ML20235S520
| ML20235S520 | |
| Person / Time | |
|---|---|
| Issue date: | 07/31/1987 |
| From: | Browzin B NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| To: | |
| References | |
| NUREG-CP-0088, NUREG-CP-88, NUDOCS 8707210681 | |
| Download: ML20235S520 (265) | |
Text
,
. NUREG/CP-0088 Transactions of the 9th International Conference on Structural Mechanics in Reactor Technology Panel Session JK: Structural and Mechanical Engineering Research at the United States Nuclear Regulatory Commission
.To be held at Lausanne, Switzerland August 17-21, 1987 U.S. Nuclear Regulatory Commission Office of Nuclear Regulatory Research Compiled by B. s. Browzin l
/ # ~*%>
NOTICE
-These proceedings have been authored by a contractor of the United States Government. Neither the United States Government nor any agency thereof, or any of their employees, makes any warranty, expressed or implied, or assumes any legal liability or responsibility for any third party's use, or the results of such use, of -
any information, apparatus, product or process disclosed in these proceedings, or represents that its use by such third party would not infringe privately owned ' rights. The views expressed in these proceedings are not necessarily those of the U.S. Nuclear Regulatory Commission.
Available from l Superintendent of Documents U.S. Government Printing Office P.O. Box 37082 i Washington D.C. 20013-7082 and National Technical Informc, tion Service Springfield , VA 22161
NUREG/CP-0088 Transactions of the 9th International Conference on Structural Mechanics in Reactor Technology Panel Session JK: Structural and Mechanical Engineering Research at the United States Nuclear Regulatory Commission To be held at Lausanne, Switzerland August 17-21, 1987 Manuscript Completed: June 1987 Date Published: July 1987 Compiled by: B. S. Browzin Division of Engineering Office of Nuclear Regulatory Research U.S. Nuclear Regulatory Commission Washington, DC 20555 l
,. -. 5, 6
_ _ _ _ _ . _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ . _ _ _ _ _ _ _ _ . . . _ _ _ _ . . _ _ _ .___________.____.___o
i INTRODUCTION The purpose of this session is to inform the technical community about the U.S. Nuclear Regulatory Commission research directed toward resolving safety issues and facilitating licensing, and to provide information on related industry research. The session is oriented primarily to seismic research. These Transactions of the JK Panel Session include preprints of papers which are listed in the Second Announcement of the' 9th International Conference on Structure Mechanics in Reactor Technology, JK Panel Session:
Structural and Mechanical Engineering Research at the United States Nuclear Regulatory Commission. Transactions provide participants with papers selected for this Session. Some of the papers will be read by the authors. Other papers will be only discussed. Four papers have been withdrawn from the Session: JK/6, JK/8, JK/11 and JK/13, and one paper added JK/14. The paper numbers remain the same as provided in the Second Announcement.
Boris S. Browrin iii
}
.)
AN HISTORICAL NOTE 1
~
1
.The first JK Panel Session was held in 1977 at San Francisco, SMiRT-4 International Conference. . This first JK Panel Session has.been proposed i j
jointly by the United States Nuclear Regulatory Commission officers and Dr.
John Stevenson of Stevenson & Associates with scientific support from Dean Bruno A. Boley of- Northwestern University.L This proposal for;a JK Panel Session was accepted and enthusiastically supported by Professor Dr. Ing Thomas-
)
A. Jaeger, organizer and the first Ceneral-and Scientific Chairman.of the first five SMiRT Conferences, as well as the founder of the International Association for Structural Mechanics in Reactor Technology (IASMiRT).. As an homage to the l i
-late Professor. Thomas A. Jaeger, the next page provides a short "Memoriam' arld portrait borrowed from.the introduction to the 6th SMiRT Conference Program.
- l One decade of JK Panel Sessions, held in San Francisco (1977) in Berlin (1979), in Paris (1981), in Chicago (1983), in Brussels (1985) and this one in'. I Lausannce (1987) has proven-to be of continuous interest to the profession, f i
Papers have been presented by U.S. Nuclear Regulatory Commission contractors.;
by U.S. Nuclear Regulatory Commission staff and by those closely related to the U.S. NRC research_in general. Proceedings for each of the five JK Panel Sessions, including discussion for the first four, were published in special issues of-the journal, " Nuclear Engineering and Design," 50(1978),59(1980),
69(1982),79(1984)and94(1986). Besides the Proceedings, Transactions, including papers to be presented at the JK Panel Sessions, were published by Thomas A. Jaeger for the Berlin Conference and by U.S. NRC for the Brussel's Conference, as well as these Transactions published for the Lausanne Conference.
.l The decade 1977-1987 of JK Panel Sessions at the International Conferences of Structural Mechanics in Reactor Technology is regarded as fruitful and may inspire continued activity in presentation and systematic publication of U.S. NRC Research in Structural and Mechanical engineering. ,
/ L @ d'. & @ -
i iv
-enm-, swegywg ,
,j ' f, 1 V h 2
- )
A
'i
- s IN MEMORIAM Herrn Professor Dr.-Ing. Thomas A. JAEGER, der am
- 21. August 1980 leider viel zu froh verstarb, widmen wir diesen 6. SMiRT-Kongress.
Herr Professor Th. A. JAEGER habilitierte sich im Jahre 1970 auf dem Gebiet der Reaktorsicherheit an der Technischen Universit8t Berlin. Seiner Initiative ist es zu verdanken, dass der 1. SMiRT-Kongress im Jahre 1971 in Berlin stattfinden konnte. Er festigte den Erfolg dieser Kongressreihe bis zum 5. SMiRT-Kongress in Berlin. Durch ihn von der Bedeutung der zunehmenden Anwendung der Begriffe der Struktur-rnechanik in den Bereichen Reaktorsicherheit, -zuver-lessigkeit und -leistung Oberzeugt, werden wir alle i Anstrengungen unternehmen, um den 6. SMiRT-Kongress voll im Sinne seines Begr0nders zu organi-sieren. Im Hinblick auf die grosse Wichtigkeit des dort stattfindenden Meinungsaustauschs sehen wir in dieser Veranstaltung die beste W0rdigung dieses aussergew6hnlichen Mannes, unseres unvergessli-chen Freundes.
Y l
1 _ _ _ _ _ _ . _ _ _ __ __
F -
0 L
TRANSACTIONS OF THE 9TH INTERNATIONAL CONFERENCE ON STRUCTURAL' MECHANICS'IN REACTOR'
' TECHNOLOGY EPFL - Ecole Polytechnique Federale de Lausanne, August- 17-21, 1987 General and Scientific' Chairman: Folker H. Wittman Ecole Polytechnique Federale de Lausanne Lausanne, Switzerland
. Scientific Co-Chairman: Sergio Finzi Commission of the European Communities Directorate General XII-JRC Brussels, Belgium Organization Chairman: Olivier Mercier Swiss Federal Institute for Reactor Research Wuerenlingen, Switzerland Advisor General: Bruno A. Boley Northwestern University Evanston, IL USA' JK-PANEL. SESSION PAPERS STRUCTURAL AND MECHANICAL ENGINEERING RESEARCH AT THE UNITED STATES NUCLEAR REGULATORY COMMISSION Panel Session Organizer and Chairman Dr. Boris S. Browzin Office of Regulatory Research U.S. Nuclear Regulatory Commission Washington, DC 20555 USA Panel Session Moderator Dr. John F. Stevenson Stevenson and Associates 9217 Midwest Avenue Cleveland, OH 44125 USA l
vii
CONTENTS Page JK/1 Status of Structural and Mechanical Engineering Research at the U.S. Nuclear Regulatory Commission.
B. S. Browzin and J. E. Richardson, U.S. Nuclear Regulatory y Commission, Washington, DC, USA............................
JK/2 Seismic Margins Review of Nuclear Power Plants: NRC Program Overview.
R.C. Murray and P.G. Prassinos, Lawrence Livermore National Laboratory, Livermore, CA, USA; M.K. Ravindra, EQE Inc. ,
Newport Beach, CA, USA; and D.L. Moore, Energy Inc.,
Washington, DC, USA ........................................... 21 JK/3 Seismic Capabilities of Containments Including Pressure Effects. i M. Amin and P.K. Agrawal, Sargent & Lundy, Chicago, IL, USA, D.B. Clauss, Sandia National Laboratory Albuquerque, NM, USA; and J.J. Ahl, CBI Na-Con, Inc., Oak Brook, IL, USA ............ 51 JK/4 Vibration Experiments at the HDR - German /U.S. Cooperation.
C.A. Kot, Argonne National Laboratory, Argonne, IL, USA; L. Malcher, Kernforschungszentrum, Karlsruhe, FR-Germany; and J.F. Costello, U.S. Nuclear Regulatory Commission, Washington, DC, USA ............................... 89 JK/5 Validation of Seismic Soil-Structure Interaction Analysis Techniques with Lotung Experiment Data - Overview of EPRI/NRC Research Program.
Y.K. Tang, R.P. Kassawara, H.T. Tang and I.B. Wall, Electric ;
Power Research Institute, Palo Alto, CA, USA; and '
M.G. Srinivasan, C.A. Kot and B.J. Hsieh, Argonne National Labo rato ry , Argonne , I L , USA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127 JK/6 Using Component Test Data To Develop Failure Probabilities and Improve Seismic Performance.
G.S. Holman and C.K. Chou, Lawrence Livermore National Labo rato ry , Li ve rmo re , CA , US A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . withdrawn JK/7 EPRI/NRC Piping and Fitting Dynamic Reliability Program.
D. Guzy and G. Weidenhamer, U.S. Nuclear Regulatory Commission, Washington, DC, USA ... . ................................. ... 145 JK/8 Probability of Failure in BWR Reactor Coolant Piping.
G.S. Holman, Lo Tin-Yu and C.K. Chou, Lawrence Livermore National Laboratory, Livermore, CA, USA ............ .......... withdrawn JK/9 Latest Research Results on Seismic Fragility Data of Nuclear Power Plant Equipment.
K.K. Bandyopadhyay, C.H. Hofmayer and M. Kaffiz, Brookhaven National Laboratory, Upton, NY, USA . . ....... .. .......... 157 ix
l l
l CONTENTS (Continued) )
I P_a.ILe
.JK/10 Determination of a Modal Interaction Correction for )
Narrowband Fragility Data. j l D.D. Kana and D.J. Pomerening, Southwest Research Institute, j San Antonio, TX, USA .......................................... 181 <
1 i I Los Alamos Buckling Research Summary - Analysis and JK/11 Design Recommendations.
J.G. Bennett, W.E. Backer and T.A. Butler, Los Alamos National Laboratory, Los Alamos, NM, USA ...................... withdrawn JK/12 An Experimental Investigation of Reduced Stiffness at Seismic Working Loads in Reinforced Concrete Shear Wall Structure.
C.R. Farrar, J.G. Bennett and W.E. Dunwoody, Los Alamos National Laboratory, Los Alamos, NM, USA ....................... 207 JK/13 Results of the First Large Scale Low Rise Shear Wall Experiment Applicable to Category I Structures.
J.G. Bennett, W.E. Dunwoody, C.R. Farrar and P. Goldman, Los Alamos National Laboratory, Los Alamos, NM, USA ....................................... ................... withdrawn JK/14 INEL/USNRC Pipe Damping Experiments and Studies.
A.G. Ware, Idaho National Engineering Laboratory, EG&G Idaho, Inc., Idaho Falls, ID, USA .............................. 235 i
1 i
l X
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . . _ _ _ _ _ _ _ _ . _ _ _ _ _ . _ _ _ _ ._. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . . _ _ _ _ _ _ . _ _ _ _ _ _ . _ . .}
JK/1 STATUS OF STRUCTURAL AND MECHANICAL ENGINEERING RESEARCH AT THE U.S. NUCLEAR REGULATORY COMMISSION I
B. S. BROWZIN and J. E. RICHARDSON
' Office of Nuclear Regulatory Research, U.S. Nuclear Regulatory Commission, Washington, DC 20055 USA.
ABSTRACT This paper ~ discusses the research conducted by the Office of Nuclear Regulatory Research, U.S. Nuclear Regulatory Commission, in structural and mechanical engineering for light-water reactor nuclear power plants. The list'of projects and the description of the objectives include newly initiated and ongoing efforts reported at the Eighth International Conference on Structural Mechanics in Reactor Technology held in Brussels in 1985.
Since the Eighth International Conference on Structural Mechanics in Reactor Technology in 1985,~ the U.S. Nuclear Regulatory Commission's (NRC's) research program in structural and mechanical engineering has remained at approximately the same level of funding.
As-in the past, the research during this period focused on the safety of nuclear power plant structures and components. Principal events that cause hypothetical unsafe conditions are normal operating conditions, severe accidents, and extreme earthquakes. The objective of current research is to establish the safety margin against severe accidents and earthquakes and to identify where safety margins must be increased or, alternatively, may be relaxed.
During this period, there were no new applications for construction permits for nuclear power plants in the United States. Partially for this reason, and partially because power plants are aging, research during this period included determining the degree of safety of existing plants, which were designed for less severe loading conditions or with less stringent criteria than currently required by the NRC. Some work addresses the degradation of structures and components due to aging as the nation's reactors grow older. The research 1
reported in this paper is conducted by the Structural and Seismic Engineering Branch of the Division of Engineering of the Office of Nuclear Regulatory I
Research.
The NRC organizational chart, which reflects the NRC reorganization of 1987, may be of interest to the general reader and to those interested in locating the Office of Nuclear Regulatory Research in the overall NRC fuctional schedule (fig. 1). Since the research reported in this paper is conducted by the Division of Engineering, the organizational chart for this division is shown on fig. 2.
The Division of Enginee,ing plans, develops, and directs comprehensive research and standards development programs pertaining to (1) nuclear safety, including the design, qualification, construction, inspection, testing, operation, and decommissioning of nuclear power plants and fuel cycle facilities and the management of nuclear waste and (2) materials safety, including materials char-acteristics, aging, and the seismic and engineering aspects of facilities and materials. It ensures the resolution of reactor engineering generic safety issues and unresolved safety issues. It establishes or recommends policy, planning, and procedures for the research and standards development programs as j required to carry out the functions of the division and coordinates these pro-grams with other Office of Nuclear Regulatory Research divisions and NRC offices to ensure that the programs are responsive to their needs. It provides funding guidance to NRC contractors, U.S. Department of Energy laboratories, and other government agencies within the division's budget. Consistent with NRC policy and to the extent overall agency needs exist, it maintains liaison with and provides technical information to other Federal agencies, the American National Standards k stitute, professional societies, international agencies, and other organizations in assigned areas.
According to the NRC Organization Chart of April 1987, the five branches of the Engineerirq Division perform the functions described in Items 1 to 5 below.
(1) Materials Engineering Branch This branch develops, recommends, plans, evaluates, and manages research programs l
l and develops standards pertaining to the design, qualification, construction, 2
Il!llli ll l l l NS T N3g T NS T L S D NN NN O OD MN E A T v mP E0 v On EO C Y IR FO DN f t C T A A V5T T Ta MT S TS T
aA Uf t H eU ,WA A 64 g 5GFAR I
NA
- S P *E FE F FE Dg A C FET iU AP FA E C L oJO QF DG A AE gAYR t T N O OF A Ag AR N OFCAE DG 0A T YL OTE
- E N# NS R5 NS A N NS Rr NS Y
A NSJ AE NN5 TA r D a R RF EE TR F A eP OR A A' L O 0*R I
F C O A OA4 FP I CPC O L' sR SO FS O ,YN T S SO Fg Og VN t sRR SOP FS O SaMW FE I A EV N A* vO MDM t
r MTC N h ,yCD E V' L
- FS S M MTC N, A NE
$A OE A P iT R CD S MTCD CG N DC DA ov pRA D DA ,y U RA DAN 0v DM.A
MI T O N E R OD R OD V R' OD N
OS G # MF SN M mF SN N 'v eAF TN E R
O DS A EA R E pga EA R O DS 'E R
A A E G R I
C E G I
R " . . E #
= " _
L " R ' = "
l L
L E _
E ES l DN HN MS T 4S T NA TU S D 0D MN AP FO o, R FO YD TR FEA I
OC EL CA O
G A
T C
E O
pE OF T
Y A
T AiU DG AE A
F O
[,
A9 S
A FET OFC Y
TS AE t*R A
IU DG AE A
FG OA T Mf FA t
R -
AO FiR C NS RF NNg L L NSJ RFA NNS SB F F ON e N OiR A 0A A OO OAI .
Cei E H SO FS 1 S D tRR fS IMtN S
- N G C Ot SOP O t
V EM T
Nh I
M1 MC NE M V MTCD CD ON T E S
FE ON t
m C DA OV D I N DAN DY N ERA r AC EN N A E
E R
tSE T U OD a -
I L I CO O R mFA I G v' F iA SN h F*S G oS E DS EA c 5# E R P 0EP R -
R . ,
O T
l a
r N
OS D U
f n E D FA o I R i TA OD t TU I EN CE a N M' EG VF mR z O ,*
OA CS OO T i S
I YR C
S a I SCA N A g
1 T T T r M ME rR A
DF nN OA D V
R N
A L FG Y S FD N
E o M OAW L FSICA U
T A
F O Pg NDy Om f F OPNUIO OIS NCST C! -
O I ES?TN C aC FLH OUC ONg tAy NE O
TE NTT OAA OAL -
C S iY FLuT EGR s g SN SILC IUI t
SRA n N O FN aA At
,l CEA tR V
s O g, tI VG VG P L MON TA o Y F0s AvRE PE T' R SE FRE DTC D' E N f D5P RA DC A
i R OT OAR A s EA O E O I CG S E
L E
R R s 5a T C i T FT FS OE C U N m A V F
OO E
J m u m u m L
U S
R W ER CP m u
m i l
u m m C o
E FIL G N N FA E A 0 OIC y M
E P r R R 1
S S DN o -
R I S tGFO S t -
A AtA AH TT S N a A H MI M TzGNtI S
T S T N
O I l E C M SAiR, t
C T C ET TY CCS u L FV T FEDF FVN S g O l OCUg OJNJ OF E rFN gT F F C C E ASr V
NRAOO R NlM M OGN e
U CSSC ME OPVP OAS ,RmD gP F E R
t I NS FEWND I
SR VL S OE $, N YA R
N F OD SA tOIA vV T I MTISS ,OEP yI r
. NT A OCmC DA DTDE a A E RA ANR S.
E E
P S P E P
9 DA P
l e
O U A R
OS R c
- T N u C
EO N
. RT e
. DA R I 6
1 V
N E
C .
. EE i
G N S DO OS
- ?VP TO T NL ID T
A S.
UR C FAO CR F MYN U
- CO E
R O E.
l O N GH EA PU ORTIO Ot
. E V
F C N R OREt NE ORUA OAO ET
,P SFT sF RA FOU S P L 1 EPA C
IR L gO R
tESY V' N OS T
tvEDA OENV f
. F A U T C OtGE C D SAE .
F E O
G E A NT ES AN
- A N
E g
T IP E Y C i
. C R R WEN S f
t 8L F VV N a TE
. u m pl A t
TS Mp AG tA 1TN SA m
m e
m m
u m "
fi MM vA l US NF F DE Y =A f
N AC R W vC O .
A OIL F TI vU tO f
ecS E
H GC OA ES S DN Ct YyU g EDG F U T
C ED CE D NO ST 5AS' A N tNN OP FIR FAT CtL N E D CU FD L R OSA DA N 0 WNi TO CN rO F
ON AA L EACv I
0 t F S ON A RIGU F E DT RR UCRT FAEF wW E N
0LFY S F S t
A mA O O F T OA MAO S F VFA G 9V8 C SA UP OtaVs A
Z I EV$
C
'v GS NCO N MEGWLA O N IRD twFN A M A DWN nDC t
F A *AN OEA FAR i SED G, M PAR O ST p IMD R
L Tv MYNA D
LT C NE F a A
S - m M s e
m w
fII;Iii'
h DIVISION OF ENGINEERING INCLUDES FIVE BRANCHES:
I I I I STRUCTURAL ENGINEERING WASTE ELECTRICAL b MATERIALS ENGINEERING ENG EE ING
' ^ ENGIN E N f
Fig. 2 Division of Engineering of the Office of Nuclear Regulatory Research -
organizational chart I
4
inspection, testing, and operation of nuclear power plants, nuclear reactors, and fuel' cycle facilities with emphasis on the materials and chemical engineer-ing aspects of the primary system coolant boundary. Specifically, this branch has the responsibility for materials and chemical engineering research and the development of standards pertaining to inservice inspections to determine struc-tural integrity, corrosion, fracture mechanics, thermal shock, effects of envi-ronment on materiais, hydrogen control, water chemistry, and decommissioning programs as well as the nondestructive examination program, which includes the qualifications of inspection personnel, procedures, and equipment. Generally, it provides assistance to other branches within and outside the Division of Engineering related to their materiais-related needs.
(2) Structural and Seismic Engineering Branch This branch develops, recommends, plans, evaluates, and manages research and standards development programs pertaining to the design, qualification, construc-tion, inspection, testing, maintenance, and operation of nuclear power plants,
-nuclear reactors, and fuel cycle facilities with emphasis on the structural and seismic engineering aspects of structures and components. Specifically, this branch has the responsibility for these engineering aspects including the effects of general and site-specific natural phenomena (e.g., seismic effects and effects of tornado missiles), load combinations and associated design limits, vibration, competence of soil as a support material, and soil-structure interaction. This branch has the lead responsibility for the coordination of activities associated with the American Society of Mechanical Engineers Boiler and Pressure Vessel Code (ASME Code), Sections III and XI [1].
(3) Engineering Issues Branch This branch provides full-time dedicated task management of generic engineering safety issues and unresolved engineering issues with a focus on engineering and scientific issues associated with design and operations. It coordinates, directs, and reviews contractor and staff efforts, technical findings, and regulatory analysis. It is responsible for developing draft and final resolutions for each issue based on a combination of technical, risk, cost, and value-impact analyses. It develops technical bases, regulatory requirements, and proposed policy guidance related to reactor aging.
5
(4) Waste Management Branch This branch develops, recommends, plans, evaluates, and manages research programs and develops rules and standards pertaining to the design, qualification, con-struction, inspection, testing, operation, and closure of radioactive waste disposal facilities and the overall performance of such facilities, with emphasis on the characterization of phenomena, methodologies, and performance related to the management of radioactive waste. It evaluates factors and phenomena affecting the public health resulting from routine and accidental releases caused by nuclear fuel cycle facility operation and waste facility system performance. These factors and phenomena include external factors such as geological, hydrological and meteorological conditions that affect facility safety; institutional and physical factors that affect routine operations and the consequences of accidents; and the operating, engineering, and system per-formance factors that affect waste isolation and the containment.
(5) Electrical and Mechanical Engineering Branch This branch develops, recommends, plans, evaiuates, and manages research and standards development programs pertaining to the design, qualification, con-struction, inspection, testing, maintenance, and operation of nuclear power plants, nuclear reactors, and fuel cycle facilities with emphasis on the mechan-ical, electrical, and fire protection engineering aspects including the quali-fication and survivability of components, effects of aging and inservice degra-dation (including the effects of environmental stresses and wear on components),
classification of components, inservice testing to determine functional adequacy of components, and structural adequacy of piping systems and spent fuel casks.
Research conducted by the Structural and Seismic Engineering Branch is reported in this paper.Research conducted by the other branches is not included.
Table 1 provides data on the research conducted to determine design loads and response design spectra for systems and components. Table 2 provides data on the research conducted on methods of analysis and structural behavior. Research ;
conducted on material properties and material engineering was reported for the I previous conferences in papers similar to this one but is not included in this report because it is extensive and independent and is conducted by the Materials Engineering Branch and other branches. It is also reported in various divisions of the SMiRT 9 conference.
6
\
l l
l Essentially the same order of presentation for NRC structural and mechanical engineering research was used at the SMiRT 4, 5, 6, 7, and 8 conferences reported in refs. [2-6]. The project titles and the objectives, and in the case of table 2, references to NRC publications pertinent to the project, are listed in the first column of tables 1-2. The abbreviated names of contractors I are given in the second column, and their full names are provided in the l Appe'ndix. '
A summary of funding for the research projects listed in tables 1-2 is provided in table 3. The comparative amount of funding for previous years also is given. ,
l Table 3 shows that, after a substantial increase in funding from about 6 million dollars in 1980 to about 10 million dollars in 1981, the level of funding for the next 3 years remained essentially stable and increased to about 12 million in 1985. In the following 2 years the funding increased by 1 million dollars each year and reached about 14 million dollars in 1987.
NRC research results are reported regularly in NUREG/CR reports that may be obtained from the NRC/GPO Sales Program, U.S. Nuclear Regulatory Commission, Washington, DC 20555, USA, for a nominal price.
7
l i
1 Table 1 Research on determination of design loads and response j design spectra for systems and components l Project titles and objectives Contractors e Seismic load design and spectra at the ground level Perry earthquake. LLNL An earthquake shook the Perry Nuclear Power Plant on January 31, 1986. The response records showed that some high-frequency spectral valves had exceeded the operating basic earthquake and safe shutdown earthquake Regulatory Guide (RG) 1.60 [7]
spectra. The major concern was whether or not functional damage could have resulted from this low-energy event.
The objective of this project is to provide assistance in addressing seismological and seismic design issues that will result from the January 31, 1986 seismic event at the Perry Nuclear Power Plant and to address generic technical issues that may arise from this event.
- Response design spectra for systems and components Nonlinear piping response predictions. HEDL Generally, nonlinear dynamic response results in lower inertia loads than are predicted by the traditional linear analysis methods used in design.
Load redistribution at very high load levels may invalidate failure locations predicted by linear methods. Sophisticated methods are available, but these are complicated and expensive to use.
l The objective of this project is to evaluate simpler i methods for predicting nonlinear results using test j
data as benchmarks.
1 Margins in piping response analysis. LLNL A new analysis technique will be compared against baseline time history analyses (presumably "best- I estimate" analyses). The baseline time history analysis in both the LLNL study investigating f Pressure Vessel Research Council damping and peak ,
broadening techniques and the BNL study investigat- l j
ing the independent support motion method assumed 1
9 8
Table 1 (continued)
Project titles and objectives Contractors that RG 1.61 [8] damping was best estimate. Through the pipe damping study at INEL, it is now known that these damping values are conservative.
The objective of this project is to reevaluate the response margins in piping spectrum analysis methods by incorporating new damping data into the baseline time history analyses, identify and quantify response margins and uncertainties in time history piping analyses, and provide a basis for piping damping criteria to be used in time history analyses.
Combinational procedures for piping response BNL spectra analysis.
Currently accepted enveloped-spectra methods, as described in the Standard Review Plan [9], introduce conservatism to the design process that necessitate the use of more snubbers and restraints than would be needed if more accurate procedures were used. The overdesign of piping systems for dynamic events (as opposed to normal operating conditions) was a major concern of the NRC Piping Review Committee.
The objective of this project is to provide the NRC staff with the information necessary to evaluate licensee submittals using less conservative (but generally more accurate) spectral analysis techniques for the dynamic load design of piping ;
systems. J l
i 1
l i
l I
4 I
9
l.
l Table 2 Research on methods of analysis and structural behavior
. Project titles, objectives, and.NRC Contractors publications.
- Testing of physical models Containment integrity under extreme load. SNL Severe accidents exceeding the original design basis that lead to core melt have been postulated.
The containment building provides the last barrier to the release of radioactivity resulting from such an accident. There are large uncer-tainties in predicting the leak integrity of containment buildings under severe accident pressures and temperatures.
The objective of this project is_to provide a basis for the reliable estimation of containment perform-ance during severe accidents. Pretest predictions for the reinforced concrete model test were made in FY 1986 by the Electric Power Research Institute and foreign entities. The evaluation of predictions and comparison with test results as well as the planning of experiments necessary to resolve questions about the possibility.that an earthquake, large enough to initiate a severe accident, might degrade containment performance will be completed in FY 87. NUREG/CP-0076, Aug. 86; -4141, Sept. 85; Paper JK/3 at this session.
Experiments on containment models. ANL Three 1/32nd scale steel containment models with l varying features were pressure tested in FY 1983-84.
The experiment on a 1/8th scale steel containment model with scale equipment hatch, airlock sleeves, and piping penetrations was completed in November 1984. A contract for constructing a 1/6th scale reinforced concrete containment model was awarded in January 1985. Model construction was completed in June 1986. Instrumentation was completed in May 1987. Testing will be performed in June-July 1987. A cooperative effort (with the Central Electricity Generating Board) on a prestressed containment model is planned for 1988.
NUREG/CR-4216, Dec. 86;-3902, Feb. 87;-4913, July 87.
The objective of this project is to implement tests on scale containment models to upport the 10
J
)
Table 2 (continued)
Project titles, objectives, and NRC Contractors publication.s development and validation of methods for assessing the capabilities of containment buildings under conditions exceeding their design basis to permit licensing staff evaluations of licensee estimates of containment performance.
Validation, Heissdampfreaktor and Taiwan Reactor. ANL The objective of this project is to provide, through domestic and international cooperative efforts, information and experimental data that can be used to validate and improve predictions of the behavior of nuclear power plants (including soil-structure interaction and the nonlinear behavior of buildings and the piping systems) subjected to earthquakes larger than design basis. The predictive methods to be validated are used both in probabilistic and deterministic calculations and in particular may be used as part of the seismic probabilistic risk assessments for nuclear power plants. Papers JK/4 and JK/5 at-this session.
Seismic Category I structures. LANL The shear walls, as part of Category I buildings, were tested. The experimental program plan was developed with the foreknowledge that scale model testing of reinforced concrete structures is a somewhat controversial issue, particularly when the structures are loaded into the inelastic range.
The models of the shear walls were subjected to static and dynamic tests. Monotonic and cyclic quasi-static tests provide information on strength, stiffness, stiffness degradation, ductility, and general load-deflection behavior up to the ultimate load. The dynamic tests yield information on fundamental frequency of the building; accelerations, displacements, and floor response spectra at various elevations in the building; and fundamental frequency, damping, and floor' response spectra changes as a function of increased earthquake motion.
The objective of this project is to provide analytical
, and experimental data that are used to predict how buildings will respond to earthquakes larger than the design basis and to modify and develop criteria used 11
Table 2 (Continued)
Project titles, objectives, and NRC Contractors publications in the design of buildings and the safety-related piping and equipment located within them.
NUREG/CR-4274, Sept. 85; -4474, Jan. 86; Paper JK/12 at this session.
Piping and fitting dynamic reliability, EPRI/NRC ETEC, ANCO, and research. GEC Inertial loadings from dynamic events such as earthquakes vary with time and have limited durations and energy content. Cross-sectional plastic collapse is assumed by the ASME Code [1]
to be the dominant failure mode for these loads.
However, an increasing amount of analytic and test data has shown that piping inertial loads behave differently than do static loads. The margins-to-failure are greater than are predicted by current theoretical limit-load estimations, and ratcheting and fatigue appear to be the principal failure modes. If these new insights into dynamic piping failure can be demonstrated systematically and conclusively, then significant changes can be made with regard to how the ASME Code sets limits on inertial stresses. This would dramatically change the nature of piping system design and could in turn reduce the number of snubbers used in nuclear power plants.
The objective of this project is to identify failure mechanisms and failure levels of piping components and systems under dynamic loadings; to provide a database that will improve our predic-tion of piping system response and failure due to high-level dynamic loads; and to develop an improved, realistic, and defensible set of piping design rules for inclusion into the ASME Code.
NUREG/CR-4859, Feb. 87; Paper JK/7 at this session.
- Design and analysis methods for mechanical systems and components Seismic design margins program. LLNL This project has the same major considerations as those of the following project, " Trial seismic margins PWR plant review"; that is, new seismological information and other considera-tions require this study.
12
i i
Table 2 (Continued)
Project titles, objectives, and NRC Contractors publications c
The objective of this project is to provide a practical, engineering-oriented procedure for assessing the ability of nuclear power plants to withstand earthquakes above their design-basis level and to provide a means for effectively and efficiently identifying plant vulnerabilities leading to potential seismically induced core melt damage. NUREG/CR-4334, Aug. 85; -4482, Mar. 86; Paper JK/2 at this session.
Trial seismic margins PWR plant review. LLNL New seismological information coupled with uncertainties in defining the seismic hazard sometimes results in the need to consider increasing the level of earthquake magnitude or intensity used in the original design of a nuclear power plant. The performance of conventional power plants in past earthquake and recent risk studies such as the Seismic Safety Margins Research Program (SSMRP) indicates that nuclear plants generally have high margins against seismically induced failure. A sound, practical seismic margins program using margins-to-failure analysis and seismic probabilistic risk assessment techniques will serve to minimize the need for licensing actions as estimates of the seismic hazard and system response change. In addition, seismic margins studies can provide a basis for identifying key elements (components and systems) that make the plant vulnerable to seismic events.
The objective of this project is to assess an actual plant's capability to withstand a specific earthquake level greater than the safe shutdown earthquake, to demonstrate the use of the expert panel approach, to provide a basis for upgrading the expert panel approach and guidelines if improvement is shown to be needed, and to provide a benchmark for possible future seismic margins reviews. Costs and times for each step will be documented along with the technical data.
NUREG/CR-4826, Vols.1, 2, and 3, Mar. 87; Paper JK/2 at this session.
Extended high confidence of low probability failure- Future Resources based seismic margins review. Association, Inc.
13
1 1
Table 2 (Continued)
Project titles, objectives, and NRC Contractors publications Full-scope seismic probabilistic risk assessments i- are complex and costly. The seismic margins approach is a much more efficient way to identify plant strengths and vulnerabilities, but this method (purposely) does not yield information '
in terms of risk.
The objective of this project is to enhance the
'information obtained from a seismic margins review. NUREG/CR-4334, Aug. 85; -4482, Mar. 86; Report by Future Resources Association, Inc.,
R. J. Budnitz, Berkeley, Calif. Mar. 87.
ASME Section III, technical assistance program. ORNL Section III of the ASME Code [1] provides the design criteria for the piping, nozzles, and pressure vessels used in nuclear power plants.
As our knowledge of component loadings, response behavior, and failure mechanisms increases, there is a continuing need to revise these criteria.
The objective of this project is to evaluate propo:;ed revisions and/or additions to the ASME Code, which provides rules for the design and construction of nuclear power plant components, to ensure that the quality of new components is adequate.
l Fragility data acquisition and evaluation. BNL l The objective of this project is to establish an ;
experimental database on component fragility through !
domestic and foreign cooperative efforts. This database will provide the licensing staff with a basis for assessing the earthquake levels at which individual components and generic classes of components fail to perform their safety functions.
NUREG/CP-0070, Aug. 85; NUREG/CR-4659, June 86; Paper JK/9 at this session.
Fragility prioritization and testing. LLNL The objective of this project is to identify those components critical to public health and safety that are susceptible to seismic damage and to provide an 14
Table'2 (Continued)
Project titles, objectives, and NRC Contractors publications experimental database with which to realistically assess the earthquake levels at which individual components and generic classes of components fail to perform their safety functions. NUREG/CR-4899, June 87; -4900, Vols.1 and 2, June 87; Paper JK/6 at this session.
Damping studies. INEL Damping information from both in situ and proto-typical tests has been gathered and evaluated.
New tests have been_ designed and completed to evaluate.how various design parameters influence damping.value parameters including support type and-spacing, insulation, load level, and input frequency. High-level and~high- frequency tests have been performed. A world database for damping is maintained and new statistical evaluations are made.
The_ objective of this project is to provide information with which to evaluate licensee requests to increase the level of damping for plant piping systems and thereby reduce the number of snubbers without reducing the level of plant safety.
NUREG/CR-4529, June 86; -4562, June 86; Paper JK/14 at this session.
Design criteria for shipping containers. LLNL Shipping containers are used to store and transport spent radioactive fuel. There is concern over the ability'of these containers to properly protect the public health and safety under postulated storage and transportation conditions.
The objective of this project is to develop licensing criteria for the design and fabrication of spent' fuel shipping containers and to develop simplified thermal stress analysis procedures. NUREG/CR-4363, Feb. 86;
-3760, Jan. 86.
- Design and analysis methods for civil engineering structures Structural margins-to-failure - containment buckling. LANL 15
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ ..1
l l
Table 2 (Continued) l Project titles, objectives, and NRC Contractors publications Uncertainty exists in the design and analysis of steel containments subject to seismic load with respect to failure that occurs because of buckling. Because of the complex geometry of.the containment and the large number of opening and reinforcement methods, analysis can be verified only by experiments. A number of experiments pertaining to the following have been conducted which were included in the first phase of this project; buckling of steel cylinders with circular cutouts reinforced in accordance with ASME rules, buckling of ring-stiffened cylindrical shells under asymmetrical axial loads, buckling of ring-stiffened cylindrical shells with reinforced openings under asymmetrical axial loads, and buckling of steel containment shells under time-dependent loading.
In the second phase of this project (the period between the SMiRT 8 and 9 conferences) the experiments resulted in the following research and subsequent publications: A Study of the effects of penetration i framing on steel containment buckling capacity, an assessment of loss of-containment potential because of knuckle buckling for 4:1 steel containment heads," l and investigation of steel containment buckling l from dynamic loads. NUREG/CR-4829; Mar. 87; -4889, Mar. 87;-4904, May 87.
Integrity of containment penetrations under severe SNL accident loads. i The objective of this project is to develop an experimental database for assessing the leak integrity of containment penetrations under severe accident conditions and for validating existing methods used to assess these penetrations, in order to provide a basis for regulatory decisions regarding continued !
operation of existing facilities, and for identifying containment penetration features, which if improved, would significantly increase containment capacity.
Standard problems for structural computer codes. BNL There is a need to review and to determine the ranges of validity of the analytical methods used to predict the behavior of nuclear safety-related structures under accidental and extreme environmental loadings including soil-structure interaction effects associated with seismic loads.
16
1 Table 2 (Continued)
Project titles, objectives, and NRC Contractors publications The objective of this project is to identify problems with experimentally known solutions (benchmarks) to be used by the licensing staff to validate licensee methods used to calculate the transmittance of earthquake loads through the soil to safety related buildings, systems, and components. NUREG/CR-4588, Apr. 86; -0054, Dec. 86.
I l
i 17
Table 3 Cost of research in thousand dollars Organizational Fiscal years units 1980 1981 1982 1983 1984 1985 1986 1987 Structural and 6,200 9,800 10,010 11,810 9,900 12,500 Mechanical Engineering Branch Structural and 13,240 14,350*
Seismic Engineering Branch
- The Structural and Seismic Engineering Branch took on the responsibility of structural research when NRC was reorganized. Mechanical engineering research conducted by the Electrical and Mechanical Engineering Branch will be funded under an independent budget.
l 18
At the-SMiRT 9 conference, NRC contractors are presenting papers in various divisions. Only selected papers are included in the JK Panel Session. The papers selected for'the JK Panel Session are those that provide the most significant results on major NRC research projects and those from former con-tractors closely related to the NRC research activities. At this JK Panel Session papers dealing primarily with seismic structural problems are included.
Appendix Abbreviations and fulI names of contractors J ANC0 Anco Engineering ANL Argonne National Laboratory BNL Brookhaven National Laboratory -
EPRI Electric Power Research Institute ETEC Energy Technical Engineering Corporation ,
GE General Electric Company HEDL Handford Engineering Division Laboratory INEL Idaho National Engineering Laboratory LANL Los Alamos National Laboratory LLNL' Lawrence.Livermore National Laboratory '
ORNL Dak Ridge National Laboratory SNL Sandia National Laboratory Acknowledgment.
NRC staff members, J. F. Costello and D. J. Guzy, participated in the preparation of this paper. J. A. O'Brien reviewed this paper and provided valuable assistance. The authors greatly appreciate their cooperation.
l 19
References
[1] ASME Boiler and Pressure Vessel Code.Section III and Section XI. q American Society of Mechanical Engineers, New York, N.Y.
[2] B. S. Browzin, Nucl. Engrg. Des. 50 (1978) 23-32.
[3] 8. S. Browzin and L. C. Shao, Nucl. Engrg. Des. 59 (1980) 3-13.
[4] B. S. Browzin, Nucl. Engrg. Des. 69 (1982) 143-148.
[5] B .S. Browzin and W. F. Anderson, Nucl. Engrg. Des. 79 (1984) 119-124.
[6] B. S. Browzin and J. E. Richardson, Nucl. Engrg. Des. 94 (1986) 3-8.
[7] Design Response Spectra for Seismic Design of Nuclear Power Plants, Regulatory Guide 1.60, U.S. Nuclear Regulatory Commission.
[8] Damping Values for Seismic Design of Nuclear Power Plants, Regulatory Guide 1.61, U.S. Nuclear Regulatory Commission.
[9] Standard Review Plan for the Review of Safety Analysis Reports for Nuclear Power Plants, NUREG-0800, U.S. Nuclear Regulatory Commission (July 1981).
20 l l
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ J
l 4.
JKl2 SEISMIC. MARGINS REVIEW OF NUCLEAR POWER PLANTS - NRC PROGRAM OVERVIEW *'
- R.' C. Murray, L P. G. Prassinos ,' M. K. Ravindra, D. L. Moore *
- l The U.S. Nuclear Regulatory Commission is sponsoring a research' program to develop and demonstrate a method for assessing snfety margins at nuclear power plants with respect to seismic events. This program is called the Seismic Design Margins Program and was started.in 1984 at the Lawrence Livermore National Laboratory. The need for this research results from the changing perception of: the seismic hazard in certain localities which could require reassessment of the seismic design adequacy at some power plant sites. It is generally accepted- that nuclear power plants are capable of withstanding earthquake motion substantially greater than their design basis, but' methods are needed to systematically demonstrate this. An " Expert Panel on' the Quantification of Seismic Margins" was formed to develop such a margins assessment method, and the method was; applied to the Maine Yankee Atomic Power Station, a Combustion Engineering, three-loop pressurized water reactor located near.Wiscasset, Maine. This paper highlights the methodology and emphasizes its application to Maine Yankee.
1.- INTRODUCTION In 1984, the U.S. Nuclear Regulatory Commission'(NRC) initiated the Seismic Design Margins Program to address regulatory needs and a changing perception of 'the seismic hazard. The NRC formed the " Expert Panel on the Quantification of Seismic Margins" and charged it to work closely with an NRC
- This work was supported by the U.S. Nuclear Regulatory Commission under a Memorandum of Understanding with the U.S. Department of Energy.
- *R. C. Murray, Lawrence Livermore National Laboratory, P.O. Box 808, L-197, Livermore, CA 94550; P. G. Prassinos, Lawrence Livermore National Laboratory, P.O. Box 808, L-196, Livermore, CA 94550; M. K. Ravindra, EQE, Incorporated, 3300 Irvine Avenue, Suite 345, Newport Beach, CA 92660; D. L. Moore, Energy, Incorporated,1851 South Central Place, Suite 201, Kent, WA 98031.
21
staff Working Group on Seismic Design Margins to provide technical guidance on the assessment of seismic margins. The overall goal of the program is the development of a methodology and guidelines that can be readily used by the NRC and industry for assessing the inherent quantitative seismic capacity of nuclear power plants [1].
l The development of a soundly based, efficient and effective method for the assessment of how much margin actually exists in important components, systems, and the plant will serve to minimize the impact of changing regulatory requirements and licensing actions as the estimates of seismic hazards change. In addition, a seismic margins assessment can provide a basis f- cor.- Jence in the capacity of nuclear power plants.
The most impor6 nt regulatory need and the focus of the seismic margins effort is stated as follows:
"There is a need to understand how much seismic margin exists at nuclear power plants. This seismic margin is to be expressed in terms of how much larger must an earthquake be above the safe shutdown earthquake before it compromises the safety of the plant."
The Expert Panel and its technical support personnel studied the available in."ormation on the quantification of seismic capacity of nuclear power plants and other industrial facilities. The results of several seismic probabilistic risk assessments of nuclear power plants were reviewed along with the behavior of industrial facilities during earthquakes. These studies were used to develop a margin review approach that involves both the screening of components based on their importance in preventing seismic core melt and their inherent seismic capacity.
l The seismic margins review approach, developed by the Expert Panel, has been documented in the report, "An Approach to the Quantification of Seismic Margins in Nuclear Power Plants" [2]. This document formed the basis for the development of guidelines for performing seismic margin reviews. These 22
guidelines are given in " Recommendation to the Nuclear Regulatory Commission on Trial Guidelines for Seismic Margin Reviews of Nuclear Power Plants, Draft Report for Comment" [3].
The performance of a trial plant review is needed to verify and test the review methodology and guidelines. For the trial review, the NRC negotiated with and selected the Maine Yankee plant. Once the trial plant was selected, two analysis teams (Systems Team and Fragility Team) were chosen. These teams were selected based on their technical approach and team composition. In addition, a Peer Review Group was selected and charters were developed for both this group and the Expert Panel. The seismic margin review was then organized allowing for participation by the plant owner, the NRC Working Group on Seismic Design Margins, and the appropriate NRC program managers. Results of the trial plant review on the Maine Yankee Atomic Power Station are I documented in three volumes [4, 5, 6]. Volume 1 gives a summary of the results, Volume 2 provides the the details of the systems analysis, and Volume 3 gives the component fragility screening and analysis results.
)
The overall objectives of the trial review were to assess the seismic margins of a pressurized water reactor, to test the adequacy of the seismic margin review approach and quantification techniques, and the guidelines for performing these reviews. There are four related objectives of this study:
o To demonstrate the use of the Expert Panel's approach (NUREG/CR-4334) ;
and guidelines (NUREG/CR-4482) for seismic margin reviews.
o To provide a basis for revising and upgrading the approach and gui delines .
o To provide a benchmark for possible future seismic margin reviews, including an understanding of the level of effort in performing a l seismic margin review. l 1
o To provide an assessment of the plant's capability to withstand a specific earthquake level greater than the SSE.
l I
1 23
The role of the Expert Panel during the Maine Yankee study was to review the procedures being used at an early stage in the process and to assure that the methods and techniques being employed are consistent with the Panel's guidance and are relevant for performing a seismic margins review. When the trial plant review is completed, the Panel will study the results, examine how the review was implemented, and evaluate the overall effort.
The Peer Review Group was chartered to review the technical adequacy of the study including participation in the plant walkdowns. The objective of the Peer Review Group is to assure that the seismic margin review is executed in a fully competent and professional manner, uses appropriate methods, and follows the guidance established, q
i
- 2. SEISHIC MARGINS APPROACH i Insights gained from the results of seven published probabilistic risk assessments (PRAs) were use in the development of a screening approach that .
combined systems insights and fragility information to simplify the margin review process. This approach is directed at reviewing a specific plant at a selected carthquake acceleration level greater than the SSE.
A general definition of " seismic margin" is stated below:
" Seismic margin is expressed in terms of the earthquake motion level that compromises plant safety, specifically leading to melting of the reactor core. In this context, margin needs to be defined for l the whole plant."
The adopted measure of margin is the earthquake level for which there is a High Confidence, Low Probability of Failure (HCLPF). The HCLPF is a conservative representation of capacity and corresponds to the earthquake level at which it is extremely unlikely that core damage will occur. From the mathematical perspective, the HCLPF capacity is approximately equal to a 95 percent confidence of not exceeding about a 5 percent probability of 24
l failure. HCLPF capacities for specific types of components are derived from a combination of test data or data from real carthquake experience, and engineering analysis.
2.1 Systems Screening The review of the available seismic PRAs indicated som6 key trends and insights useful in developing seismic margin review criteria. Chapter 4 of
[2] discusses the details of the screening approach. These trends and insights were obtained for pressurized water reactors (PWRs), for which there were seven available PRAs.
The Expert Panel review indicated that systems screening must be performed at the functional level due to the diversity of plant designs and system configurations. This led to considering the general plant safety functions normally performed in PWRs:
- 1. Reactor Subcriticality.
- 2. Normal Cooldown.
3 Early Emergency Core Cooling (injection). *
- 4. Late Emergency Core Cooling (recirculation).
- 5. Containment Heat Removal.
- 6. Early Containment Overpressure Protection (inj ection).
- 7. Late Containment Overpressure Protection (recirculation).
Examination of PRA results indicated that the dominant plant damage states to core melt for seismic events are generally early core melt with early containment f ailure. In addition, these plant damage states generally l 1
involve core melt induced by failure of the first three functions listed above followed by loss of containment integrity early in the accident progression.
This insight led to the division of plant safety function into two groups:
Group A - Functions 1-3 Group B - Functions 4-7.
I i
i 25
The systems screening approach is based on this insight. The dominant plant damage states are caused by failure of the Group A functions and these l plant damage states are also characterized by f ailure of the Group B f unctions. Therefore, we consider core damage to occur whenever there is a f ailure of the systems that provide the initial shutdown of the nuclear reaction and cooling of the reactor core. These f ailures are followed by failure of the systems that provide the Group B functions so as to preclude the mitigation of consequences by providing containment protection and cooling. The normal cooldown function is not considered in our screening criteria because a loss of offsite power is assumed to occur in all seismic PRAs as a result of " Icv capacity" switchyard components.
The relationship between the Group A and Group B functions indicated that they are highly coupled so that the combined failure of Group A and success of Group B (or the combined success of Group A and failure of Group B) is virtually precluded since their weakest links are coupled. This insight indicates that only those. systems and components needed to perform the Group A function must be considered in a seismic margins review.
The Expert Panel also discovered that seismic core melt can occur due to the f ailure of specific unique features. This implies that it will always be necessary to perform some kind of plant walkdown to look for any unique features requiring a margin review.
The insights gained from our systems review of the PRAs and the development of a systems screening criteria to simplify the margin review process has resulted in the following conclusions:
- 1. It is possible only to come to conclusions regarding the relative importance of plant systems and safety function for PWRs for which six plants were studied, j
- 2. For PWRs, it is possible to categorize plant safety functions as belonging to one of two groups, one which is important to the assessment of seismic margins and one which is not.
26 1
i 1
l I
3 The important group involves only two plant functions that must be considered for estimating seismic margin. These two functions are shutting down the nuclear chain reaction and providing cooling to the I
reactor core in the time period immediately following the seismic event, t
- 4. It is possible to reasonably estimate the seismic margin of the plant by performing a margins study only involving the analysis of the plant systems and structures which are required in order to perform those two safety functions.
Using the systems screening criteria and combining it with fragility insights, we can establish guidelines for margin reviews. These guidelines show that it is possible to perform a reasonable seismic margin review by concentrating on those f unctions (and associated systems) which are required in the early part of the seismic event and eliminating from the review those functions which are not required until later. Further, depending on the level of earthquake for which it is desirable to define a margin, certain initiating events would not have to be considered (e.g., large LOCAs). This reduces the level of effort and scope of the analysis.
Event trees need only be constructed up to the point of determining whether or not there is early core damage. Fault trees would only have to be constructed for those front-line systems (and their support systems) that appear on these abbreviated event trees. By combining a nonseismic f ailure probability and seismic fragility screening criteria with these systems models, it would only be necessary to include those components which have not been removed during the screening process.
2.2 Fragility Screening l
The available fragility information that was reviewed and assessed is based primarily on the detailed analysis of nuclear power plants performed for PRAs. This PRA information was supplemented by recent systematic investi-gations of historic earthquakes. The information includes past earthquake performance data for eight classes of equipment obtained by the Seismic 27
Qualification Utility Group (SQUG), and reviewed by the Senior Seismic Review and Advisory Pane] . Work is ongoing to document the historic earthquake performance and qualification data for additional components such as piping, valve operators, penetrations, diesel generators, battery racks, and electrical equipmer.t. This work is being conducted by SQUG, the Electric Power Research Institute, and the American Society of Civil Engineers Dynamic Analysis Committee [7]. Additionally, the collective knowledge of the Expert Panel members, performance test data, and other analyses were included in the assessment.
These available sources of fragility information were used to arrive at conclusions about which components should be assessed from a seismic capacity standpoint. In making statements about the need for capacity assessments for each component, three ranges, stated in peak ground acceleration (pga), were used: (1) less than 0.38, (2) 0.38 to 0.5g, and (3) greater than 0.5g. Each type of nuclear power plant component was assessed to have a generic HCLPF capacity within one of these ranges. This resulted in an extensive table of components indicating at what earthquake level each component will require a margin review or be removed from the review process. This categorization of components is based on the available information, and the fragility screening resulting from the use of this table should only be performed with consideration of the caveats, limitations, anti assumptions presented in Chapter 5 of [2].
During the process of assessing the HCLPF capacities for the various nuclear power plant components, an extensive fragility information base was developed from the available seismic PRAs. This fragility information base is available on a diskette for use on personal computers, and is documented in
[8].
2.3 Performing Seismic Margin Reviews The combined insights gained on plant functions and component f ragilities were used to develop an outline of an approach for performing seismic margin i
t 28 j
revi ews . The review approach consists of eight steps developed in L6] and outlined in Section 2.4.
It is important to point out the assumptions and limitations of the approach.
- 1. The systems screening part of the approach presently applies only to PWRs .
- 2. The review approach focuses on earthquakes that could occur in the eastern part of the U.S., specifically east of the Rocky Mountains.
3 The assessment of component HCLPF capacities is limited to earth-quakes of less than a magnitude of about 6.5, which are characterized by 3 to 5 strong motion cycles with a total duration of 10 to 15 seconds.
- 4. The effects of undiscovered design and construction errors are not covered.
- 5. Possible vulnerabilities in hydraulic systems associated with sensors and pneumatic systems are not fully covered.
l
- 6. Electrical and control systems are incompletely covered because unrecoverable relay chatter and breaker trip is not adequately treated at this time.
7 Evaluation of the effect of wear and aging on equipment f unction is not fully covered.
- 8. Possible adverse human responses caused by earthquake-induced stress are not explicitly covered.
The first three limitations are based on the data from PRAs and industrial facilities that were used in the development of this approach and present true limitations on the methodology. Some of the remaining items represent limitation on our knowledge of how to adequately address these issues, while others require considerable effort to include.
2.4 Guidelines for Seismic Margin Reviews The objective of the seismic margin review guidelines is to provide guidance for determining whether a plant can resist with high confidence a 29
specified earthquake level greater than the SSE. fo accomplish this objective, analyses are performed on. components, systems, and the plant to determine the HCLPF capacity so that it can be compared to the specified earthquake level. Plant failure is defined as the onset of ccre damage.
A flow chart of the margin review process is shown in Figure 1. This process involves the screening of components based on their importance to plant safety and their seismic capacity. Inspection of Figure 1 indicates I
that Steps 2, 5, and 7 are primarily concerned with plant safety functions and systems, and are performed by a team of systems analysts. Steps 3, 6, and 8 ;
are mainly concerned with capacity assessment and are performed by a team of l fragility analysts. Step 4 is performed by both teams of analysts. The l entire process requires close cooperation and interaction between the two teams of analysts and the utility.
The initial step in the review process is the selection of the margin review earthquake level. The margin review earthquake level selected for Maine Yankee had a peak horizontal ground acceleration of 0 3g and a 50 percentile Newmark and Hall spectral shape.
In Step 2, plant information gathering, review, and analysis is performed to determine those plant systems and components that are important contributors to plant safety and thus allow focusing of the effort cn the components requiring a margin review. Also performed during Step 2 is an l identification of the relevant seismic initiating events and the development !
of preliminary event trees that oescribe the systematic behavior of the plant following these initiating events.
The team of systems analysts reviewed the plant information and determine those systems that perform the +wo f unctions important to plant safety, l reactor suboriticality and early emergency core cooling. Examples of these systems are the reactor scram, emergency boration, high pressure safety injection, and auxiliary feedwater systems. The support systems to these 30 l
)
Start Time axis ')
1 Select an earthquake review level 'l
.i Gather information .;j.*../, Gather information on I on sys* ems and sort *. the plant. Determine which I
. Group A functions. /* .,.. * *.
- broad classes or groups of I Use information ***2*.*. jntegcjon
[/((
[/ 3 / components have HCLPF l on Table 2.3 and **,.".*. values greater than the review -
/// /
Ref.1 C*.: '; * * * *. // // level. I'ossibly identify plant-unique features.
I i
First plant walkdown: -"
Concentrate on identlilcation of problems.
...4'*. *,
Emphasize systems interaction. Confirm .
applicability of screening tools. Complete
, . . Identification of plant-unique features.
v i KEY: Task is performed by:
- f .;. ; . " . Revision of systems I
.***.
- l* *,;****/ :,.. relationships established l M.:
Systems analyst ;l*
, ,.
- ll .5.,.***,..In Step 2.. Develop fault
.;*.".',,.**I.*.'
trees and svent trees.
Fragility analyst v
Both Second plant walkdown:
Primarily fragility analyst for checks.
6 Collect specific data (size and other physical characteristics) or components j requiring detailed analysis.
1 I
- ; *l Determine minimal * *.**.* *; *. *. ". ***.-
- Finalize HCLPF value for cut sets for end- I .* * .*.
- 7 * ,*,*
- 8 components in final cut sets point core melt. !.***.*,*;,,*.,.,.' ,
(components not screened out).
- : *. . *.". .' ; * * ., , /
v ,
i Margin assessment complete Figure 1 Graphic representation of the screening operations.
31
t I
l front-line systems are then determined. Examples of the support systems include electrical power, cooling, actuation, and control. j Once these systems have been determined, the components that make up these systems are listed. This list is shared with the fragility analysts and these components become-the focus of the first plant walkdown. Using the identified front-line systems, preliminary systemic event trees are developed for seismic initiating events.
For the Maine Yankee review, the initiating event considered was a seismic induced loss-of-offsite power (LOSP). This event assumed a loss-of-coolant accident (LOCA) to occur as a result of the earthquake and the LOSP.
Only a small LOCA was considered. A large LOCA was not considered because large RCS piping and their supports have generic capacities above the review earthquake. level and were screened out. The break size of the small LOCA was assumed to have an area equivalent to a 3/8-to-2-in.-diameter pipe and requires the operation of the high pressure injection system for mitigation.
The event trees developed in this step were preliminary. They were revised following the first plant walkdown and finalized after the second walkdown.
Concurrently, in Step 3, knowledge gathered about the plant and the inherent capacity of components is used to sort the components developed in Step 2 into two groups, those with a generic HCLPF capacity larger than the review earthquake level and those that have smaller HCLPF capacity.
For the Maine Yankee review, the culmination of Steps 2 and 3 resulted in the identification of structures, block walls, equipment, and areas of the plant that needed to be inspected and reviewed. The first walkdown planning included organizing the walkdown teams, developing procedures for the review of the various components, developing data sheets for recording the findings, and making arrangements with the plant for the necessary health physics counting, badging and training.
32
A first plant walkdown is performed in Step 4. This walkdown was performed to inspect the plant and confirm that the plant's configuration is such that the rules developed for conducting a margins assessment are applicable und that components can be screened out based on the generic evaluation. During this walkdown, any system interactions, system dependencies, and plant unique featurr.s were identified along with confirming the accuracy of the system descriptions and configurations.
During the first walkdown of the Maine Yankee plant, the analysis team members formed groups and inspected components and plant areas that were identified during the previous steps. The Peer Review Group and NRC personnel also formed groups to walk down the plant. The walkdowns were performed with various levels of detail depending or, the requirements of a particular group. Arrangements were made to have team meetings at the beginning and end of each day. Meetings were also arranged with knowledgeable plant personnel to discuss details about the plant.
Following the completion of Step 4, many of the components identified as belonging to the two important plant safety functions were screened out based on the inspection and their generic HCLPF capacities being larger than 0 3g.
In addition, plant information that was gathered during this review was used to revise the plant models and perform a conservative evaluation of the remaining component HCLPF capacities. Those components that were evaluated to have a HCLPF capacity larger than 0.3g were also screened out.
During Step 5, the information and understanding of the operation of the plant was used to review and revise the event trees developed in Step 3 Fault trees were developed for the front-line and support system that perform the two important saf ety f unctions.
For the loss-of-coolant event, it was possible to screen out the PORV failures and the pump seal LOCA because their components were estimated to have HCLPF capacities greater than the review earthquake level. Small pipe 33
ruptures, however, could not be screened out. This was due to radioactivity concerns, because components within the Maine Yankee containment structure were unaccessible for review during the walkdowns. In particular, instrument impulse lines that form part of the RCS pressure boundary could not be inspected or reviewed. Therefore, these lines could not be screened out and were assumed to be a source of a small LOCA. This prompted the development of two event trees. One that considered a small LOCA concurrent with the LOSP and the other that considered the LOSP with no LOCA.
At the completion of Step 5, the systems f ault trees were " pruned" by renoving those components that had been screened out in the previous steps.
Care was taken when pruning the f aul't trees that the paths from the remaining lower level components were lef t intact. These lower level canponents represented the possible failures for the systems under consideration.
A final plant walkdown is performed in Step 6. This walkdown is used to obtain additional specific information for determining the HCLPF capacities of the components that remain in the analysis. In addition, the systems models are verified for accuracy and any additional information needed to complete them is obtained. During this second walkdown of the Maine Yankee plant, the fragility analysts collected detailed information about those components for which a complete evaluation was necessary.
During Step 7, the systems models developed in Steps 3 and 5, and finalized in Step 6, are analyzed to determine the Boolean expressions for the seismic-induced core damage accident sequences. This step involves the analysis of the event trees to determine the accident sequences that lead to seismic core damage and the analysis of the f ault trees to determine the Boolean expression for each system failure.
A plant level Boolean expression can be derived by logically combining all the core damage accident sequence Boolean expressions. The initiating events for these plant level accident sequences are assumed mutually 34
exclusive. Therefore, the occurrence probability for each initiating event has to be determined befo"e the accident sequences can be combined.
This probability can be compared to the fraction of the time one initiating event occurs with respect to the occurrence of all the accident initiators. Since we consider a fraction for the occurrence of each accident sequence, we multiply accident sequences by a factor called a split fraction. The sum of all the split fractions is 1 since we assume that an initiating event always occurs.
The plant Boolean expression is then derived by multiplying each accident sequence by the appropriate split fraction and logically combining all the sequences. For the Maine Yankee review, two plant level accident sequences were eventually developed. One for a LOSP initiator concurrent with small LOCA and the other for a LOSP initiator with no LOCA.
The final step in the margin review process, Step 8, is to calculate the HCLPF capacities for the important low-capacity components, accident sequences, and the plant. The HCLPF capacities are finalized for those components that appear in the single, double, and some low-capacity triple member cut sets of the Boolean expression derived from the above systems analyses . These HCLPF calculations required detailed structural / mechanical analyses based on available information. The fragility curves for the components are then used to quantify the Boolean expressions for the accident sequences and the plant.
There are two methods available to calculate the HCLPF capacity of components: the conservative deterministic f ailure method (CDFM) and the fragility analysis method (FA) . For this trial review, the fragility method was used to calculate the HCLPF capacities for these components. This method was employed because the fragility analysis team has a detailed understanding of its application and use. In addition, the fragility method also allows the inclusion of nonseismic failures into the overall plant HCLPF and accident-sequence HCLPF calculations.
35
For the FA method, a component's fragility is represented by a simple-model using three parameters: median capacity A m , and logarithmic standard ~
deviations B and S grepresenting, respectively, randomness in the capacity R
and uncertainty in the mecian value. Using a double lognormal' model, fragility curves like the one shown in Figure 2 are developed. The.
median, BR "" OU .are estimated using design-analysis information, test data, ,
earthquake experience data, and engineering judgment. The HCLPF capacity is -
expressed using this fragility model as:
2 HCLPF = A , exp [-1.64'(BR *O U
- The unavailability- of the components (f ailure per demand)'is determined by combining its random f ailure probability with its unavailability. due- to normal test and maintenance. Equipment unavailability also included both .
planned and unplanned maintenance and repair. The unavailability. due to-random failures considers the time between normal test and maintenance, or between scheduled plant outages. A component's unavailability due to human error is considered as a separate event.
The component f ailures are combined following the rules of Boolean algebra and a discrete probability distribution numerical procedure. For.this analysis, the fragility curves are not truncated in either the lower or upper tail.
3 RESULTS T THE MAINE YANKEE SEISMIC MARGIN REVIEW An objective of this trial seismic margins review is to assess the.
capability of the Maine Yankee plant to withstand a specified earthquake level greater than the SSE. The results of this review consist of:
)
o Boolean expressions for each seismic-induced core-damage accident sequence. I J
l a
i 36
=____-_-__-___-_
10 95% confidence W curve m
\
.5
~
0.8 -
m Median fragility O
curve y 0.6 -
.=D j 0.5 - - - - - - - - - - - - - - - - - - - -
0.90g 2 I a 0.4 - '
~
m c ,
~
O l l
E 0.2 -
O o HCLPF 5% confidence I
curve
.s
-- 7-f , ,
l , ,
0 0.170.30.4 0.8 1.2 1.6 2.0 Peak ground acceleration (g)
Figure 2 Example fragility curves for a structure.
37
l l
1 1
o The dominant component controlling plant seismic safety.
o An assessment of the HCLPF capacities for important components, accident sequences, and the plant, o Insight into the seismic behavior of the plant systems required to fulfill the safety functions of suberiticality and early emergency core cooling injection.
The HCLPF capacities for the important components are logically combined as indicated by the Boolean expressions to estimate the ;ICLPF capacity for each core-damage accident sequence. Each of these accident-sequence HCLPF capacities represents a plant HCLPF capacity for the particular initiating event and plant systems response.
A plant level Boolean expression can be derived by logically combining the accident sequences after they have been multiplied by their respective split f raction. An overall plant HCLPF capacity can then be determined from the plant level Boolean expression.
3.1 overall Results The components with HCLPF less than 0.3g were used in the development of the event trees and fault trees for the seismic-induced core-damage accident sequences. The event trees and f ault trees were analyzed to determine Boolean expressions for each accident sequence that could lead to core damage. The component f ailures that are significant to these Booleans are given in Tables 1 and 2. Table 1 gives the seismic induced f ailures along with the fragility parameters used to quantify their HCLPF capacities. Table 2 gives the ronseismic f ailures and their unavailabilities. Note that the component items and nonseismic failure events are numbered consecutively; the missing numbers represent the items that were screened out in the final pruning of the event and f ault trees.
38
S 0
Table 1 Component seismic fragility parameters.
HCLPF Item No. Item A,(g) S " I (8 R U i 4 Transformers 0.84 0 30 0 32 0.30 !
(X507, X508) l 7 RWST 0.45 0.20 0.25 0. 21 (TK-4) 8 DWST 0.36 0.20 0.26 0.17 20 Circulating Water 0.69 0.24 0.27 0.30 Pumphouse 21* PWST 0.57 0.20 0.26 0.27 (TK-16)
HCLPF less than 0.3g, but does not appear in the plant Boolean expressions.
39
-Table 2 Probabilities for nonseismic f ailures.
Median Unavailability Error Item No. Description (per denand) Fa ct or*
10 Operator Failure to Close PCC . 8.0E-02 2 Isolation Valves 11 Random Failure of DG-1B 4.2E-02 5 12 Random Failure of DG-1 A 4.2E-02 5-13 Operator Fa!. lure to Place AFW .1.5E-01 2 Pump Train B in Service Locally l l
-14 . Nonseismic Common Cause 1.6E-03 5 )
. Failure of DGS l l
I 15 Nonseismic Common Cause 1.2E-04 5 Failure of AFW 16 Operator Failure to Refill DG 8.0E-03 3 Fuel Tanks by Opening Valve or Running P-33A.B ;
17 - Operator Failure to Place AFW 4.0E-02 3 Pump Train B in Service from MCR 22 Random Failure of the Turbine 3.0E-02 5 Driven Aux Feedwater Pump
- Error factor = 95% Confidence Value/ Median Value.
1 i
i 40 1.
L_______________ i
Component 4 has been upgraded following the review during the two plant walkdowns . This component is a General Electric (GE) station service transformer (4160 V to 400 V) that supplies power to pumps and components needed to perform the seismic safety functions.
i Another component appearing in Table 1, with a HCLPF less than 0.3g, is Number 20, the seismic failure of the primary water storage tank (PWST). This tank provides an alternate supply of water to the auxiliary feedwater system. The PWST has a HCLPF of 0.278 This tank does not appear in either core damage Boolean expressions.
There is a component, not listed in Tables 1 or 2, for which there was insufficient data to determine the HCLPF capacity. This component is the lead-antimony batteries used at the Maine Yankee plant. There is no seismic qualification or test data available on these batteries to estimate the capacity of their aged lead-antimony plates. Maine Yankee will replace all batteries during the next year with lead-calcium type.
Seismic qualification and test data on the new batteries indicates that they have a HCLPF capacity greater than the review earthquake level.
Subsequently, the station batteries were screened out and eliminated from further consideration.
1 The analysis of the two event trees resulted in two Boolean expressions that lead to seismic-induced core damage. One of these expressions is the logical combination of three accident sequences that were initiated by the seismic-induced LOSP concurrent with a small LOCA, The othtr expression is the logical combination of two accident sequences initiate i by the seismic-induced LOSP without a small LOCA. These two Boolean expressions are given below:
Small LOCA Core Damage
- (SL) [4 + 7 + 20].
41
rw~- ----v -.-,-:-- _- - - _ _ __, _
-1 1
j i
No LOCA Core Damage
. 1 (LOSP) [(4 + '20) * (8 + 13 + 15 + 17 + 22)
+ 8 * (14 + 16) + 15
- 73, 1 I
< where the numbers in the. expressions correspond to the. f ailure of the components given in Tables 1 'and 2. Entries with designators 4, 7, 8, 20, are seismic-induced f ailures and given in Table 1. Entries with designators from 10-17 and 22 are nonseismic f ailures and given in Table .2. _The missing.
numbers in Tables 1 and 2-are. component screened out during the' final s creening . The terms SL and LOSP_ represent the small' LOCA. and loss of offsite power initiating ~ events, respectively. In the above expression, the "+"
notation denotes probabilistic addition (union) and the "*" denotes probabilistic multiplication (intersection).
Impulse line failures were assumed to be the source of a small LOCA at Maine Yankee. This conservative assumption was required due'to the tremendous
. number of hours which would be required to walk down each of these impulse lines during an outage and assess potential system interaction problems.
- These lines _ originate from the primary pressure boundary inside containment (i.e., RPV, steam generator, pressurizer, primary coolant loop piping, etc.)
and are field routed to instrument racks inside containment.
Inspection of the small LOCA core-damage Boolean expression indicates that 'the dominant components are the three singletons 4, 7, and 20. The singleton component with the lowest HCLPF capacity is the refueling water storage tank (RWST, number-7) with a HCLPF capacity of 0.218 - Failure of this tank results in no coolant being available for reactor vessel injection following a small LOCA. The other singleton components have HCLPF capacities of 0 30g. The capacity of the transformer (number 4) is estimated based on the' upgraded condition.
l 42
The HCLPF capacity for the no LOCA core damage Boolean expression is estimated to be greater than 0 30g. The higher capacity for this sequence, compared to the small LOCA sequence, is due to the absence of singletons and no low capacity doubletons in the expression. Although the DWST, i .e. ,
component 8, with a HCLPF capacity of 0.17g appears in this sequence, its failure has to occur simultaneously with one of the higher capacity components, i .e., the transf ormer or the circulating water pumphouse.
The most important nonseismic f ailure is number 15, a common cause failure of the auxiliary feedwater system caused by steam binding. This f ailure results in the inability to cool down the reactor coolant systems using the steam generators. This nonseismic f ailure is the most important because it appears in a majority of the doubleton cut sets.
To account for the fact that we could not quantify the small LOCA initiating event, we developed an overall plant core-damage Boolean expression by logically combining the two Boolean expressions accounting for the split fraction p between the small LOCA and LOSP initiating events. This overall plant core-damage Boolean expression is given below:
Core Damage
- p [small LOCA core damage] +
(1 p) [no LOCA core damage]
In this case, the split fraction accounts for the fraction of the time a small LOCA will occur along with the LOSP event.
32 Overall Plant Core Damage HCLPF Capacity To account for the two Boolean expressions in the overall plant core-damage HCLPF capacity, a sensitivity calculation was performed accounting for 43
a variation in the split fraction between the two accident-sequence initiating events (small LOCA and no LOCA). For different assumed split fraction values, the overall plant core-damage HCLPF capacities were obtained as shown in Table 3 For comparison, during the SSMRP study of Zion, the small LOCA initiating event for a 0.3g earthquake was determined to occur about 1% of the time.
The conclusion regarding the dominance of RWST f ailure in the HCLPF capacity estimation (displayed in the small LOCA accident sequence) is a function of the split rraction assumed. If the plant HCLPF capacity needs to be increased, it is not necessary to concentrate only on RWST. A walkdown and review of small impulse lines within the containment may be performed to estimate their fragilities in order to assign a realistic HCLPF capacity or split fraction. By this procedure, the plant HCLPF capacity may be shewn to be higher without the necessity of any upgrading of the components.
3.3 Effect of Nonseismic Failures The overall plant HCLPF capacity was calculated using the Boolean expressions for the core-damage accident sequences which contained both seismic and nonseismic f ailures. Since this is a seismic margin review and the interest is only in the seismic capacity of the plant, one may choose to ignore the nonseismic failures in calculating the overall plant HCLPF capacity . In the small LOCA Boolean expression, there are no significant 1 nonseismic failures. Not including the nonseismic failures on the HCLPF capacity of the no LOCA Boolean had no effect on the plant HCLPF.
- 4. CONCLUSIONS l The trial seismic margins review of the Maine Yankee plant was conducted '
with the concerted effort of all parties involved. The analysis teams worked l together closely and followed the guidance on performing the review to estimate the overall plant HCLPF and the HCLPF capacities for the accident sequences that lead to seismic-induced core damage. The Maine Yankee utility 1
44 1
Table 3 Summary of plant level HCLPF capacities.
Case Des cription HCLPF Capacity (g) 1 Small LOCA 0. 21
- Independent Seismic Failures 2 Small LOCA 0. 21
- Dependent Seismic Failures 1 3 No LOCA 2 0.30 l
- Independent Seismic Failures with Nonseismic Failures 4 No LOCA 2 0.30
- Independent Seismic Failures without Nonseismic Failures 5 Core Damage
- Split Fraction p = 0.01 2 0.30 p = 0.10 0.28 p = 0.50 0.23 and Yankee Atomic Electric Company provided invaluable assistance in performing the review. Without their efforts, this review would have been much more difficult. The Expert Panel provided an initial review of the approach early in the project. The Peer Review Oroup provided guidance and a critical examination of the process and interim results at each stage of the revi ew . The NRC assisted in the definition of the scope of the review and licensing issues involved. I l
l l
i 45 1
This seismic margin review has been performed with the following assumptions and limitations:
o The review earthquake level was specified by the NRC as the NUREG/CR-0098 median spectrum anchored to 0.38 o- The structural models and the in-structure response spectra generated lj l
.by Maine Yankee have been' judged to be adequate for the purposes of this margin review.
o Since the: Analysis Team could not perform the walkdown inside the
- containment, the seismic capacity of components inside the .
l containment was not determined. We could not confirm the absence of )
' potential system interaction effects that may make the impulse lines
~
inside the containment vulnerable to earthquakes and lead to a small LOCA.
o In keeping with the Expert Panel's philosophy, the screening of components was performed using conservative procedures. For the screened-in components, the seismic capacities have been calculated using conservative methods. In all cases, the factors contributing to the seismic margin and-their variabilities are identified and ;
quantified using procedures within the state-of-the-art, o The HCLPF capacity of the plant has been determined based on the. '
seismic capacities of components in their existing or proposed modified conditions. Maine Yankee has proposed that modifications or replacements would be made for station batteries, transformer internal core / coil assembly anchorage, vibration-isolation supports for containment spray fans and air conditioners, anchorage of diesel day tank, and a block wall near the containment spray f ans.
o The results of this seismic margins review represents the best estimate analysis of the components and the plant following the proposed modification. No effort was made to account for the effects of future aging. 4 The conclusions from the trial seismic margin review include:
o The plant HCLPF capacity was determined to be 0.21g. This capacity is dominated by the small LOCA-accident sequence with the RWST being the dominant component. There was no effect on the plant HCLPF capacity when dependence between component f ailures was considered.
Assuming an arbritrary ten percent (10%) probability of occurrence of the small LOCA-initiating event as compared to the occurrence of all other possible initiating events in the plant, the HCLPF capacity 1
I 46
increases to 0.28g, while it is greater than 0.3g if 1% probability of occurrence is assumed.
o We found that careful plant walkdowns are essential to successful l seismic margin reviews.
l o The maintenance of hot shutdown following a seismic-induced initiator l l
was considered by performing a thorough walkdown review and analysis of the components needed to perform this f unction. This led to upgrading of Fans FN-44A, B and the adjacent block wall, o Important components and f ailure modes identified during this review were:
Station service transformers (4160 V to 480 V) require a review of their internal configuration. l Lead-antimony station batteries may f ail due to the f ailure of the plates within the battery casing.
Consideration must be given to the location and possible systems interaction from threaded fire water piping.
o Consideration must be given to modifications and upgrades identified ,
during the seismic margins review process.
o The small LOCA-initiating event had to be considered for this review because of the difficulty in performing a walkdown inside the containment building.
o The walkdown review of closed-loop component-cooling systems may require considerable effort.
o The components that affect the reactor subcriticality function, in particular the control element drive mechanisms and the reactor internals, need to be considered early in the review due to the long lead time in data gathering.
A number of components were identified as having a low seismic capa city. These components are planned for upgrade by Maine Yankee and are listed below:
o Important station service transformers (4160 V to 480 V).
o A block wall near the HVAC equipment (Fan 44A & B) needed to cool the containment spray pump enclosure. This enclosure houses the long-term cooling equipment.
47
I a
o The component cooling water and heat exchangers (E-4A/B, E-5A/B).
o Station batteries 1 and 3 will be replaced during the March 1987 l
outage. Batteries 2 and 4 will be replaced in the 1988 outage, j
o Upgrading the anchorage of both diesel generator day tanks.
l The plant HCLPF capacity af ter the planned upgrades was estimated to be 0.21g. This HCLPF capacity is governed by the RilST and represents a conservative estimate of the seismic capacity of the plant. That is, given an earthquake producing this ground acceleration and specified spectral shape, there is high confidence (95%) that there is a low probability of core damage (occurring only approximately 5% of the time). After completion of this review, Maine Yankee decided to also upgrade the RWST resulting in an increased HCLPF of 0.27g.
The NRC staff has reviewed the results of the seismic margin review and Maine Yankee's committment to upgrade identified canponents. They concur with the findings and have issued a safety evaluation report which concludes that the plant has an adequate seismic design [9,10].
I l
i l
48
REFERENCES
- 1. Cummings, G. E. , J. J. Johnson, and R. J. Budnitz, "NRC Seismic Design Margins Program Plan," UCID-20247 (October 1984) .
- 2. Budnitz, R. J. , P. J. Amico , C. A. Cornell, W. J. Hall, R. P. Kennedy, 1 J. W. Reed, ar4d M. Shinozuka, "An Approach to the Quantification of Seismic Margins in Nuclear Power Plants," NUREG/CR-4334, UCID-20444 (August 1985).
3 Prassinos, P. G., M. K. Ravindra, and J. B. Savy, " Recommendations to the Nuclear Regulatory Commission on Trial Guidelines for Seismic Margin Reviews of Nuclear Power Plants," NUREG/CR-4482, UCID-20579 (March 1986).
4 Prassinos, P. G., R. C. Murray, and G. E. Cummings, " Seismic Margin Review of the Maine Yankee Atomic Power Station - Summary Report," NUREG/CR-4826, UCID-20948 Vol . 1 (March 1987).
- 5. Moore, D. L. , D. M. Jones , M. D. Quilici, and J. Young, " Seismic Margin Review of the Maine Yankee Atomic Power Station - Systems Analysis, NUREG/CR-4826, UCID-20948, Vol. 2 (March 1987).
- 6. Ravindra, M. K. , G. S. Hardy, P. S. Hashimoto, and M. J. Griffin, " Seismic Margin Review of the Maine Yankee Atomic Power Station - Fragility Analysis," NUREG/CR-4826, UCID-20948, vol. 3 (March 1987).
- 7. American Society of Civil Engineers (ASCE), "The Effects of Earthquakes on Power and Industrial Facilities and Implications for Nuclear Power Plant Design," Draft UCRL-93320 (1986). To be published by the ASCE.
- 8. Campbell, R. D. , M. K. Ravindra, A. Bhatia, and R. C. Murray, " Compilation of Fragility Information from Available Probabilistic Assessments,"
Lawrence Livermore National Laboratory, Livermore, CA, UCID-20571 (September 1985).
- 9. Sears, P. M., " Maine Yankee Atomic Power Station (MYAPS) Seismic Design Margins Program," Letter to J . P. Randazza, March 26, 1987.
- 10. U.S. Nuclear Regulatory Commission, " Safety Evaluation by the Office of Nuclear Reactor Regulation - Maine Yankee Atomic Power Company, Maine Yankee Atomic Power Station - Docket No. 50-309 - Seismic Margins Program", prepared in March 1987.
49
JEU 3 SEISMIC CAPABILITY OF CONTAINMENTS INCLUDING PRESSURE EFFECTS M. Amin and P. K. Agrawal Structural Analytical Division, Sargent & Lundy, 55' East Monroe Street, Chicago, Illinois 60603, U.S.A.
T. J. Ahl CBI Na-Con, Inc., 800 Jorie Boulevard, Oak Brook, Illinois 60522, U.S.A.
Abstract A methodology for determining seismic capacity of containments subjected to prescribed pressure and thermal loads followed by a seismic excitation is presented. The application of methodology is described for four containments, for which a number of limit states are evaluated. Effects of soil-structure interaction, material nonlinearity, and uplift are considered.
The evaluated capacities for two of the containments, in terms of peak horizontal ground acceleration, show that for all pressure and thermal loading conditions eva,7.uated seismic capacity is significantly in excess of design safe shutdown earthquake level. The evaluation of the other two containments and refinements of the results for the first two containments are in progress.
51 1
- - _ _ - - - _ - _ _ l
L 1. Introduction Most of the analytical and experimental investigations to date on the performance of light water reactor containments for loading beyond design basis have considered l only the effects of static overpressurization. Seismic capability has been considered in connection with several probabilistic risk assessment studies; however, the limit states considered in these studies have not been exhaustive. The effects of soil-structure interaction and the limit states related to foundation failures have not been
-included in these studies. Also, the behavior of containments subjected to aftershocks acting in combination with pressure and temperature loading following a severe main shock, has not been evaluated.
Sandia National Laboratories (Sandia) is sponsoring an analytical investigation to determine the seismic capacity of containments for various conditions of internal pressure and temperature. The investigation is being performed jointly by Sargent & Lundy and CBI Na-Con, Inc. The study is funded by the United States Nuclear Regulatory Commission (NRC).
To gain a perspective on the problem, the study includes the analytical evaluation of four different containments.
The paper summarizes the methodology used to obtain {
i containment seismic capacities in terms of free-field peak horizontal acceleration, the limit states considered and i
52
1 i
i criteria for their evaluation, and the results obtained for ;
two containments. Also discussed are some preliminary results for the remaining two containments.
- 2. Containments under study Table 1 lists the four containments being studied. The table also gives information on the type of these containments, their foundation material, the design safe shutdown earthquake (SSE), and design accident pressure, P 'd and temperature, Td. The study considers two concrete and two steel containments. One of the concrete' containments is prestressed. Two of the containments are founded on soil and two are founded on rock.
- 3. Methodology A common approach was developed for seismic analysis and evaluation of the four containments before performing any analysis. The project is organized into six tasks. Tasks 1 and 2 dealt with the development of the methodology described below. Tasks 3 through 6 deal with the evaluation of each of the four containments for four specific load cases.
3.1 Load cases Table 2 lists the four load combinations considered.
The load combination of Task 3 considers dead load plus the earthquake effect. This represents the occurrence of an initial shock (main shock) during the normal operation of the 53
L 1
power plant. Task 4 considers dead load, design accident pressure and temperature, and seismic loads. This represents a condition when the containment does not fail under the main shock, but, due to some component failure, an accident is initiated and followed by an early aftershock. Values of desion pressure and temperature considered in Task 4 are l listed in Table 1. Task 5 considers more severe pressure and thermal loads in combination with a late occurring after- )
{
shock. Table 3 lists values of T y and P y used in this task. Values of T y were specified from a general review of upper temperature conditions which have been calculated by Sandia for various severe accident scenarios in each containment type. Values of P y were calculated in this project to correspond approximately to the initiation of yielding in the containment.
For the Fermi containment, P y corresponds to the pressure at which yielding at the cylinder to cone junction near the drywell head begins. For Clinton, first yielding is calculated in the outside hoop rebars in the containment wall. For the Zion containment at Py, full hoop tension cracking of concrete has taken place, and prestressing tendons are from 80% to 90% of their yield point. For the Sequoyah containment, first yielding occurs in the shell )
below the cylinder to dome junction. !
i Temperature and pressure values for Task 6 are the average values of those for Tasks 4 and 5. Task 6 was not l
1 1
54 1
I
_ _ _ _ _ _ _ _ _ _ 4
evaluated for Fermi and Sequoyah because the difference between P y and Pd'is small for these containments.
3.2 Seismic input The aim of the project was to estimate earthquake levels for which containment integrity may be potentially threatened. To determine this capacity, a series of time-history analyses were made on each containment using progressively more intense motions. These analyses were performed by first generating two statistically independent ground motion acceleration histories, i'o (t) and y'(t) o for horizontal and vertical motions, respectively. These motions were generated to have 5% damped response spectra consistent with the NRC Regulatory Guide 1.60 spectra for horizontal and vertical motions anchored at peak horizontal ground acceleration of 0.25 g. To cbtain motions for other values of the peak horizontal ground acceleration, AH , the motions i'o (t) and y' (t) were scaled by A H/0.25. l o )
Since Regulatory Guide 1.60 spectra are currently used J
as standard input for nuclear plant design, it was felt that stating seismic capacity in terms of motions consistent with the Regulatory Guide spectra would enable application of the conclusions of this project to other containments as a standard index. Also, the rank ordering of various limit I
states considered for the four containments studied will be facilitated with the use of one standard input. ]
1 l
i 1
35 L _ _ _ - - - _
3.3 Seismic analysis model For each containment, a separate plane frame _model was developed. The features' common to'these models are: (1) different vertical beams to represent the containment, i internal, and adjacent structures; (2) strain dependent soil I
properties for. soil foundations, i.e., soil spring and dashpot constants that are a' function of AH; (3) vertical' !
one-way foundation springs andTdashpots-to permit i consideration of uplift; (4) horizontal-beams to represent the basemat; (5) hysteretic hinge' elements at the. ends of beams representing the containment to subject these elements
\"
to moments that are consistent with the state-of yielding in the material; and.(6) reduced shear-stiffness in concrete elements to account for shear cracking under seismic loading.
Figure 1 shows the seismic analysis model used for the Clinton containment. Models for other containments are not-shown in this_ paper, but they had features similar to those shown for Clinton. The horizontal beam elements between points A and B in this figure represent the basemat. The horizontal beam elements between points C and D are rigid, massless elements that transfer the containment base moment to the basemat. Rigid, inclined truss elements are used to transfer the base shear to the basemat. The use of rigid girder C-D in the model is to consider the stiffening effect of containment wall. Similarly, the rigid girder E-F is used to transfer the reactor-control building base moment to the f
56
basemat; it also simulates the stiffening effect of the shear walls of this building on the basemat.
The soil spring and dashpot constants for Clinton considered strain dependency of shear modulus and damping.
Through a series of deconvolution studies using the computer program SHAKE [1], soil properties compatible with various values of Ag were determined. The properties thus determined were then weighted according to thickness of soil layers.
Weighted average properties were then used in half-space equations assuming an equivalent rigid circular basemat to determine total foundation spring and dashpot constants.
Finally, the spring and dashpot constants used at the basemat nodes of Figure 1 were obtained from the total foundation values using the tributory area of each basemat node. Table 4 lists the total foundation spring and dashpot constants as a function of A H. The subscripts H and V denote the horizontal and vertical components, respectively, of the soil stiffness and soil damping.
l In addition to foundation dashpots, the dynamic model l included viscous damping providing 4% to 10% of modal 3 1
critical damping in the linear range of response. {
The containment vertical beam elements between points G I
and H in Figure 1 have hysteretic hinges at each end. The j moment-curvature (M-Q) diagrams of these hinges were f developed by considering tubular sections bending about their i
diameter. Stress-strain diagrams of concrete, meridional I reinforcing steel, and the liner were used in this 57
1 l
l evaluation. For Task 3, this evaluation considered the actual material stress-strain diagrams. For Tasks 4 through 6 the steel stress-strain diagrams were adjusted to consider meridional strains which exist due to pressure. Figure 2 shows the M-p diagram used for the element between points G and I of Figure 1 for various tasks.
For a post-tensioned concrete containment, the M-@
diagrams used also considered the effect of post-tensioning. For steel containments, seismic loads introduce I
a biaxial membrane stress in the shell, i.e., meridional stresses are accompanied by circumferential stresses. The extent of bianial action depends on shell geometry. The effect of this biaxiality'on the M-p diagrams was considered in different elevations of the Fermi and Sequoyah containment shell. Ratios of meridional to circumferential stresses in these containments were estimated using a constant horizontal load acting on the shell. From this information the uniaxial steel stress-strain diagram was modified to~give correct initiation of yielding based on the Tresca yield criterion.
This modified stress-strain diagram was then used to construct the M-@ diagram.
For a prescribed value of A H, the time-history analysi; of the seismic model was performed using a step-by-step integration procedure. The computer program NONLIN2 {2) was used to perform this analysis.
For each containment the direction of the horizontal seismic excitation was selected in such a way to maximize the l
l 58 I i
.j
possibility.of basemat uplift. This determination was made based on engineering judgment after reviewing the foundation plan of the containment and adjacent structures in each case.
3.4 Limit state and evaluation criteria Prior to any analysis, a detailed review of the containments was nade to identify the limit states that have potential for controlling containment capacity. Table 5 lists 15 limit states selected for evaluation and their applicability to the four containments. The table includes limit states that are either directly related to the containment pressure boundary, herein called direct limit states, or the limit states whose realization may indirectly lead to conditions under which assurance of containment integrity cannot be provided within the scope of this study. These latter limit states are called indirect limit states. For example, for the Clinton containment, failure of reinforcing bars in the containment wall is directly related to the containment pressure boundary and is therefore considered a direct limit state. On the other hand, liquefaction of foundation soil under basemat is considered an indirect limit state because containment performance under >
potentially large displacements that may be caused by liquefaction cannot be assured within the scope of this study.
l 59
Table 6 lists the strain , stress , or force-based evaluation criteria used to define realization of the fifteen limit states. All criteria are intended to be realistic, credible, and computationally feasible. For example, the l
strain limit of 0.02 for mid-thickness strain in steel containments can be inferred from the posttest analysis of the 1:8 scale containment model pressurized to failure [3],
based on the fact that results being calculated in the present work do not consider strain intensification caused by the as-built features.
To estimate the containment capacity, seismic time-history results were obtained at several values of AH. Peak responses, thus derived, were then used to determine the capacity margin factor at each A H f r each limit state under various tasks. The capacity margin factor m is the ratio Capacity according to failure criterion m = applicable to limit state (1)
Load effect for load case applicable to the Task For certain limit states, Equation 1 could be evaluated I directly using the maximum responses from the time-history l
analysis and data in Table 6. For some limit states, it was necessary to make a shell analysis for peak ceismic forces.
These analyses used an elastic shell of revolution analysis 60 obtain forces caused by seismic loads, which were then combined with the forces from pressure, thermal, and dead weight loads. For concrete containments, the shell element forces thus derived were applied to shell elements in a 60
computer program that considered concrete cracking and rebar and liner yielding. To assess the influence of this simplified analysis for concrete containments, a finite element analysis of Clinton containment is being performed using the computer program ADINA [4). This analysis applies dead, pressure, thermal, and seismic loads to the nonlinear model. Again, the seismic load is the peak seismic load derived from the time-history analysis.
- 4. Discussion of results and effort in progress Time-history analyses of all containment buildings have been completed. Capacity margin factors are under evaluation. Also in progress is the three-dimensional analysis for the Clinton containment. Currently available results and their significance are discussed below.
4.1 Clinton containment Clinton seismic response was evaluated for values of i
AH = 0.25 g, 0.45 g, 0.75 g, and 1.0 g. Seismic evaluations showed that because of the large extent and continuity of basemat with other foundations in the two horizontal directions, uplift does not occur even at AH = 1.0 g. Figure 3 shows the vertical displacement time-history of the basemat node nearest to the containment, indicating that uplift does not occur at AH = 1.0 g.
The extent of soil-structure interaction at high acceleration levels is demonstrated in Figures 4 and 5, where 61
the basemat horizontal response spectrum is compared to the free-field spectrum at AH = 0.45 g and AH = 1.0 g, respectively. The soil-structure interaction attenuates the higher frequency input. This effect becomes more pronounced at higher values of AH because at this level of excitation soil softens more.
Tables 7 and 8 list calculated capacity margin factors at AH = 1.0 g for direct and indirect limit states, respectively. This information shows that the seismic capacity is less than 1.0 g for only one limit state, i.e.,
liquefaction of structural fill under the basemat (Table 8). The calculated capacity in this case is 0.83 g for all tasks. For all other limit states evaluated, the seismic capacity for Tasks 3, 4, 5, and 6 is greater than 1.0 g. The liner is either in compression because of thermal load (Task 6), or it does not yield (Tasks 3, 4, and 5). Yielding of rebars occurs only in Tasks 5 and 6. The results reported for penetration failure utilize conservative pull-out capacity and stiffness values. The penetration size is typical of the main steam line.
It should be noted that a seismic capacity of AH = 1.0 g for Clinton containment would be four times higher than its design SSE level of 0.25 g.
4.2 Fermi containment The Fermi seismic analysis model (not shown in this paper) contained separate vertical beams for the reactor
{
62
l 1
building, biological shield, steel containment, reactor pressure vessel, and sacrificial shield. Because Fer.ni is I
founded on rock, foundation spring and dashpot constants used were taken as invariant with respect to AH* l Time-history results were obtained at AH = 0.15 g, .]
1 0.30 g, 0.45 g, and 0.60 g. These studies showed that l I
foundation uplift begins at approximately AH = 0.45 g.
Examination of lateral shears and moments'on the steel containment showed reduction of seismic forces due to uplift. However, the occurrence of uplift introduced higher amplitudes in the high frequency portion of response spectra
\
calculated from the accelerations on the steel containment. :
Figures 6 and 7 show the effect of uplift on the horizontal and vertical spectra calculated from acceleration near the drywell head.
Other than the uplift effect discussed above, because of rock foundation, no other foundation-structure interaction was nnsed for Fermi containment.
Tables 9 and 10 list the capacity margin factors at AH*
0.60 g for direct and indirect limit states, respectively.
This information shows that two indirect limit states, i.e.,
basemat shear failure under the suppression chamber and failure of the biological shield wall due to tension plus tangential shear, yield a seismic capacity of AH = 0.50 g.
This capacity applies to all tasks because these failures are not affected by pressure loading. The seismic capacity for all direct limit states is in excess of 0.60 g in all tasks 63
except the initiation of yielding at lower beam seats in
) Task 5 (see Table 9).
It should be noted that a seismic capacity of AH" 0.60 g for Fermi containment, with the use of Regulatory Guide 1.60 spectra would be seven to eight times higher than its design SSE level of 0.15 g with Housner spectra.
As stated in Section 3.3, the meridional tension / compression seismic stresses are accompanied by stresses in the hoop direction. The results reported in Table 9 considered the beneficial effect of this biaxial behavior when evaluating shell buckling; however, the effect of this behavior on initiating meridional yielding earlier than that indicated by uniaxial stress-strain diagram was not considered. Solutions for incorporating the effect of blaxial stress in time-history analysis are now being implemented. This modification is expected to lower margins related to tensile yielding in Table 9. There may also be secondary effects on other margin factors in this table.
4.3 Zion containment The seismic response of the Zion containment has been evaluated for values of AH = 0.25 g, 0.50 g, 0.75 g, and 1.0 g. Because Zion has a soil foundation and because the soil softens at higher values of A H, soil-structure interaction effects similar to those in the Clinton containment tale place. The basemat in the Zion containment is not connected to the basemat of adjacent structures.
64
u-Therefore, the rocking effect' is more pronounced, which causes greater lateral deflection than in the Clinton conta'inment.- This increased potential for lateral deflections causes more severe conditions for evaluation of
. penetrations and impact with adjacent structures in Zion.
Although the Zion containment begins to uplift soon after'A H reaches a value of.0.25 g, the'effect of uplift on the high frequency region of vertical and' horizontal response spectra is not as significant as that demonstrated for the Fermi containment (Figures 6 and 7). This-is because of the
. relative softness of soil foundation at Zion.
Capacity margin factors for Zion'are currently being evaluated..
4.4 Sequoyah containment For Sequoyah, containment seismic response was evaluated at AH = 0.25 g, 0.50 g, 0.75 g, and 1.0 g.- Because it is on a rock foundation, no foundation-structure interaction, other-than uplift of the basemat, occurs. The effect of uplift on intensifying the high frequency portion of the response spectra is more pronounced at AH = 1.0 g than in the Fermi containment. This ja because of the stiffer rock foundation at Sequoyah.
It is known that lateral forces on a structure with basemat uplift are usually less than those where uplift does not occur, except for very short-period structures founded on very stiff foundations [bj. The fixed base period of the E5
~ - - _ ._ - - - - . - _
steel containment and foundation stiffness at Sequoyah is such that somewhat increased lateral forces are computed for the steel containment. This amplification due to uplift does l l not occur for the separate concrete shield building of Sequoyah, waich also rests on the same basemat as the steel containment, but has a longer period of vibration. !
Capacity margin factors for Sequoyah are currently being i evaluated. Our calculations have confirmed that up to AH*
1.0 g membrane yielding in the steel containment does not i occur (with the effect of stress biaxiality of seismic shell response appropriately considered) even for the Task 5 loading condition.
- 5. Summary and conclusions A methodology for calculating the seismic capacity of containments under specified pressure and temperature conditions has been described. The methodology permits evaluation of realistic limit states considering effects of soil-structure interaction, foundation uplift, and material nonlinearity. The application of this methodology to four containments is described and it is shown that quantitative information for rank-ordering limit states and determining seismic capacities can be obtained. <
Two of the containments evaluated are founded on soil.
The results obtained for these containments confirm the expected intensification of soil-structure interaction 66
j effects at high levels of peak horizontal ground acceleration, A H, considered in this study.
At high values of A H, significant uplift was calculated for three of the four containments. Occurrence of uplift increased responses in the high frequency zone of the response spectra for containments founded on rock relative to I
an analysis that ignores uplift. In or2e case, lateral forces on the containment increased over the analysis that ignored uplift. In two other cases, where uplift occurred, l structural forces went down. As discussed in Section 4.4, this behavior is in accord with known facts about foundation uplift phenomenon.
The seismic capacity evaluations are still in progress, but for the containments for which calculations have been made, the estimated capacity for all limit states is significantly higher than the design basis SSE level.
Combination of pressure, temperature and seismic loads reduces the capacity over the case when ambient conditions are assumed in the containment. When behavior in a limit state is affected by pressure, capacity reduction may be significant at some of the high pressure conditions which are ,
being evaluated in this project. The evaluations made thus far indicate basemat response to be important. This conclusion is under further study in this project.
Rank-ordering of limit states for a containment type and isolation of sensitivity of capacities in governing limit states is a part of this study. This effort is in 67 l
3 progress. Cross-referencing containment seismic capacities, which are obtained using a t/siform approach to evaluation, provides a useful means towards accomplishing this objective.
- 6. Acknowledgement The reported work is being performed under Contract Number 04-3587 to Sandia National Laboratories and it is funded by the U. S. Nuclear Regulatory Commission, Office of Nuclear Regulatory Research. D. B. Clauss has been the Sandia Technical Representative and J. F. Costello is the NRC Program Manager. Their support and encouragement are gratefully acknowledged.
- 7. References
[1] SHAKE, "A Computer Program for Response Analysis of Horizontally Layered Sites," Based on Earthquake Engineering Research Center Report No. EERC 72-12 by Schnabel, P. B., et al., University of California, Berkeley, 1972. Modified and maintained at Sargent & Lundy, Program No. 09.7.119-3.3.
[2] NONLIN 2, "A Program for Nonlinear Dynamic Analysis of 2-D Structures," Sargent & Lundy Program No. 09.7.075-1.1.
[3] Clauss, D. B., " Comparison of Analytical l
Predictions of Experimental Results for a 1:8-Scale Steel Containment Model Pressurized to Failure," NUREG/CR-4209, 68
l l
SAND 85-0679, Sandia National Laboratories, Albuquerque, NM,
-June 1985.
[4] ADINA, "A Finite Element Program for Automatic l
Dynamic Incremental Nonlinear Analysis," ADINA Engineering Report AE 84-1, December 1984. Maintained at Sargent &
Lundy, Program No. 09.7.199-3.0.
[5] Chopra, A. K. and Yim, C. S., " Simplified Earthquake Analysis of Structures with Foundation Uplift,"
Journal of the Structural Division, ASCE, Vol. 111, No. ST4, 1985, pp. 906-930.
I l
l I
i l
i 69
_ _ _ _ - _ _ _ _ _ - _ _ _ J
_ t
_ c
_ e p
s 0
6
_ e r 1
_ u e
_ t d a)
_ rF 1 5 1 0 i u
g(
e 7
2 8
1 8
2 2
2 G y
T r o
t
_ s a r
ee t r) l u
eug g msi -
e ass 7 5 6 8 R rep 4 1 5 ar( 0 e PP 1 h t
n ,
g )
i * *
- sa r 0 r r z D er e 6 e e H E Dtecp n n n T s 1 s s 0 A ey u '- u u 2 U pT o G o o L S H R H H 'o t
A V 2 E
5' 5 8 (
7 E 2 1 1 S 1 n
l' NT E E SM S 0 0 0 0
i t
o L M a B NI c A A n i T T o f .
- N i i a l r O t pt C a l l k c
k c mc d i i aep F n o o o o O u S S R R o ts n
T F ar S ce I
L in
, f s r i u e no s gH n r i
, e o se I d t h I , n c ft d I I o aro e d c eo n
,s k e k rt ea yse rce r e c gh ret art a c rant d re Moe M i eea t rrr
,t sc r ,fncr ,e l
,e l
a - e Re R e w r yh TmW R en P pc ro Ri n Weo B rc Wt Bs Wt P s ecg dt ni eaeh zt vu i q%
rge0 unr0 t sif1 n sl eieo e h roht m
r n a PBt i o y %
a t i o = = n0 t n n m u q RR I 8 n o i r o i l e e WW C Z C F S PB*
lE
TABLE 2
. LOAD COMBINATIONS Task Combination 3 0+E 4
0+Pd+Td+E 5 0+Py+Ty+E 6
0+ PAVE + TAVE + E E = Seismic load 0 = Gravity and other static loads based on containment type i P,Td d = Design accident pressure and temperature, respectively P,T y =y Yield pressure and temperature, respectively PAVE = 1/2(Pd+P) y TAVE = 1/2(Td+T) y l
i 71 L l l i <
TABLE 3 VALUES OF CONTAINMENT TEMPERATUREy T AND PRESSUREyP FOR TASK 5 Containment Ty (D)F Py (psig)
Fermi 550 74 Clinton 400 45 Zion 360 93 Segaoyah 360 24.8 i
1 72
!~
1 i
' TABLE 4 FOUNDATION SPRING AND DASHPOT CONSTANTS FOR CLINTON 4
i C C
. KH K ikip H -(kip ysec) ft sec)-
(kip /ft) (kip V/ft)' '
ft 0.25 5.10 x 10 6 7.24 x 10 6 4.31 x 10 5 9.04:x110 5 0.45 3.30 x 10 6 4.69'x 10 6 3.47 x 10 5 7.27 x 10 5 0.75 1.98 x 10 6 2.82 x 10 6 2.69 x 10 5~ 5.64 x 10 5 i 1.0 1.43 x 10 6 2.03 x 10 6 2.28 x'10 5 4.78 x 10 5 l
l 73
h a
- y
- s s o s s s s s s s u e e A e e e A A e A e A e e o q Y Y N Y Y Y N N Y N Y N Y Y N e -
S ti s s s s nm er
- A s
e A c
e s
e s
e A A e A e e e o o me nF N Y N Y Y Y N N Y N Y Y Y N N i
a t
n o
C n s s s s s s s o s s s A e e e e e A A e e o i e e e A Y Y N N Y Y N Z Y Y Y N N Y Y Y R
I ES HT TN E n
- DM Nl t o
s s s s s s s
- s Al n e e A A A e e e e A A A e o e A i Y Y N N N Y Y Y Y N N N Y N Y DT l '
EN C
' RO 5EC D
EI O LST BN s AOY t TCT l n s s I l e n t SL e t m o a EI h a n i e s TB s m i t s t AA e a c r t TC t s t e m o n SI n l a n nd a p e L e l b o na e p m TP m e / e c oe b u n IP n h l t ch s i MA s i s l e e e t a l I
r n a a r t dl t a r t i L~ e o t t w c e el u e n n o n d n n n r t e l l b o o s s n o e n o c l w p l m i c i
l r e c m i c n oy e a t n a t n o br s h h a f o l b l i e f c d s s c d o i n
e g e a r o d e n g t
a e g n e t u n ed r t u t n i t n l g i nn t ns o o n d s i s s o i n oa t et i f i n c s c a i e i u mr s d u f r e f f h r sh b no s f i o o o r o l s u nc ip e o l f f t e r u l et t ap r s e n s e e a r i t a a t u p e f r i e r t e c a eh ns p r o o u e r u s h f r e o u u t l r p l s l pt r ct s l n n i i f l n n u s i e o o a f f a o e a o f e l f e f a u i i f o o f s w i omp i a
oh c
o f d t c
t g r - t a
p e e e e n e h a ei f e e g e i l r r l i v g r ru re r n r f r i u u i l s u t uq r uc u i u e c s l l s k n o e l e a l i l r l u n c a r n e a i q s n i i e
i i ar i
a e a e e a a e u r h at h F B F i
L D T F F T B T T P Fa S F o e
d 0 2 3 4 5 o 1 2 3 4 5 6 7 8 9 1
1 1 1 1 1 1 C
_ f
k_
' TABLE 6-CRITERIA FOR EVALUATION OF LIMIT STATES-Limit State Description Evaluation Criteria
-1. Tensile failure of steel liner Mid-thickness strain >0.02 Surface strain- >0.06
- 12. Failure of reinforcing bars Strain >10 x yield strain
- 3. . Failure of-prestressing tendons Strain > yield strain
- 4. Tensile failure'of. steel containment shell Mid-thickness' strain >0.02 Surface strain >0.06;
- 5. Buckling of steel containment shell Membrane compression > buckling stress. determined from theoretical elastic buckling analysis x l knockdown factor
- 6. Transverse shear failure in wall /basemat Nominal shear stress > capacity per ACI 318-83 Section 11.3'
- 7. Through-wall crushing of concrete Average through-wall' compressive strain >0.002
- 8. Penetration failure in concrete containments Force'on penetration > anchorage capacity of penetration
- 9. Failure of pretensioned bolted connections Shear at connection > shear resistance at equipment hatch and drywell head
- 10. ' Shear failure at buttress plate Longitudinal shear flow > frictional resistance
- 11. Failure of containment shell at beam seats Initiation of yielding according to I or ice chest supports Tresca yield criterion
~
- 12. Failure of suppression chamber supports Scress >1.60 AISC allowables or 1.0 stress limits, whichever is less
- 13. Bearing failure of foundation Average pressure on contact area > l ultimate bearing capacity of {
foundation material l
)
- 14. Failure due to sliding of containment Horizontal force > frictional '
resistance plus side pressure r differential
- 15. Liquefaction of foundation soil Average cyclic shear > average cyclic shear capacity 75
6 >
2 3 7 > I 0 5 7 0 0 0 5 * . 1 1 1 2 2 4 > > >
s
_ k
_ s a
_ T -
9 3 0 0 0 0 4 1 1 1 1 2 4 > > > >
g
_ 0 1
5 0 0 0 0 9
= 3
- 1 1 1 3 7 > > > 4 H
A T
A n
i
/t T . pe N p ie E a k r M C t N 0s I~
A) r 6n TS ar .e ) 1i a
NE ea b ~ T OT h e ao + =M l Sh CC P a ;
CA S 6 K' 3 p P, T n l tE 0 c NS o an uM + n7 / c 7O i ro oS o7 y 3 TT r ui A ' c #
ENI e xt l c S. d . r 8 LIM ec ev P
- BLI ACL T
t i
r l e FS lr ue -
PP K
(
se al BE 0
SI RD O( l T r l d C a e e A .
w n t F n s i l N
i t r l o n a bt I
e e b l n G r t m e de R
A u a n g e em M l i n t nn i l a i l s oi Y a l t c l ia f a n r a f st T w o o w o nn I
C r c f eo A a t n t e t ch e n n i n r e c P h e i e e u rtt A s m r m l paa C n t nt n n t i h o e a ia o f i a a f n i s m am i o a m e
f oot t r e t e t t in p e s ns a e n s e ete i v a oa r r o a l rcm uep r s B Cb t u C B i c n e l s l ni n
s a n i i nu aoq e r e a e
_ D T P F T F ce
_ e
_ d o 2 1 9 C 6 8 m
e t
i 1 2 3 4 5
_ yo
c snir e semog b emsp rnie0 5 l pi er ae xas 1_
cr et s s iu . nei=
k ts eoh s i s h ct3 g
- a re t - - A T crp esk nhi st
=
t ot - aa 4 of rT -
N o df o r eot ru coc s
anaf c b oF o i n etxil ra augri gl 0
3 1 l aw
> e a 0. m t l v be1 yn a t e t y=im t - cn si Hai n icA pa h a at i s as t pncn aa o re t n nc- ec Sk i th
~ cn dth
- eat st i a
.io ,hh
)
2 phi eat 0m t m o T. s ett s r i yl t
n 'n o
0.C 0
le l r otl b nn pl paia o i eap eumt h C r gW m r i et
( e sr e a o ssti ur 7 t rnC i ee ruohl e r vh n oanti n E C Awi tc ao L ceonf o B s 5 aBdi A f n -
T -
e s2oh e l n .r i c t b i eomt u e a gl t eam r T rbct r c aaaI t r n n mtf eu o i ysnoec rnc c r tiifpo f e ih g
,el o b m
ct r a aehl g u pf ml ti n n ao p i c emt h n nh aat s o emt xhi u
r i
t h u tl ,et r c a osrno c t cnoai l i a oF ev n l f h
t 'n t i ma o a i t oa .t h
e i w n di rt ob t -
p h e erenn r i g d t ed e r u i oti msa c o nisnee s r e rnion e h t ecoadi D T a b' ct l t nsn s
- d eroio l h ach n t ut e T i o ne o
- m h rih
.t e i l
sel t d n e d
o t nof 0. u C 7 o I uno7d t
s :
r 1 e
m f e e e t t R o I 6
- N O
g 3
t8 a
0 0 = 0 6 9 3 H 1A 8 7 5 g
s 3
) k t8 D s a E a 0 T T 0 = 0 6 9 O 4 H N 1A 8 7 5 S
A g 3
T t8 P a E 0 C 0 = 0 6 9 X 3 H E 1A 8 7 5
(
g o s
0 l a
1 e
= e l
. r b j pu a h)S as Cs t
T E T e s gr NA E c c nP i
h T .r t M
N S n i .i l pl i
r . pa I o cac ae ae o A T I i yC y ev Ch t T M r C C BA S s
N e r r r 5 O
I L t .a .a .k a . e C i eeee ta ex e i T r vh vh l e h a l l p
N C C ASAS UP SM b O E a p T R T a N l I
I l n 7 L D a i C IN w e l r l R( l s
l e b a
O i n l b F f o n e m T i o w u S l t i y n f R a a t r o O r d c d n T u n e o 1 C t u n t i A c o n a t e F u f o a t r c h c o N t f c i N I s o d t f G e d a i o R f e t a h t t A o r l e n M t u o h t e r na l b n d a Y
T n
o om i e i
a f l
l e
m i l i
I i t s f o e p e m C t ca w i t i A p ab g e y u a s P i f n r r q t A r er i u D E s t C c ue r l n s qd a i t e e in e a i m D Lu B F m m i o l c o A t
e s :
d r 1 o 5 3 e C 1 1 9 f e
_ e t R o N
_ m
_ e
_ t I 1 2 3 E
t g a
8 9 5 5 5 0 2 04 1 4 7 0 1 1 . 1 .
) > 1 10 2 1 1 > 3 D '
E T
O N
S '
A -
4 6 6 4 s4 0 4 2 4 7 0 0 9 T k 1 1 P s > 1 1 1 1 2 > 5 E a .
C T X -
E
(
'g _
2 3 2 7 0.
6 3 0 1
7 6 2 3 0 1
0 0 1
> 2 2 1 1 > 7 >
2 0
. =
r H a n
A) o f f T S i / / S A E T r c ) ) "
A e / m c t 2 f ( ( x
- TN _T S i 0 / r r h a r y c c 9E T 5g C 0 f f f EN I
I B TA LM L
I AN T O T C C E t d n
I t e M IR e n t R D m e g l E n m n o F( -
i n i b l a i l R e t a k d h h O e n t g c e c c)
F t o n n u n t td s c o i b o a) ae S - s c l i hd ht R - f ft k l s e a O ol oa l c a n tt t e T l e e u i e na ns C ee es e b t t d ee en A rh r t n e a ms mu F
l us um l a s l a
e r
rt pa h e p ie p
ie N n i t i e f n e ur ur I o an ab o o f f n l qu qu G i f e f i m oo l es es R t m t r r g d u i e s s A p en ea e e n i c et w 9e Se M i l i l p w i r r rc y ' r ' r r ia i l p o l l e i ue r 2p 3 p Y c st sl U L kl M C l n D 1( 1(
T s nn ne ce i n I e eo eh uh ao C D Tc Ts Bs F c A
- P A
C e
_ d
_ o 1 C 4 1 5 9 m
_ e t l 2 3 4 I
y
5 5 05 4 43 1 8
s4 00 5 k
s 44 1 a
T 2 o 3 15 2 s
. l 22 2 a e
l l a b t a n o t i
o ,t s
) r c c i e // h t
t t 26 f i 00 /
n r y o o C 00 f t C
( 5 s e
9 e i l l E b p L a p B T a A
T l n 7 l i e e h r l s e b b a t n m T n o u e i n f m s o n s n i e o 1 a r i t pt t e nn pr a t oo uo sp c o ci i N n t p f o
o f a f u i i
t ort os t n
t p ee er e r i rn re d a r ue ub i l c l p l m i s i i a e m e at ah t i D F a F c a s t
s t n
t e i m
- m m e i o d l c o 2 C 4 1 o A t
s :
r 1 e
m f e e e t t 5 6 R o I
- N 8
- 0. =H 0. =H 1A 3 1A
)
D E
T O g g N
0 0 S 4 t5 t5 A - a a 0 0 T s 0 = 9 0 =
P k H H E s 1A 3 1A C a X T E
(
g g g 0 0 0 6 3 t5 t' 5 a a 0 0 -
0 9
= 0. =H 0. =H .
1A 3 1A o y' -
s A)S l E a T T 9 e AA "y g e h ae l T T m t rr b NS n a eu a o er vs 1ET 0
N I
i r
a Ba e p As e /
c 3 t s
EI M e e .h a kr y i I
LA L t N xS C aP c h BT : i g a l e t AN CT r E M U P 0 TO C .' 1 o C E .
t
~ R I I s M D e R N i E l F (I p l 5 p R a O .
e F n l 7 i b S a e R e T l O r b T u n a C l l i T A i a F a .
c r f f f i e o N
r o g b I o m 1 G a e l u R e r o n e A h u i t M s l b n o n i l o N Y o e an fl i T i s r
f o oa t o I t p
i w a t C et gt e c A i va na rd i r P
c r sm ne id rn ul f a A l e i l C s as au ii t i e ra eo ah n m E Tb Bf F s e i d s i
t t n i e e m m
' i m d
o l o 3 c L 6 1 2 o t A s :
r 1 e
m f e e e t t R o I 1 2 3
- N S
Elev.
904' <> Containment H
862.5' q) 852' q, O
Reactor and control 832' q, buildings O
Reactor and control buildings Drywell wall Containment tp Drywell wall 772' OO 4 i RPV and RPV pedestal
/ ,,
Direction of horizontal
, motion _
Y U d Plan of Buildings in Model 732' I O4 >
717' ' " ' " ' " " " ' """'"t""""""'
F E -
A f \
N' w /nr 712' - : [
g ,, ;;;, ;
s
- ; b;
- :
$$$$$$$$$$$444G TX735NTXT373UTUTT3GTT373UTT373 y L L L L t. L f. f. L R4
& L a. t.
'I 20' 129'l 36' 136' l 29'I' 20'13.533.50 28 28 28'l 2823.5D3.51 Figure 1: Seismic Analysis Model for Clinton Containment C1210.001/M 05-87 82
I l
l l
A I4 Task 3 Task 4 g12
.e Task 5 e 10 E -
8 5
c 6 -
E S4 ~
2 i e e e i .
y 2 4 6 8 10 12 Curvature (x 10-5 rad /ft)
Figure 2: Idealized Moment-Curvature (M-$) Diagrams for Element Between Points G and I of Figure 1 C1210.002/M
i
-0.10
-0.15 -
C O 0.20 -
7 lE
\
h h 0.25 -
5 i A V 0.30 -
l I I I I ! '
-0.35 O.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 Time (sec)
Figure 3: Vertical Displacement Time-History for Node at End of Basemat (Point A) of Figure 1 (AH = 1.0 G)
C1210,003/M 05-87 l
l 84
Fr:qusney (eps) 50.0 20.0 10.0 5.0 2.0 1.0 0.5 t i i t liti t t it ti II I I II I I fIII II tt t t 10.0 (s i e isisi i s 'sas sie iiis inni aisi sisi i If i sai ins siis s i s i- 10.0 l 8.00 _
- 8.00 6.00 [ ~
6.00 5.00 _
5.00 4.00 _
_ 4.00
~
l
_. 3.00 [
[
3.00 c _
" 2*00 Free field ground spectrum 2.00 1.50 ,g......... #
b lir"'" '"""""" '-
1.50 o ,,,il,,,,,,,,. f"s ,
j 1
ti ""
)
gay i
'at;s,
~
j 1.00
\
,,,r p- ig i - 1.00
~
' 2 8 0.80 g-... . , , , ,., Basemat spectrum _ 0.80
< : ',,,,,,ad' s .[ 1 :
0.60 ........-
j ,jy A,,,,, . 0.60 0.50 ,. 1 6 ,,{-" 0.50 6
0.40 _
0.40 1
~ I 0.30 .",0.30 0.20 lI l' ' ' l Ill 'II' ' 18 ' I' ' ' ' ' ' ' l ' Il O.20 0.02 0.03 0.04 0.060.08 0.10 0.15 0.20 0.30 0.40 0.60 0.801.0 1.5 2.0 Period (sec)
Figure 4: Effect of Soll-Structure Interaction on Horizontal Spectra at C1210.004/M 05 87 l
)
l 4
l l
85 '
Frequency (cps) 50.0 20.0 '10.0 5.0 2.0 1.0 0.5 I I i If fI I1 01 1 I t f f 8 I t tt t9iI tI f f f 1 10.0 g ==
f
== - -
I
- - - - = ==== - - = =- == = = . - 10.0 8.00 - - 8.00
= ::
6.00 _ . 6.00 5.00 ,
5.00 4.00 - Free field ground spectrum - 4.00
- l (,,,.......,,oaan
. .'""""i.aI"'li, 3.00 ,.e ,
3.00 J%
S -
llI,,,,
c 3
~
.t" I
- Am J r li I.
I m 3 '"' "
g 1.50 g.,gni,,
ai ,
l j
Basemat spectrum
( f 1 50
...u'j j mi.,ll -
iii 7
> j g I.I
.; 1.00 ,
(. ...,, - 1.00
~
o 0.80 _ 0.80
=
, 0.60 L.20 _
. 0.50 0.40 _
. 0.40
~
~
0.30 .,
0.30 0.20 O.20 0.02 0.03 0.04 0.06 0.080.10 0.15 0.20 0.30 0.40 0.60 0.801.0 1.5 2.0 Period (sec)
Figure 5: Effect of Soll. Structure Interaction on Horizontal Spectra at Clinton (Damping a 5%, A g a 1.0 g)
C1210.005/M 05-87 86
Frequency (eps) 50.0 20.0- 10.0 5.0 2.0 1.0 0.5 10.0 ' . . . 10.0
- ,,.'n., =
8.00 - 1 - 8.00 y,yr- q
~ ~
6.00 6.00 5.00 .
'j r
g ,
5.00 4.00 - With uplift . 4.00 b , b 3.00 .
. 3.00
.~ . .
5 2.00 -/ I
,...ani.ii f -
2.00 7 , , , ,, , , , , , , , , .,,,,,j sy % .
3 7,,.p....., . . . . . .......
c 1.50 1.50
,o : Without uptlft :
g . .
- 1.00 -
- - 1.00
. : \ :
0.80 8 0.80 kg.
4 = =
0.60 . 0.60 0.50 .
3 0.50 0.40 . . 0.40 0.30 [ [ 0.30 0.20 O.20 0.02 0.03 0.04 0.060.080.10 0.15 0.20 0.30 0.40 0.600.801.0 1.5 2.0 Period (acc)
Fir,ure 6: Effect of Foundation Up!!ft on Horizontal Spectra at Fermi, Near Top of Drywell (Damping a 5%, AH s 0.60 g)
C1210.006/M 05-87 i
1 l
1 l
87 l
i 1
i Frequency (eps) !
50.0 20.0 10.0 5.0 2.0 1.0 0.5 1 10.0 g' . . . /. . . .' 'l '. ' . '.' / / . .'.'.'!...
. . . !. . . . '..'.'.'!.'. .' ' . . ' .' . . 1 0 . 0 8.00 -
8.00 i
~ ~
6.00 6.00 5.00 5' 4.00 A -
4.00 With uplift -
p 3.00 _
- ) _ 3.00 5
2.00 \ A / r k h 2.00
, N g g 1.50 _ ,J y _
1.50 d'
ii
,fi ,,, .
b 1.00 ,,,,,,,,,,,,,
Without upilft . 1,00 o 0.80 _ . 0.80
= -
0.60 _
g ,
0.60 0.50 0.50 0.40
. 0.40 N
~
0.30 0.30
- 0. 2 0 O.20 0.02 0.03 0.04 0.060.080.10 0.15 0.200.30 0.400.600.801.0 1.5 2.0 Period (sec)
Figure 7: Effect of Foundation Upillt on Vertical Spectra at Fermi, Near Top of Drywell (Damping = 5%, AH = 0.60 g)
C1210.007/M 05 87 88
JK/(
VIBRATIONAL EXPERIMENTS AT THE HDR:
GERMAN-U.S. COOPERATION C . A. Ko t , Argonne National Laboratory Argonne, Illinois 60439, USA L. Malcher, Kernforschungszentrum Ka rlsruhe D-7500 Karlsruhe, FRG J. F. Costello, U.S. Nuclear Regulatory Commission Of fice of Research, Washington, D.C. 20555, US A Preferred Mailing Address: Dr. C. A. Kot Argonne National Laboratory MCT Division - Bldg. 335 9700 S. Cass Ave.
Argonne, IL 60439 U.S.A.
89
VIBRATIONAL EXPERIMENTS AT THE HDR:
GERMAN-U.S. COOPERATION C. A. Kot, Argonne National Laboratory Argonne, Illinois 60439, USA L. Malcher, Kernforschungszentrum Karlsruhe D-7500 Karlsruhe, FRG J. F. Costello, U.S. Nuclear Regulatory Commission Office of Research, Washington, D.C. 20555, USA i
)
ABSTRACT .
As part of an overall ef fort on the validation of seismic calculational methods, the U.S. Nuclear Regulatory Commission, Office of Research, is collaborating with the Kernforschungt.entrum Karlsruhe, Federal Republic of Ge rmany (FRG), in the vibrational / earthquake experiments conducted at the Heissdampfreaktor (HDR) Test Facility in Kahl/Maf,n, FRG. In the most recent experiments (SHAG), high-level excitations were produced in the HDR f acility by means of an eccentric-mass coastdown shaker capable of developing 1000 tons of force. The purpose was to investigate full-scale structural response, soil-structure interacti,n, and piping and equipment response. Data obtained in the tests serve to evaluate analysis methods.
In the SHAG experiments, loadings of the HDR soil-structure system approached incipient failure levels as evidenced by high peak accelerations and displacements, local damage, nonlinear behavior, soil subsidence, and wall strains that exceeded estimated limit values. Also, the performance of different pipe hanger configurations for the Versuchskreislauf (VKL) piping system was compared in these tests under high excitation levels. The support configurations ranged from very rigid systems (struts and snubbers) to very flexible configurations (spring and constant-force supports). Pretest and posttest analyses for the building / soil and piping response were performed and are being validated with the test data.
91
l l
- 1. Introduction The Heissdampfreaktor (HDR) Test Facility (Fig.1) in Kahl/ Main, Federal I i
Republic of Germany (FRG), has been used since 1975 to perform vibrational, l thermal hydraulic, blowdown, and other experiments related to the design and safety of nuclear power plants. During the first phase of the HDR Safety Project (PHDR) vibrational / earthquake investigations, the HDR building and equipment were subjected to many low- and medium-level mechanical excitations (shaker, buried explosives, snapback). In the second phase, the HDR system is being tested at high levels of excitation. The centerpiece of these investigations is the high-level shaker tests (SHAG), which were performed at the HDR facility in June and July 1986. Their purpose was to investigate full-scale structural response, soil / structure interaction, and piping and equipment response under strang excitation conditions, i.e., under excitation levela that will induce significant strains in the structure and soil and produce nonlinear effects in the soil / structure system and piping. As with all HDR experiments, the primary intent of the SHAG tests is to verify and validate calculational procedures and analysis nethods. At the same time, the !
l experimental data provide direct information on the response and performance of structural systems, piping, and equipment under high dynamic loading; such information may have direct applicability to understanding the behavior of nuclear power plant systems.
The SHAG experiments were performed as part of the PHDR tests conducted by the Kernsforschungszentrum Karleruhe (KfK), and we(e supported by the FRG government and the U.S. Nuclear Regulatory Commission, Office of Research (NRC/RES). The NRC involvement relates primarily to a program on the validation of seismic calculational me thoc's conducted by Argonne National 92
Laboratory (ANL/ for NRC/RES. Additional participants in the SHAG experiments included the Electric Power Research Institute (EPRI), German as well as U.S.
industries, and the Idaho National Engineering Labora tory (I NEL) , which represented the equipment qualification interents of NRC/RES.
In this paper we describe the SHAG experiments, including the shaker design and test performance. This description is followed by some results of the SHAG tests. Finally, plans for future cooperative efforts between KfK and NRC/RES at the HDR facility are briefly outlined.
- 2. Shaker Design and Operation The shaker used in the SHAG experiments is a very large eccentric-mass coastdown shaker designed by ANCO Engineers, Inc., of Culver City, Calif.
[1,2]. Most of the design and functional calculations of the shaker's dynamic behavior were performed by the Fraunhofer Institut f t!r Betriebsfestigkeit (LBF), Darmstadt, FRG [3]. Safety calculations for the shaker, its mounting, and the HDR soil / structure system were performed by the engineering firms of Zerna-Schnellenbach in Bochum, FRG [4], and Hochtief AG in Frankfurt, FRG [5].
The masses of each of the two shaker arms (which form the eccentricity) are made up of an assembly of steel plates mounted on the arm baseplate, and can be varied up to a total of 45 tons each. The shaker is capable of developing in excess of 1000 tons of force. It is mounted in a very stif f frame on the operating floor of the HDR building to provide strong excitation to the entire HDR soil / structure / equipment system. The shaker was designed to develop maximum accelerations of the HDR building on the order of 5 m/s2 and maximum displacement of about i 7 cm. The total eccentricity of the shaker was designed to vary between 4,000 and 100,000 kgm. During operation, the 93
l shaker is brought to the desired starting speed (1.6-8.0 Hz) with the two arms in a balanced condition.- One of its arms is permanently fixed to the drive shaft, while the other arm is hinged to the shaft by a slightly eccentric support. In the balanced condition, the movable arm is fixed by an explosive bolt. After the desired starting speed is reached, the shaker arms are uncoupled from the drive system. Firing an explosive bolt releases the movable arm, which swings around and couples with the fixed arm, forming a large eccentric mass that provides a variable (in both magnitude and direction) force during coastdown. As the shaker coasts down through the fundamental frequencies of the HDR soil / structure system, strong resonances occur. During the entire shaker run, strong coupling and feedback occur between the HDR response and shaker forcing, resulting in a nonlinear coupled system. Details of shaker design and operation were discussed during the last two Water Reactor Safety Research (WRSR) Information Meetings [1,2), and e picture of the shaker is given in Fig. 2.
Preliminary teste were conducted in February 1986 to check the functionality of the shaker. Five tests were planned, three with the shaker arms balanced (no loads on the HDR building) and two with the shaker arms unbalanced (providing excitation to the building). The tests with unbalanced shaker arms were completely successful, proving the design concept and operation of the shaker. In the experiments with the balanced arms, however, it was not possible to reach the desired shaker f requencies because the air resistance was much higher than expected, and the torque that could be generated by the drive system was limited. Thus, the maximum frequency l
reached with no plates on the shaker arms was 5.25 Hz rather than the desired l 8.0 Hz. This problem was subsequently overcome by providing an enclosure for 94
the shaker, which substantially lowered the aerodynamic drag. It was also found that the eccentricity of the bare arms of the shaker was substantially higher at 5,700 kgm than the estimated value of 4,000 kgm. Since this would have generated forces much higher than intended during the 8.0-Hz tests, the mass of the arms was reduced by cutting out part of the baseplates. This reduced the eccentricity of the bare arms of the shaker to about 4,700 kgm.
Just before the start of the SHAG experiments, the functionality of the shaker was tested again and all systems performed as expected.
- 3. SHAG Experiments As stated earlier, the purpose of the SHAG tests was to investigate full-scale structure / soil, equipment, and piping response under strong vibrational excitation and to validate predictive analyses. While the interests of PHDR/KfK and NRC/RES include all aspects of the SHAG testing, most other participants focused primarily on the behavior of piping systems. In particular, the response of the Versuchskreislauf (VKL) piping system with different multiple support (hanger) configurations was of interest to all participants.
The VKL piping (Fig. 3) consists of a number of pipe runs ranging in nominal size from 100 to 250 mm. The piping is attached to the HDU vessel and associated manifolds and forms part of the experimental piping system at the HDR facility. The top of the pipe runs at about 28 m above ground level, ust i'
under the HDR operating floor (where the shaker is located). The original HDR hanger system provided primarily vertical dead weight support, and consisted of 11 spring and constant-force hangers and one threaded rod. The original intent was to compare in the SHAG tests the performance of this very flexible 95
f l
I conventional support system with a typical U.S. stiff support system containing snubbers and struts. Also, as part of the NRC/RES Equipment Qualification Research Program, INEL intended to evaluete the performance of a typical U.S. gate vu'.ve during SHAG tcsting. Accordingly, an 8-in. valve was incorporated into the VKL piping system. (Fig. 4). INEL then designed a typical U.S. hanger system, adding six snubbers and six rigid struts to the VKL hanger system and replacing one of the German spring hangers with a much j higher rated hanger of the same type to accommodate the added weight of the j
\
valve.
j l
EpRI and its industry associates intended to evaluate two additional hanger configurations. The first of these, designed by Bechtel Corporation, .
consists of energy absorber supports that damp out the motion of the piping through the plastic deformation of an assembly of steel plates incorporated into the support. The other configuration uses seismic stops, designed by R. L. . Cloud and Associates, that replace the snubbers. This system allows free motion until a certain displacement is reached, at which the pipe impacts the stops, limiting further movement in the given direction.
As part of Ge rman industry contribution to the SHAG experiments, Kraf twerk Union (KWU), offenbach, FRG, designed a hanger system for the VKL piping that uses, in addition to the dead weight supports, only five rigid struts placed so as to prevent large dynamic motions of the piping. With the agreement of all participants, two additional hanger configurations were incorporated into the test program. In one case, two viscoelastic dampers designed by Gerb, Berlin, were added to the KWU hanger configurations. In the other case, the six U.S. snubbers were replaced by modified viscoelastic dampers designed by ANCO Engineers, Inc. All the alternative hanger designs
{
l l
l 1
96 i
of - the VKL piping system were motivated by the desire to replace snubbers, which have proved troublesome in nuclear power plants. Therefore, the objective of these experiments was to compare and evaluate the behavior of the VKL piping system with the dif ferent support systems under identical loading conditions. Table 1 provides an overview of all the VKL hanger configurations l used in the SHAG experiments.
3.1 Test Plans Fr om the inception of SHAG test planning, it was the intention that the loading of the HDR facility be limited not by the excitation system but rather by the capacity of the building itself. It should be pointed out that the HDR building, which was constructed during 1965-69, is not designed according to modern nuclear power plant practice and its earthquake design simply follows Ge rman standards for conventional buildings. Thus, some of the experiments were designed to provide as severe a challenge to the HDR building as possible. Other experiments were intended to test all hanger configurations described above. Nearly all tests were designed to generate nominally the same peak force of 10 4 kN, at different starting f requencies of the shaker.
Higher shaker f requencies (8.0 to 4.5 Hz) were intended primarily for piping excitation, while tha lower f requencies (1.6, 2.1, and 3.1 Hz) were intended primarily to challenge the soil / structure system.
Results of safety calculations l4,5) and the functionality tests indicated that some of the test runs would severely challenge the HDR building. In particular, the 1.6-Hz runs were expected to strongly excite the building's rocking mode (nominally at 1.4 Hz) and the 3.1 Hz tests would strongly excite the bending mode (in which the inner and cater structures of 97
the building vibrate out of phase) at 2.5 Hz. In both cases it was thought that the 'oundation slab of the HDR building would be ' severely challenged.
Also, stresses in the embedded portion of the outer shield building walls were expected to be high. Thus, in the planned test matrix, the test runs that would challenge the HDR structure were placed at the end of the test sequence. Other groups of tests, in the frequency range from 4 to 8 Hz, were to compare the response of the VKL piping system with the different hanger configurations. Some of these test sequences were to be run in hot conditions (210*C), permitting the evaluation of piping response under both hot and cold conditions. Also, under hot conditions it is possible to pressurize the VKL piping system to 70 bars and thus better evaluate the performance of the gate valve under combined vibrational excitation and hydrodynamic loading.
3.2 Instrumentation A total of 460 channels of instrumentation were used during the SHAG tests, of which 340 were supplied by the HDR Central Measurement Facility (ZMA). Acceleration and strain were the primary measurements of structure and equipment response. All essential parts of the f acility were instrumented, e.g., accelerations of the HDR building (structure) and ground accelerations in two orthogonal radial directions and at two depths were measured. Also, all piping systems of interest were instrumented, as were major components and vessels. Strain measurements in the HDR walls, foundatiori slab, piping, and vessels were provided. Some of these measurements were intended as safety checks, as were some of the measurements of acceleration and velocity at the Versuchsatomkraf twerk (VAK) installation. The VAK is a recently decommissioned experimental power reactor that shares the same site with the HDR facility and is located about 100 m from it [6].
98
I l
The HDR instrumentation was supplemented by 120 channels concentrated on I the VKL piping system and the 8-in. U.S. gate valve. This instrumentation was provided by INEL under the sponsorship of primarily NRC/RES with additional 1
support from EPR1. Again, accelerometers predominated, with some displacement, force, and strain measurements. Specifically, acceleration of the HDR walls at the points of VKL pipfng support attachment were measured.
These measurements were intended to serve as input to posttest piping analysis calculations. Also, the detailed response of the piping was measured, as were all operating parameters of the valve.
3.3 Test Performance The main series of SHAG tests began as planned on 2 June 1986 with Test 34 (see the test matrix in Table 2), which was run with the minimum available eccentricity of 47Q0 kgm at a 6-Hz starting frequency. While the peak forces, accelerations, and displacements were close to the values predicted by LBF [3], the duration of the shaker coastdown was much longer
(~ 100 s) than predicted (40 s). The reason for this is the use of the enclosure around the shaker. Its effect is to reduce aerodynamic drag both during shaker spinup and coastdown. Because of the long test duration, more ene rgy is radiated to the surrounding soil, causing stronger-than-expected vibrational effects at some distance f rom the HDR f acility, e.g., at the VAK installation. However, the measured accelerations at the VAK were at least an order of magnitude lower than deemed acceptable. In order to avoid possible permanent damage to the VKL piping, the original HDR dead weight support system was modified by retaining the two rigid struts of the U.S. support system, which are located adjacent to the spherical tee (Fig. 3).
99
l
, i Additional safety calculations by Hochtief AG [5), as well as a "best i
1 estimate" soil / structure interaction calculation and detailed stress analyses performed by Weidlinger Associates, Palo Alto, Calif. [7), indicated that the 1.6, 2.1, and 3.1 Hz tests, when run at full load (104 kN), would provide a more severe challenge to the HDR structure than anticipated. Reexamination of.
the structural drawings also indicated lower capacity in the embedded region of the walls of the shield building. This led to a reordering of the test sequence, with the high-f requency tests for the piping being advanced in the schedule.
A test delay occurred at the end of the first week when Test 37 (at 2,1 Hz and less than half of full load) resulted in considerable shifting in a-section of the outer shield wall (around the equipment hatch), which was made of. concrete blocks. The concrete blocks had to be secured by a steel structure. Towards the end of the experimental rune, it also became necessary to separate the outer containment from a foundation' slab of the equipment tower in order to avoid severe local damage to the containment.
At the same time, the concern for the global integrity of the HDR structure necessitated a detailed evaluation of crucial test data (accelerations, displacements, and strains) after each test. These measurements were compared with predictions and estimates of structural capacity to guide the progress of experiments. All these factors caused the test period to extend to nearly 8 weeks. The tests actually performed and their sequence are listed in Table 2. As indicated, only the 8.0, 5.6, and 6.0 Hz tests were performed at or near full load (104 kN); all other tests were performed at reduced loads. Only the 4.5-Hz runs were performed with hot I
conditions in the piping system. All tests at 3.1 Hz and 2.1 Hz (at full l l
m
l load) were dropped to avoid challenging the walls of the outer shield building, which experience their most severe strains in the out-of phase bending mode. The 1.6-Hz tests, which involve the rocking mode, were limited to a maximum eccentricity of 67,000 kgm or about two-thirds of full load. The risk of producing permanent global damage to the HDR containment precluded the extension of the tests to the next higher level of 85,000 kgm. All testing was completed on 22 July 1986. Figure 5 compares the planned and achieved excitation levels for the SHAG tes ts. Also shown are the loads reached in Phase I shaker testing.
Ambient response measurement (RAU) tests were carried out before, during, and after the SHAG experiments. The RAU tests provide a measure of the changes in dynamic characteristics that occur in the HDR soil / structure system due to high-level excitation produced by the SHAG testing. The combined SHAG and RAU results also allow the investigation of nonlinear effects.
- 4. Highlights of SHAG Results The SilAG tests were planned to provide the maximum possible loading f or the HDR soil / structure system and the piping without inducing global soil / structure failure, which would endanger the integrity of the contain-ment. As indicated above, safety considerations for the HDR building necessitated some curtailment of the test plans (Fig. 5). Hence, the full capabilities of the large shaker could not be employed.
In spite of the limitations imposed on the testing, the overall goals of the SHAG tests were achievede Peak accelerations and displacements in the HDF building were quite substantial, reaching maxima of 0.4 g and 5 cm, respectively. Nonlinear behavior of the soil / structure system was clearly 101
observed. Much local damage occurred, such as concrete cracking and interior masonry wall collapse. Substantial amounts of energy were transferred to the surrounding soil, particularly during experiments challenging the rocking mode (1.6-Hz runs). This is evidenced by the high accelerations measured in the soil, cracking of soil (circumferential) away from the building, separation at the soil / structure interface, and soil subsidence ( ~ 10 cm) . Impact occurred between the HDR building and the equipment tower as well as the connecting ]
bridge to the office building. Strains in the walls of the HDR shield building approached or exceeded their estimated limit values. -Accelerations and motions of the VKL piping measured in the SHAG tests are comparable with values expected during strong-motion earthquakes. Settling measurements made af ter the tests indicate a maximum differential settlement of the foundation basemat of 8 mm, which corresponds to a horizontal displacement at the top of 1 the reactor building of 20 mm.
During the 25 SHAG experiments, about 460 channels of instrumentation were recorded for periods up to 200 s with sampling rates from 125 to 216 Hz. Thus, the total data accumulation was about 450 million data values. Data analysis is still in progress at all of the participating organizations and will continue for some time. However, preliminary results on the response of the site and structure, soil / structure interaction, and the behavior of the VKL piping system have already been reported [8,9] and additional results will be presented in the near future [10,11].
I i
4.1 HDR Site and Building Response Of major concern during the experiments was the response of the spent I
fuel-storage pool at the VAK. Figure 6 shows the decay of the peak vertical 102
_ _ _ - _ _ _ _ _ _ _ _ _ - _ _ _ _ _ I
acceleration in the soil with radial distance f rom the HDR containment. The peak value of 0.031 m/s2 at 100 m (VAK location) is approximately one-thirtieth of the acceleration measured adjacent to the HDR building and is more than one order of magnitude lower than the stipulated safety limit.
To illustrate the behavior of the HDR building during the SHAG experiments, a schematic cross-section of the building is shown in Fig. 7.
Three measurement locations (bullets) are indicated. Location A is at the top of the outer concrete shield building, location B refers to the top of the inner steel containment, and E designates the shaker location.
A typical sequence of events during an 8.0-Hz test with the smallest shaker eccentricity of 4,700 kgm to shown in Fig. 8. The shaker frequency versus time plot in Fig. 8a shows the shutoff of the drive system at about
-0.65 s, the start of the experiment due to the release of the movable shaker arm and the impact closing of the arms at 0 s, as well as the subsequent coastdown of the shaker. The decay of the shaker f requency is primarily due to energy transfer to the buildirg; air drag and bearing friction play a secondary role. The decay of the shaker force, corresponding to the shaker frequency, diminishes in 100 s from a value of 118,000 kN at 8.0 Hz to about 250 kN at 1.2 Hz (Fig. 8b). In the first 20 s, at frequencies above 4 Hz, the l
L acceleration response of the building is dominated by global torsion modes.
This is followed by a relatively quiet period until at 60 to 65 s the shaker traverses the out-of phase bending mode of the building (at about 2.5 Hz).
Finally the rocking mode (about 1.4 Hz) is reached after 90 s. This is seen from the response of the outer containment (point A in Fig. 7) in Figs. 8c and 8d. It should also be noted that both the characteristics and the amplitude of the response in the two horizontal directions show significant differences, indicating asymmetries of behavior.
103
The sequence of events for the SHAG experiment with the largest eccentricity (67,000 kgm) and smallest starting frequency of 1.6 Hz (Test T40.13) is quite dif ferent. Immediately af ter the start of the test and even 1 before the two arms are closed at time zero, the building responds with strong vibrations in its rocking mode.- After 10-12 cycles, most of the energy is dissipated through the damping of the building, and the shaker passes through j the resonance and subsequently coasts down with slowly diminishing f requency
[9].
To investigate the dynamic characteristics of the HDR building (i.e., )
modal f requencies and damping), it is necessary to eliminate the effects of the forcing, which varies as the square of the shaker frequency. Thus, transfer functions between the excitation force and responses in the building were constructed. Figure 9 compares the transfer functions for the top of the outer concrete shell (Fig. 9a) and the top of the steel containment (Fig. 9b) for a series of test runs with dif f erent shaker eccentricities and starting frequencies from 1.6 to 8.0 Hz. The results clearly indicate the nonlinear behavior of the building, particularly in the rocking mode, which is dominated by soil / structure interaction. Both shif ts in f requency and variations in modulus are indicative of nonlinear response.
Further investigation of the building behavior showed that the maximum horizontal responses did not occur in the direction of the global coordinates x-z in which the measurements were taken [9,10] (see Fig. 7 ). Depending on the test, the principal horizontal responses occurred in two orthogonal directions, x'-z', that were rotated from the global coordinate system by an angle that varied from 35* to 55*. Recasting the measured values into this principal (x'-z') coordinate system provided consistent results for both the j 1
104
frequency shift and damping of the rocking mode. This is illustrated in Fig. 10, which gives the rocking frequency and damping as a function of shaker eccentricity, i.e., load at resonance. The resonance f requency in the x' -
direction is consistently lower than that in the z ' -di re c tion . Its value drops from 1.35 Hz at the smallest loading to 1.05 Hz at the maximum load.
The observed rocking frequencies at minimum loading are consistent with values observed in earlier experiments conducted with forcing levels of the order of 500 kN [9]. The damping values shown in Fig. 10 indicate no specific differences between the response in the x'- and z'-directions. In general, the trend is as expected, i.e., damping increases with loading. However, the I
scatter of values is large with the highest damping value at minimum load exceeding the lowest damping value at maximum load. It should also be noted that all the damping values are significantly higher than those observed in earlier experiments where damping values from 4.4 to 5.5% were observed. The results presented here fully confirm the nonlinear nature of the HDR soil / structure system response during the SHAG experiments.
4.2 Pipe Response Investigations As mentioned earlier, the VKL piping (Fig. 3) was tested in a series of test runs with up to seven different pipe hanger configurations (Table 1). A schematic of the VKL piping is shown in Fig. 11 indicating typical measurement locations (bullets): two on the walls adjacent to the piping (A and B) and three on the pipe (C, D, and E).
To investigate the dynamic behavior of the piping, acceleration transfer functions between the piping responses and the wall responses are constructed. An example of this is shown in Fig. 12, which gives the transfer i
105
function between point C (x-direction) on the pipe to point A (z-direction) on the wall in the low-f requency range f rom 0 to 8 liz. The results are for the 8.0-Hz experiments with all seven pipe hanger configurations (see Table 1 for numbering). In each of the three parts of Fig. 12, the results for two different hanger systems are compared with those for hanger system 3, i.e.,
the stif f U.S. snubber configuration. It can be seen that the latter has a mode around 5.6 Hz with a damping estimated to be about 3%. Eliminating snubbers (Configuration 2) or exchanging them for other supports (Configurations 2, 4, 5, 7) leads to more response amplification in this section of pipe and to peak broadening of the principal mode in the 0-8 Hz frequency range. The very sof t Configuration 1, on the other hand, shows very I different behavior, with a multiplicity of resonances in the low-f requency range f rom 2 to 5.5 Hz. However, all of these modes exhibit less amplifica-tion than the peak amplification for Configuration 3. No resonance is observed for Configuration 6, because one of the viscoelastic dampers was attached in the vicinity of this measurement location.
No general conclusions on the relative merits of the different pipe support systems can be drawn on the basis of this single example. Response amplifications at other locations of the pipe may be very different. More importantly, pipe stresses and strains may follow a very different pattern.
This is illustrated in Fig. 13, which shows the maximum longitudinal, bending, and torsion stresses at two pipe locations (points D and E of Fig.11). These results are all for the 8.0-Hz test runs with an eccentricity of 4,700 kgm.
Note that the stresses for the relatively soft Configuration 2 are not significantly worse than those for the stiff Configuration 3. Also, stress levels for the very flexible Configuration 1 are quite acceptable. The 106
installation of viscoelastic dampers in Configuration 6 lowers the stresses at one location but increases them at the other. The replacement of snubbers by other support systems (Configurations 4, 5, and 7) does not seem to offer any significant advantage for reducing the stresses at the locations under consideration. In fact, Configuration 4 with plastifying energy absorbers exhibits the highest measured bending stresses.
l 1
- 5. Calculational Efforts Pretest and blind posttest calculational predictions were performed by a number of organizations; other calculations are still in progress [11].
Besides the already mentioned safety [4,5] and best-estimate [7] calculations for the HDR soil / structure system, other computational ef forts were concerned with the applicability and limits of approximate methods for nonlinear soil / structure interaction, quantification of safety margins in seismic design calculation and load determination of plant components, and the applicability of probabilistic structural analysis for seismic design and load determination of plant components. Some of these computations included the modeling of the
- coupled shaker-building response while others started with the measured force l
inputs of the shaker. Evaluation and comparison with measured values are nearing completion.
Similarly, computational efforts were also undertaken with respect to predicting the response of the VKL piping system. The calculations included static design calculations, quantification of the safety margins of the linear methods used for design, evaluation of the effect of different hanger configurations on the stresses in the pipe system, and validation of a piping code with multiple support load input. Again, evaluation of the results and comparison with the SHAG test measurements are in progress.
107
- 6. Conclusion In spite of the limitations imposed on the SHAG testing by the safety considerations of the HDR building, the experiments were an unqualified success. The wealth of data, which will require substantial analysis efforts, should provide insights into many aspects of concern to the nuclear industry. Specifically, the results will contribute to a better understanding of the nonlinear behavior of soil / structure systems under high-level excitation. The data lend themselves to the investigation of load transmission in buildings. The effect of different hanger configurations on pipe behavior at loadings equivalent to strong motion earthquakes can be evaluated. Most importantly, the data will serve to validate and verify analysis procedures for piping and soil / structure system response, including typical linear design methods, simplified techniques, and state-of-the-art nonlinear computational procedures. l In the SHAG experiments, the HDR building and equipment were repeatedly subjected to the same excitation (see Table 2). At the same time, multiple predictive calculations were carried out for single expe riment s. This combination of experimental and analytical results provides an opportunity to investigate the variabilities in response that are due both to parameter variability and modeling uncertainties. Finally, the SHAG data also provide information on the behavior and response of specific equipment, e.g., valves and snubbers, under earthquakelike excitation.
Because of the very fruitful collaboration in the SHAG testing, the PHDR/KfK and NRC/RES are already planning further cooperation in the upcoming HDR-SHAM experiments. In these tests the VKL piping will be subj ected to 108 j
l l
direct multiple-point excitation (using hydraulic actuators) at extremely high 1cvels. Pipe plastification and/or failure may be expected. The objective of these experiments is to investigate piping behavior at extreme loading l 1
(including the efiects of different pipe hanger configurations), to establish )
I seismic margins for piping, and to investigate possible failure modes in an j l
in situ piping system. Other goals include the investigation of equipment l (valves and snubbers) operability and fragility at ext reme excitation i I
levels. Also, the response and fragility of various pipe mountings and !
supports will be evaluated. )
Acknowledgment The collaborative ef forts of many individuals at a number of institutions l in the Federal Republic of Germany and the United States contributed to the success of the HDR SHAG experiments. The authors wish to thank all of their colleagues and coworkers, especially Dr. H. Steinhilber, LBF; Dr. D. j Schrammel, KtK; Mr. H. Wenzel, PHDR; Dr. H. Idelberger, LBF; Mr. R. Steele, INEL; and Drs. M. G. Srinivasan and B. J. Hsieh, ANL.
i 109
References
- 1. L. Mal c.he r, H. Steinhilber, and C. 3 Kot, "Heissdampf reaktor Phase II Vibration Tests," Proceedings of the '. S . NRC 13th Water Reactor Safety Research Information Meeting, NURE 1CP-0072, Vol. 3, pp. 15 9-17 8, NBS-Caithersburg, MD, October 22-25, 198A
- 2. L. Malcher and C. A. Kot, "HDR Phase II Vibrational Experiments,"
Proceedings of the U.S. NRC 14th Water Reactor Safety Information Meeting, NUREG/CP-0082, Vol. 3, pp. 295-312, NBS-Gaithersburg, MD, October 27-31, 1986.
- 3. H. Idelberger and H. Steinhilber, "Re chne rische Simulation des Shakerverhaltens bei exzentrischer Lagerung eines Shakerarms," LBF, Darmstadt, FRG, Bericht Nr. 5681, PHDR Nr. 4269/85, 1 August 1985.
- 4. "Sicherheitsrechnung Zur Lastabtragung der Shakerkrufte bei hohen An regu ngen ," Zerna, Schnellenbach und Partner, Bochum, FRG, PHDR Nr. 4285/85 July 1985. l 1
- 5. H. Werkle, W. Weber, and G. Waas, "Sicherheitsberechnungen Zur Standsicherheit von Baulichen Anlagenteilen des HDR bei Shakerversuchen T40," Hochtief AG, Frankfurt, FRG, PHDR Nr. 4.298/86, July 1986.
- 6. H. Idelberger and L. Malcher, "Auslegungsbericht -
Shakerversuche im Reaktorgebuude HDR-Versuchsgruppe SHAG," Vs.-Nr: T40, PHDR-Arbeitsbericht Nr. 4.280/85, October 1985, Revised March 1986.
- 7. D. K. Vaughan, R. Mak, and T. Piland, " Analysis of the HDR Containment Structure Subjected to High Level Shaker Loading," Weidlinger Associates, Palo Alto, CA, Report R866-3DV, March 1987.
- 8. L. Malcher, H. H. Wenzel, and C. A. Kot, "Erf ahrungen und Gebuudeverhalten bei den Erdbebenversuchen mit grossem Unwuchtshaker (SHAG)," 10.
Statusbericht PHDR, Kernforschungszentrum Karlsruhe, Arbeitsbericht 05.27/86, Beitrag Nr. 1, pp. 41-69, 4 December 1986. !
l
- 9. H. Steinhilber, H. Idelberger, and D. Schrammel, "Schwingungsverhalten der l Maschinentechnischen Anlagen bei den Erdbebenversuchen (S HAG) , " 10.
Statusbericht 05.27/86, Beitrag Nr. 2, pp.71-100, 4 December 1986.
- 10. L. Malcher and H. Steinhilber, " Earthquake Investigations at the HDR- l Facility," 9th International Conference on Structural Mechanics in Reactor ;
Technology, Paper K9/1, Lausanne, Switzerland, August 17-21, 1987. i l1. B. J. Hsich, C. A. Kot, and M. G. Srinivasan, " Vibration Testing and Analysis of a Multiple Supported Pi ping System," 9th International Conference on Structural Mechanics in Reactor Technology, Paper K17/4, Lausanne, Switzerland, August 17-21, 1987.
i 110
TABLE 1 VKL HANGER CONFIGURATIONS CONFIGURATION VKL SUPPORT DESCRIPTION OF SUPPORT NO. SYSTEM CONFIGURATION
.1 HDR Spring and constant-force hangers + two rigid struts (flexible system) 2 KWU Five rigid struts; simplified design concept 3 USNRC Six snubbers and six rigid struts (stiff system) 4 EPRI/EA Four Bechtel-designed energy absorbers 5 EPRI/SS Six Cloud-designed seismic stops 6 GERB Two viscous dampers designed by Gerb 7 ANCO Six modified viscous dampers 111
i I
l l TABLE 2 SHAG TEST MATRIX: RUNS PERFORMED 1
RUN TEMP., VKL SUPPORT ECCENTRICITY, STARTING MAX. FORCE, TEST NO. C SYSTEM kgm FREQ., Hz kN WEFl 34 20 USNRC 4,700 6 6,600 35 20 USNRC 4,700 8 11,800 1 36 20 USNRC 8,200 5.6 10,100 37 20 USNRC 27,800 2.1 4,800 40 20 EPRI/EA 4,700 8 11,800 20 20 KWU 4,700 8 11,800 2 60 20 GERB 4,700 8 11,800 3
50 20 EPRI/SS 4,700 8 11,800 70 20 ANCO 4,700 8 11,800 4 10 20 HDR 4,700 8 11,800 30 20 USNRC 4,700 8 11,800 31 20 USNRC 6,450 6 9,100 41 20 EPRI/EA 6,450 6 9,100 21 20 KWU 6,450 6 9,100 5 11 20 HDR 6,450 6 9,100 51 20 EPRI/SS 6,450 6 9,100 52 210 EPRI/SS 8,200 4.5 6,500 32 210 USNRC 8,200 4.5 6,500 42 210 EPRI/EA 8,200 4.5 6,500 6 12 210 HDR 8,200 4.5 6,500 22 210 KWU 8,200 4.5 6,500 12.1 210 HDR 8,200 4.5 6,500 7 14 20 HDR 33,000 1.6 3,300 16 20 HDR 54,000 1.6 5,400 8 13 20 HDR 67,700 1.6 6,800 112
le >
/
/:-
f.
, , /
,+>;'
/ p
. j
, 'r g
k' ,
5 n.
) Q , ' ['
9 u
[ f'", , ,
( ,. t g .;
p f.v I;h(h;a, C,. -f
- a
}:
3.,,
', . < L. i { - ! 17
"]
- 9 e f;( ' . , . e.
Y a
. ,;..,. 4 '
. . hMp,, '
f Figure 1. HDR containment building 113
?,{
.A
$5*' ma [' /' ,&#
E. ,:(. ,
t kl Myt s% g
'k j, .
i I n" ,
3 3 -
~f Uk$_ ~
';~
d' ,k 5 , .
h; i
s, 'j e s, z a -4
- ,, e _
Yf '?l
\;
. :w s..
ll' ?'
^
f _.o - '
~
y i ,,
l i 5.,[N i
,1 l :; j' 3 $ ;
s . n:t ,
'% \ % .g m t
- 9 e4 N
- l.,, $p, ' 'j\- k y . ..
Tl:, Dg I' ! .g. -
g
- Tf
/t
,% ^
'f
~'".;,
I e
l Figure 2. Large coastdown shaker mounted on the ilDR operating floor (30-u level) 114
DF 16-5"3 ~6 s
Flow -kKll y . K % tlO S.f/
tu
. ' HIE s.'
-- l i
f \d 7"8 FIow .
W.?
l? /- 8 {
HDU y b -
g I
b Valve detail
/ h Il $
Threaded rod See valve hanger e detall- " \"8 Strut _
n( y y Spring goe hanger- d Snubber . y ho*
9
%ky,/ s A
10" y 4"
l Figure 3. VKL piping system with U.S. stiff supports 115
4,) i *
- s. p= >
Q
[\
d d :: ~
% l.
hh ]
=4 $
t of W}
- 96. j l-f-[( l' 9 1
- . 4 i
p.
- I ~
' !l k '
s)
- 1
( ' 'I ,
( -
d I' l l i s
f
" g it
=
c.
.,,J..-.
lI
- j. sj
)
.. 1 - ,
Figure 4. U.S. 8-in. gate valve installed in the VKL Pi ping system I
116
Shaker Design N N N N N N
' II II I I (1981) oe ,- o o o 10,000 s
'- "***8 ,
s.
. . 3
/ ll Excitation Range jff
/>s Phase II (T40) , ,gjj! j . .: W (1986) 3'
- g-jg f8 p
3 41/g;:g$;;.'
c
'l.i .
- ts 8
8 u 1,000
~~ f: 4
,9 j p 4;3/ = : --
- - 2fy j} y g f;; ll' jf y
.yl?
f ~: :j - 500 kN (D f::_O ' :
w.
Sgh '. ;;:: I~
./ j
.e c,, :: - .
f
[f :hh'I '
Excitation Range kI;f-A-g g',: A Phase I (V63) Q Pf;g;jF;.# (1979)
- [j'hyk,;'
' 8" '
@)
100--
0.1 0.5 1.0 5.0 10.0 Frequency (Hz)
Figure 5. SHAG excitation region (test design and actual) f 1
117
t i )
m 0 t n
i
.!!:: i l 0 e L 1 m e
y r u
t s f
e l a
e a m S l a
c K l i
t A r e
V (
v l
l i
o s
R D
l
) H e
(
m h t
0 e n l
5 c i e
n d t
a u t
! i s i l
p D m a
n o
i t
a r
b i
v f
o y
a 3
c 0 e D
1 9 6 3 6 0 0 0 e r
u g
_ - mM N E E 9. e u 2 8 o < i F
- m
9^
[
_c lB h - , N '
i
. ses-N! T t~ f =t 41 !
- I'tF a
w
- h , a ol ! jo I
?
1 !
l Shaker
/ r i,l n
j3 I
M d E
l'j I
?
M '
l }!j wc 1 i a 5 %
L' i 15 IIki 5 i
- l. b.
W i.
l . , (4.".
=== if e .
Y,lg,i l .gp Y g[~ r4 4 fb +
~; ; e 7 y- i - , . J/
'n '
a M U p i _S O 4g n.
a y_ ;, ; ,
[ h>ihMNk
^
14
- wn i e 5 1 11F ;-
...u
[
4 L.. A l
h
$s
- j
} ,
~
~ A U 1 n.. l _,
mm.f ! it IM 9 g rm.
r <;<
r i
, M, II4} .
l
-_ 4 ,--h
_r. .
, i I J A $)
Z X l l
Figure 7. HDR containment cross-section with measurement locations 119
I' ll> l 0 0 1 1 b d 8 b f
o 6 b p 0 0 o 1
- 1 t e
h t
) )
s s t
(
4( d e e a m m i
T i
T s n
o i
t -
a
,2 b r e
l e
c c _
a .
,0 b l a
t n
! o 2
6 o 6 2
0 0
0 0
z 0 i 1 - 1 0 5 0 5 0 r
- 3 3 1 1 o
,S' E6 e22 - N h0 1
d .
n0 a4 0 0 ,T 1 1 et cn a c re om fi r
re
.8 ep b
kx aE h
sni yg
,6 cn b ni 0 0 ed 1
1 ul qi eu rb f
)
s
)
s R
,4 (
b ( rD e e eII i
T m i T
m k ae hh St
,2 b 8 e
r u
g i
F
,0 b 6 3 0 0 0 0 0 0 0 5 0 5 0 3 1 1 3 E5 c3?t i - -
- x wNO i l' (Ill 1'
20 - --
/.
Point A ,'
s v ,. : ~,
s t
[
o f, , \ ',
r 15- ,
l l;;
l I
z x
- s[ 10-04 ,
,,l,
,0 's, s
v>
s ,., 1.I , i E
- lQ\n,',',\ ', I
\ s
,,a. ,o l \,
, + ,*: , ;. r 5- . ' .'l . . ,I '\s j ,,i ,ll ,,Y,'U.. , f.3.ww .
x
! ,.Tf . _ -- N i / -
~~- s
\
\
,l',s. :l l ,7 l
%,. ,#w; i, ,' -
0 . . . . . .
0.00 0.50 1.00 1.50 2.00 2.50 3.00 Frequency (Hz) 20 Point B
/\
v ,i ti i
o !\
r 15- ! \
! \
,! \'
z x i +; A ,-
- r. , , o ;,. ,
m 110- !! y :
, ,a
,v.
i c) :ti r ,, if s
E lll Vt '..
ill it a's 't s
- s,
- r .,
.e . s .
/ s' 5- !/>/ . %
j s
jf ,f j N :,.,
l lll!/,l 'Q
.\'~,,
.\b
'N,
' ~ ,'
llll,;i s' ',,
', Ii
,-,,,> .; ,y,-
g- .-
u ., \,'
0 .
0 0.50 1.00 1.50 2.00 2.50 3.00 ;
Frequency (Hz)
Figure 9. Transfer functions for the HDR building response for six different shaker tests; z-direction 121 !
i.
1 1.5 -
1.4 - ik _
X' - Direction g o Z' - Direction b 1 .3 -
U 5
51.2- 8 Q) tt 11-4 1.0 -
0.9 i > > , , , , -
0 10 20 30 40 50 60 70 Eccentricity (kgm) *1000 )
15-r
@ 10-
.E' g -
E e
O o X' - Direction 5- o Z' - Direction
=
0 . . . . > > >
0 10 20 30 40 50 60 70 Eccentricity (kgm) *1000
)
l Figure 10. Variation of rocking frequency and dumping with shaker excitation 122 1
_ _ - - - - - i
I i
B cc '] ,
.?<"
'-- ' l f '
l
(
l !
l I
l C !
! I o
i i
E O
i i
i D l i
i i
i i .
i a
Figore 11. Schematic of VKL piping with measurement locations 123
50 40-2
/
- fi "a 30- ij E
I
- r '\ ..' 7 ?
%E 20' ) ri b3 ll ;\
- , ;T
, il' ' -
l 10- j
/
' .)
,l ' . \,
\ \ 'A
- ..~.. J - .- ' '
. _,,, g.f . .... . - " '.~ .\ .
, ,g 0 , , ,
O 2 4 6 8 Frequency (Hz) 50 fN 40- !
);
- h. \'#
\
"l30- 6 I\ \
E e !' : '\
~~ ! \
I h20- / ,' g I \s 10-f l ll
&' \\
3 l t.
i'
/ ' t
., h ';
O , , ,
0 2 4 6 8 Frequency (Hz) 50 i-7 40-
!I
$ 30- '
E t -
L'3 m'
f20-
. l 10- ,..'
6 .
0 , , ,
0 2 4 6 8 Frequency (Hz)
Figure 12. Acceleration transfer functions: pipe response to wall response (Tests T40.*0) 124
-A Point D
-1 HDR
~
2 KWU 3 NRC 4 EPRl/E.A.
" 5 E EPRl/S.S.
E 50- 6 GERB 3
s 7 ANC0 -
z
~
2
~~
~
1 E
'5~~
05 _
1 1y-
?
5
~~
L 61 34 1
Tension / Bending Torsion Compression A Point E m
^
m 6 E
E -
1 3 50- 3 m .
5 7 b -
1 0 ~
1 6
5 --1 2_14~
1 1 ~
51 7 L l Torsion Tension / Bending Compression Figure 13. Maximum stresses in two pipe cross sections: Tests T40.*0 for hanger configurations 1 to 7 125
JK/5 )
VALIDATION OF SEISMIC S0IL-STRUCTURE INTERACTION ANALYSIS TECHNIQUES WITl! LOTUNG EXPERIMENT DATA - OVERVIEW 0F EPRI/NRC RESEARCH PROGRAM Y. K. TANG, R. P. KASSAWARA H. T. TANG I. B. WALL Electric Power Research Institute, Palo Alto, CA 94303, USA M G. SRINVASAN, C. A. KOT, B. J. HSIEH Argonne National Laboratory, Argonne, IL 60439, USA A large-scale seismic experiment facility has been constructed in Lotung, Taiwan, by the Electric Power Research Institute (EPRI) and Taiwan Power Company (Taipower). The objective of the experi-ment is to obtain actual seismic response data to validate soil-structure interaction analysis techniques. Eighteen earthquakes have been recorded at the site since the completion of the facility.
A method validation program sponsored by EPRI, NRC (through Argonne National Laboratory, ANL), and Taipower has been initiated. The program involves blind prediction and result correlation of recorded responses for forced vibration tests and selected seismic events using various existing SSI methods. The results of the program expected to provide basis for a more realistic SSI practice for the industry.
- 1. Introduction Seismic soil-structure interaction (SSI) is a complex phenomenon which involves the intSrplay of building structure and its surrounding soil medium during a seismic event. SSI effect is particularly considered important on massive and embedded structures such as nuclear power plants. In the past years, various seismic SSI analysis techniques have been developed and used for nuclear pcwer plant design and licensing. The methods vary from simple spring and dashport representations [1] to more complexed finite element and substructure impedance approach 12,3,4,5]. However, due to the lack of well-controlled actual earthquake data, relatively little effort has been devoted 127
to validating the assumptions and approximations employed in developing these analytical techniques. Such lack of method validation has created constant debates among the SSI analysts on which method to choose and it, in turn, has led NRC to adopt an overly conservative approach in plant licensing review [6]. ;
In view of this, EPRI, with cooperation from Taipower, has undertaken a large-scale seismic experiment in Taiwan. Two scaled concrete containment models with full instrumentation have been constructed near Lotung where strong motion earthquakes are known to occur frequently. Since the completion of the model and instrumentation in October 1985, eighteen earthquakes have been recorded. With this database, a cooperative SSI method validation program was initiated in 1986 by EPRI, U.S. NRC, and Taipower.
In this paper, essentials of the Lotung experiment, including model configuration, instrumentation setup, and site soil testing, are described.
The forced vibration test and earthquake data are discussed. The objective, approach, and scope of the method validation program sponsored by EPRI and NRC are presented.
- 2. Description of the Lotung Experiment Facility The experiment model of the facility consists of two scaled concrete containment structures (one 14 scale and one 1/12 scale) located 50 m apart inside one the Taipower's substations. The site (Figure 1) is also located within an existing strong motion seismic array (SMART-1 array) sponsored and operated by the National Science Foundation and U.C. Berkeley [7] as part of the National Earthquake Hazard Reduction Program.
The configuration of the 1/4-scale containment structure (Figure 2) is based on the containment building of Taipower's Maanshan nuclear station. For easy construction and forced vibration test, the hemispherical dome of the ct,ntainment structure is represented by a 1-m thick flat concrete slab.
Inside the 1/4-scale model, a mocked-up steam generator and a simple piping are installed to monitor internal component responses. The design configura-tion for the 1/2-scale model is similar to one of the models in the SIMQUAKE test series (8]. This would allow direct comparison of SSI response data to study the effect of different types of ground wave motions. The 1/2-scale model is not included in the current scope of the SSI method validation program.
128 l
- To establish a well-controlled SSI database, a total of 130 channels of instruments are-deployed to record the earthquake responses. The instruments-tion consists of in-structure, ground surface, and downhole triaxial acceler-ometers as well as interfacial pressure transducers. The ground surface accelerometers consist of an array of three arms (Figure 3), each containing five stations. .The downhole stations consist of two vertical arrays (Figure 4) located one at near-field and one at far-field to the 1/4-scaled containment building. The deepest downhole accelerometers are placed at a depth of 47 m below grade. The in-structure instrumentation of the 1/4-scale containment consists of 10 triaxial accelerometers (Figure 5) and 13 interface pressure transducers (Figure 6).
- 3. Site-Exploration and Soil Testing The soil at the site mainly consists of saturated sandy silt and silty sand with shear wave velocity in the range of 150 m/see to 300 m/sec. Soil exploration and testing have been conducted by Taipower to provide detailed information needed for the SSI validation program. The tests were carried out in two phases with 12 boreholes drilled at the site (Figure 7). Both field tests and laboratory tests were conducted. The field tests include standard penetration tests and cross-hole and up-hole measurement of soil wave velocities. The laboratory tests include both undisturbed and remolded sample tests to obtain soil index properties as well as engineering properties.
Resonance column and cyclic triaxial tests were performed to obtain all the dynamic properties of the soil needed to describe the strain-dependent parameters used in the SSI analysis.
- 4. Forced Vibration Test of the Model Structures The objective of the low-level forced vibration tests is to define basic dynamic characteristics of the soil-structure system in an as-built condition before the system underwent strong earthquake shaking. The results of the tests are to be included as part of the database for the SSI method validation. They are to be first predicted by the analysis and subsequently to be used for refining the predictive models.
The testing program, partly sponsored by NRC and partly by EPRI, was conducted by ANCO Engineers in two stages. In the first stage, the tests were 129
l l
1 I
performed on the 1/4-scale containment model with .only its basemat com-
-.pleted. 'In the second stage, the teris were executed when both containment models were completed and foundatiore soils were backfilled. A single 1 eccentric-mass shaker was used to generate steady-state vibrating motions.
For the basemat test, the shaker was placed on top of the mat,.while for the complete model test,-the shaker was placed on top of the building roof slab.
~
For. each directional excitation, the eccentricity of the shaker rotating i masses varied from test run to test run to generate a sufficient wide frequency range of excitation (1 Hz to 30 Hz). The measurement locations were selected so that essential response modes of the system can be determined from the test data. Figure 8 shows the shaker location and the placement of the 20 accelerometers on the 1/4-scale containment structure during the second stage test. Additional sensors .were also placed on the internal piping-steam generator model during the test. Table 1 summarizes details of the test matrix for the second stage test of the 1/4-scale model structure.
Before transmitting the database to the SSI analysts, ANL conducted an extensive study to investigate the quality of the test results as well as to provide proper interpretation for the test data. An experimental modal analysis of the test response data was also performed to assist the SSI analysts in identifying the predominant modes and mode shapes of the soil-structure system.
- 5. Recorded Earthquake Data The Institute of Earth Sciences of Academia Sinica in Taiwan is responsible for maintaining, collecting, and reducing the earthquake data for Taipower and EPRI. The reduced earthquake data are stored in a digitized time history form for analysts to use. Eighteen earthquakes with Richter magni-tudes from 4.5 to 7.0 have been recorded since the seismic instrumentation was in place. The sensors are activated by a common seismic trigger with threshold acceleration level set at 0.01 g. The earthquakes recorded to data consist of both near-field and far-field events with epicentral distances varied from 4.7 km to 78 km. The database also contains earthquakes having hypocentral depths from 1 km to 79 km. Table 2 summarizes the eighteen events.
130
Among the earthquakes, events 4, 7, 12, and 16, which occurred on January 16, May 20, July 30, and November 14 of 1986, respectively, are the most significant ones. The peak accelerations recorded at the free-field ground surface are 0.25 g, 0.20 g, 0.18 g, and 0.2 g, respectively.
In the current scope of the cooperative program, only one earthquake event is to be selected for SSI method validation. The May 20, 1987 earth-quaxe, which oroduced the most complete recordings, has been chosen for the f study. Depending upon the budget situation, additional events may be selected. For the selected event, the earthquake records are provided to the analysts in two stages. First, the triaxial accelerations recorded at the free-field ground surface at the far-field downhole location are used as input control motions to predict the SSI responses. Second, after this " blind" prediction is completed, the entire set of recorded responses (building, downholes, and free field) for the event is released for correlations of result and method evaluation.
- 6. SSI Method Validation Program The scope of the current EPRI/NRC cooperative program focuses on validating U.S. industry practice of SSI analysis using existing method-ologies. The methods selected for evaluation include the lLmped-parameter spring-dashpot approach, direct method using finite element, and sub-structuring impedance techniques. Computer codes used for the validation study include FLUSH, CLASSI, SASSI, and SIM. Each method will be used to make blind predictions (recorded responses not available to predictor) of the site, building, and equipment response to the given input. The program utilizes a round-robin approach so that independent assessment on various methods can be provided by more than one analyst. In the program, EPRI is sponsoring four industiy practitioners including Bechtel, Sargent & Lundy, Impell, and EQE; while NRC, through ANL, is sponsoring three university groups involving original method developers from U.C. Berkely, U.C. San Diego, University of Southern California, and City College of New York.
The validation program involves simultaneous independent effort by each '
analyst to devise models, perform blind prediction calculations, compare predictions with measurements, and finally to evaluate and assess the methodology. The essentials of the programs consist of the following four major elements:
131
- Given soil, structural, geological, and geophysical information typically available to a nuclear plant design, construct a best-estimate analytical model and perform blind prediction of response to forced vibration tests (in the NRC/ANL program, two models were developed using different levels of available soil information).
- Compare the predictive calculations with the forced vibration measurements and develop an improved soil-structure model by correlating dynamic response characteristics between test and analysis.
i
- Given earthquake control motion at free-field ground surface, perform blind prediction of site and structural responses by using common industry SSI practice, which considers soil parameter variations and vertically propagated shear waves. At least two models are to be used in these predictive calcu-lations. One is the original best-estimate model, and the other is the forced-vibration correlated model.
- Compare prediction results with recorded earthquake response data for both original and modified models. Conduct an engineering assessment of modeling technique and analysis method used.
Details of the EPRI and NRC/ANL programs are shown in Figures 9 and 10, respectively. Although there are some differences between the workscopes of the two programs, the approaches undertaken to validate the existing SSI methods are consistent.
- 7. Concluding Remarks The round-robin SSI method validation program provides a unique opportunity for the industry practitioners as well as method developers to
]
evaluate and assess various SSI methodologies in a systematic and independent fashion._ Since blind predictions are essential to this effort and are 132 1
_ _ _ _ _ _ _ _ _ _ . _ _ l
currently being performed by the program. participants, no measured data and analysis results are reported in this paper. ~A workshop is planned in which all the results and findings will be presented and discussed. The results of the program are expected to form the basis for evaluating.various SSI methods as to their conservatism and sensitivity -to various assumptions and proce-dures. Ultimately, the conclusions of these assessments will form the basis for modification to the USNRC Standard Review Pfan and hence a more realistic SSI practice and improved plant licensing procedures.
References
- 1. Whitman, R.V., and Richard, F.E., " Design Procedure for Dynamically Loaded Foundations," Journal of the Soil Mechanics and Foundations Division. l Proceedings of the ASCE, 1967.
- 2. Lysmer, J., Udaka, T. , Tsai, C.F. , and. Seed, H.B. , " FLUSH - A Computer Program for Approximate 3-D Analysis of Soil Structure Interaction Problems", Report No. EERC 75-30. Earthquake Engineering Research Center, University of California, Berkeley, November 1975.
- 3. Wong, H.L., and Luco, J.E., " Soil-Structure Interaction: A Linear Continuum Mechanics Approach (CLASSI)", Report ' CE79-03, Department of Civil Engineering, University of Southern California, 1980.
- 4. Lysmer, J., Tabatabaie-Raissi, M., Tajirian, F., Vahdani, S., and Ostadan, F., "SASSI - A System for Analysis of Soil-Structure Interaction", I Report No. UCB/GT/81-02, Geotechnical Engineering, tlniversity of California, April 1981.
- 5. Miller, C.A., and Costantino, C.J., " Soil-Structure Interaction Methods -
SIM Code", Brookhaven National Laboratory, NUREG/CR-1717, Vol. III, September 1979. I
- 6. U.S. NRC, " Standard Review Plan", NUREG-0800, July 1981.
133
I l
.7. Tsai...Y.B., and' Bolt, B.A., "An Analysis of Horizontal' Peak' Ground Acceleration and. Velocity from SMART-1 Array Data", Bulletin of Institute Earth Sciences, Academia Sinica, 1983.
l- 8. Higgins, C.J...et al., "SIMQUAKE II: A Multiple Detonation Explosive. Test l to Simulate the Effects of Earthquake-Like Ground Motions on-Nuclear Power-Pla'nt Models",.EPRI Report NP-2916, October 1983.
l:
l l
)
l l
134
e u
q e
r 0 0 0 U 0 0 0 u F 8 7 6 7 0 7 0 6 8 6 7 0 6 7 4 7 N g 9 5 6 0 0 6 5 a 1 91 2 4 4 4 4 2 nt I d n
.E e
d t eA t
t i
py sc A n e
eu cq r e or FF 5 5 5 5 5 5 5 5 8 8 9 3 2 2 3 9 g 8 8 3 1 1 1 8 3 n 4 4 6 1 1 1 1 6 i
t r
a e t r S u t t
- A c
- u r
3 t
f S o . 7 9 1 q 6 8 l 7 7
. e 7 7 F 5 9 7 4 7 d or NF e
t e
. l y p c
n m
e o C
u 7 0 q z 2 6 7 0 2 1 6 eH 5 0t 0 0 0 0 1 0 f r . .
o F n 0 0 8 0 6 6 0 8 3 3 1 1 1 1 gi n n
i o d i n t E a z t H i n
c gi x
. n E 0
1 i , 0 0 0 0 0 0 0 t y 0 0 0 0 0 0 0 0 r c .
an 5 5 9 5 I l 5 9 t e 1 1 1 S u q
e e r l F b a
f T o
ng on ii t c I 2 cr R T R R R T t T d e o e rF c i a D p a
y
. l y l t a i a u
) q c -
i g 5 5 0 0 0 0 0 0 t e rK 5 5 0 5 5 5 5 0 ) s t h et n n 0 0 2 1 8 8 1 2 t W o ei 1 2 2 1 u - n c ot c S s e E - a r hE a t(
r r n e 2 2 3 3 3 3 3 3 ol e k 1 1 1 1 1 1 1 1 N at a - - - - - - - - (i e h K K K K K K K K t n S M M M M M M M M l ne aeu i gq d n e n . aa r uo R Tf R N I i 1 i i i i i
= = e h
RTT t .
s o . . .
eN I 2 3 4 5 6 7 8 T 123 ww
l1l
)
m k
G (
N U A T 7 1 9 7 5 4 2 2 0 1 0 2 7 9 T L A A A A O E 5 9 6 3 8 1 3
6 6
9 6
5 6 6 5 /
N 4 / 7 / /
N 7 N N L D 4 2 1 2 T
A )
g T e N d E (
7 0 6 1 3 7 3 6 1 1 0 5 5 3 M . 9 9 4 3 4 1 5 6 2 4 1 0 7 9 I M R I 6 4 0 1 9 4 4 1 6 2 0 1 A 8 A 3 A A 3 6 5 7 9 9 4 6 9 3 / 1 / 7 / /
E Z A
0 1
6 1 1 1 1 1 1 1 1 N 1 N 1 N N P
X E
)
C M I (
M G 3 3 5 5 7 4 5 2 5 5 0 2 A 9 A 0 A A S A / / / /
I M 5 5 5 6 4 5 6 6 4 4 5 6 N 4 N 7 N N E
S E )
m L k A (
C H 9 5 0 2 2 9 2 4 4 8 1 5 8 4 r S T 9 1 0 2 3 8 8 8 1 8 0 5 2 9 e.
- P A A A A tr E E 3 1 9 0 0 0 5 1 1 0 2 1 / 2 / 6 / / ne _
D 1 1 2 N N N N et G 7 1 1 cn _
R ie pc A
L e _
l oe Y td o
B rm e
D te E nl ea D cc R s O l -
e4 C d/
E 1 4 1 7 6 1 8 9 7 4 1 3 0 5 o1 R 0 0 3 9 6 5 5 6 2 8 5 5 7 1 m m
S 4 9 7 1 4 8 9 1 6 1 3 7 1 1 eo 2 3 1 3 1 5 4 3 2 3 3 4 3 0 l r E : : : : : : : : : : : : : 5: :0 : 3: :1 6 0 8 af c
K 1 0 5 4 7 4 5 2
7 3 2 5 0 3 2
3 8
3 5 2 4
0 1 se A 0 3 2 0 1 1
- : : : : : : : : : : : : : : : : : - c U 5 3 5 3 7 2 5 5 8 3 0 1 1 1 0 1 3 0 4 n Q 1 0 0 1 0 0 0 0 1 2 0 1 1 1 0 2 2 0 /a 1t H E M 0 6 7 5 9 8 0 0 1 6 7 0 0 0 5 4 4 5 s T I 2 2 0 1 2 0 2 2 1 1 1 3 3 3 0 1 1 1 mi R T 7 1 1 od r
A N 9 0 1 1 0
3 0
4 0
5 0
5 0
7 0
7 0
7 0
7 0
7 0 0 8
0 1
1 1 1 fr E I 0 1 1 e
G 5 5 5 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 ht I 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 t n R 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 ue -
O 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 mc 1 ii zp E AE f
L B T :A A N E MT T V 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 I L -
E 1 1 1 1 1 1 1 1 1 ZE AD s
s
i i ;j
~
[% ,
f d.,.
Q -
_ , A jf
'g
~
- ~
' ~j m&
Y.~t
. ,1 n o
i t
1 a
- c
- o L
1 l T -
e e R Td Ro t
i fp S
A AM M t M S s e
S 'A T 1
g n
u e t r o u L g i
F
. g i n ep u
@ia t l
a T T g a n n u . i h h c g C
l ,'
a n f T u o i
e n h s c a i n o l i a b a K p
u T
e e e R
_ y i i'l.
[:'?.
I L
1 1
I l' M 1
( 16'-3" >
3'-6'
/
[
- /
/ I /
' /
' /
/ /
' /
' /
' /
' /
/
'.1 '
l- Main Stekm Line $
/
l' ./
j tean Generator ,
l- .__t L
/
50' % /
' /
- i /
' /
/
l- _ _ 17'-6"- $
I ,
T PWAW/j .
s NvWA%
' /
- /
15' ; "
a
/
- i 5' $
v k k $
v v u
///f/4 V // V////////f)
/
4-- --+
l b I Steam Generator Containment Figure 2. Cross Section of 1/4-scale Containment Model 138
!: n }lN Do.e- ARM 1 m.
s tom
-- m l s.io-l g 5 accelerograph 3 06m!l 1
- Figure 3. -The Surface Instrumentation Array
- & .s., - ;;
DHA DHB N
/6\/V(. 4-m -.
'A%\
te. Ls F
a uT. e a nr. a e !!" - a Figure 4. The Downhole Instrumentation 139
_ e n N E c o a i
_ t f t n ur a
\
e s t n
_ m e e
_ e d m
\ s u
dd a is o o b r e t
_ r h s e t n I
n n
- u o l s s a i
e g g e c
- a e g g f a
r
- e e e
%u r t
r u n
, s s I s s
/o e er e r h r p p T e e
/ t o
n e
d d e
o n
t 6
e r
e o u g
i F
L E
D O
M E
4 n
/
o N 1 i
t a
t n
e m
j s u r r
- N N a
t e
e t
s n
m I o
r e l a
I I lc r c u g_ a t e c 1Il t
u o r t
n Qd e S e e a h f /f T W
[S 5
e r
u g
i F
^
_ L
I s
e s lo l
e h o e r
h o e
r o B B e s e a i I S $ s h ha P h m O ,d P d g i i
t t n a i n s o c
r w _
U i F S e
\
_. ge W\
'6 l
H B
4-K4 4
e d p o 0 M m 4 s u 1 e
/
2 1 S l 1
8 o 0 L h 4 e r
o N 3 8
B 1 l 3- i o
g 4
y"H B 3
S f
3 o 8 n l
e 5 o d i o t M4 m u
a c
- 5
/
3 S 2 o
- L g
2 9 1
0
'2 .
2- j 1
') 5 7
e
'y, H T B
1, 2 3 0 r
u
) . [ (3, L N6L 7 i
g W. j 0 F 0 \
2 G g 7
(
i
- ? ~ -
L 1T 1
H B -
- {
s
,l
=l t
I 4
'k Containment Model 1 N Driving Point N O
' s. /,7 Sh
.-e MN7 0 3
's J i 2
, Half-height
[ofCyl.' Wall ~ ~~~ ~
/ !. -
\10 9 j ,16
- I
- IS 4 1F" C *";" d N Total of 20 gf N 20 6 -
Accelerometers 176
/(19 bl3
" 18 '
-b 12 11 l!
Figure 8. Vibration Test Instrumentation of Structure
).
l 1
l i
142 l E_____ _---
Z Z M WC t-OMM y
==t-- J - t- LaJ Q E
W-=
QW LaJ 3 M
L&J LaJ z .-MO 3Z4H M z ga MHW <
m Ego E 6- O
- eEWNW -
sh y
mm
==E OJM E-E m E .== E L4J E
~
gg A L4 M C t== LaJ LaJ O t- LaJ 6-= N g gg ^
L.J V M > L.J L.J cg a:t M M M LaJ - :'D C --DDJ ,,,
ce @ E 644 E M CMM C M J LaJ g .
EOM b WC W%A zu E W5CO e
AEWE
>O y
en AE-bL ft "
4 E LaJ t- Q.
Ld < 1AJ wO O
Cb < =) 44 LaJ h OM #
E< 4 M >=- xz c D> 64 c:t t- O as Ln eX MC MA g a:f, O LaJ 6- <M ,
> LaJ LaJ g E W M E
X LaJ U
W E Q d~ "
E 3c "W J . J .
W LaJ Cd LaJ 6-= J CE E E C3 8-- O Z C3 < LaJ < < 7 LAJ Q OOOM ZOO U L.J > c M- Z == E - Et-set -
J 6- 3 LL tJ lll3 2 6- J J gr3 L.J .7 c Cf Cf - W OU J V dlr -- % Z D LaJ C Q A- CD-%D *O--M>M ,,,J MC O - LA. MW LaJ LaJ W 46J g CC CE >M O CM e e LaJ g6 e e O A L&. A e e 4
IL JL N .m,-
I C y n .
h e a:C l
$ D; t_ .- ._
de Ox
-a MW " J ., ,
O t-- MC % M LaJ === W -
[.
- O M M W =
b z cb LaJ 2 wO
- h D CC === A.L-
.O t - D 6-OO W LAJ ZE ct M t-E CC O 7' 2 CD n ct J O LaJ C) - 6-=
- >>A La.J Ld
. CE === -
t a t-- O >- LD A D- t- - M LD C ==.-
Z #
CL z L.J Z O LaJ D LaJ 2- D- LAJ J LJ t- t- La.J w ct w >= C Cz' M O ct aaJ 6- LaJ =W Q LJ J LA.
O M C3 M LD W t-143
1 I' '
I !5 1 E'$,.d IIj
{e Ill ..
I,.
- , la : :i[s, '1 .a ;1 1 b t ,1
, c ,tg* g}r. ;! ,I): ;5 e e
,I !ta*
n.
tsgf Ij. 'g ,.: . ! i,1
!,ls : ,1
- (
Jiliti19k al d;" c21;fil :a III ;},!> fl E
. ,, ,. .. ,. ,, y
- sp : : ;t it ;;
1 1{
MP g
. .[ g - ..3 a jq }.l . 31
}-};5; tj,h 3 => >> .a , n f p s ,y -
i il i ::,
': i: e t ,:
g fE
h$$ I =!!!ll[ll!iill }}.!alll lll
}}d,lll E
' s 1 o 5[ 'i E
- i I E g
a i . 9: 1 A i_ l17gs I
,i l :[
g11
- 1. l.lf 31i; C
Ili ' lg ;: li..Il 3 11. . a il d1 II I i ' 1l <! .!<l
;I Nr !
fiti rm l gl !. il y l {'
- a p:p i{11]
y-i s, laaj! 1.ti 305
;t c'
is I: :[3I al; r y J3 ri.; fe- : did) e5 !
> a, t J;!p t
a g c -un -
}
o ** E " 5 i o a
- E iN E 1 I I I I i 144
l l JK/7 l EPRI/NRC PIPING AND FITTING DYNAMIC RELIABILITY PROGRAM DAN GUZY j ttSNRC, OFFICE OF NUCLEAR REGULATORY RESEARCH, WASHINGTON, DC. 20555 I Abstract ) I Until recently, NRC and industry efforts to improve the overall " balance-of-design" of piping have concentrated mainly on reducing conservatism associated with dynamic input loads and response calculation. Now it . appears that revision of the ASME Code design criteria offers the best future alternative for improving piping design. The EPRI/NRC Piping and Fitting Dynamic Reliability Program will provide the principal justi-fication for these changes. This paper gives en overview of this research ; program from an NRC perspective, and notes briefly how it relates to other NRC piping research and standards activities.
Background
The USNRC Piping Review Committee (PRC) was formed in 1983 with a charter to perform a comprehensive review of NRC requirements in the area of nuclear power plant piping. The PRC completed its mission in 1985, presenting recommendations for both regulatory changes and research in the five volumes of NUREG 1061 (References 1 through 5). A major concern of the PRC dealt with the overdesign of piping for seismic and other dynamic load events. After reviewing the data available at that time concerning earthquake experience (References 6 through 8), dynamic testing (References 7 through 12), and analytical studies (References 13 through 17), they concluded that the existing nuclear piping seismic design criteria and practices were very conservative. In fact, too conservative in light of the the negative impact these had on plant operation and overall reliability. The PRC summarized their concern in Reference 5: 145 l _ _ _ _ _ _ _ _ _ _ _ -
" Seismic design criteria of nuclear power plant piping evolved over a period of years- through a series of often discrete regulatory actions without an overall. assessment of . their col-lective effect on the actual systems constructed. Some criteria were established without an adequate data base. The existing requirements, along with prevailing industry design practice, i generally result in inherently stiff- piping systems because of '
the increased use of supports, including snubbers. Because . stiff systems increase thermal stresses and nozzle loads, they i may be more adversely influenced by construction . and operation l errors, including maintenance and inspection . errors. In addition, snubbers may suffer degradation or aging during operation, which ! may increase piping stresses because of snubber freeze-up." .; y PRC recommendations in NUREG 1061 Volumes 2 and 4 form the primary basis 1 for the piping research activities sponsored ' by the NRC/RES Engineering Branch in the late 1985 to 1987 timeframe. NRC plans for these. are: discussed in References 18 through 20. Plans for future piping research 1 1 will be presented in Reference 21. While most of the regulatory recommendations from Volumes 2 and 4 of NUREG-1061 concerned redefinition of piping input loads k.g., decoupling OBE from SSE) or improved dynamic response calculative methods - (e.g. , higher damping values), the PRC recognized the potential benefit that could come from improving ASME Code piping design criteria ~ for dynamic loads. Unfortunately the data and information available at the time of writing NUREG 1061 were too lir.ited to justify specific regulatory re-commendations in this area. The data were limited in the sense that ] while it gave indications that the margins-to-failure were much greater than assumed when the current ASME Code rules were written (see References 22 through P4 for background on these rules), the true margins- to-failure
]
were not well cuantified. In order to develop improved and realistic l design criteria, the actual level of failure and the mechanisms that contribute to this failure should be understood. The earthquake ex- ; perience data and dynamic test data available to the PRC showed mostly !
"non-failures" and provided only lower bounds on failure margins. The PRC thus did not recommend immediate changes to the ASME Code rules, but I
146 1 l
instead recommended as a high priority research activity that tests be undertaken .to identify seismic design margins and failure modes for typical piping systems. The NRC's participation in the joint EPRI/NRC Piping and Fitting Dynamic Reliability Program is a direct result of this reconinendation. Current ASME Piping Design Criteria 1 ASME Code piping requirements now evaluate inertial loadings from dynamic events the same as gravity and other sustained loads. by Equation (9) of Subsubarticles NB/NC/ND 3650 of Section III. Cross-sectional plastic collapse is assumed to be the dominate failure inode for all these loads. Equation 9 only evaluates the highest calculated stress that a dynamic response will produce; it does not consider cycle effects, load reversals, or the fact that most dynamic loads have limited durations and energy. content. An increasing amount of analytic and test data have shown that piping inertial loads behave differently than static loads. Not only is the the margin-to-failure greater, but the principal failure mode is different. Hara and Shibata (Reference 26) were among the first to recognize dynamic ratcheting as a piping failure trechanism. Analytic studies (References 17 and 25) showed that if these new insights into dynamic piping failure can be demonstrated systematically and conclu-sively, then significant changes can be made with regards to how the ASME Code sets limits on inertial stresses, and this could in turn reduce the number of snubbers used in nuclear power plants. Piping and Fitting Dynamic Reliability Program To develop and justify changes to the ASME Code piping criteria, the NRC and Electric Power Research Institute (EPRI) are cooperating jointly in the Piping and Fitting Dynamic Reliability Research Program. The objec-tives of this program are: 147
1 j i
.i o 'To id'entify failure mechanism and. failure levels of piping components l and systems under dynamic. loadings, o To provide 'a data base that will improve our prediction of piping.
system response and failure due to high level dynamic loads. o To develop an improved, realistic and defensible set of piping. design rules.for inclusion into the ASME Code. The results to date of this program are detailed in. paper K14/8 of this 9th SMf RT Conference (Reference 27), and in other. recent pub-lications (References 78 through 30). The following discussion will only ' highlight general findings and outline the programmatic aspects of: this program. The. joint EPRI/NRC program began in the Spring of 1985, and will take approximately 3 years to complete. General Electric of San Jose is the main contractor anc' performs the overall management, coordination and integration of the program. General Electric has worked with consultants to develop test matrices, configurations, data acquisition procedures, and plans for analytical studies. Analysis is performed mainly by General Electric of San Jose and testing is performed by subcontractors. EPRI and the NRC approve program planning ana review results as they aevelop. The NRC's contribution to this program is primarily through funding of piping system and component tests. Piping component tests are being conducted at ANC0 Engineers (see Re-ference 31). The objective is to systematically obtain dynamic data i failure for components under severe (but characteristic) seismic, and high , frequency loadings. Elbows, tees, reducers, support connections, nozzles f and lugs are being tested. The test specimens consist of both carbon and stainless steel 6" piping components with various thicknesses and internal 148
r
-pressures. Both in-plane and out-of-plane loadings are applied to the components; driven at one end by hydraulic actuators and having weights at the other end. High level time history inputs are repeated (usually more-than . twice) until the components rupture. In addition, there have been two static load failure tests of elbows that help demonstrate differences between static and dynamic failure mechanisms. The 40 component tests will be completed. in late 1987, with the NPC and EPRI each funding 20 tests. Roughly half of the tests have been completed and are presented in Table 1 of Reference 27. Systematic evaluation of these tests will take place after this task is completed, but some preliminary conclusions can be made from the tests performed so far:
- 1. Failure levels are much higher than previously believed - typically 15 to 25 times higher than elastically calculated Level D allowable limits, and measured moments are higher than theoretical static collapse moments.
- 2. Ratcheting and fatigue, not collapse, are the observed failure mechanisms for all dynamic load failures.
- 3. Piping components bulged and cracked. They did not collapse and crimp. Thus " loss-of-flow" functionality failure did not occur.
- 4. At failure load levels, equivalent lincer damping values were computed to be much higher than design values (on the order of 30% of critical damping).
As a separate task in this EPRI/NRC program, piping systems are being tested under . simulated earthquake, hydrodynamic, and water hammer load-ings. The NRC is principally funding the seismic and hydrodynamic load tests of 6" carbon steel and stainless steel piping systems at ETEC. These pressurized systems are constructed with components similar to those in the ANC0 component tests and also include branch lines and intermediate 149
J . l supports. The carbon steel systems should be tested to failure in May 1987, and.the stainless steel system test will be completed in the fall of
-1987. EPRI will sponsor' piping system waterhammer response tests at ANCO Engineers.
The basic phenomena of fatigue ratcheting is being studied at General i Electric of Schenectady using laboratory specimens. (EPRI sponsors this i task). Since at the present time there is no standard laboratory test specimen which addresses this failure behavior, specimen designs had to be developed. Both ' uniaxial and bending loads will be applied. The in-fluence of different material and temperatures will be studied. The outcome of this task will be the basis for evaluating ratchet effects under dynamic fatigue for ASME Code pipe materials. The results of the three types of tests discussed above will be analyzed and synthesized to form the basis of failure criteria for the combined static and dynamic loading of nuclear piping components. These criteria will be design-oriented and applicable for ASME Class 1, 2, and 3 piping. Analytical studies will be made to develop and- justify the alternative piping design rules. A strong liaison is being maintained with the NRC licensing staff, the PVRC, and the ASME Section III Code body which ultimately makes the rule changes. Other Related NRC Piping Research A separate but related NRC-sponsored pipe system test was conducted at ETEC in 1986 (References 32 and 33). The objectives of this test were to demonstrate the feasibility of failing a representative piping system under a high earthquake-like load, and to provide information and insights needed in the test planning of the main EPRI/NRC program. The 6" piping system withstood a 309 seismic input without rupturing the piping, al- l though ratcheting was observed. This input was an in-plant bu11dir.g response that was scaled up 15 to 20 times the elastically calculated i level that would reach the Level D design limit for that system. The I
\
L 1 150 l
system was eventually failed by applying sine-wave inputs at the system's resonance frequence for a limited amount of cycles. Strain concentration
.and ratcheting'were the cause of rupture.
l Two NRC sponsored projects have already made use of the test results dis-cussed above. HEDL has used the ETEC test as one benchmark in a study to investigate simplified nonlinear- response prediction methods (References 34.and 35). INEL has used the ETEC system and ANC0 component test data in a study to quantify equivalent damping for nonlinear piping response (References 36 and 37). Future NRC projects that will use the upcoming ETEC system tests as benchmarks are currently in the planning stage. Regulatory Uses of Research Results At this time (April 1987), the development of actual ASME Code piping criteria revisions are only in the . conceptual stage. It should be re-cognized that there is much more testing, analyses, sensitivity studies, and star.dards development that needs to be (and is planned to be) performed as part of the EPRI/NRC program. Changes to the ASME Code it-self will begin in 1988. In the interim, however, the results-to-date can be used in a qualitative ! sense to support regulatory decision making. For instance, I believe the test data can be used now to justify revising the NRC's piping "func-tionality" criteria. Inertial loads do not appear to cause piping failures that result in loss-of-flow. Also, the test results provide useful background information for the ongoing ASME Pipe Damping Task Group effort. The data show that although a particular pipe system's damping may be less than seme specified design value, the system will introduce much greater damping as input levels reach the failure regime. The results of the EPRI/NRC Piping and Fitting Dynamic Reliability Program have been used recently to support the development of a new ASME 151
l 1 l Code case that essentially removes the piping inertial stress evalution for Level B (OBE) louds. The NRC's endorsement of this code case would come through a future revision to Regulatory Guide 1.84. l l 1 n I I i l i l 152 l \ L----_------- .
References
- 1. USNRC Piping Review Committee, Investigation and Evaluation of Stress Corrosion Cracking in Piping of Boiling Water Reactor Plants, NUREG 1061 Volume 1 (August 1984).
- 2. USNRC Piping Review Committee, Evaluation of Seismic Designs - A Review of Seismic Design Requirements for Nuclear Plant Piping NUREG 1061 Volume 2 (April 1985).
- 3. USNRC Piping Review Committee, Evaluation of Potential for Pipe Breaks,
.NUREG 1061 Volume 3 (Hoveniber 1984).
- 4. USNRC Piping Review Committee, Evaluation of Other Dynamic Loads and Load Combinations, NUREG 1061 Volume 4 (December 1984).
- 5. USNRC Piping Review Committee, Summary - Piping Review Committee Conclusions and Recommendations, NUREG 1061 Volume 5 (April 1985).
- 6. J. D. Stevenson, Summary and Evaluation of Historical Strong-Motion Earth-quake Seismic Response and Damage to Aboveground Industrial Piping, NUREG 1061 Volume 2 Addendum (April 1985).
- 7. LLNL, Equipment Response at the El Centro Steam Plant During the October 15, 1979 Imperial Valley Earthquake, NUREG/CR-1655 (October 1980).
- 8. R. L. Cloud, Seismic Capability of Nuclear Piping, (May 1979). (Review performed for Stone and Webster Engineering Corp., Boston, MA).
- 9. ANC0 Engineers, Inc., Laboratory Studies: Dynamic Response of Proto-typical Piping Systems - Final Report, NUREG/CR-3893, (August 1984). '
3
- 10. ANCO Engineers, Inc., High-Amplitude Dynamic Tests of Prototypical Nuclear Piping Systems, EPRI NP-3916, (February 1985).
153
L
- l. i i
l i
- 11. Teidoguchi, Experimental Study on Limit Design for Nuclear Power Plant I Facilities During Earthquakes. 1973 JPNRSR-5 (USERDA Technical Infor- l mation Center, Oak Ridge, TN); Part 2.2, Vibration Tests of the Distri-bution Piping System."
- 12. INEL, A Survey of Experimentally Determined Damping Values in Nuclear ]
Power Plants, NUREG/CR-2406 (November 1981). J l
- 13. LLNL, Reliability Analysis of Stiff Versus Flexible Piping - Final Project ,
Report, NUREG/CR-4263 (May 1985). l l 1
- 14. SMA/LLNL, Response Margins of the Dynamic Analysis of Piping Systems, j NUREG/CR-3996(October 1984).
- 15. BNL, Alternate Procedures for Seismic Analysis of Multiply Supported I Piping Systems, NUREG/CR-3811 (August 1984). )
- 16. LLNL, Impact on Changes in Damping and Spectrum Peak Broadening on the Seismic Systems, NUREG/CR-3526 (August 1984).
- 17. Impe11 Corp., Conceptual Study to Develop Revised Dynamic Code Criteria for Nuclear Power Plant Piping, EPRI NP-4210 (August 1985)..
- 18. Edsmic Safety Research Progam Plan, NUREG-1147, Appendix 0, (June 1985)
- 19. D. Guzy, Piping .% search Overview,1 transactions of the Thirteenth Water Reactor Safety Research Information Meeting, NUREG/CP-0071 (October 1985).
- 20. D. Guzy, Piping Research Overview, Nuclear Engineering and Design, Vol. 98 Issue No. 2, (December 1986).
- 21. NUREG-1222, Piping Research Program Plan (to be published).
154
- 22. E. C. Rodabaugh and K. D. Desai, Realistic Seismic Design . Margins of Pumps, Valves,andPiping,NUREG/CR-2137(Jur.e1981).
j 23. E. C. Rodabaugh and S. E. Moore, Evaluation of the Plastic Characteris-tics of Piping Products in Relation to ASME Code Criteria, NUREG/CR-
'06211978).
- 24. S. E. Moore and E. . C. Rodabaugh, Background for Changes in the 1981 Edi-tion 'of the ASME Nuclear Power Plant Components Code for Controlling Primary Loads in Piping Systems. Journal of Pressure Vessel Technology, Volume 104,-(November 1982).
- 25. Quadrex Corp., Guidelines for Reducing Snubbers on Nuclear Piping Systems, EPRI-NSAC-104(July'1986).
- 26. F. Hara and H..Shibata Ratcheting Fatigue in Full-Scale Piping Elements, Paper K'15/3, 6th SMiRT Conference, 1981.
- 27. S. i Wt, Y. K. Tang,- D. J. Guzy, E. O. Swain, Piping and Fitting Dynann ,f ability Program, (to be presented as paper F 14/8 at the 9th SMiRT Conference, August 1987).
- 28. W. F. English, Piping and Fitting Dynamic Reliability Program First Semi- j Annual Progress Report May-October 1985, NEDC-31272, (November 1985). l I
- 29. W. F. English, Piping and Fitting Dynamic Reliability Program Second Semi-Annual Progress Report, November 1985-April 1986 (November 1986).
I 3D. S. W. Tagart, Y. K. Tang, H. L. Hwang, K. L. Merz, D. J. Guzy, Seismic Analysis and Testing of Piping Components, (to ' be presented at the June l 1987 ASME Pressure Yessel and Piping Conference). l l 155 _ _ ____-- D
I i: I l l l 1
)
.q l
- 31. P. Ibanez and K. L. Merz, High Level Seismic Testing of Components (to be presented as paper K 14/5 of the 9th SMiRT Conference).
.32. W. P. Chen, V. DeVita, and A. T. Onesto, Six-Inch Diameter Pipe System Fragility Test (to be presented as paper . K 20/15 of the 9th SMiRT Conference).
- 33. W. P. Chen, A. T. Onesto, V. DeVita, Seismic Fragility Test: of a 6-Inch Diameter Pipe System, FUREG/CR-4859, (February 1987).
- 34. .L. K. Severud, Simplified Analytical Methods and Experimental Correla-tions of Damping in Piping During Dynamic High-Level Inelastic Response (to be published as paper-K 17/7 of the 9th SMiRT Conference).
- 35. L. K. Severud, M. . J. Anderson, M. R. Lindquist, S. E. Wagner, E. O.
Weiner, High-Level Seismic Response and Failure Prediction Methods For Piping, Draft HEDL Report (February 1987) (to be published as a NUREG/CR-report),
- 36. A. G. Ware, .INEL/USNRC Pipe Damping Experiments and Studies (to be pub-lished as paper JK/14 of the 9th SMiRT Conference).
- 37. A. G. Ware An Evaluation of Damping in Piping Systems at High Strain Levels, Draft Informal Report EGG-EA-7380 (September 1986).
i i 156 :
)
JK/9
. LATEST RESEARCH RESULTS ON SEISMIC FRAGILITY DATA 0F NUCLEAR POWER PLANT EQUIPMENT K.K. Bandyopadhyay, C.H. Hofmayer, M.K. Kassir, S.E. Pepper i
Department of Nuclear' Energy l: E Brookhaven National Laboratory Upton, NY 11973, USA As part of the Component Fragility Research Program sponsored by the U.S. Nuclear Regulatory Commission existing seismic fragility data for safety-related electrical and mechanical equipment used in nuclear power plants are collected and analyzed. The test results are presented for different equipment ' categories in the form of test response spectra which are useful for seismic margin' studies. The data are also analyzed further and presented in the form of a single fragility descriptor.for each equipment category for direct application in probabilistic risk assessments. The equipment categories discussed in this paper include motor. control centers, switchboards, panelboards, and power supplies.
1.0 INTRODUCTION
As part of the Component Fragility Research Program, Brookhaven National Laboratory (BNL) is engaged in defining seismic fragility levels of nuclear power plant equipment by use of existing test results. In the last two decades, numerous seismic tests were performed for development of these equipment pieces and for Determination of their suitability in 157
l i l l l postulated earthquake environments for specific nuclear plants. In many 1 s of these tests, the specimens revealed fragility information as the j vibration level was gradually increased. With cooperation from various source organizations, BKL has collected such high and fragility level test information and created a data bank for a number of electrical and mechanical equipment pieces. Subsequently, BNL has analyzed this infor-mation for several important equipment categories and determined the threshold of vibration levels corresponding to various failure modes. The generic fragility levels of motor control centers, switchboards, panelboards and power supplies are included in this paper. The method-ology employed in analyzing the data and deriving the respective fragil-ity levels is also discussed. The details of this information will be published in a forthcoming NUREG report which will also include fragility information on switchgears, relays and local instruments.
'the fragility analysis ef fort is continuing and in the future will address other electrical and mechanical equipment categories such as transformers, circuit breakers, inverters, small valves and diesel gener-ator peripherals. In all, current program plans include a total of twenty three equipment categories which will be addressed by the end of 1988. A supplemental testing program for relays as well as special studies on dynamic amplification of electrical cabinets at high vibration I
levels will also be conducted. i 158
1 2.0 DATA ANALYSIS METHODOLOGY The test data from various sources for an equipment category are assembled together to assess its seismic behavior. Test response spectra 1 (TRS) are used as a measure of the vibration level. As expected, with ! the increase of the response spectrum level, the equipment is observed to exhibit various malfunctions and structural damage. Thus, from testing of one specimen with gradual increase in the test input, a number of re-sponse spectrum curves are obtained corresponding to various failure modes. Depending upon the use of the equipment, some types of malfunc-tioning may not be considered to incapacitate the equipment. Conse-quently, the objective is to present response spectra associated with respective anomalies so that the user can select the appropriate fragil-ity level for a certain application. Test data from various test programs are thus separately analyzed and then compared for determination of the generic fragility levels. The lower-bound TRS curves are presented for each equipment category with the associated malfunction type. However, the collected data are the results of a wide variety of test programs and of ten require further processing before they can be compared with each other. Therefore, in order to arrive at the seneric equipment fragility level from the individual test programs BNL has studied the individual test reports and analyzed the results by employing a uniform evaluation technique. l 159 _______ D
Variations of' testing methods, vibration' inputs and damping' values of TRS are'the main obstacles for direct comparison of. data. Random' multifrequency biaxial vibration inputs and.TRS at the 2% damping value I have been considered standard in'the BNL study,'since a major portion of' i the collected data belongs.co this group. Test data for different vibra-I tion inputs and/or different damping values are converted to the standard I form by using a scaling factor. j For conversion of damping values, the following multiplying factors j y have been used: I
' Conversion Damping Values Frequency Range Multiplying Factor 1
;From,5%.to 2%- 1-12.5 Hz 1.4 13-20 Hz 1.3 '
21-31 5 Hz' 1.2 From 3% to 2% 1-31.5 Hz 1.2 From'1% to 2% 1-19 Hz 0.77 20-31.5 HZ 0.85-The above conversion factors have been obtained from a' study of a large number of TRS data. These factors have been used for conversion to 2% damping and are not recommended for a reverse operation (e.g. from 2% to )/ 5%). It is recognized that the above factors are approximate and may need further refinement.
.Regarding variation of the test input a multiplying factor of 0.7 has been used over the entire frequency range to transform a narrowband (e.g. sine beat, sine dwell) or a single axis response to a standard 160
.I
response. A multiplying factor of 0.5 (i.e. 0.7 x 0.7) has been used when the input is both narrowband and single axis applied simultane-ously. Engineering judgements have been used for conversion of other vibration conditions. By employing the analysis and the standardization techniques discussed above, the generic fragility level is firs't described in terms l of TRS. However, in order that the fragility information can readily be used in Probabilistic Riek Assessment (PRA) studies, a further data re-duction becomes necessary such that a single descriptor can be used to represent the fragility TRS data. To this end, both the Zero Period Acceleration (ZPA) and Average Spectral Acceleration (ASA) over a fre-quency band of interest (e.g. fundamental frequency or 4-16 Hz) have been used in this analysis to represent a TRS data set. A probabilistic analysis is then performed by assuming a lognormal distribution of the data obtained from various test programs. The median values and the co-ef ficients of variation due to randomness of the test inputs (Sr) and uncertainties of the specimens (Sc) are presented in this paper. A con-venient fragility descriptor to indicate a high confidence of a low probability of failure (HCLPF) is readily computed corresponding to a 95% confidence of not exceeding 5% probability of failure. The numerical values reported in this paper are presented to provide an overall understanding of the available data base. Further analysis of the data base is underway and final f ragility descriptors will be pre-sented in the forthcoming NUREG report. l 161
3.0 MOTOR CONTROL CENTER (MCC) The BNL data base covers test results of twenty MCC specimens manu-factured by five major U.S. companies. Ninteen of the specimens were rated 480 VAC and one was rated 250 VDC. The data base test programs were conducted during the time period 1977-85. A typical bay of the data base specimens measured 90 inches high, 20-24 inches wide and 20 inches deep. The usual weight of one bay including devices was 600-700 lbs. Most test specimens consisted of two or three bays bolted side by side. Typically, the frame and housing of an MCC specimen incorporated a welded and bolted reinforced steel frame sheathed with steel panels. i Biaxial multifrequency vibration inputs were applied for sixteen specimens in the data base. The MCC specimens were mounted on the shake table and connected only at the base. Only two specimens were supported both at-the base and on the top. Most test specimens were mounted with four bolts per bay; others were welded. The electrical cable entrance was simulated at least in one test program. Representative devices including motor starters, circuit breakers and relays were installed in all test specimens. Selective devices were monitored for ascertaining electrical continuity and detecting change of state and contact chatter. Although for most specimens, the contact chatter was monitored for a duration of 2 milliseconds (ms) or greater, for some specimens In the data base, the limiting duration varied from. 1/2 me to 20 ms. The devices were monitored for electrically energized (E), de-energized (DE) and transition (E-DE, E-DE-E, DE-E-DE) states. l 162
j l 3.1 Test Results The fundamental frequencies in both the horizontal directions were observed to be in the range of 4-9 Hz for welded or bolted (with 4 bolts per bay) floor-mounted cabinets. With the addition of bays, the fre-quency in the side-to-side (SS) direction increased, although the front-to-back (FB) frequency remained almost unaffected. The lower-bound horizontal TRS plots are shown in Figure 1. Curve 1A represents the lower-bound envelope of the highest qualification levels in the horizontal direction. The test data indicate that at about the same acceleration level, the starters and the relays exhibited con-tact chatter problems indicating that curve 1A can also be cone.idered as the lower-bound fragility level corresponding to device chatter. The starter load changed state inadvertently during several test runs between the levels represented by curves 1A and 1B. Breaking of base frames, mounting bolts and/or mounting welds initiated at the level of curve 1C for some seismically designed MCC cabinets. The corresponding level for a cabinet strengthened with seismic braces is represented by curve 1D4 The various failure modes exhibited by the AC MCC specimens are j i enumerated as follows in the order they usually appeared with increasing test levels: o Motor starter, NC* auxiliary contact - mostly DE, sometimes E o Motor starter, N0* auxiliary contact - mostly DE, sometimes E i o Motor starter, main contact l o Motor starter load change of state - E, DE (inadvertently change of state or no change of state on command)
- Legend: NC - Normally closed, NO - Normally open 163 ,
}
_____a
1 I 10.0 O.0 - - 8.0 - - 7.0 - - S C s.0- [\ j - O I O ,l l g g 5.0 - g o
'iG / r 8 ' 4.0- lf
\ /
---- 'y~ ~' ,
.., m l ,,~~....5 s.0 -
f l/ V' ,
, / \i /\ -
1D(ZPA=2.3g)' 20- - 1B(ZPA=1.79 ) 1 o ,,/ yg . . . . . . . . . . ........ IC(ZPA=1.5g) 2' ~ 1A(ZPA=0.99) 0,0 ., 1 10 ..100 Frequency (Hz) i Fig.1 Horizontal TRS @ 2% Damping - Motor Control Center 1 { 164 I .- - _ -
o Relay chatter o Loosening of screws and mounting bolts o Snapping out of self-tapping screws o Motor starter - blowing of fuse (sporadic) o Motor starter load - dropping out and erratic behavior o Structural damage - plastic deformation, cracking and tearing of base metal especially in corner members o Strucural damage - breaking of panel bolts, mounting bolts and mounting weld; severe damage of cabinet structure For the DC MCC, the contact chatter initiated at the energized 1 state. 3.2 Probabilistic Fragility Estimates For the purpose of statistical analyses, the failure modes discussed above have been divided into the following three broad categories:
- 1. Electrical malfunctions: contact chatter and load (voltage) drop out.
- 2. Electrical malfunctions: change of state and fuse blow-up.
- 3. Major Structural damage.
The data associated with each of the above failure categories have been evaluated for determination of the respective fragility parameters. Sufficient test results have been found in the data base for performance of a separate statistical analysis corresponding to each of the first two ' failure categories. However, for the third failure category, the data have been considered inadequate for mathematical computation of the 165
1 4 l l i fragility parameters. Judgements in conjuction with the experience of ) the manufacturers have been used in estimating the fragility parameters for this category. For all three cases, the respective fragility de-scriptors have been computed by use of the parametric values following the methods discussed in-section 2. The fragility parameter and de-
.scriptor values corresponding to various failure categories for AC MCC'e are shown in Table 1. Since in the data base there is only one DC speci-men which was subjected to fragility testing, no attempt has been made to provide fragility parameters or descriptors for DC MCC's.
3.3 Remarks The fragility level of an MCC is for most applications controlled by chattering of motor starter auxiliary contacts or interlocks. However, an MCC typically contains several control relays. Based on an on-going ] study of relays, it appears that some of these relays may chatter at a lower vibration level. Therefore, screening of relays is recommended in using the fragility data presented above. There are sporadic instances of relay problems in the data base. 4.0 SWITCilBOARD The data base covers test results of six switchboard specimens. Test data for both 125 VDC and 480-600 VAC, circuit-breaker-type and - fusible-disconnect-switch-type switchboards are included in the data base. All test programs in the data base were conducted between 1976 and 1983. i i 166
i TABLE 1 AC MCC Probabilistic Fragility Analysis Results Failure Mode Indicator Median Su Br HCLPF in "g" in "g" Electrical ZPA 1.3 0.17 0.09 0.8 malfunctions: contact chatter and ASA 3.0 0.21 0.11 1.7 voltage drop-out @ 2% Electrical ZPA 1.6 0.10 0.11 1.1 malfunctions: change of state and ASA 3.7 0.08 0.10 2.7 fuse blow-up @ 2% Major structural damage: a) without seismic ZPA 2.1* 0.10* 0.06* 1.6 brace ASA
@ 2% 4.2* 0.10* 0.06* 3.2 b) with seismic ZPA 2.5* 0.10* 0.06* 1.9 brace ASA
@ 2% 6.0* 0.10* 0.06* 4.6
- Based on judgement i,
167 l I
l l-l l l A typical two-bay switchboard specimen in the data base test pro-l grams was 70-80 inches wide, 40-50 inches deep and 75-90 inches high, and 1 weighed 2500-3000 pounds. The cabinet enclosure was typically con-structed of die-formed, code gage steel members bolted together using formed steel panels and utilizing steel barriers to provide dead-front construction. All test programs except one employed random multifrequency phase-incoherent biaxial vibration inputs. All test samples were attached to - the shake table with intermittent welds. ; i Representative circuit breakers were electrically monitored during l I the test runs for detection of contact chatter and tripping. Although { relays were monitored for operability and detection of contact chatter, i most test reports did not address monitoring of relay chatters. Motor i starters, wherever used, were monitored for chatter detection. 4.1 Test Results l The fundamental natural frequencies were observed to vary as follows: Horizontal 5-9 Hz Vertical 15-20Hz With the increase of the vibration input, the first malfunction observed i was contact chatter of relays and motor starters wherever these were monitored. In Figure 2, curve 2A is a TRS plot which corresponds to ; l chattering of several relays for a duration equal to or greater than 20 milliseconds. A mounting weld broke at the level of curve 2B. i l 168
10.0 .- - - \ l 's.-l s 9.0 - l \.. e.0 - l- ; f a v.0 - l l 8 t 9 v 6.0_ r.
..s ,/ \, _
c *../ \ o / V
- 2 i
- l. .
g 5.0 - i ; ,
. i o . .
. A i l \ /-
T3 g,,,, j , l i ,t ., s, i ----- .------- 2C(ZPA=4.19)
< , ri l
2B(ZPA=3.4g) l l s.0 - : * -
/ / 2A(2PA=2.5g) l e.0- / -
l 1.0 - / - 0.0 .. 1 10 ..100 Requency (Hz) l
'1 1
i l'ig. 2: Horizontal TRS @ 2% Damping - Switchboard i 3 j 169
Chattering of a motor starter was observed at the level of curve 2C. The circuit breakers maintained electrical continuity during and after all these test runs. 4.2 Probabilistic Fragility Estimates In the data base test programs, there was no evidence of the breaker malfunction, nor was there any indication of a major structural damage of the switchboard cabinet. Therefore, the fragility level of the circuit i breakers and the cabinet structures are higher than the respective levels achieved during testing. Based upon the above discussion and experience gained fromproba-bilistic analyses of other electrical panels, a conservative estimate of the fragility parameters is made as follows provided the switchboard does not contain any relays: ZPA ASA @ 2% Median 3.5g 7.5g Bu 0.3 0.3 Br 0.1 0.1 By use of these parameters, the fragility descriptor is calculated as follows: ZPA ASA @ 2% HCLPF 1.8g 3.9g I 5.0 PANELBOARD The data base covers test results of sixteen panelboard specimens manufactured by four major companies. Test data for both 125 volt DC and 120-600 volt AC panelboards are included in the data base. The data base panelboards contained circuit breakers and/or fusible disconnects, but 170
not any motor starters. The current rating of the main breakers in the data' base varied from 225 amps to 800 amps. The' data base test programs were conducted in the time period 1975-85. The box and internal frames' of a typical data base panelboard speci-l men were constructed of code gage steel, while the current-carrying parts such as bus bars and the breaker straps were copper alloys. A typical-panelboard was 20-40 inches wide, 6-12 inches deep and 40-80 inches high, and weighed 200-400 lbs. .
' 1 In order to" simulate wall mounting, the test specimens were mounted on' vertical fixtures which were, in turn, anchored to the shake table.
Bolts of 1/2 inch diameter were used for a four-bolt mounting system;: whereas, 3/8 inch diameter bolts were used for six-bolt and eight-bolt connections. Representative circuit breakers were electrically monitored during the test: runs for detection of. false operation, chattering, voltage dis-continuity and malfunction of the contacts. 5.1 Test Results l Almost all test' specimens exhibited fundamental natural frequencies
)
i in the following ranges: i Front-to-Back 12-18 Hz j l Side-to-Side 12-20 Hz ; i Vertical 20-30 Hz 1 The lower-bound TRS plot are shown in Figure 3. Curve 3A is the j s lower-bound highest qualification level in the horizontal direction. l 171
10.0 ) 9.0 - 8.0 -
)
7.0 - I\ s v c 6.0 - ,LE l Nfll<,
-O y l' a d'
- l{
g S.0 -
. l ,
s , .: : 8 4.0 I' ' 'c -
< - .)'s lj i '* \\ ,~
3.0 - ,1 ,
't / js,
- t, '.~
2.0 -
. l%' ':
's', l
'l
/ -
__.3C(ZPA=2.29)
,d .
._____ ., 3 B ( Z P A= 1. 2 g )
1.0 - 3A(ZPA=1.0g) 0.0 .. .- 1 10 100 Frequency (Hz) Fig. 3: Horizontal TRS @ 2% darnping - Panelboards 172
Curve 3B corresponds to the initiation of spurious breaker tripping. At
? -the level of Curve 3C, the attachment' structural elements and' screws for l
L circuit breakers. vibrated loose and this resulted in breaker tripping. 1 L 5.2 Probabilistic Fragility Estimates The results discussed above indicate that spurious. breaker tripping is'the.first failure mode that a generic panelboard will exhibit with gradual increases of the vibration input. If the breaker can be man-ually set such that this failure mode is considered recoverable, the next failure mode appears to be loosening of structural elements which, in turn, produces local vibration and/or misalignment, and causes breaker tripping again. However in this event, breaker tripping may not be easily recovered since it may require repairing of the structural element. Therefore, breaker tripping as a consequence of structural loosening or misalignment has been considered unrecoverable. Similarly, burning away of relay contact as observed in the data base has been con-sidered an unrecoverable failure mode. By employing the methods discussed in section 2, the fragility analysis has been performed for both the (possibly) recoverable breaker tripping and the unrecoverable failure modes. The respective fragility parameters and descriptors are presented in Table 2. These results have been obtained for panelboards which did not contain any motor starters. In addition, chattering duration up to a maximum of 5 milliseconds for the main breaker and that of 20 milliseconds for auxiliary contacts were part of the acceptance criteria. 173
TABLE 2. Panelboard Probabilistic Fragility Analysis Results Failure Mode Indicator Median Su Br HCLPF in "g" in "g" ZPA 1.9 0.31 0.06 1.0 Breaker. Tripping (possibly recoverable) ASA
@ 2% 4.8 0.15 0.02 3.7 ZPA ?,2 0.07 0.02* 1.9 Unrecoverable !
ASA
@ 2% 5.4 0.12 0.02 4.2
- Based on judgement l
l 174
6.0 DC POWER SUPPLY The data covers test results of eleven DC power supply specimena manufactured by four major companies. These test programe were conducted in the time period 1976-83. A typical power supply specimen of the data base consisted of a step-down AC transformer, capacitors and rectifiers required to convert AC input to DC output, all mounted on a sheet metal base. The overall dimensions were 19 inches long, 5-10 inches wide and 6-12 inches high, and the weight was 25-100 lbs. In order to simulate the in-service mounting condition, all test specimens were attached to the vertical surface of a test fixture with machine screws. The test fixture, in turn, was mounted on the shake table. In all test programs, the output voltage and current were monitored and any fluctuation beyond a specified tolerance limit was considered a malfunction. Although in most test programs, a tolerance limit of 1-2% was used, at least for two test specimens, output variations up to 10% were considered acceptable. Electrical continuity was monitored during the strong motion for all test items. ; I
)
6.1 Test Results ) Some bounding TRS levels are pictorially exhibited in Figure 4. I l Curve 4A of this figure represents a lower-bound envelope of the qualifi-f cation level TRS for the data base results. Output variations of less than 2% were satisfied by this TRS level. A qualification level as i 175 ; i _ _ _ _ __ ____N
l 15.0 i 14.0-l 13.0- - 12.0- 1 11.0- - 10.0- I - l I \ l m I 8 9.o - u------------------ , l c l '
~
.y a.o- , l -
I l B v r.o- 'l - 1 g / \ e.o- , -
< , - . _ _ 4C(ZPA=5.79) s.o - l,lj -
-...... 4B(ZPA=4.0g) 4.o -
f/ s.o - 4A(ZPA=3.0g) 2.0 - - 1.0 - 0.0 ., ,. 1 10 100 Frequency (Hz) i Fig. 4: TRS @ 2% Damping - Power Supply 176
high as curve 4B exists in the data base. However, a short duration out-put vo!Ltage dropout was observed at the g-level of curve 4C and this resulted in system failure. Note that for curves 4A and 4B, the test reports did not specify whether the device output was monitored for con-tinuity on the order of milliseconds. The vaious failure modes observed in the data base can be summarized as follows:
- 1. Temporary loss of output power.
- 2. Variation of the output level in exceedance of the acceptable limit.
- 3. Structural loosening.
- 4. Structural failure.
6.2 Probabilistic Fragility Estimate Although a large number of specimens were tested at very high levels and up to the malfunction levels, the acceptance criteria and the failure modes were not necessarily the same in all test programs. Therefore, the fragility estimates have been for different acceptance criteria and failure modes. When subdivided into various groups, the data base has been considered inadequate for a statistical analysis for each failure mode. Therefore, judgements have been used in processing the test re-sults and arriving at the fragility parameters. The fragility de-scriptors have been calculated by use of the parametric values. The result:s are presented in Table 3. j i 177
-_- a
TABLE 3 Power Supply Probabilistic Fragility Analysis Results Acceptance Criteria / Median HCLPF l Failure Mode Indicator in "g" Su Sr in "g" , l Output level variation ZPA 4.6 0.13 0.03 3.5 less than +2% and out- ' put continuity satisfied ASA when monitored by meter. @ 2% 10.7 0.13 0.03 8.2 Output level variation ZPA 6.0 0.15 0.05 4.3 less than +10% and out-put continuity satisfied ASA when monitored by meter. @ 2% 13.1 0.15 0.05 9.4 Output continuity ZPA 3.6 0.15 0.05 2.6 monitored by oscillograph recorder and duration ASA of power loss not greater @ 2% 9.0 0.15 0.05 6.5 than 0.5ms. l l 178
7.0 CONCLUSION
S The fragility data presented in this paper are limited for equipment types with similar physical characteristics, electrical ratings and mounting types of test specimens described above. Since the equipment is believed to evolve with time, the results are applicable for the time periods mentioned in the respective sections. In addition, since some relays have been observed to have low fragility levels, such relays should be screened out prior to use of the data presented above. The fragility research is continuing. The detailed information and final results for the equipment categories discussed in this paper will be published in a forthcoming NUREG report. The results of the analysis of other equipment categories not covered in this paper will be published in the future. ACKNOWLEDGEMENT This work is being performed under the auspices of the Division of Engineering, Office of Nuclear Regulatory Research, U.S. Nuclear Regulatory Commission. The authors wish to express their special thanks to Dr. John A. O'Brien, NRC Project Manager, for his continued support and guidance throughout the course of this program. The authors sin-cerely acknowledge the cooperation of those individuals and organi-zations who made this program possible by providing test data and expertise. NOTICE The findings and opinions expressed in this paper are those of the authors and do not necessarily refect the views of the U.S. Nuclear Regulatory Commission or the organization of the authors. 179
JKl10 DETERMINATION OF A M0DAL INTERACTION CORRECTION FOR NARROWBAND FRAGILITY DATA Daniel D. Kana and' Daniel J. Pomerening' Engineering and Material Sciences Division, Southwest Research Institute San Antonio, TX 78284, USA Abstract Laboratory tests for safety equipment operation under seismic environ-ments in nuclear power plants have typically included various motion simula-tions based on either successively-applied multiple narrowband waveforms or simultaneous multifrequency broadband random waveforms. However, only broad-band excitations are directly applicable when equipment performance is affected by interaction between simultaneously responding modes. Therefore, a modal interaction correction factor is developed so that a narrowband response spectrum can be transformed to an approximately equivalent broadband spectrum which accounts for modal interaction effects. The approach includes study of the fragility response of a simple two-degree-of-freedom oscillator for repre-sentative narrowband and broadband excitations, and relating the two resulting fragility response spectra. It is found that multiplication of the narrowband response spectrum by a 0.7 factor produces a conservative equivalent broadband response spectrum. The results are interpreted in terms of a secondary device responding on a primary support structure, or a primary structure having two resonances. The approach is useful for updating existing test results based on narrowband spectra, developing composite spectra for similar equipment, or providing more flexibility in designing new tests, and is applicable to quali-fication proof test data as well as fragility data.
)
- 1. Introduction Qualification of nuclear plant equipment to function properly under-seismic environments can be verified by means of shaker table tests which may include a variety of excitation waveforms [1]. Since 1975, the guidelines have included a requirement that the effects of modal interaction due to simultaneous multiple frequency excitation must be accounted for in some justifiable manner. However, prior to this date, many tests included narrow-181
band waveforms, such as swept sinewaves or multiple sine dwells or sine beats, successively-applied in such a way that modal interaction effects were.not
. included. Extrapolation of the results of such narrowband tests to an equiva-lent broadband test which includes modal interaction effects has become highly desirable. 'The approach would be applicable to update of existing qualifica-tion proof test and fragility test data, and provide more-flexibility in designing new tests.
l The purpose of this paper is to present evidence that a constant correc-tion factor can be applied to narrowband data to produce an'approximately equivalent' broadband result. The use of such a' correction factor approach was originally recommended by Kana and Pomerening [2], and suggestions for devel-opment of a numerical value for the correction were included. In the absence of any other supporting information, a factor of 0.7 was speculated to be appropriate [2,3,4]. However, the adequacy of this numerical value was never substantiated beyond a recognition that two independent narrowband waveforms added in such a way that the resultant is the square root of the sum of the squares (SRSS) for the individual signals. More recently, the demand for a correction factor has been increased
-because of the trend toward the use of existing test and actual earthquake experience data for qualification purposes. With no better value available, the 0.7 factor already has been employed in developing composite generic equipment ruggedness spectra (GERS) [5], composite fragility functions [6),
and in developing seismic margin assessments (7]. Nevertheless, no specific recommendation has been included in revised standards for its use in interpre-tations of experience data [8,91 Therefore,.the results of this paper are especially important in verifying the adequacy of the 0.7 correction factor. The information is condensed from a more comprehensive report [101, which addresses several areas of the use of similarity principles for equipment qualification.
- 2. Effects of modal interaction The concept of a modal interaction correction factor was developed initially in terms of test severity comparisons [3] and then in terms of 182
fragility functions [4]. Herein, the development will also be in terms of fregility functions in the form of fragility response spectra (FRS) which, of course, can be generated from fragility test data. Nevertheless, it should be understood at the outset that the results are also applicable to qualification proof test results as well, and examples for this will be given later. Typical effects of modal interaction and cross-axis coupling [4] on the fragility of an item whose performance is influenced by two dominant resonance frequencies is shown in Figure 1. Herein, we will concentrate only on those effects due to modal interaction. The narrowband fragility function envelope may have been generated by input of a constant amplitude, slowly swept sine-wave and observation at what frequencies malfunction occurred. Likewise, it could also have been generated by input of a sine dwell, sine beat, or narrow-band random excitation at successive discrete frequencies and slowly raising the input level until malfunction was observed. Subsequent transformation of
.he input data to response spectrum form is understood. Furthermore, the result is an envelope of all of the discrete point data. ;
Note the implication in Figure 1, that the narrowband fragility function envelope (for which no modal interaction has occurred) is lowered by subse-quent consideration of a simultaneous broadband input for which modal inter- ! action is included. Of course, the exact shape of the modified curve depends on the actual frequency content (spectral shape) of the broadband excitation, { as well as the precise physical mechanism (response transfer function) affected by the modal interaction. In Reference 3, an example correction factor is developed based on an identified critical acceleration response transfer function and the spectral shape of subsequent broadband excitation. Such a factor would vary depending on the given data. Therefore, a pertinent question is, "Can one conservatively approximate the true fragility function with modal interaction by multiplying the narrowband fragility function by a constant factor that is appropriate for most practical cases?" The results of this paper tends to answer the question affirmatively, and provides evidence that the factor can reasonably be estimated as 0.7 in many practical situa-tions. 183 _ _ _ _ _ _ _ _ _____________a
1 1
- 3. Simple analytical model-
-3.1. Basis for model' A two-degree-of-freedom model, as shown in Figure 2, was used to study. potential modal interaction effects between two parts of a dynamic system. This model may be construed to be analogous to several different practical physical systems, depending on the values of stiffness, mass, and damping involved. On the'one hand, it can be used to represent an equipment item having'two dominant resonances, such as an electrical cabinet or a valve.
It may also represent a device mounted on a much more massive support struc-ture. In either case, the degree of coupled dynamic interaction depends on the particular parameters specified. Two different approaches were applied to predict the responses of the two masses for two different type base motion excitations: 1) a broad-band excitation which included modal interaction effects; and 2) a narrowband excitation which did not include such effects. The results were then compared in such a way that a modal interaction factor could be evaluated. This com-parison was developed in terms of response spectra for the two different types of excitation, since most existing equipment qualification data is given in terms of these parameters, as implied in Figure 1. The first approach included a time history solution which resulted from a specified broadband displacement history for the base. A procedure which. included specified displacement was used, since the corresponding velo-cities and accelerations could be determined with differentiation techniques. This avoids the numerical problems associated with determination of the velo-city and displacement from a specified acceleration for long-duration time histories. The corresponding acceleration response spectra for the input were calculated for development of a broadband FRS. The second approach included development of narrowband fragility functions from the transfer function of the two masses. These transfer functions correspond to the results of swept sinusoidal resonance searches available in qualification reports. They were converted into response spectra for development of a narrowband FRS similar to j that described in Figure 1. ! l j 184 I l
The failure level of the system could be defined in terms of a number of parameters. These included:
- 1) Peak displacement, velocity, and acceleration of the masses; j 2) RMS displacement, velocity, and acceleration of the masses; I
- 3) Relative displacement or velocity between the masses and the support; and,
- 4) Force in the springs and dampers.
For the present study, the failure levels were defined in terms of the peak acceleration response of each of the two masses. This type of failure is
~
appropriate for a large number of items which contain a peak acceleration sensitive device located at some elevated position of the equipment structure. Failure by exceeding a specified peak acceleration on mass 2 corresponds to resonance of a secondary device located on a flexible structure. Failure by peak acceleration on mass 1 corresponds to exceeding a peak specified input to a secondary device located on a flexible structure. Each of these cases were studied independently. Thus, the problem reduces to determination of the respective excitation fragility levels which correspond to the specified peak response failure levels. By specifying the relationship between the masses and their corre-sponding stiffnesses and damping, the frequencies could be adjusted as desired. The ratio of the frequencies of the two masses were varied from 0.25 to 4.0. For this paper, details are given for a case representing a small test item, mass 2 = 10 lbs, in conjunction with a large support structure, mass 1 = 1,000 lbs. Further data are also given for two other mass ratios: mass 2 = 100 lbs, mass 1 = 1,000 lbs; and mass 2 = 1,000 lbs, mass 1 = 1,000 lbs. These correspond to coupling of a medium test item supported on a primary structure and coupling of two major modes, respectively. With the ratio of the two masses set, the problem was then defined in terms of the ratio of the undamped frequency of the top mass to that of the lower mass. The frequency of each mass, spring, and damper combination was defined independently, i.e., two coupled single-degrae-of-freedom systems. 185
l For,most of the present study, the frequency of the lower mass on its spring and damper was set at 10 Hz with a damping of 5%. The upper mass frequency was then varied from 2.5 Hz up to 40 Hz with a damping of 5%. Limited cases for other values of damping were also included. 3.2 Broadband fragility response solution The following relationships were used' in the time history solution to define the velocity and acceleration in terms'of the applied displacements. Displacement: zn = either x or y in Figure 2, with n = time step index Velocity: $n " a (1)
+z n-1 Acceleration: z n" zn+1 - 2z2 n (} )
at > where: at = time' step i z n-1
= displacement at time t - at z
n
= displacement at time t z n+1
= displacement at time t + at These were then combined with the equations of the free-body diagram given in Figure 1. For mass 1, mi x'3 + c2 ( i 2) + k,(x -x ) + c ($ -y) + k (x -y) = 0 3 2 3 3 3 3 (3)
For mass 2, mx2 2 + c 2 (k 2 i) + k (x -X i ) = 0 2 2 (4) By substituting the expressions for velocity and acceleration, i.e., Equations (1) and (2) into Equations (3) and (4), expressions for the subsequent time displacements, x3 of mass 1 and x2 of mass 2, can be computed in terms of the previous time displacements of these masses and the known input, y. 186 J
For the broadband time history solution, the following parameters were calculated and plotted:
- 1) Displacement, velocity, and acceleration of the base and the two masses;
- 2) Relative velocity and displacement between the support and mass 1 and between mass 1 and mass 2;
- 3) The forces in the springs and dampers; and,
- 4) Input response spectra for failure at each of the two masses.
The input was first specified as a simulated earthquake accelero-gram whose response spectrum matched the R.G. 1.60 horizontal spectrum for 5% damping and 1 g ZPA. Typical input displacements, acceleration time histo-ries, and corresponding acceleration response spectrum are given in Figures 3 and 4. Figure 4 shows both the required response spectrum (RRS) and a typical computed response spectrum (CRS). Note that it was not possible to match the spectrum exactly at the ZPA due to inconsistencies between the specified spec-trum and a random time history which has a peak-to-RMS ratio of 3.0 during the stationary portion of the signal. Finally, a broadband FRS was developed for each solution case by determining the appropriate excitation ZPA level that produced a 1 g peak response on each of the masses. Details of the results are given in Section 4. 3.3 Narrowband fragility response solution For the narrowband FRS calculations, Equations (3) and (4) were again used. In this case, the displacements, velocities, and accelerations were assumed to be of the form. z = Ze iet (5) z = Ziwe iwt (6) z = -Zw'eiwt (7) 187
Taking into account the phase relationship between the motion at the input and the two masses and substituting these assumed solutions into Equations (3) and (4) gives the following relationships: (-k 2-c wi)X e-10 2 2 2 + [-m 1 w'+ k,+ k + (c +c )wil X e-ie! = (k + i 2 3 3 i c wi)Y (8) 3
-ie
(-m2 w + k2 + c2wi)X e-2 2+ (.k - c wi)X ie 2 2 l =0 (9) In the present case, the value for Y was assumed to be 1.0. The magnitude of the transfer function for each of the two masses, x i and xt, then is the magnitude of the complex variable obtained by solving these two equa-tions simultaneously. This magnitude was inverted, which gives the input level required (at each frequency point) to produce 1.0 g output at the masses. This input level was subsequently converted to a spectral value by dividing by two times the damping value. The resulting narrowband FRS corre-sponds to that which would be obtained from the envelope of peak spectral values from a slowly-swept sinewave or independently-applied, discrete sine-wave excitation, whose amplitude produces 1 g peak response at the respective frequency. This approach corresponds to that which was described in Fig-ure 1. An example for each of the two masses is given as the upper curves in Figures 5a and 5b for a frequency ratio of 4.0, s = 0.05 for both masses, and i Mi = 1,000 lbs, M 2 = 10 lbs.
- 4. Development of correction factor The results of both the broadband and narrowband fragility response solutions were compared in a manner which allowed a determination of a numeri-cal ratio between the broadband excitation which included modal interaction, l and the narrowband excitation which did not include modal interaction. Sample calculations for one mass ratio are summarized in Tables 1 & 2 for several frequency ratios. For the broadband time history results, an excitation ZPA was sought which produced the specified 1.0 g peak acceleration failure level on each of the masses. This was done by first using an excitation with a 1.0 g ZPA (whose peak mass response is given in Column 3) and then inverting each of the respective response values to produce the corresponding input ZPA level 188
given in Column 4. .Thus, a fragility excitation ZPA level (Column 4) was established for a broadband FRS for which any potential modal interaction effects are sure to be present. For the narrowband calculations, the value of the narrowband FRS at its lowest point was first noted (Column 5), along with the critical frequency at which it occurred (Column 6). Then,'the corre-sponding ZPA level of an equivalent R.G. 1.60 spectrum, which has the identi-cal spectral value at the critical frequency was established. Note that the amplification for this equivalent spectrum at the critical frequency is given in Column 7. Furthermore, the ZPA for this equivalent broadband spectrum is l obtained by dividing the minimum narrowband FRS value (Column 5) by this amplification. The resulting ZPA for the equivalent broadband spectrem is listed in Column 8.. An example of this FRS matching process is shown pic-torially in Figure 5 for one set of. system parameters. Finally, a numerical ratio is determined between the ZPA level of the initial broadband FRS which included modal interaction, and the equivalent broadband FRS which is based on-the narrowband results. That is, the value of the fragility ZPA level from the broadband time history results (Column 4) was divided by the ZPA level (Column 8) for the equivalent R.G. 1.60 spectrum which matched the narrowband i FRS at the critical frequency. The results are displayed in Column 9 of the tables for both cases where peak acceleration causes failure at each of the respective masses. The final interaction correction factor (broadband to narrowband ZPA level ratio) is shown as a function of frequency ratio for three different mass ratios in Figure 6. As can be seen from Figure 6, the use of a universal correction factor of 0.7 is conservative for most cases studied, and a minimum occurs when the two masses are exactly tuned for failure at mass 1. Further-more, the correction factor appears to be necessary only for the range 0.5<f2 /f <1.5 i . Outside this range, the narrowband excitation is at least as, or more, severe than the broadband excitation. For the tuned case, the frequency band centered about the resonance frequency is controlled by the damping values and by statistics of the excitation. The assumed value of 5% critical damping gives'a relatively broad band and is representative of the largest damping expected to occur in equipment. The consequences of varia-tions from this value will be discussed momentarily. Another important cbser-vation from Figure 6 is that failures due to resonances at mass 2 require no 189
correction. Consequently, it would appear that one needs to consider the dynamics only of a primary system (i.e., failure at mass 1) in establishing fragility. As a result, the critical fragility function can be obtained solely from the primary system, i.e., from a cabinet at a point of attachment of critical devices. This is extremely important since, in the past, such transfer functions typically have been measured during resonance searches conducted prior to qualification tests. The extent of validity for this observation and its practical consequences need to be explored further. It is important to emphasize that the identical broadband time history sample was used to generate the various data for all cases given in Figure 6. If other sample broadband time histories are used, a statistical spread of the data results. An estimate of this spread was calculated by using six differ-ent time histories to generate the critical minimum value of the curve for M i = 1,000 lbs and M 2 = 100 lbs. The values ranged from 0.68 to 0.85, with a mean value of 0.789 and a standard deviation of 0.063. Thus, Figure 6 is based on the lowest value of correction factor that was obtained from the study. Note that, at three standard deviations, the minimum could be as low as 0.6. However, the low probability of this value must be combined with the low probability of the secondary mode's occurrence nearly tuned with the primary mode of a typical equipment item. Therefore, a value of 0.7 appears to be reasonably conservative. It also appears to be reasonable to assume that the statistical spread which results from detailed time history varia-tions wculd be no worse for other, off-resonance frequency points shown in Figure 6. The mass ratios selected to calculate the results in Figure 6 were considered to be representative of three different practical equipment designs. Since the M 2= 100 lb case produced a minimum for the correction factor, it is appropriate to wonder if this represents the true minimum for variation of mass ratio. This can only be answered by further calculations, should the selected mass ratios not appear applicable in a specific problem. Variations from 5% critical damping can have a significant effect on the fragility response spectra at some frequencies. To demonstrate this, one case for the narrowband FRS for mass 2 was computed for two additional values of 190
damping, and the results are shown in Figure 7. It can be seen that the off- i resonance values'of the FRS are significantly altered, while the critical minimum spectral values near resonance are essentially unchanged. (Note that narrowband excitation ZPA's do change although the spectral value does not.) This, of course, means that the ZPA for the equivalent broadband spectrum which matches this spectral value is changed only to the extent that damping affects the broadband (R.G. 1.60) FRS. On the other hand, the initial broad-band FRS which is developed from a time history is also affected proportion-ally. Therefore, since the correction factor is based on the ratio of these two broadband spectra, it appears that the spectrum matching process is essen-tially independent of damping for the types of spectra investigated. This result may be different if combinations of narrowband/ broadband spectra were involved. A final practical consideration must deal with the result of applying the aaove process to other forms of narrowband spectra rather than the extreme case of a FRS based on sinusoidal excitation. To shed some light on this, a FRS based on a 10-cycle per beat sinebeat was applied to the system whose parameters are given in Figure 7. Several FRS points based on a time history solution and 5% damping are shown in the figure. It can be seen that the results appear to produce nearly the same effect as an increase of damping to about 7%. However, the critical spectral value is only slightly increased. Therefore, the correction factor ratio appears to be relatively insensitive to a FRS based on a waveform that contains a single dominant narrowband compo-nent. However, as the narrowband waveform becomes increasingly broader, it is logical that the spectral matching process must produce a correction factor that approaches 1.0 when the narrowband FRS approaches the R.G. 1.60 spectrum. At this point, one might ponder whether the interaction of more than two modes also is a practical consideration. It is obvious that three or more modes could affect fragility in an item having densely-packed modes. However, this would appear to be a condition to be aware of, but probably is an excep- ; I tional case. 191
- 5. Examples of application The modal interaction correction factor is particularly useful for quali-fication of equipment by existing data. An example described initially in Reference [3] can be used as a typical demonstration. The upper curve of Figure 7 shows the response spectrum envelope for a 0.5 g peak slowly-swept sinewave. Consider this to be the narrowband TRS for which an equipment item has been qualified. A typical question is whether this item is further quali-fied to the lower broadband RRS. By the above procedure, the answer is affir-mative. In fact, the item is qualified to any spectrum for which no spectral value exceeds 8.75 g. This approach is similar to that of Reference [3],
which includes the 10-step procedure for change from narrowband old to broad-band new qualification test. However, the present results replace steps 2-6 of Table 1 (Reference [3]) with the single constant factor of 0.7. Should this not be successful in a given application, one could of course revert back to the more complicated approach to determine a more accurate and, perhaps, more favorable correction factor. It is obvious that a variety of practical scenarios can be developed for the use of narrowband proof or fragility data to demonstrate qualification to broadband requirements. Only one additional example will be described herein. The above example included two specific response spectra for a specific hard-ware item. However, the results can be extended to a more general class of hardware by means of similarity principles given in Reference [8]. Documented examples of this generic approach are given in Reference [5]. That is, existing qualification TRS data are collected for a class of equip-ment (i.e., such as electrical panels) that can be argued to form a physically similar set. The argument is based on evidence that all of the equipment would respond dynamically to a newly-specified motion (new RRS) in a similar fashion, i.e., with significant resonances in a common narrow frequency band 4 dnd peak respon!es of the same level. Note that significant resonances are only those which affect the fragility of the equipment. Furthermore, all of the various TRS data are reduced to a composite spectrum, which can be argued to represent excitation similarity when compared to any constituent spectrum. Any narrowband constituent spectrum is reduced by the factor 0.7 before being 192
used in the composite. Then, it can be argued that any new physically similar item is qualified to any new RRS which falls under the composite spectrum. Thus, a most useful extension of existing qualification data whatever the source (i.e.,-test, actual earthquake, or analysis) results from such an-approach.
-6. Conclusions This paper provides evidence of the adequacy of a 0.7 factor for trans-formation of narrowband qualification (proof or fragility) data to broadband data. The correction factor is given in terms of the peak response fragility of a simple two-degree-of-freedom mechanical system whose behavior is inter-l preted in terms of practical hardware in which two responding mo6es have specified mass and frequency ratios. It is important to question whether the results can be generalized to any type of fragility behavior expected to occur in nuclear plant equipment. The answer is not obvious, but can be resolved on a case by case basis. It is our judgment that the results are applicable to
-many kinds of typical equipment. This may be shown in a specific case by postulating the mode or modes of failure and describing how they are analogous to the behavior displayed by the simple model used herein. If the mass and/or frequency ratios are significantly different, then additional calculations may be necessary. Furthermore, if the narrowband fragility spectrum is some combination of broadband and narrowband components, the 0.7 factor appears to be similarly conservative. However, should doubt exist, the actual narrowband FRS could be analyzed according to the procedure used herein, and correcticn factors tailored to that spectrum could be developed.
193
REFERENCES
- 1. "IEEE Recommended Practices for Seismic Qualification of Class 1E Equip-ment for Nuclear Power Generating Stations," Standard 244-1975, The Institute of Electrical and Electronics Engineers, Inc., New York, NY, January 31, 1975.
1
- 2. Kana, D.D., and Pomerening, D.J., "A Research Program for Seismic Quali- i fication of Nuclear Plant Electrical and Mechanical Equipment," USNRC .)
NUREG/CR-3892 Vol. 3 (Recommendations for Improvement of Equipment j Qualification Methodology and Criteria), August 1984.
- 3. Kana, D.D., and Pomerening, D.J., "A Method for Correlating Severity of Different Seismic Qualification Tests," ASME Journal Pressure Vessel i Technology, Vol. 109, pp. 58-64, February 1987.
- 4. Kana, D.D., and Pomerening, D.J., " Dynamic Fragility Concepts for Equip- '
ment Design and Qualification," Nuclear Engineering and Design, 94, pp 41-52,(1986).
- 5. Merz, K.L., and Smith, C.B., " Generic Qualification of Equipment Using Test Data." ANC0 Report 1087.74 (prepared for Electric Power Research Institute), ANCO Engineers, Inc., Culver City, CA, April 1985.
- 6. Bandyopadhyay, K.K., and Hofmayer, C.H., " Seismic Fragility of Nuclear Power Plant Components (Phase I)," NUREG/CR-4659, BNL-NUREG-52007, Vol. 1, Brookhaven National Laboratory, Upton, NY, June 1986.
- 7. Prassinos, P.G., Ravindra, M.K., and Savy, J.B., " Recommendations to the Nuclear Regulatory Commission on Trial Guidelines for Seismic Margin Reviews of Nuclear Power Plants," NUREG/CR-4482, UCID-20579, Lawrence Livermore National Laboratory, Livermore, CA, March 1986.
- 8. " Recommended Practices for Seismic Qualification of Class 1E Equipment for Nuclear Power Generating Stations," Draft Standard P344-08, The Institute of Electrical and Electronics Engineers," New York, NY, Decem-ber 1985.
- 9. " Recommended Practices for Seismic Performance Qualification of Mechani-cal Equipment Used in Nuclear Power Plants With Particular Application to Pumps and Valves," ASME Draft Standard - Revision 5. American Society of Mechanical Engineers, New York, NY, September 1986.
- 10. Kana, D.D., and Pomerening, D.J., "A Framework for Qualification of Equipment by Safety Margin Methodology," Report No. SwRI-8608-001, South-west Research Institute, San Antonio, TX, November 1985.
194
nbv 5 3 1 8 6 7 7 7 7 7 oawe 0 0 0 7 9 0 0 0 0 0 ib oL td r 1 1 1 0 0 1 1 1 1 1 aarA R oaP rNZ B l h Arta ct P - Z c M)a 6 7 8 3 4 2 7 8 8 9 e 0 2 4 4 2 9 8 8 8 8 t p g 4 4 4 5 4 3 3 3 3 3 u p S u e( 0 0 0 0 0 0 0 0 0 0 In rl
)
o aV f M s E e s T S i r t 0-6.if n o l _ Y sa u v 9 6 4 2 1 0 7 5 5 5 S bv s i 1 ii 4 4 6 6 6 6 5 5 5 5 _ l( e u .lpa t F % R q 2 2 2 2 2 2 2 2 2 2 0 0 =5 E mc 0 1 1 y= S R G. R A 2 - =cng F S ES t en d HA h ui n l y TM 2gqp a a c ' i em b c n 9 2 2 4 7 9 9 9 9 9 RT sera w ei e) 8 1 1 4 6 8 8 8 8 8 OA sWFD o S ut uz . F . a qH 9 0 1 0 91 9 9 9 9 9 9 N r R li e( SO M r F ar r RI s a VC F FT r N m A e u lan DR t m ro NE e i AL m n ti e BE a cts M ean) i r 1 5 P, 2 1 2 9 9 9 9 WC a prog 0 0 1 4 1 0 9 9 9 9 OC P S ep( . . RA l s 1 1 1 1 1 1 O 0 0 0 R l ee AK e cR NA d c E o A DP M N AF s A O t P D s NS l Z g t 5 0 5 2 6 7 5 6 6 7 AM b z u r u p 2 4 5 2 0 1 1 1 1 1 l H s t uf o 0. ut BR e 4 4 4 4 4 4 4 4 4 4 DE 00 R AT p 1 0 0 0 0 0 0 0 0 0 0 O 01 d S n O RN BI 0,= 5 n a R I 1 F y= b d n t
= ng c ayr o u 1 t en o o i 1 p h ui r t t n .
E 1 gqp B sk a sI 3 5 8 5 4 7 9 8 8 8 L iem i ar s 3 2 1 3 4 3 3 3 3 3 B ag A sera H ee M - sWFD Pl 2 2 2 2 2 2 2 2 2 2 T a e e t 1 M m c i c aro T A f y c no 5 5 0 0 0 5 0 0 0 0 ei 2 7 9 0 1 2 5 0 0 0 ut . qa eR 0 0 0 1 1 1 1 2 3 4 r F y c 2 n 5 5 0 0 0 5 0 0 0 0 ef s) . . uosz q 2 7 9 0 1 2 5 0 0 0 aH 1 1 1 1 2 3 4 e M( r F gm j
oawe 9 0 9 1 3 2 1 0 1 0 iboL tdr aarA 0 1 - 0 1 1 1 1 1 1 1 RoaP rNZ B : l a h Ar P t c t .
=
Z c M) a 6 6 2 7 4 9 9 4 6 5 0 6 1 1 71 9 4 1 9 4 6 t e p g 3 0 0 1 2 2 3 3 uS p e( u - 0 0 0 0 0 0 0 0 0 0 _ In or l a
) f V
_ M s _ E e s 0 T t - i v 6. in _ S r l 1 9 4 2 1 0 7 5 5 5 _ Y sa u 1 f o - S b v s i ii 1 6 6 6 6 6 5 5 5 5 - l ( e u .lpa t F -
% R q 3 2 2 2 2 2 2 2 2 2
- 0 0 = 5 E mc - D 1 S G. R A 2 .y= R - 2 = ng c F S ES t en d y n HA
.h ui l a c TM 2gqp a -
n i em b c e) 7 9 4 4 9 9 9 9 9 9 - RT sera w ei uz 4 4 4 4 8 8 0 8 8 8 - OA sWFD o S ut qH - F a r li 2 7 9 9 9 9 9 9 9 9 - N M r R F ar e( r SO s a VC F RI r N FT e m l u A t m an DR e i ro _ NE m n ti e _ AL a cts M ean) 5 5 0 0 5 9 6 5 8 3 i BE r _ WC a prog 9 4 3 2 2 3 5 7 8 9 _ OC P S ep( _ RA l s 0 0 0 0 0 0 0 0 0 0 _ R l ee AK e cR _ NA d c E o A DP M _ N AF s A _ O t P - D s l Z gt 9 6 1 9 4 3 5 8 1 71 NS bz u ru 9 6 1 8 2 8 4 8 9 _ AM lH s p 2 3 3 3 BR DE 00 R e tuf p o 0. 1 u t 2 1 1 0 1 1 0 0 0 0 0 0 0 0 0 0 AT 01 d S n O O % RN 0,= 5 n a R i BI 1 b F _ y= d t
= c a y n o2p u
. ng o r r to
- 2 t en i n hui t 1 8 5 3 1 2 4 2 0 0 - E 1 gqp Bska sI s 3 9 9 1 0 4 0 1 6 5 L iem i ar ag . B A sera Hee Pl M - 3 5 8 1 81 5 4 3 2 2 sWFD e 1 T a e c tar M m i c o T A f y c no 5 5 0 0 0 5 0 0 0 0 ei 2 7 9 0 1 2 5 0 0 0 ut . qa 2 3 4 eR 0 0 0 1 1 1 1 r F y c 2 n 5 5 0 0 0 5 0 0 0 0 ef s) . uosz 2 7 9 0 1 2 5 0 0 0 2 3 4 q aH 1 1 1 1 e M( r F g
ste ve*t p
,'l'&gg'zumr :: + gger""t?
if ff -T, , ' Malfunction Region pwllg;jj[gegjgrj;
~
'-l' :StepiijsmityS's#M
:Ighi
~ ,,,: /gse *;~h$$ni ' Shi3~1IANNEwahe&MN?,y,I'fk" f .
c'; ';p;c 5::'
, s Idi, ,' ~4j Without Interaction ** ,, 7,jiOss" '
C 5d$'idi';'ig?jy;gs):jhj!!f
~
or Coupling
,5," '.'9M /g,,,,,,,
% %Y 9* spas'~/wsgreargwer*% r -
's,[,~~* AES '
s w$>:f ,i!l
~ , , azin[; ,,;jj;jf$.:.ih!:sf[c sy :Fe /
& \ > ~
g \. ifep G ', .,,, ; ~ '/ s,(f) n \ &W c
\\, / Developed by Slowly
- \, k if:f*
s,i!!! s
,/ Swept Sine a \ '/ '
j \ / i '/
/
\/ v t
\ /
E n \J / v w f/) w y Narrowband Fragility function Envelope L 0 x Approximate Fragility Function with Modal Interaction Approximate Fragility Function with Modal Interaction and Cross-Axis Coupling FREQUENCY. HZ Figure 1 Approximate fragility function for complex system (Ref. 4). 197
xp unus *2 *2 II If k 2$- - l::1 c 2 - k2 (*2 *1) c 2(E2-E1 ) I i l I l xi l 3 usua *1 *1 I l I f I f ky I--IC 3 k (x -y) 3 3 C(i-j) y 3 ll i i i l y suuu VEOS/SEE////A VE/NE//Off/A Danped 2-DOF Oscillator Free Body Diagram i i Figure 2 Two-degree-of-freedom model used for the analysis. 1 l l I 198
3
; i i i i 2 -
h 1 - l - s' 0 85 t il mi i ld l l l
~' ~
l s i i l
-2 -
-3 0 5 10 15 20 25 30 35 TIME, SECONDS a) Displacement time history 1.0 i i i i
-0.
-1.0 - -
.5 0 5 10 15 20 25 30 35 TIME, SECONDS b) Acceleration time history Figure 3 p time histories for RG 1.60, horizontal earthquake, A celer o 1me histo 199
l l1 l' l l p. 2 10 _ . . . i iiii e i i i e i i i_
- i
_ i BETA = 0.050 - ZPA = -1.120 _ i j 1 10 - 5 - g _ w , g - - g CRS O 10 _ _ RRS _
-1 i e , i - i,,, , , , i i i ii 0 I 10 2
10 10 FREQUENCY, HZ Figure 4 Acceleration response spectrum for RG 1.60, horizontal, 5% damping. I 200
1 1 2 10
.. i i i iiiiii , i,,,,i
- f y = 10 Hz
_ Mg = 1000'1b 7 } [ f2 = 40 Hz M2 = 10'1b / [ 8 = 0.05
/ -
/
10 I !
-- ~ ~ ~ %
/
5 - N 3 p -
~
\ / -
5
~
\
g - f . 0 -
\/ -
W 10 0 -
\! -_
ZPA = 0.389 _ 10
~I ' ' ' ' ' ' ' ' ' ' ' ' ' '
10 0 10 I 10 FREQUENCY HZ a) Hass 1 Figure 5 Narrowband FRS and spectral value matching RG 1.60 level. ; I a) Mass 1 201
2 10 _ i i i i iiiii i i i a i i i i- i f
~
ff y = 10 Hz M3 = 1000 lb _ [ fp = 40 Hz ! [ E = 0.05 M2 = 10 lb
/ a\t ~
/ l 10 i / _
- ~%g\ :
~
s : \ / : 5
\/ -
g _ g - w \ 0 V ac 10 _ __ l - - ZPA = 0.365
' ' ' I ' ' ' ' '
10-I 2 0 1 10 10 10 FREQUENCY,HZ b) Hass 2 Figure 5 Narrowband FRS and spectral value matching RG 1.60 level, b) Mass 2 l l 1 1 202
> i i i i i i 1.4- -
e I D c. A ~ ~ ~~ . _,"'- -
~
l d $)
!.1.2-y
- u. /
r - 5 - g
- g# <>-- - 0 4)
$ 1.0 _ , .
.)
g , . u - j _ 5 60.8- - 5 w - M3 = 1000 lb QM2 = 10 lb
$ fg_= 10 Hz OM2= 100 lb 0.6 6 = 0.05 OM 2= 1000 1b -
i i l , I i 0 2 3 4 FREQUENCY RATIO,pf /fg a) Failure at Mass 1 l l 1.4- , g e
- p O- - - c. ,
g f g f %,,O---- .(j g 1.2- / - l 5 7 g' A y -- b - H) 1.0- 4
^
-)
e - - E E0.8-g
\
~
fly = 1000 lb Q 117 = 10 lb i f3 = 10 Hz OM2 = 100 lb 0.6- 6 = 0.05 OM2 - 1000 lb e i i I t I i u 1 2 3 4 FREQUENCY RATIO2f /f1 b) Failure at Mass 2 Figure 6 Modal interaction correction factor for peak acceleration failure. a) Failure at Mass I b) Failure at Mass 2 203
102_ , , , , , ,,i g , , , , , , _
- ,o f, -
'~ ~
_ ,8 I
,' /
e = 2% y
-----____,,~. ,
l/ s y 6 = 5% \ ^ 10 _ ,
._ g a _
LD
$ g ,
-. \ \
j h g h 't i A
~
$ t g l g i!!all 8
o 10 g = 1000 lb Curves = Sine Dwell _ 1 M2 = 100 lb 8 = 2%, 5%, 8% _ f = 10 Hz 1 f = 10 Hz 2 00 0 = Sine Beat - 6 = 5%
~I ' ' ' ' ' ' ' ' ' ' 'i' 10 0 A 2 10 10 10 FREQUENCY, HZ Figure 7 Influence of damping and waveform on FRS for failure at Mass 2.
I 204
2 , , , , , , , , , , , , i 10 ,,,,1 BETA = 0.020 - I - l 12.5 a 1 10 :- 831_ _ _ _, -- e : EXISTIN TRS : 5 - P - I NEW RRS d 30 0 _ m 0 a
' ' ' ' ' I ' ' ' ' ' ' ' '
10
-1 0 1 2 10 10 10 FREQUENCY,HZ Figure 8 Qualification of an equipment item by existing narrowband data.
205
JK/12 AN EXPERIMENTAL INVESTIGATION OF REDUCED STIFFNESS AT SEISMIC HORKING LOADS IN REINFORCED CONCRETE SHEAR HALL STRUCTURES CHARLES R. FARRAR JOEL G.-BENNETT' HADE E. DUNH00DY Los Alamos National Laboratory, Los Alamos, NM 87545. USA L Low-aspect-ratio reinforced concrete shear walls are the structural element that provides the primary lateral load re-sistance in Seismic Category 1 nuclear power plant structures exclusive of containment. Since 1980, Los Alamos National Laboratory has conducted a large number of experiments to in-vestigate the seismic response of these structures and their shear wall elements. Results have shown that these structures respond to seismic excitation with a stiffness reduced by a factor of 4 from-the theoretical uncracked cross-section stiff-ness based on. strength of materials. The equivalent static load levels ~ at which this reduction in stiffness occurs are well below the predicted first cracking load for the struc-tures. To.further examine the response of these structures and to clarify the. discrepancies between previous static and dynamic data, a large structure with six-inch (0.152 m) thick shear walls was first tested with low-level dynamic excitation and then was statically tested to failure with cyclic loading. Results showed that, prior to cracking, both the static and dynamic response are well predicted with linear analysis. For this structure, reduction in stiffness was not observed until cracking had occurred. The relation between this test and previous test results obtained at Los Alamos is discussed along with a comparison between the observed response and current U.S. design documents.
- 1. Introduction Previous work that has been carried out at Los Alamos National Laboratory l
as part of the Seismic Category 1 Structures Program for the USNRC Office of i i Nuclear Regulatory Research has consistently measured reductions in stiffness
.of four or more in scale models of low-aspect ratio shear wall structures sub- ~jected to working loads. In this context, working loads refer to load levels equivalent to those experienced by a structure during an operating basis earthquake. These loads would produce peak stress levels as low as 50 psi 207
__-_-____a
(0.34 MPa) average base shear stress (ABSS) with maximum normal tensile stresses well below the tensile strength of the concrete. The models tested thus far have been made of both microconcrete and con-ventional concrete and have been tested statically and dynamically. 'Scalabil-ity of the models has been reported at the current and previous SMiRT confer-ences (Bennett et al. [1], Dove et al. [2):, and Endebrock et al.'[3]) along with the results on stiffness reduction, Bennett et al. [4). The reduced stiffness has been observed at significantly lower load levels in dynamic tests than in static tests. Also, quantification of the stiffness ! reduction requires that the contributions of orthogonal shear walls to the flexural stiffness of the shear wall in question be assessed. I Currently, the U.S. nuclear civil engineering community has no standard method
- for treating these flexural boundary elements. To study these issues, a large shear wall structure was subjected to low-level random excitation for modal testing and then was tested statically and cyclically with complete load're-versals. This structure is one of the largest low-aspect-ratio shear wall structures to be experimentally investigated. Results from this test concern-ing the previously mentioned issues as wC > b attempts to separate the shear and bending components ot deftmation wii) ! e presented herein. Implications of this test on past tests t W "ct 6 at Los Alamos along with implications on current U.S. nuclear civil engineering design practices will be discussed.
- 2. Test structure, instrumentation, loading j The test structure, Fig. 1, was constructed with 0.75 in. (19 mm) aggregate batch-plant concrete and No. 3 (10 mm) rebar. The 6-in.-(0.152 m) thick walls allowed the reinforcement to be placed in two layers near each face with one in. (25.4 mm) of concrete cover. The reinforcement was spaced 208
at 14.5 in. (0.368 m) on center in both the horizontal and vertical direction to provide 0.251. reinforcement by area in each direction, the minimum allowed by the American Concrete Institute's Nuclear Design Code (ACI 349-85) [53. All reinforcement detailing requirements of ACI 349-85 were met or exceeded. The concrete's average ultimate strength and modulus of elasticity were 4150 psi (28.6 MPa) and 3.23 x 106 psi (22.3 GPa), respectively. The average split-cylinder tensile strength was 357 psi (2.46 MPa), and the minimum yield strength of the reinforcement was 60 ksi (414 MPa). As shown in Fig. 1, two 6-in.-(0.152 m) thick steel plates were placed on top of the model to simulate the normal stresses that an actual Seismic Category 1 structure's shear wall would experience. The steel plates were secured to the model with thirty-six 1.25-in.- (38 mm) diameter steel bolts, and the model was fixed to a load frame with thirty-six identical bolts and 2.0-in.- (51 mm) steel plates. Fig. 2 shows the completed model set on the top of the load-frame base. Twenty-four displacement transducers were mounted at various locations, as shown in Fig. 3, and 30 strain gages were attached to the reinforcement. The displacement transducers located on the interior shear wall were used to 1 obtain relative displacements necessary to separate shear and bending compo-nents of displacement. Overall structural deformations, including rigid body translation and rotation, were monitored with displacement transducers mounted on an independent instrumentation frame. Torsional motion and sliding shear at the base of the structure were also monitored with these gages. The strain gages were primarily used to determine contribution of the boundary walls to the flexural stiffness of the shear wall. Load for the static testing was applied by means of a hydraulic actuator attached to the bottom steel plate on the top of the structure and was 209
)
transferred through friction to the top controte s 20. Force input into the structure was monitored with a load cell located between the actuator and the steel plate. After a series of low-load level tests to verify that the ! instrumentation was functioning properly, the load history, shown in Fig. 4, was followed until failure. The complete load reversals shown in Fig. 4 were intended to represent forces that would occur during a seismic event. The 1 l final low-level cycle was to provide information concerning the response of the damaged structure. Each integer on the horizontal axis (Fig. 41 represents a point for which the displacement transducers, strain gages, and l load cell were scanned and recorded. i i I 1
- 3. Analysis and experimental results Before the actual static load-cycle testing, the structure was ultra-sonically inspected for defects at 180 locations on both the shear wall and end walls. No flaws were detected during the ultrasonic testing, and the 6
concrete modulus determined from the speed of sound (3.2 x 10 psi [2.21 GPa)) agreed very well with the average value determined from compressive test specimens (3.23 x 106 p34), Next, a low-load level experimental modal analysis was performed to de-termine the initial stiffness of the structure and to characterize its dynamic properties in the virgin condition. The structure, without the steel plates, was supported by air bearings and was excited with a random signal applied through a 300-lb (1.33 kN) shaker (Fig. S.) The air bearings were used to simulate a free boundary condition and to minimize ambiguities caused by base constraints, when comparisons were made with finite element modal analyses. Coupled with the low-force level,this support system introduces negligible damage into the structure during testing. Acceleration response was measured 210
f
. at 89 locations, and with the aid of a commercial software package [6], the natural frequencies and corresponding mode shapes, up to 200 Hz, were identi- , . fi ed. Excellent. agreement was obtained with a finite element modal analysis, as indicated in Table I, that compares the resonant frequencies for corres-ponding modes from the experimental and analytical modal analysis. An example of the correspondence between an experimentally measured mode shape and the corresponding mode shape as determined with finite element analysis is shown in Fig. 6. Note that only half the structure was modeled in the finite ele-ment analysis.
The overall horizontal deformation vs load as determined from the in-terior relative displacement gages is shown in Fig. 7. This is the displace-ment at the top of these gages relative to the bottom of the gages. Since the displacement field over this region is nonhomogeneous, the displacements computed in this manner represent an average value for the wall. The method for computing the horizontal displacement is illustrated in Fig. 8 and, with the instrumentation used in this test, four values of horizontal displacement could be determined and averaged. Also, it is assumed that these displacement values do not significantly change when extrapolated to the exterior of the structure. This assumption was verified with a two-dimensional finite element analysis of the shear wall. From this analysis, the horizontal displacement at the terminal point (point C in Fig 9) of the diagonal gage had a displace-ment of 6.669 x 10-3 in (0.1694 mm) when subjected to a 100 psi (0.69 MPa) ABSS, while the exterior point, D, at the same elevation had a displacement of 6.69e. x 10-3 in (0.1700 mm), a 0.47. difference. The data from the interior relative displacement gages are independent of rigid body rotation and translation and the assumptions necessary to remove those quantities. 211
I Stiffness based on these relative displacement readings was determined using Castigliano's Theorem. By examining the free body diagram in Fig. 9, the expression for internal strain energy stored in the structure between the section A-A and B-B can be written as U - (M + Px + 6x)2 dx + + 6[ dx , [1] o 2EI o 2 AeG where U - internal strain energy, M - moment at section A-A, P - shear force at section A-A, 6 - imaginary unit load, E - concrete modulus of elasticity, i I - cross-sectional moment of inertia, include entire end wall but neglects steel, G - concrete shear modulus, A - effective shear area, and e L = length of the wall between planes A-A and B-B. Using standard procedures described in Popov [6], the horizontal displace-ment of the structure at plane A-A relative to plane B-B can be determined and the stiffness of this portion of the structure can be expressed as ; I K " T 2 3 l hL L L
+ +___
2EI 3EI AG e 212
This total stiffness may be decomposed into a bending component and a shear component yielding K B
" ' 8"d 3 2 2L + 3hL AG e
Kg - , L When the properties of this structure are substituted into the above equations, the theoretical stiffness values become K T
- 8.4 x 106 lbs/in. (1.47 GN/m),
K B
= 50.6 x 106 lbs/in. (8.86 GN/m), and 6
Kg - 10.1 x 10 lbs/in. (1.77 GN/m). The structure showed linear response through all of the 50-psi-(0.34 MPa) and 100-psi-(0.69 MPa) ABSS load cycles and the measured stiffnesses during these pre-cracking load cycles based on the average displacements determined from the interior gage readings, were K T
- 8.5 x 106 lbs/in. (1.49 GN/m),
K B
- 52.6 x 106 lbs/in. (9.21 GN/m),
Kg - 10.2 x 106 lbs/in. (1.79 GN/m). Figure 10 shows the load vs average horizontal displacement for a 50-psi-(0.34 MPa) ABSS load cycle. 213
i i During the _ initial loading _ of the first 200-psi-(1.38 MPa) ABSS load cycle,'the structure cracked at a load of.71,000 lbs-(316 kN), and cracking was again observed at a load level of 65,000 lbs (289 kN) during the reverse cycle. After completion of this load cycle, visual inspection. revealed several diagonal cracks in the shear wall. The structure appeared to behave in a _ linear manner during the subsequent 200 psi-load cycles; however, the stiffness was reduced by a factor of 2 from the uncrucked value. In terms of percentage, the loss of stiffness occurred almost equally in both the shear and bending ; components. The new. values of stiffness were 6 K - 4.05 x.10 lbs/in. (0,71 GN/m), T 6 K .23.0 x 10 lbs/in. (4.03 GN/m), and j B Kg - 4.91 x 106 lbs/in. (0._86 GN/m). The structure reached its ultimate strength during the initial loading of the first 300 psi (2.07 MPa) ABSS load cycle. At the ultimate load of 140,000 lbs (623 kN), a large horizontal flexural crack opened completely through the tensile end wall and propagated at approximately 45 degrees through the shear wall. Other shear cracks also opened, and previously existing cracks extended. During the reverse load cycle, the structure was able to sustain a load of 125,000 lbs (556 kN). Similar cracking occurred during this reverse cycle, but these cracks were not as pronounced as they were in the initial portion of the cycle. The final crack patterns on both sides of the shear wall are shown in Figs. 11 and 12. Note that the horizontal crack visible in the end wall was the one produced during the initial load increase. 214 -
Initially. .the response during the final 50-psi-(0.34 MPa) ABSS load cycle
' appeared linear with stiffness values of.
1 K T
- 5.42 x 105 lbs/in. (94.9 MN/m)~,
5 K - 48.4. x 10 lbs/in. -(848 MN/m), and B Kg - 6.11 x 105 lbs/in. (107 HN/m). However, during unloading after the initial force. increase and during the load reversal, the displacements appeared to be nonlinear. Currently, results are being further' analyzed, particularly with regard to strain gage data and to response in the nonlinear range.
- 4. Discussion of.'results Based on the results of the modal analysis and of the initial pre-cracking
. load cycles up to 100-psi-(0.69 MPa) ABSS, the initial stiffness of this struc-ture is within 96-99% of the theoretical stiffness, as determined from either a finite element analysis or a strength-of-material approach. Also, these results show that, before cracking, the entire end wall contributes.to the flexural stiffness of the shear wall. The effective width exceeds the portion of the walls that would be considered effective as based on ACI 349-85 T-beam criteria. The results concerning the initial static stiffness are consistent with those found by other investigators on smaller-scale walls with various boundary. elements, Benjamin & Williams [9] and Barda et al. [10].
.The first-cracking load of 71,000 lbs (316 kN) that corresponds to an ABSS of 130 psi (0.90 MPa) and that produces a maximum normal tensile stress (MNTS) of 171 psi (1.12 MPa) was considerably less than would have been pre-dicted from a strength-of-materials anaiysis using either the split-cylinder 215
tensile strength (SCTS) or.the modulus of' rupture, 7.5 /f'c (2.0 /f'c) per ACI 349-85. However, the first-cracking load agrees fairly well with the load that ACI 349-85 would predict for first cracking, 82,600 lbs (367 kN). A pos-sibility for the discrepancy between the measured first-cracking load and the corresponding strength-of-materials value is that the end walls are only par-tially effective at the. time of cracking. Agreement between predictions based on the MNTS and the SCTS can be obtained at this load if this assumption is made; however, this assumption cannot be verified from the instrumentation readirgs obtained during the test. I This first cracking load would correspond to the load induced by a 1.6-g-maximum horizontal acceleration earthquake with no amplification. The previous large-scale shear wall structure (similar to the structure reported herein but with 4.0 in. [.102 m) walls) that was dynamically tested and reported in [1] showed a reduction in stiffness of 4 during a 0.73-g peak horizontal accelera-tion earthquake. This seismic excitation corresponded to an equivalent static load of 32,900 lbs (146 kN), an ABSS of 91 psi (0.63 MPa), and a maximum normal tensile stress of 92 psi (0.64 HPa), well below stress levels predicted to produce cracking. Obviously, there still remains a significant difference between the static and dynamic responses of similar structures tested at simi-lar load levels. The ultimate load of this structure exceeds the design load specified by ACI 349-85 (126,000 lbs [560 KN]). However, it should be pointed out that the reference on which the ACI design criteria is based, Cardenas et al. [11] does not consider the effects of the boundary elements. The ABSS at failure, 260 ; psi (1.79 HPa), was slightly lower than had been observed in other static tests carried out on microconcrete isolated shear walls in this program, (290 216 ,
psi [2.00 MPa] ABSS). This was due in part to the relatively higher tensile strength of microconcrete as compared with conventional concrete.
- 5. Conclusions One of the primary purposes of this test was to determine whether, during a carefully monitored static-load-cycle test, a stiffness reduction of 4 would occur at similar load levels as have been observed in dynamic tests. During the pre-cracking load cycles and the low-level modal analysis, no stiffness reduction was observed, and the response of the structure was accurately pre-dicted with currently used linear analysis techniques based on strength-of-material s. These same linear analysis techniques have not adequately predicted j
the dynamic response of structures previously tested in the program even though stress levels during the dynamic tests were well below those required to crack the structure. Hence, several questions arise about previous tests conducted in this program and about the dynamic behavior of actual Seismic Category 1 Structures. In particular, the following possibilities must still be con-sidered: (1) Does microconcrete adequately simulate actual concrete in both static and dynamic response? (2) Here previous models damaged prior to testing either by handling or, in the case of smaller structures, shrinkage cracks? (3) Are there dynamic effects that cause the discrepancy between the reductions in stiffness observed statically and dynamically? (4) In all testing and analysis have the boundary conditions been properly accounted for? These questions are currently being examined in light of the latest test results. This test was also to provide information on the effectiveness of the end walls. Up until first cracking, they appear to be fully effective. After cracking, the extent of their contribution is not clear and data are still being evaluated at this time. 217
The ability to separate shear and bending components of deformation was clearly demonstrated. Loss of stiffness was shown to occur equally in each component of deformation. In addition to further studying the data in post cracking region, another model is currently being constructed that is identical to the large model reported by Bennett et al. [4]. The previous structure was tested dynamically, and the structure being constructed will be tested statically and cyclically in an identical fashion to the one reported in this paper. It is hoped that the direct comparison between a static and dynamic tests will further clarify the reduced-stiffness issue. 218
TABLE I ANALYTICAL MODAL FREQUENCIES COMPARED WITH EXPERIMENTAL MODAL ANALYSIS RESULTS Freauency Modes Finite Element (l) . Experimental Modal Q-200 Hz Analysis (Hz) Analysis (Hz) 1 36.3 37.1 2 77.8 79.2 3 86.0 88.3 4 102 100 5 111 111 6 120 122 7 130 (2) 8 136 141(3) 9 143 (2) 10 154 171.5(3) [ j 11 162 (2) (l) These frequencies were calculated with t:'e ABAQUS [7] finite element code using eight node continuum elements. P. enforcement was neglected in this model. (2) These modes were not identified. Relocation of the shaker and/or in-creased input level would make identification possible. (3) The larget discrepancies between the FEA and the experimental result is due primarily to a drop in the input power spectral density and to the associated loss of coherence at the higher frequencies. 219
i LIST OF SYMBOLS 1
- effect',ve shear area Ae E - concrete modulus of elasticity j G - concrete shear modulus ,
h - distance from applied external load to the top of the interior gages i I - cross-sectional moment of inertia neglecting steel KB - ' sending stiffness ! Ks - shear stiffness KT - total stiffness L - interior gage length H - moment at a section through the top of the interior gages I P - applied external load U - internal strain energy 6 - imaginary unit load l 220 t- _ __
REFERENCES
- 1. Bennett, J. G., R. C. Dove, H. E. Dunwoody, and C. R. Farrar. 1987. An Experimental Investigation for Scalability of the Seismic Response of
,Microconcrete Model Nuclear Powerplant Structures. Transactions of the 9th Int. Conf. of Structural Mechanics in Reactor Technology, H Session,
.Lausonne, Switzerland, August 17-21, 1987.
- 2. Dove, R..C., E. G. Endebrock, H. E. Dunwoody, and J. G. Bennett. 1985.
Seismic Tests on Models of Reinforced Concrete Category I Buildings. Transactions of the 8th Int. Conf. of Structural Mechanics in Reactor Technology, Brussels, Belgium, August 19-23, 1985.
- 3. Endebrock E. G, and R. C. Dove. 1984. Nonlinear Seismic Response of Small-Scale Reinforced Concrete Shear Hall Structures. Transactions of the.7th Int. Conf. of Structural Mechanics in Reactor Technology, K Session, Paris, France, August 1984.
- 4. Bennett, J. G., R. C. Dove, H. E. Dunwoody, and C. R. Farrar. 1985.
. Latest Results from the Seismic Category I Structures Program.
Transactions of the 8th Int. Conf. of Structural Mechanics in Reactor Technology, J/K Session, Brussels, Belgium, August 19-23, 1985.
- 5. ACI Committee 349. 1985. Code Requirements for Nuclear Safety Related Concrete Structures and Commentary. American Concrete Institute, Detroit, MI.
- 6. SMS Modal 3.0 Manual (Structural Measurements Systems, Inc., San Jose, CA, 1982), Vol. 1.
- 7. ABAQUS Computer Program Manuals (Hibbitt, Karlson, and Sorenson, Inc.,
Providence, RI, 1981), Vols. 1-4.
- 8. Popov, E. G. 1968. Introduction to Mechanics of Solids. Prentice-Hall, Inc., Englewood Cliffs, NJ.
- 9. Benjamin, J. R., and H. A. Hilliams.1957. The Behavior of One-Stor" Reinforced Concrete Shear Halls, J. Structural Division, ASCE, Vol. 83, ST3, pp. 1-49.
- 10. Barda, F. , J. M. Hanson, and H. G. Corley.1977. Shear Strength of Low-Rise Halls with Boundary Elements, Pub. SP-53, American Concrete Institute, Detroit, MI, pp. 149-202.
- 11. Cardenas, A. E. , J. M. Hanson, G. H. Corley, and E. Hognested.1980.
Design Provisions for Shear Halls, Pub. SP-63, American Concrete Institute, Detroit, MI, pp. 221-241. 221
oR gf NEIO
- 1 h
4 E NC V TT H C A K 1 e Q g A EE j 1 SE r H CR I E 0 S H I D T b A l g I 41 5/ -
/ 5 L CH L 0 /1 LNC P 80 5
M g' AI 4 W 45.A E L 1 H1E El 8 C NC E TX 6 N OA I SO 6 F XR H R I OP SAB C WP A I L EA TA LRE A I 8 l ' n s
,l z
/ h~9
-t
!*C 4 5
..}
gpic - f i i i ll F 6 ;- 0 wg5 ; j - 2 fdl pMf i h d" [i , f; 1
.,;b, H 7,
il: i _M
;/:
.I
- [
(l F[fwz,s'~ I i'I ' ii p' .l i
- .gy" f*
y
.j l
;f;J L.
.n y i' p sA
, ~
y 4 jl. Fj?
;i x, l-y- '; ,. ll ,
f
,1 '
. 9 _a ]l '}'-j+
k lI
- gi r '
j j;
- Mh aNf" k y!l;N ,- : I l
{s ~ 1 y - 6'
]':~)' h -
g:
~
= \ s! ' -
i { g- :jph I i m-Q
; ?.
y" jkM f A.!# &a37_
;?; '-
o77
;f ' ' pu,h p! pd M
,i s r, y[. i
,$: f I
wsMi sE i.% .. m'
.}-
l
; i M.?. 0
,T m .f 9
\ YB B
% E
-%N T E 6 i2 AC LA/
PL a L P1 6 d, EN EI
] 0 1
TD S
.L qis % %d nE iHG 2
5 l
h , *'- O . k
, , . *- i!%.
^ -
hJ nw na idhf}'[,d i y ,
+
.a. . g ',d}.?p);,
f'y
+
p . y'%,Jg3, ts-
.., j %.
* *' Ih,t[ i4
;gr"< 'm a ;g tp{;.jy 'j ;jgg 4
.4' y g .. -** >{ , , ,
c a.:k ?w 4
;< . k ,' *
- i
, r j
( # ( N M; '{.g y g 3 e- ; qp.; q ,;j y,g
-a. -
. 3 a
. 'i-
?
s l
- g;; g ,
p ,,f)ra;( i cQ M
'v .
i.' . [.5 s1 p ,,
+ ., e i jr :fY )
Y
</
h')gl?- ,; fl s!'llp , k 4 f * ' >p%j, .. 3$ 1
*\ - i " Y.
, 1 s
+$um-1 M1fg v g t, , ...^ np c
I N( I og adI l e *]r$f(' t g I nw 3 t 3 ) - 1 i, s t. I (' j'. { 1 , If' i 't 1 il il
'I h 'ta s
I
</.3 b i t 1 a
! \f[ 10 ,
n (- '
, .c
+^~
' y)
..7
., f ns i ;-
r . a ',( : 4a
. . p>h c 1 ,
fn <( .% .
.f ). 1, n.,
y t;tqg :dp . F , .( yJ k4 f, 3r 'Q,f? x n . a ,, f}fl f ) fAh whmaammw & yV
,- ,-T.
o - 7l , l
' F '
l (!- 1 l/ ------------q'
,4m
_.O
,o[x]*
i I :' I
+ 1 "x]
+
+
/
,i
~
.ec p u
/
c [- RELATIVE DISPLACEMENT O DISPLACEMENT, FIXED REFERENCE
- FOUR GAGES CORRESPONDING TO THESE ON THE OTHER END x FOUR GAGES CORRESPONDING TO THESE ON THE OTHER SIDE OF THE SHEAR WALL l
l Fig. 3 Displacement transducer location. 22a
O E I 4 4 4 W
~ ~
l p ~ - e - O 1 -
~~ _
o
%Q .
~
7j . y . S _ s a6 - X
~
$0< i m
9 i u Q ._ _ 9 5
<C W M O -
l: J - _ , w (> n 4 Y l ! I I O O O N O M OSd) z OL X SS381S BV3HS 3SV8 3AV 225
, t, ,
s. l' 0;?!" '
, ,Q hj h, l}I( y ',i :: '
.a Fi!' u. 4tf : i
.M.y hIh js .
[% . c yy, ' ,n o %.s g g 1.1 i
'2- i p, g
r K;-r
> 8,4
' , ' '%[
n.. : ..fyl . 2
.g ,1,,:_d
'h:/l[, ,s ,j
' r o " j ,6,
,; 's PW , :yl9 , +
%; l-lf cs? ,
- 1. b'
~
l yy'h"% ff ?%; ~ n; f. y 'y" a
- r. ..
#.J y
,} ,.
;p e'
- m ,no g W. ' s ,
.h -
/f h .[
gmy y
};
- po ,.a!> f: d yp y y) 4x}'[7 f! l. "w :
g l . ~ ( go -Ji gc;f i y'
- ay. ,
3' f;i;f )l . l 'd l , q ; > Wf _ ;, . p ... t {g4 , u; y, flf-:) '3; ,
, , ' >, , lf. l
- s. .
g
.n}
.. n , , , . .
;y i
o
; q* fQ}y} ,f ,....
, i.y[ Q* W ? ,*}. _
- m. : E ;}f' % OgP M V % ~~
yqy (g,,
, ,x:3 ,.~ . y, a.
; , ,,x , .
My "y e e y [. m r r ~
;W,~st - c o. .
4 ,S
- 4. 'n-tym.;wayj g,@> g;
' Wu h
"f, q"7vw ,.g
'L- .g
,Y.Eg*ghggg %dQ ,qh ; 2 'a *' LM QR*Wa i 9 .D (9
((
h
$_' y' s y, a, a, ~, M.. g:y l N ; ;h I n l , ,h{ N 1 - 4 'k. , -[3 , i y j y f, Tjf * , - t ;. t a g , .- ~ 3 ,l 'I d $, & .h %M m MQhy > p, .. 6 * (g [y, e s[,.K ' de ' ]"f.g(k ' f, ji j i i g ' [-NW 'g t I ; p Q. , {.
- m. ;
n'o ,~ ., y i,' .s g;;n'%p - # -. > m fqn ,n v ns .y; . f ~ sw x-1 b y iyno w a e ~j Q. b , , gw.cf f - r , apt n a me e. F1 ,, w test sete;> . [ t' _ _ - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ ) S p.; ; - : - 3 . U
- - t p; , Q A
5 BM z .AAR H ~;j. .~
- NoHT G1
}: 1
- ; 7 :
- } : : I O1 1 - }: EW D R P = g :c: - : : : OD q e c_ :: d ME E r e o m N F . F e
- I 1- -
I M m a s R e ;- c h ; - .; E t ;i { T h 7 E t i w D e p a h i s yy y D E d e o m N d e / - . n ., ., . I i M m. / - - rs R ei t s , , / E T z dl ey a y ySo.D E H 1 l l yn at t n ne a em N LY1 1 me il re E L = e pe s A xt D T q ei n ON e ni x ME F M r af f y ob nd y....p I oe R ii sn E rm ar P pe mt X oe Cd E 6 g i F g~ _ - - 2 4 G , ' R T Y , ' R ) O N T ( S I 0 H 1 X E . L , / ' 0 P S C I Y D C L D A T A O O T L , ' E R I T N .
- E 2-0 0 0 2 0 1 -
5 6(DJ_. 9 O" X wUr t ma l' 2 3 A X~ }n / / AY 3' / .'- / / Oy y>N / os L / ? / ~ / * / 4 ' " 1 '7#/#/#####/#####/###/#####//////UK4 = W = AX = (D, + APS) cosp - W A Y = (Di + APS) sin 4 - L , , _j W2 + (D, + APS)2 - (L + ARB)2- _ 2 W(Di + APS) . APS = DIAGONAL DISPLACEMENT GAGE READING ARB = VERTICAL DISPLACEMENT GAGE READING Fig. 8 The method-used to obtain hori2ontal displacement from the interior gage. 229 l Y//////////////////////SS/////////l$b V/////////////////////S///////$$/ll/b = L P ; P h = 20 in. j 9 in, n . F M=Ph i e ~ AP D C P . . yA ~h [ ; FX + + i:i8j g g as L = 72 in. 3 in.-*l !- 72in. =l 43 in. -P w w B" B1' M + PL r 9 in. . 6 Fig. 9 50-psi-(0.34 MPa) average base shear stress load-deflection curve based on interior gage readings. 230 c . - _ _ _ 5 - 3 4 . G R T . ) E N L ( I C ~ Y C 1 0 D X A i ' 0 P O S L I D _ I S L A P T 0 O 5 T _ T - S R I F . 5 3 3 0 3 0 O 0- $0& 0 9O" X Jlug I 5 l $f& }W0 ~$q@JQ5%t4$$l R'Mb
- n Ace,c+
- e. a -
a# ~ ~ ., h?" og -). + j-n., g wgw>Kn* " 7, ", i 3 m, .-.> hf(' &. ,. .- , ?e i ;' f $$fh: >ql+. .hi l5n ,.;pp* A, g; a b p ' , p .. t,. .. % w-r ;rm , v q. A - ff g'%, 1a n ry- ,vs s,.: v - y .. . 3 . 4;
- g.
l&,f". !?c ~ y-i u a ., . N .fk y- ' c y i n: ar. s. 'M 7' 4 T. %.. , o ' . v. . 3 . y na t - A /,ff'sm;w. ' . ,3 p s , ; a/p$, uk-? ,y <. - 1 k' ,, V 4 y* %'1 (' %41 - -[ - s ., < c>. "&,; gd > ,p2 9e -A . s . wt , ,f 1 W5? l ppk}i 3 S I t, 'I(v .l 9 . ~sm 3;g , - b_, y1 J . qp;%g 3M sra ,. j' - - m% .J. P6, ' ' % .1 . 1 db SQ i- .j- _Yf@%;Il Q ., . s .m . \. s z.s,c ;Cl,m ?h & ," , 4S g. > A d + e 1- - , x, 3gi* fm... 4 t . - a. o .-, .- ,-l 4,- '! t % :4 , e # ,.__.L [! s~e w L ,g % .;m,%.%
- u&[w+%;fzygg punk
, p, ,.m,,~#y: s,%e e ', J c .. . 4 "M% , ,.4<.. _,Q . 3'. ' 4 @ ]g-- jf;YQ ,, . . ,. ;;, w!; p . ~ . , .c q , ,# + - e ~;4.h;ge{. tj , 4 y w q , g y yv' .f4 ,g.g y 1 w Fig. 11 Shear wall crack pattern after failure cycle (north side). '232 p*inlm - e..# : ami;(y,is.. ..N * ~ ":W >gy,t q ha[%e 9pc % n ,,,L;7 - : er ; % p.,- :" ,; i ,y - . ,,q
- y. p qgg, y., 4suphn~%g,g % ,
q vv
- m. t j q9
' 7y$ :n % . .a.4 -4 pw $s -. f, , .
- n. - K .
e p, n. w .p g',o Gr . ig y . .hnp + ,a - th1;nc A .. p; su,. ? kyg u. V, g [. . 8 ; .. k ' ry f%
- g f
.wi n a ,c y y Qiyyy, >p - 9 .. v y, ($ i ~d.,< M _ > i_. ,Wg:, s c e. . A .c,u# &e; p ; o%, ? J e&P p M g'. e \' . ' e n .' / .,f:,- 339 q ;; ,c w y y a; x /s , r. R. _ 4.y':y%, 1.k - {fg f,mx 4# , M 'r' &g _ %. k,iYm ,N F ,4;W.p- y ~t a d+' . '\ y v. , 'q: , . s V 't'ayy G: q)y=liM w , hl j ? \ { ft -W
- f. ;'f;h %. u.y .- .
s %' %./ ,. . / c 4 4 B: 0
- ]
Y4 i k ; a Y.h ') f.%. _. ' i *3,, 4 a 's~%brW ' j, 2 x . D sl h..TE; : /f d , ,1 tr t e i '!2d_ ; m $: h t.M' It k c g, d ! ' t a -p. y%m *Aij a j - .g ig Jt ' 4 .. '4 .; .h . ' A n,L W% R:) .. qf *, % y' '/ 3.w /, 4 li b } "I . : IM 'T . e .u ; pTyp,f.e;/44 i qf, m.d. ^ &Ni f fT O ah y$[;n)y n .bm j, L4 . .,: "y: L; $Uf n ,' f m A- 7n% A %y ? fV ( 3 Qy i" M ? s,, , ,8. q .. . f y' 3 4 ._ ~a,WhF <,w9 Jf d M,., i e pd : ,#'i1 . 7 u,' >:7/e ^ v.e - 6! 64-c F W ! fan;pe /Mh @M.A i . g M5h.InM 1 1 y $gg$ r p." s g d% W7Q mnp k;4%.. o qvt g ;r ':1.p;o%e. M9)stn n;; 4;3 m9,, g., .e. AMd ;. a,p q M - v# ,,4.f r^ wafffSpR edh@>a % L E AH b+ T . @a%gf5Miy? bN ,a, n c kqA,%... . ,a.. m i g. .;+/ . a;m. .__m y y;/; wn yn f e meugu;'; , nym. 3 6.;3.< -d.A .. . % g 0%%l% _- ML ow RkGgfQ;}pAj?lf?W.gdwia#[Q~kfQj( m u - Fig.12 Shear wall crack pattern after cycle (south side) 233 JK/14 INEL/USNRC PIPE DAMPING EXPERIMENTS AND STUDIES A.G. WARE IDAHO NATIONAL ENGINEERING LABORATORY EG&G IDAHO, INC. IDAHO FALLS, IDAHO, 83415 Since the previous paper on this subject presented at the 8th SMiRT Conference, the Idaho National Engineering Laboratory (INEL) has conducted further research on piping system damping for the United States Nuclear Regulatory Commission (USNRC). These efforts have included vibration tests on two laboratory piping systems at response frequencies up to 100 Hz, and damping data calculations from both of these two systems and from a third laboratory piping system test series. In addition, a statistical analysis was performed on piping system damping data from tests representative of seismic and hydrodynamic events of greater than minimal excitation. The results of this program will be used to assist regulators in establishing suitable damping values for use in dynamic analyses of nuclear piping systems, and in revising USNRC Regulatory Guide (RG) 1.61. INTRODUCTION The Idaho National Engineering Laboratory (INEL) has been conducting a research program to assist the United States Nuclear Regulatory Commission (USNRC) in determining best-estimate damping values for use in the design and analysis of nuclear power plant (NPP) piping systems [1]. The results of this program have been used by the U.S. Pressure Vessel Research Committee (PVRC) to i propose higher allowable pipe damping values [2] than in USNRC Regulatory Guide (RG) 1,61 [3]. Higher damping values allow the design of more flexible systems with fewer supports, making the piping better able to expand and contract during thermal loadings, less costly, and will reduce inspections, maintenance, and worker radiation exposure. The PVRC proposal has been approved as Code Case N-411 to the American Society of Mechanical Engineers Boiler and Pressure
- 1. Work supported by the U.S. Nuclear Regulatory Commission, Office of Nuclear Regulatory Research, under DOE Contract No. DE-AC07-761001570.
235 r Vessel Code (ASME Code) [4], and in Revision 24 to USNRC RG 1.84 [5]. l- Comparisons of RG 1.61 and PVRC curves-for piping system damping are shown in Figure 1. The benefits of this change in allowable damping values are already being' demonstrated as.a cost-savings measure on commercial NPPs. For example, Gilbert / Commonwealth, Inc., has determined that 10 of the 23 snubbers in one nuclear piping system could be eliminated using the revised damping values [6]. Using PVRC damping and other new techniques (zero deflection criteria and-peak shifting), 18 of the 23 snubbers could be eliminated or replaced by rigid hardware. Similarly, using PVRC damping and independent support motion (ISM) techniques, 76% of the snubbers in three Commonwealth Edison piping systems (main steam, standby liquid control, and residual heat removal) could be eliminated [7]. Since nucle.. power plant piping systems. include several hundred (over 1000 per unit on recently completed plants) snubbers, the savings in design, procurement, installation, and maintenance of snubbers are. considerable. The PVRC proposal has subsequently been modified as shown in Figure 2 to include the effects of heavily insulated piping based on data from the U.S. Hanford Engineering Development Laboratory (HEDL) [8]. The first task described in this paper included conducting additional tests on a 5-in. INEL laboratory piping system and evaluating ANCO Engineers 6-in. laboratory piping system vibration data to further examine the parameters influencing damping in the seismic frequency range, and to add data to the already existing data base in support of the PVRC recommendations. I The allowable damping values in RG 1.61 and the PVRC proposals cover only the seismic frequency range, O to 33 Hz. However, there are several transients ,1 1 I 236 i e 6 i e i i s. a i e e i e i a i e i 5 PVRC proposal g- RO 1.61 SSE and OBE o alldiameters o- ' SSto D>12 in n g4 - 73 3 - - - - - - - - - - \---- - - - - - - - - - - - - - RG 1.61-p SSE. D R12 in.e OBEe D >12 in *2 - ~ ~ - - - - - - - - - -~~ f RG 1.61 OBEe DR12 in --- . .SSE -5 Safe shutdown earthquake OBE Operating basis earthquake D Pipe diameter g i I e a e e e i f f I i i f I 'I O 2 4 6 8 10 12 .14 16 18 20 22 24 26 28 30 32 Frequency (Hz) sugg 4 g,,o Figure le Comparison of RG 1e61 and PVRC pipe damping curvese ,in..n . e. .pil, mm. . w,e.sei v.i. percean eriksm pree,eae, tan e 23i pee same e en to me-ei== to sin er nu as e ir o ir + 6 30 Ng er more 2% 14.8 IA 10.8 In + 2 10- is is a to n ,,,,,, see me,,o .,ee e., In . en en n.,e inn, m.mei ,e.essend nee 6.$81 O IR e meeshes essei s,ahe (scia.i s.n .e bei as to 08 in - oss E u g6 8 4-R - o 23 E a E 2-y 8.re and hgntly mimated pipeg (IA . o.23) 0- i e i i i i i i i i , i i i i 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 Frequency (Hz) ..e Figure 2. HEDL proposal for heavily insulated pipinge 1 237 .in which the forcing frequencies on piping are > 33 Hz. These are mainly fluid l; . induced, such as water hammer, pressure relief valve discharge, and 1' i loss-of-coolant-accident (LOCA) related hydrodynamic loads. Neither RG 1,61 L nor the ASME Code provides guidelines for damping values to be used for these high-frequency loads. With no official guidelines, the usual approach has been to use the RG 1.61 values for high-frequency analyses, although these values of damping have not been specifically approved for use in this frequency range nor for fluid-induced transients. Recognizing the need for allowable damping values at higher modal frequencies, the USNRC and the INEL expanded the pipe damping research program to include a.second task covering modal frequencies in the 33 to 100 Hz range. A considerable amount of work has recently been performed by the U.S. . l Lawrence Livermore National' Laboratory (LLNL) on probabilistic risk assessment i (PRA) analyses of seismic safety margins. A risk assessment is conducted using a Seismic Safety Margins Research Program.(SSMRP) methodology that has been developed as part of this program. To actually assess the effect of an earthquake on a given NPP using SSMRP methodology, specific information concerning site and plant geometries, seismic input, structure and piping response, and component fragilities must be known or assumed as input to the j SMACS computer code [9]. Damping is one of the piping response parameters. Previously, parameter values have been sampled from assumed probability distributions according to a Latin hypercube sampling procedure. The mean values and uncertainties for piping system. damping have not yet been determined from actual vibration test data. The third research task was to conduct statistical analyses of damping data for those piping systems in the nuclear piping data bank that are representative of light water reactor (LWR) NPPs,and 238 from tests in which the excitation was of the order that might be expected during a significant seismic or hydrodynamic event. Thus future SSMRP analyses i can be conducted using piping system damping values obtained from actual test data, rather than assumed probability distributions. It was recognized by the authors of RG 1.61 that damping was amplitude dependent, and thus they allowed an extra 1% of critical damping for the higher ' level safe shutdown earthquake (SSE) than for the operating basis earthquake (0BE). Since the issue of RG 1.61 in 1973, considerably more damping tests have been conducted worldwide at higher strain levels, and our knowledge a j piping system damping at plastic strain levels has been increased. It has been noted that once the strain level in the piping system reaches a certain strain threshold [approximately the material yield stress (Sy)], the damping tends to increase significantly [10]. This phenomenon has been attributed to material damping in the piping and supports. Because of the increased damping at high strain levels, the use of linear elastic stress analysis methods with RG 1.61 or PVRC damping will result in drastic overpredictions of the bending moments and stresses in piping systems excited sufficiently to vibrate in the plastic range. This will cause a significant design margin to be present in the stress analysis predictions. Consequently, the USNRC authorized the INEL to evaluate j j high strain level damping data from recent and ongoing pipe vibration testing, and to outline ways to combine and present these data. A further objective of this fourth research task was to provide quantitative measures of the design margins in terms of piping damping values. I 239 LABORATORY PIPING TESTS AT SEISMIC FREQUENCIES The first INEL damping tests were conducted on straight s'octions of 3-in. (76-mm) and 8-in. (203-mm) carbon steel piping [11]. From these tests, the effects of various typical piping supports and amplitude ranges were investigated. -Using.the results of these essentially.one-dimensional tests, a more complicated three-dimensional S-in. (3?.7-mm) laboratory piping system was designed and a series of experiments were conducted to determine the. correlation of. excitation type (shaker and' snapback), supports, amplitude, ' insulation (insulated and.uninsulated), response frequency, and pressure'on -pipe damping [12]. A diagram of the test configuration is shown in Figure 3. Another series of piping vibration tests was conducted by ANCO Engineers for the USNRC and the U.S. Electric Power Resoarch Institute (EPRI) on two 6-in 1 (153-mm) laboratory piping systems, shown in Figure 4. An extensive series of i experiments using a variety of test excitation methods, supports, and response . amplitudes was completed by ANCO Engineers as part of this program. The USNRC furnished the INEL with the raw data from these tests to reduce and to evaluate the damping'results. In' addition to the main series of tests on the ANC0 laboratory system, the INEL participated in another, smaller, series of tests-in which ANCO supplied the excitation to the system and the INEL collected and ~ analyzed the data. _The effects of varying a number of parameters on damping has been investigated in these two series [12] and will be discussed individually below. 240 l *-- 8 f t 0 h e 4 Shaker Bl [** C A l' Horizontd R . f Snap - 45* >. , 1r , vertied Snap . f- ~~~f j 7 Y A \.,8 A .B N7 hitn ' Stra.ri 6 ft O h t 6 'n Rose es Accelerometer c 1' 6" - --f p' ~ ~ nn - nn -Figure 3. INEL 5-in. (127-m) piping system configuration. 0.7.;' seI4e *A.. ..A .A :=r s.: ~ N a .._ x== eL n ma v' %, ,eve.- , a** v.,m . - o **' ,s s x y .e 4 1 Lesetlen 4 g, " ~ '
- a. , __J M*J'T w
28 , , , ,' W ha s.ei. .a ,,e w , % ca ca 54 . reac i..e , Figure 4. ANCO laboratory piping systems. 241 _ _ _ _ .___-___________a Pressure Damping tended to decrease with increasing pressure for the INEL 5-in. {127-mm) system with end supports only, while the damping did not change with pressure when a support was added at location A in Figure 3. Overall, the effect of pressure on damping at seismic levels is considered minimal when compared to the other infidencing parameters. Supports Starting with no supports and adding a rigid strut at location C in Figure 3 (INEL tests) had little effect on system damping, while adding a support at A or B increased the damaing. A mechanical snubbar dissipated the most energy, producing the highest damping levels in any of the lower excitation level tests. During swept sine, random, and simulated earthquake excitation tests, there was no noticable support effect in the ANCO series, while in the impulse tests the cases with mechanical snubbers at both intermediate supports produced higher damping than did other support conditions. Frequency In contrast to much of the data in the technical literature which shows an inverse frequency to damping relationship, data from the INEL series with uninsulated pipe and shaker excitation showed damping was fairly constant with mode number. For the insulated pipe-shaker excitation data, damping was lowest at the first mode (8 Hz), was maximum around 15 Hz, then decreased with frequency. In the uninsulated pipe-snapback excitation tests, the only frequency dependence was for the mechanical snubber configurations, where an inverse frequency to damping relationship was apparent. It is concluded that the frequency effect often noted is really a result of the interaction of the 242 , modes with the supports. Those modes causing high loads in energy dissipating supports have higher damping. The lower modes have significant interactions with the supports for larger piping systems' common to most NPPs, while most of the vibrational motion of the higher modes takes place between supports, often at. lower amplitudes of vibration. In.the ANCO test series there was a definite trend of an inverse frequency to damping relationship for the swept sine, random, and simulated earthquake tests (Figure 5), but no such trend was observed for the impulse tests. Insulation The INEL test series shaker excitation tests showed a pronounced increase in damping.in all modes when the insulation was added (Figure.6), while the snapback excitation tests showed virtually no effect on the few modes that responded. This was probably more of an excitation method effect rather than an amplitude effect. The piping was continually vibrated against the insulation during the shaker tests, resulting'in considerable energy loss due to friction. However, the insulation moved freely with the pipe during the snapback tests, with no additional energy being input to the system during cycling, and when resistance was encountered, both. pipe and insulation changed directions in a consistent manner. Since only a few modes could be identified for the snapback excitation-insulated pipe series, the other modes may have been so heavily damped that the free vibration decay did not excite them. The shaker tests would be more representative of an earthquake, which produces random motions. 243 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ a l t 3 y , , . -> . i i i s i T o Random S *g 1 e Swept sine U l % 2le ~ I "E I s o . O e E. ! \* * , ].E 1h 1 8, N.%'*% o o 9- I . - C o O O , i , , . , , , i g 0 2 4 6 8 10 12 14 16 18 20 22 Frequency (Hz) l Figure 5. ANCO 6-in. (153-mm) system excitation method results. 10 0 Rigid strut at A or B, no support at C 8- A Support at C open, no insulation ,_ ^ insulated , T7-o " E e D 8-o 8 g 5- $ & 4 E 4-E y 3- e D 2- a g , O D A }- 0 & A D D 6 0 i i i i i i . .# 10 to 30 40 Frequency (Hz) Figure 6. Effect of insulation on INEL 5-in. (127-mm) system. 244 . Amplitude The INEL high-level test series, at dominant frequencies of around 7, 9, and 18 Hz, all showed a trend of increasing damping with strain level. Although the low-level damping was about 1% of critical for all three test sets in the series, the slope of the increase was different for each higher level case. This is probably because of the difference in mode shapes. Damping is proportional to the ratio of the dissipated energy to the total energy. Since each mode shape places an individual stress pattern in the piping and since the vibrational energy associated with each mode shape is unique from those of other mode shapes, it stands to reason that the ratio of the two will have its own value for each case. Furthermore, each mode shape will cause the pipe to impact with the supports at a different force level. An important point to note is that the same effect was noted for the higher frequency (18 Hz) case as for the lower frequency cases. Some of the data in the ANC0 series exhibited a similar trend. Excitation method Base band random excitation (excitation over a wide range of frequencies) produced generally higher damping in the INEL tests than did the swept sine tests. This is attributed to lower resolution of the data and nonuniform energy distribution among the modes for base band excitation. Since the number of lines of resolution are constant as set by the data analyzer (400 spectral lines of resolution in these tests), more data points are concentrated in the region of the damping versus frequency curve used to calculate the damping for narrow band data acquisition than for the broader base band excitation. Thus there are more data points to better define the shape of the curve for half-power damping calculations. Excitation levels high enough to produce a 245 good measure of damping were achieved by concentrating the excitation energy in only one or two modes. The zoom random (excitation of only one or two modes at a time) and swept sine methods were judged to produce comparable damping levels. The swept sine method is more time consuming and fatigues the metal in the pipe more rapidly; consequently, the zoom random was the preferred method of excitation in the tests. In the ANCO series the swept sine, random, and simulated earthquake excitation methods all produced almost the same results. The impulse tests gave similar low damping levels, but did not match quite as ) closely as the other excitation tests for an.y frequency. In particelar, the damping was slightly higher in the higher frequency range for the impulse j tests. Calculational method l For the low response level tests where the damping was fairly constant with response amplitude, the half-power, curve-fit, and logarithmic decrement methods gave approximately the same results. However, at higher response levels where damping was not constant with response amplitude, the half-power method gave a damping value between the extremes of the logarithmic-decrement computed points. Each method offers its own unique advantage. The logaritmic decrement method is good for tracking damping when it is amplitude dependent through a decaying amplitude range. It is difficult to use when multiple modes are excited, and extensive digital filtering of the data may be required. The curve-fit and half-power methods are much easier to perform; yet they provide only a single average damping value for each test, even if the amplitude is varying nonlinearly. The curve-fit method is judged to be more accurate than 1 the half-power method because it considers both the amplitude and phase of the response. In order to achieve the best understanding of the damping associated 246 L_----- with a piping system, excitation should be undertaken by a variety of methods and damping calculated using several methods. Then, based upon a comparison of results, a better overall perspective can be gained. This approach would highlight some of the inconsistencies which are often observed in damping test results. HIGH FREQUENCY DAMPING EVALUATION The plan followed in conducting this research task was to review published literature and the USNRC/INEL nuclear piping data bank for high-frequency damping results, conduct tests at high frequencies on laboratory 3- and 5-in. (76- and 127-mm) piping systems at the INEL, evaluate the data and recommend damping values for use in analyses of piping vibrations at high frequencies [13]. Relevant high-frequency piping system damping data was found in only two references. Kraftwerk Union (KWU) in the Federal Republic of Germany conducted a vibration test series on 250-mm (10-in.) NPP piping, with and without insulation. There were two high-frequency modes each with two data points, one insulated and the other uninsulated, for each mode. The General Electric (GE) Company also conducted vibration tests on BWR piping systems. The excitation was by safety relief valve (SRV) discharge transients, and these tests provided the only data in this review in which excitation was produced by bona fide fluid-induced transients. High damping values of 5.9%, 7.1%, and 16% of critical were reported at 28 Hz, near the upper frequency limit of the seismic range. 247 High-frequency damping values'in an INEL test series using a section of 3-in. (76-mm) straight pipe ranged from 1 to 3% of critical for the configuration with no intermediate supports. For configurations with intermediate ~ supports, high-frequency damping values ranged from 3% to 7% percent of critical; averaging'4.1% of critical. The high-frequency damping-values'were' generally > damping values for the seismic range. Intermediate supports supplied an additional.2% to 3% of critical damp'ing at high l frequencies. l I 1 The 5-in. (127-mm) INEL laboratory system in Figure 3 that included elbows, bends, and reducers was also tested at higher frequencies, both in the insulated and'uninsulated conditions'and at high and low stress levels. Results for the uninsulated cases with no intermediate supports showed that the j modes above 60 Hz had significantly higher damping than did the lower modes. 1 For the uninsulated pipe with a rigid strut as the intermediate support, damping increased at frequencies > 50 Hz and the higher level excitations that produced stresses near Sy resulted in higher damping than did the lower level (Sy/4) case. At frequencies > 50 Hz, results for the mechanical and hydraulic snubbers were comparable to the rigid strut results. Damping values at frequencies > 50 Hz ranged from 2% to 8% of critical (uninsulated condition). For the higher excitation levels in which pipe stress was near Sy, damping ranged from 4% to 12% of critical which is 2% to 4% of critical higher than for the lower excitation case in which pipe stress was about Sy/4. However, no definite difference was noted between the supported and unsupported condition damping values in the insulated test results. Damping in the high-frequency 1 range was scattered from 3% to 7% of critical (insulated condition). For modes _ with low damping, i.e. below 1% of critical, insulation increased the damping. 248 l _ _ . For modes > 50 Hz for which the damping was already high, insulation had minimal effect on damping. The data chosen as the basis for estimating high-frequency damping values was heavily weighted by the results of the INEL high-frequency 3- and 5-in. (76- and 127-mm) tests (Figure 7). One characteristic of the data worth noting is that damping appears to be higher when insulation is present (Figure 8). Another feature of interest is that damping levels were greater at the higher excitation levels, i.e., when stress levels were near Sy, and for systems with intermediate supports, which are more representative of conditions which would be found in actual nuclear piping systems. It was concluded that suitable representations of the test data could be achieved by increasing the PVRC uninsulated piping curve (Figure 1) from a value of 2% of critical damping to a value of 3% for frequencies > 20 Hz, and by extending the proposed insulated 16 ~ INEL 3.in. 16 O uninsuiated 15 - INE . O uninsulated 14 - + uraiea 13 - Kwu $ 12 - , NY.Yd 11 - , !,"A % ;"' " i , g 1o - A v lasuiaied # 9- A E 8~
- 9 SSIn me + Regression
{ 7* . x* using all data above 20 Hz *
- curve using g ,
o 6- . x* g 0 + a# data above 33 H v* 5- x+ g g o 4* * +A 4- 9, 3 O + 2- * * , ,o , 1- o 4 o a o 8 i i i . , , , 20 40 60 80 100 Frequency (Hz) e 3:4 Figure 7. High frequency test data. 249 V ' 10 - 2 m .o 8 ....___________________________________________.IR'-0.58- ,g insuiated test results u o g 6 -- x IR = 0.33 m h _________________________.---- IR = 0.23 ;4 - . Uninsulated test results (F = 50 Hz) Mets 8.tv.t.reu.'!s.1F, g,2,0,liz,[,,,,,,,,,,,,,,,,,,,,,,,,, a E ~ M 2-Bare and lightly insulated .O piping (IR < 0.23)
- 0 . . . . .
. . i. l 0 10- 20 30 40 50 .60 70 80 90' 100 Frequency (Hz) Figure 8. Comparison.of high-frequency results to PVRC curve. For IR = 0.23 Bare and lightly insulated incremented damping . Total damping Frstuency (IR 4 0 231 pipe damping - due to insulation percent critical 3 to Or or less 5% 9 IR' 9 IR +5 20 A2 or more 3% 8.6 1R 8.6 IR + 3 jQ- 10 H2 to 20 Hz Stral0ht line interpolation For IR
- 0.23 use bare and lightly insulated damping values O IR Insutstion weight ratio (actual value but not to exceed 0.58)
M ,o . 8 IR = 0.58 O g 6- -__ A IR = 0.33 ~ [. IR = 0.23 4-cn .5 ~ C. sare and tigntiy insuiated Em 2 -. piping (IR < 0.23) o 0- . 0 10 20 30 40 50 60 70 80 90 100 Frequency (Hz) Figure 9. High-frequency representation of test data. 250 curves (Figure 2) to 100 Hz at the levels as shown in Figure 9. This recommendation has been endorsed by the PVRC Technical Committee on Piping Systems {14], and has been submitted to the ASME Working Group on Dynamic Analysis for consideration as a revised damping position. 4 STATISTICAL DISTRIBUTION OF DATA Test data from 27 LWR type piping systems where the piping had been vibrated at levels representative of at least moderate severity seismic or hydrodynamic transients were chosen for a statistical damping study [15]. Most of these piping systems are in actual NPPs, and the lowest mode was < 8 Hz for over 80% of them. Thus a great deal of the damping data was from low frequency piping systems. The data were broken into eight groups (see Table 1) representing all data and data used in the original USNRC regression analysis, l NPP and test (laboratory) systems, insulated systems, and pipe sizes of > or < 12 in. (305-mm). A further categorization was :onducted (see Table 2), separating the data into three runs that represented all data, one data point per mode for each piping system, and the damping of the lowest mode at the highest excitation for each piping system. TABLE 1. DATA GROUPS GROUP DESCRIPTION 1 All data 2 Data used in original PVRC regression analysis 3 Data from NPP piping systems 4 Data from test or laboratory piping systems 5 Data from insulated piping systems 6 Data from all piping systems > 12 in. (305 mm) 7 Data from all piping systems < 12 in. (305 mm) ; 8 Data from NPP piping systems < 12 in. (305 mm) 251 TABLE 2-. DESCRIPTIONS'0F 3 DIFFERENT DATA SUBDIVISIONS RUN DESCRIPTION 1 all damping values 2 one damping value per mode representing the average of the damping values for that mode from multiple tests on the same system 3 a single damping value for each system representing the highest excitation of the lowest mode. TABLE 3. NORMAL AND LOGNORMAL DISTRIBUTION PARAMETERS; % OF CRITICAL DAMPING (SEE TABLES 1 AND 2 FOR DESCRIPTIONS OF GROUPS AND RUNS) ASSUMED NORMAL DIST. ASSUMED LOGNORMAL DIST. ARITHMETIC ARITHMETIC GEOMETRIC GE0 METRIC i GROUP RUN MEAN STAN. DEV MEAN STAN. DEV. BETA 1 1 3.84 3.33 2.82- 2.28 0.82 2 4.45 3.82 3.36 2.16 0.77 3 6.96 5.67 5.06 2.36 0.86 2 1 4.22 3.53 3.29 2.06 0.72 2 4.38 3.12 3.56 1.91 0.65 3 7.42 4.07 6.50 1.75 0.56 3 1 4.04 3.52 3.08 2.12 0.75 2 4.37 3.61 3.35 2.11 0.75 3 7.62 5.93 5.67 2.30 0.84 l 4 1 3.38 2.80 2.32 2.57 0.94 2 4.72 4.55 3.39 2.32 0.84 3 4.45 3.96 3.28 2.39 0.87 i 5 1 3.70 3.39 2.68 2.32 0.84 2 4.40 3.62 3.36 2.14 0.76 3 7.73 5.92 5.74 2.32 0.84 6 1 4.16 4.0a 3.05 2.27 0.82 2 4.48 3.03 3.72 1.83 0.60 3 7.47 4.67 6.10 1.98 0.68 7 1 3.72 3.03 2.75 2.28 0.82 2 4.43 4.33 3.12 2.38 0.87 3 6.61 6.39 4.44 2.61 0.96 8 1 4.41 3.26 3.55 1.94 0.66 2 4.84 4.48 3.39 2.49 0.91 3 8.68 7.13 6.13 2.67 0.98 252 In the first part of this study damping was treated as independent of frequency (or mode number). The statistical analysis showed that a lognormal fit provided a suitable approximation of the raw data (see Figure 10 for histogram and cumulative distribution plot of Group 1, Run 2). A statistical summary is presented in Table 3. Average damping values ranged from 3.38% of critical for test systems to 4.41% of critical for NPP systems < 12 in. (305-mm) in the Run 1 cases (duplicate tests for each system included). When duplicate tests were eliminated (Run 2), the averages for all groups were nearly the same, ranging from 4.38% of critical to 4.S4% of critical. Damping values ranged from 4.45% of critical for test systems to 8.68% of critical for NPP systems < 12 in. (305 mm) in Run 3, consisting of the lowest mode at its highest excitation level for each data set. The fact that Run 3 values are higher is not surprising since in many tests the lowest mode was the most heavily excited, with little of the input energy being concentrated in the higher modes. Tests have shown that damping values are relatively lower in the lightly excited modes. Thus the Run 3 values are probably the most representative of the damping that might be expected in the most heavily seismically excited modes. The Run 2 values would give the best sets of values to use over the entire frequency range, because some of the unequal weighting by the data sets with more damping values than others has been eliminated, while multiple modes have been retained. The standard deviation is fairly large in all 24 subdivisions (8 Groups, 3 Runs / Group), reflecting the generally wide scatter in the data. The geometric mean damping values calculated for the lognormal distribution fits are all lower than the average values. This is a result of the averaging 253 3 g . . 4 3 .. 4 g . n . l . < q ' i. . n . y , j _ ""!::::::::::::::!lllll!!!lfllj; (a) Histogram . . . . .... . : ,c. : . .. . - .. 8 ,,..-i........ ............ ...a ,. i.- ...a ...a f l. ...y q ...a .8 , . . . t. . (b) Cumulative distribution plot. Figure 10. Group 1, Run 2 data 254 'of.the natural logarithms of the damping values which gives a higher weighting to the lower raw data values. Most of these data were from tests in which the maximum stresses in the piping systems were $ OBE levels. Very.few of the damping values were recorded at stress levels representative of an SSE. Therefore, the conclusions reached in this report are most applicable to OBE events. Since damping values are generally higher at SSE levels than at OBE levels, the results can be considered to be lower than comparable SSE damping values. EVALUATION OF DAMPING AT HIGH STRAIN LEVELS Although the concept of damping as related to a linear elastic system is not strictly defined for elastoplastic piping system vibrations, a concept of equivalent damping was used in Reference 16 to estimate the response margins in predicted piping system behavior. The equivalent damping concept used was based on the computed linear' damping that would cause amplifications of an ' elastic system to match the available experimental data at high strain levels. The theoretical variation of damping and frequency with ductility (the ratio of the total elastoplastic displacement =to the elastic displacement at the system yield point) is shown in Figures 11 and 12. Damping is shown to increase at higher ductilities using both an energy concept and the logarithmic decrement method. The increase also depends on the change-in system stiffness due to the 4 I formation of areas of plasticity. The results are comparable to those of Lazzeri [17]. l l 255 l l l 60 ~5 # = k /kj 2 for bilinear system .e 50 - " s
- 0 gy?)
40 - O - g 30 - " ~ , Lazzeri (# = 0) g . 1/4 $. 20 - - / , { E ' m 10 - / ! O . 1 I o . . . i - i - 0 1 2 3 4 5 Ductility - 1 Figure 11. Theoretical variation of damping with ductility. 0- " ' ' ' ' ' ' ' ' ' . \ \ 0.9 - \ - p . 3f4 \ \ c 0.8 - 's N g s 4 = 1/2 N 0 0.7 - '*s C ,,'.~G ~ ~~ ' o, ' ~~.~_ 0.6 - a ., o (Epp, 0.5 . s - s s s 0 1 2 3 4 5 Ductility - 1 Figure 12. Frequency shift due to ductility. 256 High strain damping data from seven different sources are plotted on a common basis as a function of ductility in Figure 13. Data based on local strain ductility tend to follow an elastic-perfectly-plastic (EPP) material curve, while data based on global ductility is somewhat lower. At the highest observed strain levels, the damping values are estimated to be 30% to 50% of critical. This results in a response margin of actual input acceleration to linear elastic computed acceleration (using 5% of critical damping) of from 2.5 to 3 for seismic excitations and from 6 to 10 for sinusoidal excitations. 50 ; ,- / 7' l ' ,/ = 40 - s' i 3 l,' 5 !/ U l,'! - 30 - ll o ;r m,* 4 en 20 - / .E Q, I l k ----- Theoretical o 10 - Data / / 0 i i i . 0 2 4 6 8 10 Ductility Figure 13. Normalized test data compared with theoretical curves. 257 Although an estimate has been made as to the expected damping and response margins of piping systems at the time of failure, such information is not well suited to predict piping system failure by itself. The proper method is to conduct an inelastic analysis that accounts for the nonlinear change in system properties at high load levels. This must include a model of system geometry and the determination of the localized strain patterns that will occur before system failure. The 30% to 50% damping values stated in this paper would not be used in such analyses because the material hysteresis or revised plastic spectra in the mathematical model will already take into account this energy dissipation. Rather, a representative elastic damping value such as 5% for actual nuclear power plant piping systems and 1 to 2% for laboratory systems should be used, i j What does the 30% to 50% of critical damping for piping systems immediately prior to failure mean? If a linear elastic analysis were to be q performed on a system with hinges at the locations at which the cross-section had yielded, This damping range would be required to predict the response motion. However, at 50% of critical damping there is no dynamic amplification of the motion. Thus by the time just prior to failure the piping system appears to be forced into a plastic dynamic deflection shape with all of the input energy being absorbed by the system by means of changing its plastic shape.
SUMMARY
Four tasks aimed at describing the parameters that influence damping in nuclear power plant piping systems have recently been performed at the INEL In the first two of these tasks the parameters were investigated in the 258
seismic (< 33 Hz) and high frequency (33 to 100 Hz) ranges. A recommendation for pipe damping levels up to 100 Hz has been endorsed by the PVRC [14]. Results of the third task showed that a lognormal probability distribution of damping was found to give a good statistical representation of the data, and the means and standard deviations of damping for several combinations of parameters were determined for use in PRA analyses. The final task evaluated damping at high strain levels, and concluded that once plastic action begins, the apparent damping begins to steadily increase with strain level. REFEPENCES
- 1. A.G. Ware and J.G. Arendts, "The NRC/EG&G Nuclear Piping System Damping Study Program", Paper K17/3, Transactions of the 8th International Conference on Structural Mechanics in Reactor Technology (SMiRT), Brussels, Belgium, pp. 285-290, 1985.
- 2. Welding Research Council Bulletin 300, " Technical Position on Criteria Establishment", ISSN 0043-2326, 1984.
- 3. U.S. Atomic Energy Commission, " Damping Values for Seismic Design of Nuclear Power Plants", Regulatory Guide 1.61, 1973.
- 4. American Society of Mechanical Engineers, ASME Boiler and Pressure Vessel Code, Section III, Nuclear Power Plant Components, Division 1, 1983 Edition.
- 5. U.S. Nuclear Regulatory Commission, " Design and Fabrication Code Case Acceptability: ASME Section III Division 1", Regulatory Guide 1.84, Rev.
24, 1986.
- 6. L.P. Buchanan, " Snubber Optimization Results on In-plant Piping System",
Seismic Engineering for Piping Systems, Tanks, and Power Plant Equipment, American Society of Mechanical Engineers, PVP-Vol. 108, pp. 31-42, 1986.
- 7. R. Srinivasan, J.J. Marianyi, and J.T. Fox, "A Snubber Reduction Pilot Program", Seismic Engineering for Piping Systems, Tanks, and Power Plant Equipment, American Society of Mechanical Engineers, PVP-Vol. 108, pp.
177-183, 1986.
- 8. M.J. Ande mon, M.R. Lindquist, and L.K. Severud, Further Considerations for Damping in Heavily Insulated Pipe Systems, HEDL-SA-3258-SP, Hanford Engineering Development Laboratory, Richland, Washington, 1985.
- 9. J.J. Johnson et al., SMACS--Seismic Methodology Analysis Chain with Statistics, Lawrence Livermore National Laboratory, Livermore, California, 1 NUREG/CR-1780, UCRL-53021, Vol. 9, 1981.
259
- 10. A.G. Ware, " Pipe Damping", Proceeding of the U.S. Nuclear Regulatory Commission 13th Water Reactor Safety Research Information Meeting, Gaithersburg, Maryland, October 1985, NUREG/CP-0072, Vol. 3, pp. 73-82, 1986,
- 11. A.G. Ware and G.L. Thinnes, Damping Results for Straight Sections of 3 in. j and 8 in. Unpressurized Pipes, NUREG/CR-3722, EGG-2305, EG&G Idaho, Inc., '
Idaho Falls, Idaho, 1984. 1
- 12. A.G. Ware and J.G. Arendts, Pipe Damping--Laboratory Tests in the Seismic Range, NUREG/CR-4529, EGG-2447, EG&G Idaho, Inc., Idaho Falls, Idaho, 1986.
- 13. A.G. Ware, Pipe Damping--Results of Vibration Tests in the 33 to 100 Hertz Frequency Range, NUREG/CR-4562. EGG-2450, EG&G Idaho, Inc., Idaho Falls, Idaho, 1986.
- 14. " Minutes of PVRC Meetings, January 26-28, 1987", Welding Research Council Progress Reports, Vol. XLII No. 3/4, March / April 1987.
- 15. A.G. Ware, " Statistical Evaluation of Light Water Reactor Piping Damping Data Representative cf Seismic and Hydrodynamic Events", EGG-EA-7260, EG&G Idaho, Inc., Idaho Falls, Idaho, 1986.
- 16. ...G. Ware, "An Evaluation of Damping in Piping Systems at High Strain
..evels", EGG-EA-7380 (Draft), EG&G Idaho, Inc., Idaho falls, Idaho, 1986. ;
- 17. L. Lazzeri and M. Scala, " Equivalent Damping in Piping Due to Local Yielding", Seismic Engineering in Piping Systems, Tanks, and Power Plant Equipment, American Society of Mechanical Engineers, PVP Vol. 108, pp.
5-11, 1986. NOTICE . l This paper was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, or any of their employees, makes any warranty, expressed or implied, or assumes any legal liability or responsibility for any third party's use, or the results of such use, of any information, apparatus, product or process disclosed in this report, or represents that its use by such third party would not infringe privately owned rights. The views expressed in this paper are not necessarily those of the U.S. Nuclear Regulatory Commission. 260
NRC PDRM 338 U S NUCLEAR REGULAYORv COWMISBIDN ' "k PO* ' NUM8E " 'd 8'8888 8 8 'DC, 888 " " # '*F# II 841 "Z"3# ' BIBUOGRAPHIC DATA SHEET
... .Niv ucr.og ,R. Rev.R , NUREG/CP- 88 2 Tsf L4 AND $UBIITL - 3 LE AVI OLAE Transaction of the 9th International Conference on ,!
Structural i hanics in Reactor Technology l Panel Session K: Structural and Mechanical Engineering / *a*"""'c""' l Research at th U.S. Nuclear Regulatory Commission fo*'" a"
. A u , RO.,,,, l Ane 1987 ,
f . oA1 R rOR1,ssuto l Compiled by B. S. rowzin MONTH vtAR [ July l ] 1987 { 7 PtRFQRMsNG ORGANS 2AllON NAME AND totNG ADORESS tsacessete Coel R0JECirT ASKtWOHK UN#1 NUMSER I Division of Engineeri , , ,N ooR A,, N ,,,,, Office of Nuclear Regu tory Research U.S. Nuclear Regulatory ommission ! Washington, DC 20555 - Technical 10 trONSONiNG DRGANi2 Af SON NAME AND MAtatNG ADD S& ttactua I,p coaes - its YvPt Of htPORT j Sama as 7. above.
. Vt RaOD COvthtD rinerena*ue nces) 12 EVPPb tMtNT AR v NOTt5 To be held at Lausanne, Switzerland - ugust 7-21, 1987 3 A.si ,, ACi am.,,. ., , .,
These transactions of the JK panel sessi includepreprints of papers which are listed in the Second Announcement for the th International Conference on Structural Mechanics in Reactor Technology. These p s represent the body of the JK pariel session, " Structural and Mechanical Engeer g Research at the U.S. Nuclear Regulatory Commission." t i V y I 14 OOCUMENT AN AL Y5'5 - a R E VWORDR'DL5L ftiP7 0RS ' in A w Aig Ag,giy , panel session
-' ST ATEMENT research y structural and mechanical eng 'neering Unlimited 15 BEC%RITY CL AS5es IC ATION ITha p.get
. .o Nvi. i R oP N .Noeo r M. Unclassified
- a. .,
Unclassified o NvM. R o, Ac.cs le Pfit(L ____ .____ _ _ _ _ _ _ - _ _ _ _ - - - - - _ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - ~
UNITED STATES sn ciat ,o w t+ctass u n NUCLEAR REGULATORY COMMISSION *5'*GjUP5'^* WASHINGTON, D.C. 20555 ,g'5t% 1 OFFICIAL BUSINESS PENALTY FOR PRIVATE USE,4300 120555078877 1 1AN US NRC-OARM-ADM 0 I u i BR-POR NUREG W-53T DC 20555 WASHINGTON
\
l l C _ . _ _ _ _ . _ _ . _}}