ML20003E678
| ML20003E678 | |
| Person / Time | |
|---|---|
| Issue date: | 02/06/1981 |
| From: | Advisory Committee on Reactor Safeguards |
| To: | Advisory Committee on Reactor Safeguards |
| References | |
| ACRS-1809, NUDOCS 8104080240 | |
| Download: ML20003E678 (6) | |
Text
- .
CERTIFlB
~
c=
M/7/8/
MINUTES OF THE ACRS SUBCOMMITTEE MEETING ON THE COMBINATION OF DYNAMIC LOADS WASHINGTON, DC DECEMBER 16 & 17, 1980 The ACRS Subcommittee on the Combination of Dynamic Loads held an open meeting on December 16 & 17, 1980 at 1717 H St., NW, Washington, DC to l
continue its review regarding the use of dynamic load combinations as a i
design basis for nuclear power plants.
Mr. E. G. Igne was the Designated Federal Employee for the meeting.
Documents provided to the Subcommittee are attached to the Office Copy of the minutes.
ATTENDEES:
The following,ACRS members were present:
M. Bender, Chaiman, H. Etherington, and C. Siess.
The ACRS consultants present were:
W. Gall, T. Pickel, E. Rodabaugh, and Z. Zudans.
The peer review group re ; a red by the NRC and contracted by Lawrence Livemore National Laboratory (LLNL) were also present at the meeting. They are E. Rodabaugh, Chaiman, G. Apostalakis, G. Irwin, M. Shinozuka, and W. Stoddart.
The principal attendees are listed below:
NRC LLNL Industry J. O'Brien C. K. Chou D. Landers j
M. Vagins C. S. Lu J. McInerny J. Knight M. Schwartz R. Bosnak R. Larder J. Richardson R. Mensing C. Serpan C.. Cornell, Consultant l
J. Kelly, Consultant M. Ravindra, Consultant R. Kennedy, Consultant D. Harris, Consultant D 04 080 %Q
~
C0ftBINATION OF DYNAMIC LOADS 12/16&l7/81 INTRODUCTORY STATEMENT BY THE SU8 COMMITTEE CHAIRMAN:
Mr. Bender, Subcommittee Chairman, convened the meeting at 8:30 a.m. and stated the purpose of the meeting. He indicated that the Subcommittee did not receive any comments from members of the public.
TECHNICAL PRESENTATIONS:
December 16, 1980 C. K. Chou, program manager, LLNL, discussed the overview of the Load Combination Program. The program consists of two phases: Project I - Large LOCA Earthquake Event Combination Probability Assessment and Project II -
Load Combination Methodology Development. He stated that it is difficult to design nuclear components and supports when considering the combination of asymmetric' blowdown and SSE loads. He also stated that it is difficult to reach a decision on which loads need to be combined and how to combine these dynamic loads realistically.
The objective of Project I is to detennine the probability of a double-ended (DE) guillotine break of PWR primary piping induced by earthquakes for a selected PWR plant - Zion Unit No.1. ' Key results determined for Project I are listed below:
The estimated probability of a large LOCA for the primary loop piping of Zion Unit No.1, resulting from f atigue growth is on the order of 10-12 per 40 year plant life.
The probability of leak is several orders of magnitude greater.
The objective of Project II is to develop load combination requirements based on _ component reliability methodology, in order to provide NRC with tools to evaluate, (1) the reliability of existing nuclear components, (2) load combina-tion criteria that assures acceptable levels of safety.
f
-Preliminary results obtained from Project II, which is a long-term project, are as follows:
' A methodology based on component reliability has' been developed to evaluate the reliability of components and to rationally detennine (1) which loads b
COMBINATION OF DYNA'i!C LOADS 12/16&l7/81 need to be combined, (2) how to combine responses and, (3) use of appropriate service levels or design allowable.
The developed methodology has been demonstrated on a simple beam problem and on an essential service water system piping design.
C. A. Cornell, Src Consultant, LLNL, discussed current potential limita-tions with the methodology developed in Project II. Some of these limita-tions are (1) limited data, (2) lack of separate treatment of uncertainties i
)
and (3) effects of engineering or other errors not accounted for.
R. P. Kennedy, Sr. Consultant, LLNL, commented on Project II.
He stated that the load combinations methodolooy provides a powerful technique for evaluating current design criteria and for developing rational relia-bility based design rules. He further stated that near future activities should include data collection and performance of case studies to aid in licensing decisions.
S. C. Lu, project leader, LLNL, presented an overview of Project I, large LOCA and earthquake event combination probability asessment. The objective of this project is to estimate the probability that a large LOCA and an earthquake will occur simultaneously. The result of this work should provide technical input which is needed in reassessing the requirement that earthquake and large LOCA effects be combined in nuclear power plant design. ' The current assessment focuses on the combined effect of a large LOCA and an earthquake.
Zion Unit No. I was considered for demonstration for the following reasons:
Zion represents the size of future PWRs about.1100 MW plants.
j Both utility and architect-engineer are cooperative.
The plant is also the demonstration plant for SSMRP wort by LLNL.
The structural models to calculate the stresses in the primary pipe accounted for dead weight, pressure, thermal expansion, nonnal and abnormal transients, and seismic motions.
It was found that vibrational and residual stresses at the welds were insignificant for growing fatigue
~
COMBINATION OF DYNAMIC LOADS 12/16&l7/81 cracks.
The effects of SCC were not considered in the model because no evidence of this phenomena has occurred in the PWR primary piping system.
A LOCA indirectly induced by earthquakes was studied. The intent was to identify sources that could cause a LOCA and to establish the ground work for a more thorough evaluation in a subsequent project.
The stress analysis results support the conclusion that seismic loads will introduce a negligible probability of a LOCA. Best estimate seismic stresses using the most advanced procedures available are very low.
D. O. Harris, consultant to LLNL, discussed the fracture mechanics model of piping reliability used on Project I.
Assumptions in the model de-scription are listed as follows.
Failure occurs only as a result of fatigue growth caused by defects intro-duced during fabrication.
The defect growth rates can be predicted from laboratory fracture mechanics tests.
Water chemistry and plant operation as anticipated.
No design errors have occurred.
The probability of failure of the primary piping, within 40 years i: about 10-12 The probability of a leak is about 10-6 Large uncertainties in some input, e.g., initial crack distribution, introduce large uncertain-
. ties in some results.
The peer review group provided an oral concensus report at the meeting.
Under prevailing thinking that the perceived risk of important structures must be measured in terms of probability of failure, the LLNL study is worthwhile. Within this framework, the specific objective of the LLNL program is important.
The panel has not yet obtained complete documenta-tion of the LLNL study, but based on information received, the following technical comments are provided.
C0"SINATION OF DYN V.!C LOADS 12/16817/81 Greater emphasis should be placed on identifying and evaluating unccriainties, for example, uncertainties in the assu ed crack pc:u-lation in the fracture cechanics model.
The influence of design and manufacturing errors on the results should be given more consideration since these have not been included in the analysis.
The reported results of :ne decoupling study shew two answers which are of najor interest.
13 The increase of large LOCA probability by incidence of earthquake loads is extrecely small for the cases studied, and
- 2) The likelihood of through-wall cracks prior to occurrence of a LOCA is very large supporting the claim that " leak-before-b wak" will enable LOCAs to be forestalled before they become major events.
There are plausible reasons to suggest that these conclusions might not be changed by further study, although, the extremely low probability of earthquake-induced large LOCA of aceut 4 x 10-I3 per plant life is deemed to be subject to large uncertainties.
Nevertheless, the low value does not account fcr many factors that might dominate the resulting values and these need to be considered (e.g., engineering errors, defective materials, fabri-cation faults).
The amount of activity planned for Project II will not satisfy NRC's senedule for attaining the project goals.
J. E. Kelly, consultant, LUR., discussed the prcbability of LOCAs indirectly induced by earthquakes at Zion Unit No.1.
Results of the study indicate that
~
external sources, i.e., overhead crane, spray header pipes, and equipoent sup-ports could structurally fail and may have a higher probability of causing a double-ended guillotine break of the primary coolant loop than directly induced breaks. Further work is needed to assess this finding.
O
COMBINATION OF DYNAMIC LOADS 12/16&l7/81 M. K. Ravindra, consultant, LLNL, discussed the methodology to calculate the probability of structural failures that could cause a large LOCA.
The analysis used component fragilities or resistance due to seismic events to calculate the probability of structural failure.
Results of the best estimate value for the probability of failure of structural components is about 6.4 x 10-8 per year and 1.6 x 10-5 per year with
'a 95% confidence level. He stated that since the probability of external structural failure that may lead to LOCA could be relatively high and further in-depth study of structural failure modes and pipe break mechanisms needs to be performed to evaluate these modes of failure.
The following observations were offered by LLNL:
The requirement to design for the largest LOCA blowdown force warrants reconsideration.
The requirement to design for the combination of a large LOCA and SSE does not apear to be warranted on probability grounds if the results of this study are valid.
Mr. Vagins, NRC, discussed NRR effort in the area of leak before break.
Currently, NRR is reviewing Westinghouse WCAP-9558, Rev.1. The evalua-ti.on report will be completed by March 1981.
l i
t i
e
.w
--,s.
-,