - Run 3 (which is a uniaxial random test using the same horizontal excitation source as Test 6 - Run 2), it is apparent that essentially'the same test run has been applied. These results are also use-ful for demonstrating the repeatability of the simulator and data acquist-tion system. Sample results for a sine beat run (Test 9 - Run 5) and a sine dwell run (Test 10 - Run 6) are shown respectively in Figures 6.15, 6.16, and Figures 6.17, 6.18. The response spectra of Figure 6.18 show that some vertical motion was present for the horizontal run, although it was relatively small, as can be seen from the alz, trace in Figure 6.17. More importantly, the response spectra clearly show that a pure sinusoidal motion was not applied (i.e., response peaks occur at harmonics of the excitation). This type of result is typical for a motion produced by a hydraulic actuator, since harmonics are generated by friction in the actuator and in the table support system as well. The presence of the harmonics generally has no in-fluence on the final outcome of a given test. ,. l,

i 4-83x-8 0 ' ' ti' F ; J

i "3y-8 0

, ;. s. y.- ' p.4 'j.-ae 'vd e' q..J.'.e. 'c.['lej J .g., ; l..'i..l4'p',<;h j.t i 4{ l;t j, -j.. i p,.+ :.,d T j, 41 J g-j. {. l. 4 .: i 4

i

+ a3x-g 0l ,k i .,, - L.1 1 2-u .,.,.L p ; ; 3.1 ..p. ,e o..., i,. o q,, 4 3. - t -.~'3 -y ).- j. j...l~ 3.

7. ' 3" ' 7 1

'l 3.1.- -.i t,.

t.,..

.4.

3... 7 hri x

0 4 ~~ 4.

'.j

.J I,* 4 f

4. ta.
.. l 0l

!.~,'l. \\ i N.+ l !!- 1 l' p! 'fl 1. [ l. -]. U -h f ; I+ .,g...

t.,

v ? .. L.. a2y-g 0 ._." l. ' " 4 '., ; j;.].,-Q d p. ';j' j '~ j P. .~, g. .;I a <. t -i j e g *f;^,,p [,.j . Q h.*} '.'i ~1' O 4-4 s. . t

K p u~ e t W -t.,.4.I '"'d.

.J. Z.4 3 A-t4,* p' ...,- e '~~ ' ' ypr. s.t ' <; < 1 a --g 0 2 ~ * ~ ~w 2z c-c..g.., '_'g g,,,, ~ ~ ~ ~ ' ~ ~ E' i~.I,.f. N.l D E ' /,, ;.d, 200 ' . I.l ' ;.

l... s' 3 ' i. 1~.." !.

.3.. . D.',1, %. ' I .j - Lp q -- :. ?- 4 :kg :~p. f r o t .:y-y ;. -, f - . r~:s ; ?.. -... ~- m, m-c c -pc 0 - ~., - w 1 - m,.,,.._. _,' ".,..,,y, 7..,.,.,,_t. 4, 5 } y, __L 4y .r 7 t ..., i... ,... ) * 's e .g ....1 f-a ~y 200-rr-- m. ,i se; ga 4 .g./ e u.g g.-- s. ..t. f. b,.I c -pc ^ ^ ' ^ ^ ' ~ ' " " ' " " " ^ ' " " " ' * ^ ^ ^ '.

t. C.. b.,. t. l,..

q.* t 3 ,t. l e it - M'- ... 4 t ' [ '7,.' 2 c#. j yz 0- " '"' ' '] "T" f.' gi.MT"r"g.r5'Mr7y ~ ~ ~ " * ^",^^ ^" 3r7_.,..,. " ^ - - 7 m r'. p r' F. r y , + ' - 3.) I' + -8 1'lTi.f.,1 & i ;Qd..j.",'_],0:p.t i,.i,, ' a} r ]' 4[ . i i ; 'l ' F 'l ' 8 " 'l

. :p t L a 7x-g

. -... ".W"'."..'v,~,~ p'.',,'idppV,,,,,~,",, 'y' ".,,',..' ^., *y ' 1

' ' ^ ' ' ' ".. - ' ', '.

0- ^ v., y y.,. . tz i 1 -i , j, i s. i 4 4r e t } ff I i I a.l- .j-t-I 'l' i .l l . l.? I - ( ._L t j t. . _........ _ r J i.* _ _.;.. 1 .__...'.t_._.,_.*J_' "1z1 0'- l j . 0 i l .. ( i .l. __ l i I I l, 1 .l l l.,_,t.l. . l. .. l. j. }. .._.-l. .l., ..-j.. .j.. . i l i 1 A i 0 5 10 15 20 25 30 Time, Seconds - Negligible Test 8 - Run 3 cyx, ayy 1 1 FIGURE 6.13. RESPONSES FOR UNIAXIAL RANDOM FLOOR LEVEL TEST - X-EXCITATION 46 ) a

l i l l l t 100 Test 8 - Run 3 - e s 4 =,_ =g= w3 y=- 4 y===m =2;---=== p=4 =- 3-p Test ZPA = 1.20g - _ ; -.7 = - -== d = =ii=== =y=-qrt=@- -i -~ B = 0.010 in ~ 2 4

==# i==4&1: - M. ~t tw

==

==y %=='=-

.4 =1=;=Emj= s= = = i = = =.. L _. _ _. ~~ ~ ~~'

T.~~ ~ ~:

Mi ~ ~ ~ ~ = T Lit =. =t=~ ' =f-:-a >= d.... 10 -- _.. _. _ i. m_. _ n 4+ (- M=-55E~==UN 51 i j "I J = = =i-i:--15= li' = = y

===- i~ M -E::5 I= '=7t '- - + 1 si:.. iZ ~ - ~ ~ -^ifl- ?b tf ~ ~ ",%le-9#ifit-- o =-iM - -a = ,-.g. e -+.- t. =m-3 w,ji . 3=.yw; -neg;n y -= = =- u - = $ =- % -e q = e: 0 ._J. EE Ar= Ei=ff ~_ . _ u_ . ' _1 ~ ' I ???CC' ' ~. ~ 9ffE TE o a f, _ u

w.

o . c-4: w ie" m 1.0 , 7 ~ .8 - g jn :1_ ;;.; _ - - - _x .- 1-.n ^^ *-- ^

:.~_._T C

- ~ ' ~ ~' --'^ ~ ^^~ ~-^ ^ ' e-w m 5= m 3.-j 3 _ a ____ _ _ _.y .g e #-#g=-s-7=e g g..s=- +.5 -_.a==. . = = _

h ;i -... -,.. -. __=__

_.-.=. =- _-. _ - .= = s =w = = = =- = =.--.. .s m l! J I I 0.1 f.,

i..

i 0.1 / 1.0 10 100 Frequency, Hz l i i FIGURE 6.14. RESPONSE SPECTRUM FOR UNIAXIAL RANDOM FLOOR LEVEL TEST, X-EXCITATION (aix) L_

4' a3x~R 0 m rfew, w e Tyre vF"VYve e 4 ,,4

4 a

d3y' O r roweegs-e6 eMP I ' l.. '. '*I*!,'I 'll - l'., 4-I ' i i ( s -, a t j l-i 1 - e 4. ( ^3z'R 0 ---o e-

f
4.: j -

1 ; i a c l 4-t ). d i 1 t l l ; ; *j,4, -j {fi. p [m _.L.q-.,. j_.{r i 1 ii.I '- r e r - *- -P ' - .._.3 j .m f..j h .. 4 -pp.c[! _ ;,--f-f f T-k 1 h U A 7 ;p Y 4, ;,- - - - r i l. ; t-t.. c :...'{]. .,l j,4 4 2y-R 0 8 ..~;4 .. i - s.. y r.p 3

p

,.3. ., 1 4 j ,.4 , $. },, f. 1, if}]. b3 dn f.,.J,. 3. 7,d.f.. },.'. f,,*. m[**.d )b'J,h_,,'* ' kI I 4'

i. -

t+ .a '..'-,-*.,-*.L-g a2g-g n ...r...,,p., . i. r 1,1. ; ,yg,ng ga npg,_ g t.*g;,,

4. ;.. 3 - i g h, h.1 -

J'*. .. g. 200 -. _,3. g, g t m g igd..,._ t r ?, a.- 4y-DC 0 MMMM 0 200 - - -. }...., ! '. 4..- ..-,. '. -+1... -. -. J--.g- +.' D.M .'e r' ..-A-0 sw 4 ~ }.-1:*,h;.g-ty[t. bJ.,1-ts a;dj b ' -te (Ot $ + -d .i;e.t t "1, . t ' t 'P l.! - [ y~ ' ci,-pc 0 C..$.' i,d - i' WM,.,'.'.,'.' d' - v .d l ~ I -I T 4 ; i j.-} fI '. gt,{:. [c} j j.i[h%.:cm.I~ %..h_ w.Y ..a...a ...n., w ci7 bT;?,*e.L 3: { d '.1.- i 4-l }Q2 j : ! gj.rj :j agg-g 9 ,e ,e e,,- w.w w w-w w,- w w w w,w w,w,- =

a-ww 4

. y W.y..g.{.l tg.i ..l - w ..;ll z_ . f I-4 -[ g i { i. l.. }. 't

., j

., j s: c. .a a.

s. -.

.......wu.. l I 6 j s 7

f. $. $ i,s

l. a.-.

s 5.1 L. i: Y n 0 5 10 15 20 25 30 Time. Seconds cyx, agy - Negligible Test 9 - Run 5 l l FIGURE 6.15. . ESPONSES FOR BIAXIAL INDEPENDENT SINE BEAT FLOOR LEVEL TEST - XZ EXCITATION . -,- o .) E 40 lJ PJ M" ~Lm 0 =

l 4 o 9 -.e. - s' f a$ ' t---+ . [ g g --7j. ! f[...9 n a n i . w ~ ~' TTY .m e q v 3 ,o 1 a o g.

ij

'"? I, 4 i ~T f^ m o y o s 4 ;g ..[.J. [ an .n - p I o v 7.) p; j i a e {h 4 a.w 4" a N p., m N 4e1s1 ..4 4 3 c(',.4ai (( ! l m 1-A..,1,11 4 ,q4 a, j} yp .i.- q; o- }'.: i i I'.P.A'*.27.'i!! T.,i',f7 E lh .!/AE v N. 4 I.h 4.t c s ! l ^ ..._Nl.. f. i..lh, h.b.i . f .~ .m _o, h H .N., ..-e e <a ..u + --," = N ir y 3 f.H '+~Y +~.--A@t j*$@dNM j I i ' E a M O i' ac g- @ p

.in 3

l ..p-.i ain ai j I m $n:f1l: % .iir 3i M[ d 'j Eli: t ti ' i 't : ..4t!'. ? !!: J i x!;;. 7 3 l H W.; Qi i ih4 ......m,.;.,i - Q,. i ap '}i '"t .d 5,.m ". ' a.. " 4 i d 1 r .r m 1 !. z;.].!. h 'I .'1,.. '!.T**a*'.'*.'".(.$.N w l j i; - l .q'.ihl i,U.lh.! I ! Ih.. . -b 7 ! Sl.h.. I l 4h I 1... o, 'l[ e n. u q 1 < 1 i n ,o n m g I x .a e O 3 9 .\\ i .,,,m. .... j., m}. s 4 w.- p m, _1.L. .m y yl .l -1 n

.P: i

. 1 i p in N

o. IT+

11 ! o -:J. s a W,. wg Au e+ 1p i!.tni dy, l .1 !y a::pi dd .'_[ 4.!. m ;ii "1 h! . L g s

==- i .J. $qu 4:g;d. {l 1 ..A;h 24i.! W;11i 34y y w i --f--a _ .y

ay l

-.h h q 3 4 .q l r 4 a n yn e 1 2x f h,.:-.^

U m.

-f; 1.q,.%,7 l! 4: ]. .H: !ij ,1..W.. .;,d..g6, ..i ,a.i, f,.....l!. .,e: r.... a qJ, i h. ? r: l ll'

j q

$, yi a b,1! 4..! l.' t d t a,.., c.4 n, H 4 . gt L o m s . rt i i : .-d =H a M o o. 04 8 s N O AW g 2 'uo;3e2ataaay Nl E a HUM C O m om. = i i , yy

kg k

C .y: o 4 i m Z I

1 i

...-.e .a-r. ji t e

4 g

s t d O .g .i { >r on %aI. ._4. i t"o.g .,n A e m n 1 ,; p-.~:m t p,<,. a q w j (.Lq .i i u a e p n s a .u y r ./ .,!.:h...i 2 4 dll 1,3-._' % 'l. if

a t i

.y. t m a.mqp i $:!i 1 ,+ .. e,..$. .nl ..i.n...e,g A4l1 $ $ -t.;. , p m

ot @-

N e N u

p.: n a:

o

[.d % l0 r'

,h ,I[h v.a4 ~ g;U7.f,- 4 3 ./,l 'i;jh. rf u... H.t}l-x m e &(([ + r 1-y,1:t 'y! illi %-J e e 4 I u x - y !l H H

%...)!.: y t

.g I i og ,~ ~ ~ '.~ l !i g. - *4 .., +t. e: . r _e 3 1. i t i. o x w M N u .c 4'- a 11 '*.Q t. N 3 .e. r' t I i di! if t 'id .L..I'.J.4 44;;Qd.'.1L r .r. . '.NU lill. '. f .br .i a pi I q g i rl r. p,

t.. 3 I

a biti c .,,,,i. 4 o 1 'Ci llA.ji: $ M l / A d! is ! ll [ - fj

i H y l P

% { + f,. J!- .2M !j - 4S dh 2. l 1 L.f . ill 9 ..f! ((?,dd, 1 $ egl e p q mi e n: j ( At

H ;:

p" .e 3 i. M

s. q

.,I"u 3 t g, \\$ hh o i I [ f Y.

l I

k.!b h (h l < Pr 9 gg y 7 qidMlpllNl - 2 N b ! Il ] r.

l O!.N. I..

Sin f f $% l - l i ....+ --*a - i -+* 9 + ^ m ii, i t t ~ O 1j4l I .d. + 6 .o .+ + haL I

o.,.

4: i . )I... p a p,

i

+ 1 a/. 1 # i.- n p 1 ..+ .+ o i 4@ M l l l d @ ; 5

{l 1 lU d

Ul pli .!..L f. e hh f ! T h bT ] f. l f : N M. - e.g. . p' ' ..+- ib h i i ? 07 b 0 fl p ..i .?.p.J!j i y I a 4.- A 7l['i!' Si i ! l 0 ji! ? ', ! p ,,,'.,f.'h! ' ' h lh I l . i..... w o 8 .O.< O. e4 w o S 'uoT3e2ata33y r I 49 J:2 a H A l dd a b e, m

. - - ~. _. -. - _ i l I 1 I i i l 4-e s g a3x-g 0 4-t

i.,...

..f l a3y-E O 4-i

I i

's t/ La '. I 15 a3z-8 0 .f i -j j !! 'f'g' t e 'g-E 4-e m, _ _ ~ , -m. m 82x'E O 4. j. _. u s=u,_ u. ~ .., a i- . a w .~' Q ? ;='. m" .- ! i t

c. su ;. g

~, - m y *)(,. I ~ ]l t. pt. g p s 1 I. [ f> U.0

I ',$.

j t ,y r[.' O ;; 6 - 3 , ?'8'f, " A 8#

9'
1 r

.. te 52y~K 0 4' l1 [ f it Y-h '.h

i h *1

O.hr

'.hSM*Eff I ' - I' Ilsk; h-01- -- g 4 ', 5W,qil, ..h {, w.4 f.'.i i.. [F 8(i W M h g **_.'TM, e' E h. h . L 11 '.; p. *,' i 2

r...

_ ~4 .;d W;E, ...y a4 9 200 N I' - i 'Y YS YYb Y. Y5 $*A : 'h. " -* f . ( C4y-PC 0 u,; u cup, .r.. i,- c. u w, 4 s a -..,. -- p... s -..

i t a

. _.4 , J 4 7 _ i_ g.e;..oi w . g_, a L _pC..:,.,. e . N.1., t - a 2 a. 200 1 t.'t e.-..a,, y -$ *9 l E, g,.j. *,

i. :
-w-

.~.x,., c gg

p+ * -.. :.L.

. v. x i. .. f re 6 1 gfr /,- 4 ( T -t cix-oc o 4-

-ae+

e - .u- - ei.. f.=,< t k g , ;g

- f

! f M N di 8.. { M D f 6 i M 9 M. ;-, e$f-;Gydd M llk % f i 5 ,1x_, 0 s g , g l t - 8 '. i, ; l' - l g r-v, - 4-r' - s-

r. p v

1. J-u. . ~ t-. j. c 8 ! t g r l ...1:. _, ~ <g - - ag l;, .2 m.. 0 5 10 15 20 25 30 35 Time, Seconds Egx agy - Negligible Test 10 - Run 6 FICURE 6.17. RESPONSES FOR UNIAXIAL SINE DWELL FLOOR LEVEL TEST - X-EXCITATION ,.s s Ig ' ! i f: ' bl'O b(' J 'i i ? rh c c k Uhl I il N.r. r / 50 1\\ m a

100_ 100 -~~ if. Test 10 - Run 6 _4 + Test 10 - Run 6 mx====- ; ,==; r=-- - !!I fM -2i[ Test ZFA = 0.19g 2 i. ,-p q.m - 7.. =q w Test ZPA = 0.36g . 2[NMWW-

n=#

- =-Aq w :: t+i-s-.._ =e

==c _ M:l 8 = 0.010 l Qi] ~- ~ Mj=.

i --i.:-j g. 0.010 l

~ ~" p

i::

7 ~. s c;-=:_ _'"T -x w =

(

^ bT+ M+ K-w ui=H t =% K '~ . par g4- . y,,

7

-=5 !. T""'=: Y ? 20 ~ = =- M E WV s n ? En-~ ~ = = c 'T r.asiE=WK -s: 14 gg( ,= ti M I-- =_Y=. = = t4 ps t m_p .= E 3 7 - i+El-ii f f ~. -yj = . = - 5__13j_:u__ = - + t-

.u=

10, 10, --z ~~ ^ ' - ' ~ ~^ . _L.L i. m ~~ en e e e ao . ~ .h5 q 1-,j Ihi [f - 2 '-J =;riE: -i:

5

'S7 i-@ i- - M' V FM > = -G: 5 Ir i f- 0) i-M4 W 2 -=- j@sj = -4: -s. - [q+,;4.-.' 43r -E.y!p 7 [ E 4g .c= u=u

g7_

= - g:.gr :-;q. q: 2. _:'.t -- ~ ~ ~_E-E _.e{' fiiE EE=;== __=.. =. __=. _ -.. O, . =..

==-:- i : O,. de-E._i:=_i._=._=____. _. = h:; ,e4 EE? 3

E;;;.;+1 =d y+

- r'& g d:E _yi _f.-y-Kf *M E4C be [d-M M-;!Q--@ } _j _a. Mi be t M- -d -~ [Fyl-5 2 5i_5 I ]!E5-f 2 n Z = - - =.t i3 W =E-tu =-eiW iEE:am . r._te- - =

==

2-- -4 gy=g IEE- = y-N $g L_ 4, jg g

-:=

N - a. __t . p.- + tn, _.. / 1.0 2_. _ 1.0 .... 1..,;. r w = r ..._._e-..- .w w + y.g m--- -. + g((g E3 g1, . 1.: _ _ -Jr j' _..g gg ' izj4,-E gid'i JK -t.]4 1- + . ' ' g' . 's-.Q:s.: [ ] -.j..g. --.,3 :pQpr!g = = - -:.a1 g.;3 j1 _g..- l

sf.-

q ;. 9-c.. - Kj pi di_i. .x - pgif p di5:.1i.- y'-:.rgy-4.hi-y.) pi g ^ = I-I':--!dbj Y'

^

' IIF ~ d M'Mkk;hd . E-- k.N.:;i' 'l tire== dddhe.is5N$-5 ~ j i' ~ -=-- ^ g-e

q.m3: 1 e

( _. g m n,g,-- y_yy .-E. .g y r pm_ g .= .9,512 E =.= g4-s j 1 ((f N ; iEg g gj5 i;= @ E r-

D_ -j-E-

j .;. :-j+W . ) 1 Ef=- 5EiE:- E =. f" __.._,- --.:::-f _._=, .u_. _ __=,_. _. . Ay 4pq -.mer..: = -n=.. = 1 = u. r _% u. 2='-'~,. ~ ' ' I,_. 'a pa m w2._ .. r _ :'.~::= eI e .e .s e ea. esI e o e a e,ee I se m e en a m os j s a s e ; as a e e oe eo ao es en wa=== 0.1 1.0 10 100 0.1 1.0 10 100 Frequency, Hz Trequency, Hz a 9

a. Horizontal (agx)

b. Vertical (agg)

FIGURE 6.18. RESPONSE SPECTRA FOR UNIAXIAL SINE DWELL FIDOR LEVEL TEST, X-EXCITATION EEED

=

Esa r'

7.0 DERIVATION OF DATA CORRELATION IARAMETERS Up to this point, we have presented information about the test speci-men and apparatus, procedures developed for the tests, and only sample pre-liminary data in a form which are part of the standard requirements for typical Class IE equipment qualification tests. In the remainder of this report, we will embark upon a detailed analysis of all of the data that has been acquired in order to meet the broad program research objectives out-lined in Section 1.0. In order to provide a bacis for the analysis, s.ome mathematical background development is first appropriate. This background is developed in this section. The firnt development on response spectra contains no new information, but is presented in a form which is most useful to the problem at hand. The second development, which deals with test damage severity factors, contains significantly new concepts that have been formu-lated under this program. 7.1 Response Spectra Relationships The development of this section is based on information provided in References 7 and 8, which can be consulted for more details. Here, we merely summarize several relationships in terms of the notation used in this study. For a generalized structural system which is free at the top and ex-cited at its base with a harmonic displacement x1h f frequency w at the r r-th resonance frequency of the system, the relative displacement response amplitude uir at some upper point i can be written in terms of the follow-ing matrix equation f"x1h, ~ If j 9r 4 (X1h ( s where the 1mft side is a column matrix, each element of which represents the transfer function of each point i when the sys; tem is excited at its r-th resonance. ($] is a square matri which relates physical coordinctes uir to a set of generalized coordinates g, and {qr/X1h I 18 "C 1"*" " '#I* r which consists of the set of generalized coordinates which describe the mode shape for the r-th mode of the structure. Equation (7-1) includes the assumptio tha the system is lightly damped, and the normal modes are f , l 2, { } 52

l If we concentrate on the absolute acceleration response of a single upper point (E ) of the structure, we can write for the magnitude at the i l r-th resonance: N /E "Dr E &rj mj/(2S ) (7-2) Ulr/51hl"lE lh i r ih j=1 $r and c j are the indicated elements of the matrix [$], mj is where 1 r the mass of element j of the structure, and S is the critical damping r ratio for the r-th mode. The left side of this equation can easily be in Figure 5.6a) measured at a resonance frequency (such as for a2y/ayy and will be useful to us in a moment. dow consider the same structural system excited by an earthquake transient acceleration E In this case, the generalized coordinates for tt. the peak response in the r-th normal mode are given by T [c1 fml E it qr. = (7-3) (W -W +I20 W W) rr s r Furthermore, by definition of the response spectrum, we have E /(W -u + 12Br r )

S (Wr)

(7-4) WW d lt r S (*r) is the relative displacement response spectrum value at w where d r for a singic degree of freedom oscillator of light damping B, which is r subjected to the same base transient acceleration E Thus, we have lt. T I&l Iml S (Wr) (7-5) < qr d and for a singic mode which is an element o' chis vector, we have N S (Wr) ( -6) gr " E Drj Sj d j=1 Now recall that the actual physical peak transient response displace-ment at point i as a result of the r-th mode is given as N u{t"Dr'lr*&r E @rj "j S (W ) (7-7) d r i l j=1 Furthermore, we can relate S (Wr) " S (Wr)/w and Eit " "it w (7-8) d a 53

where S (Wr) is the absolute acceleration response spectrum for the single a degree of freedon Oscillator. Thus, combining Equations (7-2) and (7-8) into Equation (7-7), we have EIt=2B E /5 S (Wr) (7-9) r ih 1h a or in terms of the notation used for the instrumentation identified in Section 2.2, a$t=2B IIll(Wr) air (7-10) r where S is the value for the acceleration response spectrum at w, and ir r Ilgy(w ) is the acceleration transfer function for the structure at e r r under harmonic excitation. In words, Equation (7-10) allows one to calculate the peak absolute acceleration at point i by using the value S at the frequency (w ) Of i r the absolute acceleration response spectrum S (w) which has been computed a for the base transient. Note that the computation must be for a damping value of Sr, which also must be valid for the mode of the structure being investi-gated. Also, one must include the magnitude of the harmonic transfer func- / which can be measured during a resonance search. tion aih alh If the structure is governed by one ' dominant mode in the frequency range of interest, then Equation (7-10) is sufficient for prediction of the peak acceleration response. However, if more than one mode influences the response, then the most likely peak response can be obtained by E(28 air /^1r alr\\ (7-11) aftP= r .r 4 that is, an SRSS peak response. These relationships will be useful for correlating the mechanical behavior of the cabinet in a lat r section. Note in doing this, that different applications of the equation nust be performed in each the Y and X directions, since the cabinet has different response characteristics along each axis. Furthermore, the above relationships have been developed for acceleration responses. Of course, they are also appli-cabic to strain or any other type of response, providing that the appropriate transfer function has been determined under harmonic excitation. 54

=. i 7.2 Development of Test Severity Factors One of the major objectives of this' program has been the development of some means of comparing the severity or damage potential of various seismic qualification tests. The basis for this comparison will be developed in this section. Ilowever, it is first necessary to introduce other concepts which will be used as ingredients for the severity factors. We start with a dis-cussion of time-average responses in systems subjected to nonstationary ran-dom processes. The general basis for this development has been given in i Reference 9, and has been applied, in part, previously to the problem of seismic response of liquid slosh in a cylindrical tank in Reference 10. We consider the response at point 2x of a linear system subject to an excitation at point 1x by a nonstationary random process having the nonstationary power spectral density G1x(w,t). We can predict the non-stationary response power spectral density as G2x(w, t) = lIl2x1x(WI C1x(w,t) (7-12) 4 where ll2x1x(w) is the linear harmonic transfer function for the system. If we now consider earthquake or simulated earthquake type transients, we will also average these quantities over the duration T of the transient. e Thus, we can write the time-averaged relationship as l U2x(W) " II2x1x(W) Ulx(w) (7-13) Now we consider classes of excitation transients in which all samples have the identical normalized time average power spectral density shape (as a function of frequency), but the magnitude is proportional to the time-average mean square of the acceleration. Note that this type of transient classification is consistent with the general nature of earthquake ground 4 motion transients, and in fact, is analogous to the response spectrum en-velope curves specified by the NRC RG 1.60. Thus, we can write 2 G_1x 60) = G_k1x(w) alx (7-14), where Gk1x(w) is the normalized power spectrum for a given type of L tran-sient k -(i.e., k = 1 may denote earthquake ground level, k = 2 may denote carthquake floor level, k = 3 sine beat, etc.). Furthermore, atx is the 55

l time average RMS value for the acceleration. We now substitute Equation (7-14) into (7-13), integrate o'er frequency and take a square root to obtain 1 a2x

Ak2x atx (7-15) l where Ak2x "

JI' ! H2x1x(W) Gk1x(w) du (7-16) 2 _o Thus, Ak2x is a constant for a given response point on a given specimen i or structure, and for a given class (k) of excitation. Then, Equation (7-15) states that the RMS time average response acceleration is propor-tional to the RMS time average excitation acceleration. This assertion will be checked with the data obtained from the present experiments. We are now in position to develop relationships for a test severity factor D. It is recognized that RMS vibration levels are useful.for deter-mining effects of sustained vibration on failure such as fatigue. However, it is also recognized that peak acceleration levels are useful for deter-mining the occurrence of threshold type failures, such as fracture or opening of electrical relays. Furthermore, time duration of exposure must play a role in damage that will occur. Therefore, on the basis of physical reasoning alone, we define a damage or severity factor according to D = a* a T (7-17) e It is recognized that this is only one possible way that the definition could be postulated. That is, each term might appear with some exponent (or frac-tional exponent). However, this is a detail which must be left to future work. With the present development,.we will provide at least a means of relative comparison of severity. Thus, it_can be seen that with this relationship, we can define excitation severity ix=afx aix-T (7-18) D e and response severity D2x " ^2x a2x,T - (7-19) e 56

= ~ l l and the ratio D2x/D3x is a measure of the tendency of a structure to amplify or attenuate the severity of the excitation. It is, therefore, desirable to i obtain a relationship between the excitation and response severity so that l this characteristic of the structure can be determined. We assert that for a given type (k) of transient excitation process, the ratio of peak value to time average RMS value is a constant. That is, for a sine dwell longer than about six seconds, a /5 u 1.414 and for stationary Gaussian random process, a /a = 3.0 at 99.9% probability. i For a nonstationary earthquake type transient at ground level, the ratio j should be even greater than 3.0. At this point, we simply assert that the value can be considered constant at a given probability IcVel, without deter-mining its exact value. Thus, for a given type of process (k), we have i ) (a{x/alx)k =Bk 1 Now, for linear systems, the response has the same probability distribution as the input, hence (a2x/a2x)k " (" x!"lx}k " B (7-20) k Now, if we square Equation (7-15) and multiply by Bk as expressed by Equation (7-20) as well as multiply by T, we have e D (7-21) Dk2x " Ak2x k1x where T )k (~ } Dk2x = (a2x a2x e ( Dk1x " (" x ayx T )k (7-23) c and the latter accelerations are understood to have occurred for a given type of process (k). Hence, Equation (7-21) says that for linear systems, .he A Provide a measure of the tendency for a structure to amplify or k l attenuate the input damage severity. For nonlinear systems,. Equation (7-21) must assume some more complex' form. '57 l

_q 1 i I We now have a basis for computing the severity of a given run (j) ( l for a given type of test (k). For a test which includes several runs, the total severity becomes 1 (7-24) -Dk= Dkj The above discussion has ignored the existence of cross coupling of i responses between input axes, as well as the possibility of multiple inde-pendent simultaneous excitations. These problems can probably be handled on an SRSS basis, and are left to future work. The present cabinet'speci-men ".n be analyzed with the expressions as developed. At this point, it l la more important to use the correlations with experintental data obtained I to determine whether plausible relationships result. l i 4 I I i 4 i i e I t i t i i i I j 58 ~_

8.0 ANALYSIS OF CORRELATED RESULTS Having acquired all data from the series of representative seismic qualification tests, and outlined the preceding mathematical background d,evelopment, we now develop an analysis of the data which forms the real meat of the results for this pro:; ram. Discussions will be presented for a series of topics which f all under the original objectives outlined in Section 1.0. 8.1 Mechanical Behavior of Cabinet It 'is first appropriate to establish the general dynamic behavior of the cabinet. Certain aspects of its response characteristics have al-ready been noted in Section 5.0, i.e., transfer functions for harmonic ex-citation, and in Section 6.0, i.e., sample response time histories. Herein, we investigate certain response data from all runs in a form which allows an initial comparison of the. peak and RMS values for individual test runs. Most of the respoases at various locations are governed by one dominant mode, so that a comparison of measured and predicted responses can be obtained by applying Equation (7-10) repeatedly. First, actual measured peak values af are plotted against peak spectral values E lr of the excitation response spectrum at the appropriate resonance frequency. Then, predicted peak values are developed by using the same spectral values air, a damping value of S = 0.05, and the transfer function Hit (e ) r r at resonance, as presented in the data of Section 5.2. Note that for these correlations, all floor level response spectra for the excitation a had to be recomputed at a damping value of 8 = 0.05 (since initially, values for all floor level tests were computed at.8 = 0.01). Also, the measured damping fer the cabinet modes were~slightly different from 57. in some cases, but'this was considered negligible. It can be seen that the resulting com - parison determines the validity of applying Equation (7-10) to prediction of cabinet responses, as well as provides a basis for comparing peak re-sponses for individual runs of all tests. Figures 8.la and 8.lb show results for the paak responses at the cabinet top forLY-excitation and X-excitation, respectively. Experi-mental data for the various test. runs are labeled in separate categories, '59

12 12 i i i i i i i i o Ground Level A,+ Resonance o Ground Level A,+ Resonance o Floor Level o, a Earthquake Source o floor Level a, a Earthquake Source 10 A Sine Beat a Uniaxial Randra 10 A Sine Beat a Uniaxial Random O Sine Dwell O Sine Dwell ? T .f' 8 =f 8 a' - 1.23 2 5 a 3 3g - 1.28 3, 3 I 6 I 6 5 5 a sa a 5-4 E 4 E E o A O 2 o 2 A l + t i i e i i i i l 0 0 l 2 4 6 8 10 2 4 6 8 10 13.0 Hz-SPECTRAL ACCELERATION, 3 g 9.8 Hz-SPECTRAL ACCELERATION, 2 -g 3y 3 (a) Y-Axis excitation (b) X-Axis excitation Figure 8.1. Peak Acceleration Responses at Cabinet Top

as indicated by the symbol key on each figure. This includes a separate notation for runs which utilize the earthquake source, identify the uni-axial random test run for the respective excitation direction, or emphasize those sine beat and sine dwell runs which are applied at a resonance fre-quency. Finally, the appropriate numerical form of Equation (7-10) and its associated theoretical line are presented in each case. It can be seen from these figures that the experimental data, in general, do form a single correlation line, but it is somewhat different from the predicted one in each case. Furthermore, there is a significant separation of the results for the sine beat-and sine dwell resonance runs from the rest of the data, except at the lowest amplitude. In fact, the resonance points tend to fall below the predicted line in each case. The deviation is more pronounced.for X-excitation, than for Y-excitation. Much of this behavior can be attributed to the nonlinearity in the response, as f was described in the transfer functions in Figures 5.8. The values of 1131 N) at 13.0 Itz and at 9.8 Hz, respectively, represent the slopes of r nd a3x curves at an input value of a1x = 0.10g. These slopes the a3y diminish for larger amplitudes at resonance, and demonstrate that their use in Equation (7-10) would provide a better correlation with the experimental data which resulted in excitation of larger responses at resonance. At the same time, however, the' tests which include a more random or earthquake type of motion were not affected by the nonlinearity, as also were not those sine beat and sine dwell tests which were applied at a frequency off resonance, regardless of the severity of the test. In general, the actual measured peak value correlati*n line tends to be higher than that predicted by Equa-o tion (7-10). Figures 8.1 provide a measure of the severity of_ peak responses for the various tests. Of course, the peak spectral value Ely atx pro-or vides a measure of the ability of each input transient to excite the respec-tive cabinet mode. Likewise, a37 provides an indication of the peak severity for the response point a3 In view of these assertions, it would appear that on an absolute basis, the individual ground level test runs were the most severe, and the severity of other runs fell below in an order ac-cording to their_ positions in the plots. However, at this point, the 61

information can be misleading, since the sine dwells at resonance were con-ducted at significantly reduced amplitudes, as was explained in Section 3.3. Th'is, a better total basis for comparison will be given in the next section, 3 I in terms of the severity factors developed in Section 7.2. A similar set of data is presented for the cabinet interior panel re-sponse a2, in Figures 8.2. licuever, the response for Y-axis excitation in Figure 8.2a is now influenced by three modes, one each at 13.0 liz, 23 Itz, and 27 liz. In this case, the predicted correlation line is based on the use of Equation (7-11). Experimental values are plotted at corresponding points as well. It can be seen that even more deviation occurs than for the single mode case; values for sine beat and sine dwell tests at resonance still deviate the most. Finally, another set of data influenced only by single modes is pre-sented for the strain c4y, in Figures 3.3. The behavior here appears to be similar to that described for the top acceleration a3 That is, the order of the points is similar. This result simply says that the strain at the cabinet base and the cabinet top acceleration are similarly correlated, while the interior panel acceleration a2 is not. This result is consis-j tent with previous observations of transfer function data. In view of all of the above data, it appeara that the general form of Equation (7-10) is valid for predicting the results observed, but its accuracy is very sensitive to slight nonlinearities. Furthermore, transfer function values at the smallest amplitudes appear to be most useful in these equations, when applied to random type excitation, regardless of the severity of the tests. This conclusion may not be valid if such teuts had produced more bending motion of the cabinet. The consequences of this statement will be explored further in Section 8.5. A second type of comparison is now considered, in the form f time-average RMS respc,nses for individual runs. This, in.effect, produces evi-dence which supports the validity of Equation (7-15) as applied to specific excitation orientations and individual test run:. Data correlations are presented for res.,ponses at those points whose peak values were discussed above. They appear in Figures 8.4 through 8.6. Several observations should be made about these data. In Figures 8.4 and 8.5, the rigid body line 62

i C9 MEASURED PEAK ACCELERATION, a*y-g 2 l o m u o = o m I i i i i o>0 o wmm a 5'5'8, 8 m o c_ m Q E?WE m "%8; o .e 1 o@s O O > m + 0 em+ i to _x. n o a_. = .x. - cm o2 m on C o m m m F

u. 5 3W m

m c Xn x o %an 21 = > x c % 2 = m o a M = 5-o= m gg B o z q-a, ~ n m. a m ~ = x,o <= o c = g o R2 f f f f f N N 5 3 .x E" PEAK ACCELERATION, a*x-g 2 a mm o N h 00 o N m 6 4 i I 6 ,n. o C>O o E. p a n CD 00 v M M m a 2 ~ - + F 55ga_ 2 E o o e, c s N 9 o en r-3 W G CD TO G l ~ ~N5 $h S ~Q C 8 m ~ m xu 3 X y 1 e r-d> 0 > -. n n E D + to m ~ ~ = cm x-x = m rso 5-d a o o o

u. o a os

= - z g - g a xr no_ => w , _ = > m o 8[ X, = = nm l f f f f f a

1 300 300 o Ground Level o Ground Level A,+ Resonance a Floor Level O Floor Level o, o Earthquake Source A Sine Beat A Sine Beat e Uniaxial Random 250 O Sine Dwell 250 - O Sine Dwell ~ 200 200 W E' = 32.5 3 w-c' - 30 3 s o 4y ly 4y lx a 'u# G y 150 $ 150 E 5 o o M x 5 6 a. 1 o 100 100 CD O o A,+ Resonance o 50 G,G Earthquake Source - 50 ^ A ,A e Uniaxial Random 9 0 0 2 4 6 8 10 2 ~4 6 8 10 13.0 Hz-SPECTRAL ACCELERATION, 3 -g 9.8 Hz-SPECTRAL ACCELERATION, 3 g 3 3 (a) Y-Axis excitation (b) X-Axis excitation Figure 8.3. Peak Strain Responses at Cabinet Base

59 AVERAGE ACCELERATION, I - g 3y RAG w .N .~ .o o w a w a w a 4 I i l i O>ao o TM2Q ~ \\ N 55k8 ~ s \\ N N om = \\ \\ N 5225 x x N =~12 $P \\ 1 w U m~ N o \\ p# N, o> o \\ < p \\ ec+ p i> zo N 5. >w cm x - m a a + m C'

m..s 8 2 3 s

o Ez n EEE = FS W E *' b \\O 5 o T ox, \\ $w a 8 3 3 g o P o d. 'w ~ qo R S o i- \\ c>", \\ 'm P ,t e ng AVERAGE ACCELERATION, 'il -g 3x RAE g E P P r .N .N .w g-o u o w o w e O 4 6 4 4 6 El 3 E O>ao B 0 _ e r 8, o 3 .o _ t s a e e C .m U o CD e O G m o m < r er m h'* x o b@ $P E_ mN w g \\ c> x m o m 5* # 'W ~ C Dx ~ w b. eke 3" W C C "x g E*g ~ g,*, l p ~ g. w O 3 x, s P P O '\\,O ne G P 4 w 0 \\ ,o e i f Ch

3.0 3.0 i i o Ground Level A,+ Resonance o Ground Lev 21 A,+ Resonance a floor Level o, a Earthquake Source o Floor Level o, a Earthquake Source a Sine Beat a Uniaxial Random A Sine Beat a Uniaxial Random 2.5 ~ O Sine Dwell

2. 5 - O Sine Dwell SD E

/ }SB / 8 SD [2.0 [/[SB [2.0 / n n p SD p 5 4 / / 5. / SB f1.5 SD 1.5 // y l / >/ 23 Hz l 8 y 8 f 8 / / 27 Hz FL g SB FL g 1.0 / / @ 1.0 GL / c' a a / 0.5 D[# 0.5 1/ DQ p D / / t i t t I f f 1 1 I O 0 0.1 0.2 0.3

0. 4 0.5 0.6 0.1 0.2 0.3 0.4 0.5 0.6 AVERAGE ACCELERATION, i - gR MS AVEkAGE ACCELERATION, i ~9RMS h

(a) Y-Axis excitation (b) X - Axis excitation Figure 8.5. Tirre-Average Acceleration Responses at Cabinet Interior Panel

28 a Groun evel SD SB o floor Level 24 A Sine Beat 24 ~ SD SB O Sine Dwell / o FL ~ ~ ~ Q FL GL f 5 m = 'x o 'x if 16 12 16 W O W <c 0o E E e m 12 / " 12 A ~ u ( 0 0 o 5 W W o 8 ^ 8 A,+ Resonance o Ground Level A,+ Resonance 4 ~ 4 0 Floor Level o, u Earthquae Source-G, u Earthquake Source ~ a Uniaxial Random A Sine Beat a Uniaxial Random o Sine Dwell 0 0 O.1 0.2 0.3

0. 4 0.5 0.6 0.1 0.2 0.3 0.4 0.5

0. 6 AVERAGE ACCELERATION, a g - gR MS AVERAGE ACCELERATION, a,- gR MS g

(a) Y-Axis excitation (b) X-Axi excitation Figure 8.6. Time-Average Strain Responses at Cabinet Base

represents equal acceleration response and excitation. Any points above this line represent an amplification over rigid body response. Individual straight lines have been drawn through certain types of test runs, i.e., ground 1cvel, floor level, sine beat, and sine dwell. Note that all data for ground level (CL) and floor level (FL) tests based on random or earth-quake sources, respectively, fall oc ccmmon lines. On the other hand, data for sine beat (SB) and sine dwell (SD) runs widely differ, depending on the corresponding frequency relative to a resonance. As indicated by the cross-hatched areas in Figures 8.4a, 8.5a, and 8.6a, similarities exist between sine beat and sine dwell runs which excite a common resonan:e. Furthermore, the slopes of these data are much higher than those of the other type test runs, which are somewhere between resonance and rigid body motion. The slopen of the individual lines of 'hese figures clearly repre-sent values for the A and A k3y L3x constants in Equation (7-15). This being aa, it can be seen that there is an increasing severity of response per unit excitation, as one passes from ground level tests to floor level tests, to sine beat and sine dwell tests at resonance. The latter two types of tests are much less severe when applied off resonance, as can be seen for results of this type of test run also. It is recognized that the above correlation has been established with rather limited data. M>re data needs to be acquired at intermediate test levels in order to establish these curves more accurately. However, several such additional runs were made for the ground level tests in the X-direction (see Figures 8.4b, 8.5b and 8.6b), and the straight line correlation is reasonably established. We assert that all indicated curves could readily be established similarly, and that the sine beat and sine dwell curves at resonance would include significant decrease of slope at higher amplitudes because of the nonlinearity which has previously been discussed. Finally, it may be observed that the strain curves of Figure 8.6 are similar in pattern to those of the top acceleration in Figure 8.5. This result is consistent with the observation on peak values, that both are an indication of overall bending in the cabinet. All of the above results are extremely important in establishing a comparative basis for severity of various seismic qualification tests, as 68

l l will now be demonstrated. Furthermore, there are even more far-reaching implications of these results for use in the development of improved speci-fications for qualification tests, as will be demonstrated in subsequent l sections. t 8.2 Comparison of Sever _ity for Various Tests The severity factors developed in Section 7.2 will now be used as a basia for comparison of all of the qualification tests defined in Table III-1. In this regard, it must be recalled that each complete test is comprised of two or more test runs. Therefore, excitation severity factors for each separate run were computed by the use of Equation (7-18) and response sever-ity factors were computed by Equation (7-19). The total respective input and response severity factors for each complete test were then obtained by using Equation (7-24). Thus, data for ten different representative qualification tests (which were defined in Table III-1) were developed. In this development, it must be recalled that some of the sine dwell runs were performed at reduced levels, as has previously been explained in Section 3.3. For the latter cases, the full level results of these runs were estimated by a linear scaling up of the excitation and response values to full test requirements. This will re-sult in a conservative estimate for the present system, which suffers reduc-tion in response at higher levels. Final results of the above computations are given in Table VIII-1. Excitation damage severity is given by ED, while response damage severities i are given for three example respon/e points a2, a3, and c4y. These absolute excitation and response severity factors for each test are useful only for a qunlitative comparison of one test with another. The most useful comparison can be made in the last three columns, where the ratios of the damage factors-are given, liere, it can be seen that all ground level tests, regardless of absolute level, fall within a small range of this ratio (i.e., 1.6 to 2.04 for the accelerations a2 and a3). The floor level random and earthquake-tests are the next most severe'with a range of 4.99 to 6.36 for the same accelerations. For a2 the sine beat test is about twice as severe,'while the sine dwell test is again twice as severe as this. Similar quantitative . comparisons can be made for each' response parameter of consequence. 69

Several general conclusions can now be made by studying the results in Table VIII-1, along with results that have been presented in Figures 8-1 through 8.6. The effects of using different random samples for various tests which match the same response spectra are very small. Furthermore, the abso-lute damage severity can be used as a gage whether a sufficiently identical matching of the response spectrum has been achieved in the two cases. For example, from the values for ED, it can be seen that both the dependent i random (Test 3) and the independent carthquake (Test 4) ground level tests were more severe than Tests 1 and 2, but all'four tests fit the same category as seen from the severity factor ratios for all ground Icvel tests. The fact that the inputs were more severe resulted from an overmatch of the response spectrum for Test 4, and the additional exposure to two more runs in Test 3. By looking at the results for individual runs in Figures 8.1 through 8.6, it is cicar that whether one uses an earthquake. source or random source, all points fall near the same correlation lines, so that the choice of motion source is immaterial. The uniaxial random floor 1cvel test (Test 8) appears to be of similar severity as the independent axis tests (Tests 5, 6, and 7). This conclusion must be qualified, however, since these tests required very little vertical excitation and there was negligible cross-coupling between the axes of the cabinet structure. I At this point, it must be emphasized that the above results have been obtained for a single specimen of electrical equipment, although it is thought to be representative of such equipment in general. Therefore, one must be very careful in generalizing the above conclusions. Obviously, additional work must be performed to provide a better understanding of the generality of the results. One might.also ask how the entire above described approach can be useful for estimating the effects of different types of tests on another piece of equipment? One general approach is now described in a sequence of required steps. There are two essential ingredients that are required for the esti-mation; (1) the detailed test specification for the different tests, and (2) the liucar transfer functions for thel specimen. _The steps of the pro-cedure to be applied to each test to be compared, are as follows: t l I l 70 i

l TABLE VIII-1. COMPARISON OF DAMAGE SEVERITY FOR SEISMIC QUAI.IFICATION TESTS lED/ED E C C Test Test A D B F ED /ID ED /ED No. Type ED1 ID2 ED 3 Ec4 3 1 g47 i 2 1 7 1 BIRC 131.7 258.5 268.2 119780 1.96 2.04 910 f 2.02 1.91 971 2 BIRC 170.9 345.3 326.8 165880 l 1.71 1.74 860 3

nRC 285.2 488.8 496.5 245220.

l 1.99 1.60 795 4 BIEC 263.8 524.9 421.1 209700 ll 5 31EF 9.5 54.0 52.0 32800 5.67 5.46 3452 6 BIRF 14.8 75.9 73.7 43370 5.14 4.99 2938 7 BIRF 10.7 58.2 68.0 37034 5.45 6.36 3467 8 URF 13.8 82.7 73.6 l 39890 5.99 5.33 2825 9 SB $0.7 585.8 237.2 149120 11.56 4.68 2944 10 SD 46.2 977.3 521.7 261586 l 21.13 11.28 5657 8 - Siaxial C - Ground Level 1 - Independent Axes F - Floor I.evel D - Dependent Axes U - Uniaxial R - Random Source SB - Sine Beat E - Earthquake Source SD - Sine Dwell 71 L

(1) Obtain transfer functions for the equipment specimen for response locations that are significant with respect to potential failure. This can be done by resonance search tests or by analytical development. (2) Develop a test time history for the appropriate excitation. In the case of the specification including a required re-sponse spectrum, compute the TRS such that it matches or envelops the RRS. (3) Compute the time average power spectrum G1x(w), the normal-ized power spectrum Gk1x(w), and the RMS excitation accel-eration li all associated with the developed time history. 1x, (4) Compute the constants A by means of Equation (7-16). k2x (5) Calculate the time-average RMS response li by means of 2x Equation (7-15). (6) Estimate peak values of response a2x by using Equation (7-10). Trancfer function values H (w )and peak spectral r values a obtained from the TRS are required ingredients. (7) Compute the damage a:everity factor by menna of Equations (7-18), (7-19), and (7-?4). The above procedure is applied to each type of test to be considered. The results then allow a determination of which test is more severe. Thus, it can be used as a design tool, or may also be used as a basis for com-paring the aeverity of previously-conducted tests on. equipment that is al-ready in operation, with predicted results for tests whose specifications are based on more recent criteria. 8.3 Floor Versus Simulator Natural Modes The results of the resonance search tests in Section 5.0 have demon-strated that some differences in modal response may be experienced between floor mounted tests and simulator mounted tests. These differences occur because of dynamic interaction between the specimen and simulator, whose impedance is large, but can never be infinite. The imnediate question is, what infi m nce this difference may have on the outcome of a given qualification 72~

l l l test? The results of the present study indicate that the outcome of a ground level test, in which the input motion is sufficiently broad to encompass any shifts in the resonances, would be influenced very little. On the other I hand, in a floor level test where a concentrated energy exists (such as in the present tests), the results may vary considerably, depending on whether the shifted resonance has the proper relationship relative to the excitation energy. In the latter case, one should consider running resonance searches with both a floor mounting and on the simulator, and if differences exist, the excitation motion (i.e., response spectrum of the simulator) should be modified so that the concentrated energy of the excitation will match the f resonance for the simulator mounting. In this way, the correctly amplified responses will be achieved in the specimen, although they will occur at a somewhat different frequency. It should be obvious that a similar type of adjustment should be made for a sine beat or sine dwell test. 8.4 Component Excitation In electrical cabinets, it is not unusual to qualify the cabinet structure with only dummy components attached. During the procedure, it is then necessary to provide sufficient measurements so that subsequent qualification of the components can be performed on an individual basis. The usual procedure involves development of a new test response spectrum from the motion that was measured on the cabinet at the appropriate mount-ing point during qualification of the cabinet itself. Examples for the pre-sent ground level test (Test 1 - Run 2) are given in Figure 8.7. These spectra are based on the response signal a2 at the interior panel of the cabinet. They should be compared with Figure 6.4, which gives the original input response spectra for this particular test. As one would expect, there has been some amplification of the motion and, in particular of the ZPA, at the response point. Ilowever, there is a large amplification at about 6 Ilz (Figure 8.7a), and no apparent amplification at the known reso-nance of 9.8 Hz, as one might expect. These differences obviously can af-feet the required RRS for subsequent tests on components attached.at this l point. This apparent discrepancy led to further investigation which relr.tes to whether or not the enveloping of an PRS by the TRS (as required by Refer-ence 3) is a sufficient condition for defining the various tests being con-sidered. More on this matter will be discussed in the following section.

l: if &f' L c I'

f~,;p@l;:n. ;,;;l p .73

__ g ...g... .,m .~

r. t 9

f. -+ .c. g e , m,.

,p--b..t

- _. _.,+,- , ps.- .,i f .tp. .v g n .,-.m 44 ,g ~4 .n j l!,[ ..l I ) hl! hW l !fl 4' 1 l _.i. '.:.I. ill 'l .f e ...i. .'1 ddW. ]. j!,. h.1 g: L, m Vd i G l 4 @ 6

El o

'1,, a y 1 s i 3 o p 3 gg, y 7 q m -- ggy 1 -~ t g r ~e %m w$..d y n

e. M_i ' _.i,.m(1H d

'.d l < ..dly .i . l.~t...! .4+. i l ? ! A A _ o,4 l e 4, .r .._..q.4 4,... ..&.+4.. 77,'-.j e ie p 4 ~ f.,.- 4.. ...L9.,. !T.. .+ L t.L.: .+4 q.

.. a4

's' ..n. y .t** d '..L. .,.a#- j ... y i i 4 f. .u o. 3 ..t

i t

o n ..3

n

,o.o.W l ... 4 i,. n i.. 4 +w - g

, s%-

~ 2 g: - m. , u nn c :, m e 4 a u m w .u,;n.. 2 c 4c 1-g +. a e,. O.. ,,+,p .Q't J %, 4 14.._ J J l ]',.'.h 1H { 5-U 1.1% t I >[t. J.4 I;4 l 4; k

, jj,'i g

E pl.k h } W i a. I d ,,)...,(.! a.. 4p ,.p g ..,...s g M ;7 73l 3 i ! a lL.y 'J 'D q ll 1 ly o m .lI , i .,in.,. . _o e an;w a 2.:a.. ..,.y t._z. e.n,..g - 7. a.,_. _.,.. g . ~ g~.,, r 7.,. .f "r**"-'t ...144 41 O, ...?_ o --+++*t'..

- -'t r-

- -1

r**

Ttt" i m m r _.ita .,h .,p,t.;,1.Iy

. o 44 L

q l i o p-l!1 ;W r4 II_..: i n.nu s til go i as WM ! l i.,! @ t n Tl ! u.J_ o TE.plm} g t: .._i i s 4.

m.,y yn

+,s. v w. q l ..y.;. .9 H _r r 6 9, $ g.t..i. _,.j e d t m, __-W : g, dy _1,m:.. y 1n,g. L : t

j g i

g j l ,g a y 9, [ g .n 1 o g a .c e nd d. 8 4M o o ~ 4 d

g 3 *uope2aga33y g d a-hJ

'T 7 ,. j Q 4. .;. h. 'll' h.,,,.. 'Ty Z. l__ " i. m u. ,7 4, 7 ec g +4 a.. .g

g.g

.q; w g ...:.g q g M ,g a @4 3 i M W gli 1 h M $. i ' . g.,. 1,.. g ,f % p.W { { % 4 ,i 0; i 3 ..u.. x.2. . a _ : -g J'"I;hi'gl/I

_.w!'.

.*r *.l l' l d ' lie i / 7: 1 k in ,,,.e9 ,-p' ' .,,[ 'l j i .'.d%

aj

+- H-4 w u Z 3 li l dp ^ O.1 ..[I

j l h' '-.l se in l

f '.g ,I if

ll i

r*T dl '! ; 1 l< = e'> 3. y ,,..g .a.t L H e 4. o M g N 79 4-- 1. -~ .4 +- e- .4 y r. r. 4 44..., y . i $1w. 4 , -[. ..4 OJ n'.% u t 1-2. 7 4 e. 1.- '.p..'. ',.A, 9' mp i i, t a t + 7 7

n e.

p -1.12 .?. as h a.L.q,l!L. . j i ? t W .H 2 a i @ c y 41 -~ , t-- e ?. ? d M y n 19-l D 8 ~.;. i ') c .a 4' I m,,1 +, gaC t s c wil l.< ..,i,.m .It. ~ 4.i, .m:

y

..a....,i m, s .it .y 3 m. 44,. ..Ij".'.u T n T dl"]hji l lI L{ ,4 l' T. i

%.6.d.

>c;il. ] a i l se , e + - !.m.; v I li,* o', i _.;1 o u., <. + m. - j .-+ --+. r ..p+ g .-4., %,.7

p' i.""

j o' ', ^, 'n '88 88.1,-* Q " * *4'0".4- ...1. O

,4,;,tg,.

T .h n I 1.. T .g i-4;i3g i .p I t '* w I .,i.i.M e nn v.;i ! h j q c r m q r o n3 J 4 L 1 } ' 0 ht h,JI 1 . a. ,u; h p+ gn.j!k 1 l L.,!JM L I 4 .. N i i i J -+;- + u m;- y' .in.a. : ap. 3 +- t t p t +1

,h y Li

.q .a %, i i y ia i m j m s i........ - f C O O. O e4 O U 'UOJ192a{a33y 74

l l l 8.5 Effect of Excessive ZPA l The apparent discrepancy just described leads directly to another I problem area, a determination of how an excessive test ZPA influences the results of a qualification test. The presence of the excessive ZPA results from a lack of sufficient matching of the TRS with the RRS, a procedure which at times can be quite time-consuming. It can also result from the presence of resonaraes in the simulator shake table system itself, so that no amount of effort can remove the effects. Thus, in the test lab, it is not unusual to have the excessive ZPA occur. This has been ascertained from discussions with personnel of various test labs as well as published evi- ' I'I dence. Most test specifications based on the guidelines described in Reference 3 simply require that a TRS envelop the RRS, and no maximum may be given. Of course, one might expect that an unlimited ZPA cannot be condoned, as the large peak excitation must eventually produce an overly conservative test. As long as no failure occurs, one may not be concerned with this. However, if a failure does occur, how can one argue that it did or did not result from the excessive ZPA? Furthermore, can the presence of the excessive ZPA produce other problems which may inv.111date a test in other ways than overconservatism? The latter consequence will be discussed first, as it relates back to the discussion of Section 8.4. In the present work, the earthquake derived ground motion test (Test 4) appears to be the most severe with respect to excessive ZPA. Therefore, some additional analysis of excitation and response signals for Run 1 of this test will be developed. Figures 8.8a and 8.9a show re-sponse spectra for the excitation ayy and interior panel a2y, respec-tively for this run. It can be seen that significant motion amplification a2y, but it does not give the appearance of excitation of any occurs at dominant modes. Recall that a similar observation was made for the re-sponse a2x in Figure 8.7a. At this point it was becoming obvious that less modal amplification of the response was occurring than one might ex-pect. This seemed peculiar, for the presence of the first mode at 13.0 Hz for the Y-excitation, and 9.8 for the X-excitation, surely should be strong-ly felt in the response for the required motion. Note that the RRS has approximately an amplification of 4 to 1 over the ZPA at these frequencies. J5G 3 75 0 l

100,- _-. _. _ _ .... _ _ _ _ _. _ _ _. - -._ _.. _ =._.__ _ ___._. _. .~; Test 4 - Run 1 O.10 w.._ 5 ~ h _h B=0.{0 g7; -- ,l _7_7 bb' E ~ z = - =g m n & n & =,-kd+1%. LLinwu We-*"Y=' LL '. M ^ s ~k21 +.1 - m-an ~ I- =- a m 3:_ _- --=-_ s i-5 O.: 3 [1 q.. 2 a =.s w sv e: =_9 _.+- 1 Q-y j;pg.,4.[___.z 5 f. f .f __-p__. ] [p,_j}-l-f[- _ ' _ifM#E4 ?i.~~_'__-' _ _ 5_:.M F:3- ""__==-15 _- f,._E-_;-.. .,iE

1J E

gy; = .r M5t_ i _f-h(j p.=.;.Q$.y @ ,d t..]ft g [' g gy 4.gg g -s c.._._._ I.- E r.4r - - -fit;j. An j t;14f - -- --r-q==. '~"'i q._ - y 0.01, ,I*M. %g g 10 ; as ..11._ _; r-J:L: 'W: n.c __ g,._r :. .p e- . - _ _. -. - +.. _._... _. +. u _._n. x.-- m-w g gi-g e y.m w s,

_ r 3 3== m ww n_n-g.

4w_: ,,u j; p ,p ,g 8. =.- . =_y ; _ m,.wm g_ .;_;- ; ;. g_m. m n %_.g z 4 w g g m m,._ .= _,_~ -A g _=_ =, 4 _ 4u=. =.mnt-&c m..

d= # 4 a

= =

==~=-

-,=mqw;

e 5 M isii=a

==r m=,_m _m o _- i-JW f V~-T p-a'N-- ^.,.=g i =yage - gmy. m. e ,o W==paa5= ---= -s;c g_ egg = we n g

r_r rt r._- -t __

..q ..? g J T 1.0 - T 0.001 . M h. ..y ...r. ~ y .o + - - -. _... u u u p. y = F- 's = Ic y l - -e t =-> .=*-11 f, 4 .._y-bw er a .t F#.ilN-sw I- = m L fih bzi41 - 4 E 9 E i Mf;= W:. e-4 t;- - t -+rt 7 -ErtdrJ#- 'G -t-- f Wi== 5-=' ji MJ5 =d=E- =i-=. =.: =':- . = '4=;-=~-. .f-.--+ t};gy --. L{3 .s.l . _y--

j 3.L___

jdF'

. -.JC Er5:M=- 9 I I L

= L=l?=-9E vi:-- J. :t- :1: M= -4 g]5-'gww . u=qw' L p '{E. Test 4 - Run 1 ~ =y y l }. =~==m ug - ~L-r .[ ."f9 _n 1 -=m t-n=: Test ZPA = 6.30g - - -l - *- .j m._;___._ 0.1 0.0001 gp gl.l .l ....i ......) .....j lt 100 0.1 1.0 10 100 0.1 1.0 10 Frequency, Hz Frequency, Hz W, .. Re.,en.. s,ect,u. ,.T1.e-A.e,..e. e s,ect, IIGURE 8.8. PARAMETERS WHICH LESCRIBE CROU1D ACCELERATION (ag ). Y-Z SXCITATION y c.e:2 s. 3 \\. =* o.

100 0.10 i 1 Test 4 - Run 1 _ l L L ' _ ___. .,,r _.6 i.., - .y 4 i i'.! ? B - 0.10 Hz t.1-i 1_l I- + *- * + '.3 ; t r__ -3 t}. _,,,. g ; p_ g _ ;., p g"q - g g_ 6.-=. 6..< - t 4 a y.4._ _., 7g 8 - O A50 l j g.4_ g 4, g1.q _g y. g. g f.__. :.. .qgre .., g u. i -- q 3__f}M.t;g f A_p7;.'~q g.3, _ ; g. _. _ . - i + _. , 4-r -. 7 21 - p } 3_.-.. ,q.31 3;..q g j' _,) TQm t _q. ;. _4 p -W 4 1-Mr. - Gi=~71W ' rsf-\\1 hjw : u& .Ms; n-g gn it hpmam k. .s - + 5_ u = n2 =ta_ "3__E. iyE_l q. 7-rEEM_,.E__E, ).E.j LP'al=i=p q, a.spja_ =__., _ p i - - - + - r :tr .:E = 4 i ~ z y l _ g.. 1 ___,f..!. . l. ; :. es i. 4 m.-.. ___g _._.__.m._4_ g._.__ !1 l _l? g - 4 < L,g ].-{ _p_ f at r., E t ' 1 f.~- . %/4 + ' - -t Si 'f } } '=4 '4 i- _.N u 'i c_ + t g. m. -j .g _, r g gt Wb .Tij , -- -.;--- g j ,.;-,+q w g j 44 g ,-4-- r tM //N Ei M b-M-4.ElEl'4142?- k M=. -+.:g

j i I.&u.-.=r

--) j - ] h. ':.~[ c f@ +2 ~ 7 _g 7 m_ 85 @b .-y E-.= .<, - - - - # t.-C-t-y -- y ;. y.., i, t, . _.. ; ic i ..a - ) ip ;Mh - r ; pp,d .y37 r=; a 3 l -

j. y gl

- =.m : 50#" Alt .j .Wiii: o w,- -1_ . __g_- g d_._ 7 [ MN. r.-=d-HE--r 2/cli.-i diWH9-: ut=-Jt\\:*-1r T 141^ d=h' :: = _4 g . n__. nd_- . -+ - $ 0.001, - - ~ " ~ n .. feH -~ -"(_:..

~~'

A "--~~~-' ~ [ ___.._..i_ i_ - -i N_._ 1. -i. 3 's ~ '.___.4... u . _. ~ -.. ,4 g ..i-p -.,--4m

p -m 3 2_

, i., q.., ..t t i m .-22 4 r t-5 - l.. =_:4 a v. J.= t. c.ay -

4 i-ei p,,s

n. 6 i

.t h-~- g .r

gp--} p r

g _j - rp _m _ 7 l .j' y_ l .j f jh _ pd.t .:n it-j p.) l. 5 _Qi} i_-:d.. Q J.N i [ Y ~ ' - liNHTi% j. 3h ~.i _D~ _q M M m @J: 1 &j .__ { f r

)

_ t-h nd. Test 4 - Run 1 { l' J .21:3 LE L -M

- ~ - -: .=

Test ZPA = ll.56g i -.p- --t_.u,.. = u=-4.- =.. . _=_ u -= = 2: y x.- } 0.1 -. r r: h- :j-l i .r - _tr.cr 0.0001 .....i i a ......l i 0.1 1.0 10 100 0.1 1.0 10 100 Frequency, Hz Fregnancy Hz

a. Response Spectrum

b. Time-Average Power Spectrum r---m FIGURE 8.9.

PARAMETERS MIICH DESCRIBE INTERIOR PANEL ACCELERATION (a2y). Y-Z EXCITATIOA 'J5ED U=='

On the other hand, the test ZPA of the response in Figure 8.9a is 11.56, while the actual input ZPA in Figure 8.8a is 6.3. The actual amplification was less than 2 to 1. The above observations suggest that less energy was included in the excitation near the 13.0 liz frequency than may be appropriate. In order to demonstrate this, time average acceleration power spectral densities were run for the excitation and are shown ayy and the response a2y, respectively in Figures 8.8b and 8.9b. From Figure 8.8b it can be seen that, indeed, very little energy exists above 9 Hz in the excitation. It is not surprising then to note in Figure 8.9b that there is no significant energy in the a2y response at any of the 13.0, 23, or 27 Hz frequencies, compared to that below 9 liz. The power spectra clearly confirm the suspi-cion that an insufficient excitdtion of the structural modes has occurred, even though the TRS severely envelops the RRS for the excitation in Figure 8.8a. This result is enormously important! That is, in spite of the fact that an apparent overtest has been indicated according to the present cri-teria required by Reference 3, in fact, an undertest has occurred, insofar as response'in these modes is concerned. The discovery of this discrepancy occurred just near the end of the present program, so that further analysis of the data could not be performed. In view of the above diecovery, it is immediately obvious that the adequacy of the presently-accepted test criteria is in question, and requires immediate attention to resolve the discrepancy. Although the ZPA's indicate that an adequate, and in fact, ev.cessive maximum peak response has. occurred, the distribution of lesser peaks will not necessarily be adequate. At this point it appears that it may be necessary to use some combination of re-sponse spectrum and time-average power spectrum to assure that both the cor-rect peak values and the correct frequency distribution of the excitation energy have been developed properly. Furthermore, the question of over-conservatism of the ZPA and its effects on the test is so in arrelated to. this discrepancy, that its resolution must be developed joi..tly in additional work. It may very well be that the discovery of the discrepancy outlined in thic section is the single most important result of this entire research program! I

av0Clw dUWbj 78

9.0

SUMMARY

OF CONCLUSIONS AND RECOMMENDATIONS A series of brief conclusions and, in some cases, associated recom-mendations will be provided in this section. The order presented bears no significance to the relative importance of each item. A plausible method has been developed for comparing the severity of various types of seismic qualification tests. The use of the severity factor method as a design or development tool should be giver careful consideration. At the same time, additional research work should be performed to determine the utility of the method to various types of nuclear power plant equipment. Both analytical and experimental efforts are in order to determine whether the exact postulated form of the severity factor may need modification, as well as whether it should be incorporated into guidelines for conduct of qualifi-cation tests. It has been demonstrated that some differences in resonance behavior may be encountered during floor mounted tests and simulator mounted tests. Consideration should be given to adding floor 1cvel resonance searches for those cases that may be influenced by simulator compliances. For subsequent simulator mounted tests, widening of the response spectrum to include the en-tire range of resonance frequency shift is recommended. An extremely important discrepancy has been revealed in the use of matching or enveloping an RRS with a TRS as a criterion for the qualifica-tion test. Since this criterion is the basis for most current test speci-fications, it is imperative to provide specific modifications of the guide-lines which will climinate this discrepancy. The matter of further investi-gating the effects of excessive ZPA should be pursued. This recommendation is the most important one offered in this section. Because of the structural independence of one axis from another on the specimen utilized in this investigation, the riatter of cross-axis coupling and its influence on the outcome of various tests has not beca re-solved. A similar series of tests should be applied to another specimen, such as a valve with attached motor drive, so that significant cross-axis coupling will be present. The applicability of the developed methods to this type of specimen can then be determined. 79

The choice of random source or earthquake signal source for running tests appears to be immaterial. If anything, the use of a Gaussian noise source is much more convenient, and incorporates the desirable random char-acter to the motion. The order of severity for the various tests indicated by the damage factor ratios is a most significant result. It is especially important to recognize that a numerical reir*.ive damage potential has been established by these results. Note in particular that the damage severity of both sine beat and sine dwell tests is greater than that of the random type, as long as excitation exactly on resonances is included (which usually is the case). The obvious implication is that more often than not, these tests have inflicted over conservatism into testing at an almost unconscious rate. It is further currently being emphasized in test specification of the so-called RIM (Required Input Motion) type, which employs a 4.5-g sine sweep through the complete frequency range for line mounted items. It would appear that a swept narrow band random test at the same RMS level would be more than adquate for such a test, would even be a more realistic simula-tion from a physical reasoning point of view, and in fact would be signi-ficantly less severe at the same time. A comparison of the results'for-these two tests using the damage factor criterion should be investigated immediately. Development of a standard normalized' power spectrum which is analogous to the Reg. Guide 1.60 response spectrum should be considered carefully. This would of course be done similar to the procedure utilized for develop-ing the standard response spectrum. Furthermore, the use of power synctral density techniques as a design tool which is complementary to response spectrum techniques should be considered as well. This development would be essential if the damage severity factor is to be utilized seriously. Tnroughout this study, emphasis has been placed on measurement of: . mechanical effects in the specimen. However, it shoul/ ret ts overlooked' that influences on electrical behavior in subcompor?ste Attacned to the cabinet are also implied. For example, chattering si .'3ys can be related to both frequency content as well as peak acceleration leveis at the point of attachment. Furthermore~,'the damage-severity factors which were developed can readily.be expanded to incorporate a threshold-tyre of failure severity. 80

i The matter of mechanical fatigue can also be included in a special manner. The concept of damage sevecity under seismic excitation obviously only has been introduced, as a result of this study, and deserves much further research consideration. 81

4 i 10.0 ACKNOWLEDGEMENTS ] .The authors wish to express their sincere appreciation to several i i individuals who made significant contributions to this study. First, it must be recognized that the electrical cabinet specimen rsed in the study was contributed at no charge by the llailey Meter Company of Wycliff, Ohio. Mr. W. B. Fellner and Mr. M. E. Menard were especially instrumental not only in seeing that the cabinet was provided, but also in lending discus-sion and suggestions throughout the program. Several staff members at Southwest Research Institute provided aid in different capacities. Mr. D. C. Scheidt provided many suggestions for equipment use during the conduct of experiments and data analysis. Mr. F. R. Pittman helped in the conduct of all experiments, Mr. G. E. Ransleben provided cuggestions with regard to specimen design, and J Dr. J. F. Unruh provided discussion and comments on mathematical develop-ment, as well as seismic analysis concepts, in general. 1 l i 1 i e 4 r i i 5 82-4 = e m,

a

11.0 REFERENCES

1. Seismic Qualification of Electric Equipment for Nuclear Power Plants, U.S. NRC Regulatory Guide 1.100 (Preliminary Issue), March 1976. 2. IEEE Standard 323-1974, Standard ft.: Qualifying Class lE Equipment for Nuclear Pos - Generating Stations, Feb. 28, 1974. 3. IEEE Standard 344-1975, Recommended Practices for Seismic Qualifi-cation of Class lE Equipment for Nuclear Power Generating Stations, Jan. 31, 1975. 4. Bessey, R. L., and Kana, D. D., "Some Research Needs for Improved Seismic Qualification Tests of Electrical and Mechanical Equipment," Special Session at SMiRT-IV Conference, San Francisco, California, August 1977 (To appear in Nuclear Engineering and Design). 5. Kana, D. D., and Scheidt, D. C., "A Broad Capability Seismic Simu-lation Facility," Proceedings of 22nd Meeting of the Institute of Environmental Sciences, Philadelphia, Pa., pp. 195-199, April 1976. 6. Menard, M. E., Bailey Meter Company Seismic Qualification Test Procedure, December 1976. 7. Clough, R. W., and Penzien, J., Dynamics of Structures, McGraw-Hill Book Co., New York, 1975. 8. Hudson, D. E., " Destructive Earthquake Ground Motions," pp. 1-34 of Applied Mechanics in Earthquake Engineering, edited by Iwan, W. D., AMD-Vol. 8, ASME, New York, November 17-21, 1974. 9. Bendat, J. S., and Piersol, A. G., Measurement and Analysis of Random Data, Chapter 9, John Wiley & Sons, Inc.,1966. 10. Kana, D. D., " Seismic Response of Flexible Cylindrical Liquid Storage Tanks," Final Report, Project 02-9184, Southwest Research Institute, May 1, 1977. 4 11. Horlacher, W. R., and Haslinger, K. H., "A-Biaxial Seismic Simu-lation Test System," Test Engineering and Management, June / July 1977, pp. 8-11. 12. Skreiner, K. M., Fitzgerald, E. M., and Test, L. D., " Seismic-4 Qualification of Class lE Equipment Today," Journal of Environ-mental Sciences, January / February 1978, pp. 19-23. + L 83

OHM M U.S. NUCLE AR REGULATORY COMMISSION (7 77) BIBLIOGRAPHIC DATA SHEET NUREG/CR-0345 4 TsiLE AND SUBTITLE (Add Volume No., of moropriate) 2 (Leave blank) An Evaluation of Seismic Oualification Tests for Nuclear Power Plant Equipment 3 RECIPIENT'S ACCESSION NO.

7. AU T HO H IS)

5. DATE HEPORT COMPLE TED Daniel D. Kana, Robert W. LeBlanc "N*

I " ^a Fahrinary 1978 'J PE Hi OHMING OHGANilATION N AME AND MAILING ADDRESS (/nclude lip Codel DATE REPOHT ISSUED Southwest Research Institute l n4R Mosm 6220 Culebra Road September 1979 San Antonio, TX 78284 6 (teave bena*> 8 (Leave blank)

12. SPONSOHING OHGAN12ATION NAME AND MAILING ADDRESS (/nclude Zip Codel (l.S. Nuclear Regulatory Commission Office of Nuclear Regulatory Research

11. CONTR ACT NO.

Washington, D.C. 20555 13 TYPE OF REPORT PE RIOD COVE RE D //nclusere dates) Technical Report September 1976 - August 1978

15. SUPPLEMENTARY NOTES

14. (Leave b/m4;

16. ABSTH ACT (200 words or less)

A series of seismic qualification tests has been conducted on a typical nuclear power plant electrical cabinet. Acceleration and strain responses were measured for four ground level and six floor level specifications. The test types include resonance search, biaxial independent random, biaxial dependent random, uniaxial random, sine beat, and sine dwell excitations. Tests involving random motion were derived both from a random generator and earthquake signal source. Response data are initially presented in terms of transfer functions, time histories and response spectra. Then, analytical parameters are developed for correlation of the data in terms of peak responses, time-average RMS responses, and a new parameter defined as a damage severity factor. Typical sine dwell and sine beat tests are found to be far more severe than biaxial random simulations. The developed damage severity factors indicate this result vividly, and also provide a useful design tool for comparison of test severities before the tests are conducted, so that a choice can be made. e

17. CtEY WORDS AND DOCUMENT ANALYSIS 17s. DESCRIPTORS Seismic, Qualification, Tests, Electrical, Cabinet, Resonance, and Response.

171x IDENTIFIERS /OPEN ENDED TERMS

18. AV AILABILITY STATEMENT

19. SE CURITY CLASS (This report /

21. NO. OF PAGES i

Unlimited Had m ifiaa 83 4

20. SECURITY CLASS (This pagel 22, PRICE S

NRC FORM 33S (7 77) )

l LJNIf t D ST ATi s N Let le A M HIGULATORY COMMISSION I ) W ASHINGT ON D. C 205 $"a POST AGE A ND F f f % P A OD ()F i lCI A L B USI N E SS US NUC LE A R RE GU L A T O R Y Pt N A L T V F OH PHIV A T E USE, $ 300 COMMstssoN t J l \\ \\ D*#}D j P &Jud k t,g l.Wdau \\ _ _ _ _ _ _ _ _. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _.}}

ML19309G506

Contents

Text

SUBJECT:

REFERENCES:

Contact:

1.0 INTRODUCTION

2.0 DESCRIPTION

SUMMARY

11.0 REFERENCES

1.0 INTRODUCTION

2.0 DESCRIPTION

.u=

==

=

==~=-

-,=mqw;

SUMMARY

11.0 REFERENCES

Navigation menu

ML19309G506

Text

SUBJECT:

REFERENCES:

Contact:

1.0 INTRODUCTION

2.0 DESCRIPTION

SUMMARY

11.0 REFERENCES

1.0 INTRODUCTION

2.0 DESCRIPTION

.u=

==

=

==~=-

-,=mqw;

SUMMARY

11.0 REFERENCES

Navigation menu

Search