ML20133N381
Text
- - . - -
.[pa %q{og UNITED STATES
@O%
TO .
i 3 g NUCLEAR REGULATORY COMMISSION
- y u sm warow.o.c.nossa 1 f ooooo AUG g 9 gg i
Power Engineering .
ATTN: Mr. Frederic C's Olds Senior Editor 1301 S Grove Avenue Carrington, Illinois 60010 ,
Dear Mr. Olds:
The article in the July 1979 issue of Power. Engineering entitled, "No Cause for Seisr.ic Shutdowns" was read,with great interest. That editorial stated that the "NRC's actions had been somewhat precipitous" and " drastic.. . . . . . .
when the potential. risk to the public.....is remote". You also indicated that the basis for NRC's actions was not easily perceived. After reading the article I believe that it would be appropriate for me to make some observations and to provide some comments to hopefully clarify some misleading information.
For several months prior to the March 13 Show Cause Orders, the NRC had been investigating a significant difference in computed pipe stress at Beaver Valley when different computer codes were used. On March 8,1979, during a meeting in Bethesda, Maryland at NRC Headquarters, the licensee for the Beaver Valley plant (based upon information provided by Stone and Webster
- Engineerirg) reported that the-difference in pipe stress were due to differences in the method of modal summation in the computer codes and that the code SHOCK II used an algebraic summation of modal responses. (Enclosure A is a detailed chronology of events leading up to and following the March 8, 1979 meeting.)
Following the March 8 meeting, NRC personnel were sent to Stone and Hebster Engineering Offices in Boston to review the results of reanalysis of additional pipe stress problems which were originally analyzed-with SHOCK II. The results of the reanalysis showed piping systems which were overstressed by as cuch as three times code allowable values. In some instances the seismic stress increased by 3 to 6 times the original SHOCK II values. Additionally it was determined that systems whose failure would cause an accident, as well as systems designed to mitigate the accident and that systems necessary for safe shutdown were involved. The NRC's basic " defense-in-depth" philosophy for reactor safety was compromised in that a single event (in this case an
. earthquake) could cause an accident, could prevent mitigating systems from functioning and could disable systems necessary to safely shutdown.
8508130472 e50703 301 PDR hRMA. . .. . .. . . ~ . . .
-r--- -~-v---
l f
Mr. Fredric C. Olds :
l Based upon the information then available and not being able to demonstrate l the adequacy of the plants design, the Beaver Valley licensee shut the plant down on March 9,1979. Based upon the results of reanalysis of SHOCK II problems between March 10 and 12 and the extent of the safety systems involved, the NRC issued Show Cause Orders to the affected plants.
Therefore, the bases for the Orders is quite simple:
(a) The licensee and his contractor informed the NRC that safety-related piping (including shutdown and emergency cooling piping) might be greatly overstressed should a major earthquake occur .
at a site.
(b) .The licensee and his contractor could not provide adequate assurance that the plant could safely operate and therefore shut down the Beaver Valley facility.
(c) The contractor attributed the calculated overstressed condition to the une of the algebraic sum method.
(d) The contractor st'ated that four other nuclear facilities had similar problems.
(e) After a few days of recalculation, the licensee and Stone and
- Webster provided additional calculations which again confirmed the -
highly overstressed con ~ditions.
Based on this information from the licensee and Stone and Webster, the NRC no longer had assurance that the public health and safety was being protected.
The editorial states that "other responsible organizations used the same technique for seismic pipe stress analysis in perhaps as many as 20 other nuclear power plants". ,It also referred to "the fact that the same computational method was in common use by others, and was accepted widely -
as good engineering practice at the time.....". -
To the best of our knowledge about 20 operating reactors and four reactors still under construction used algebraic summation. Enclosure B (attached) is a plant-by-plant summary of each of these plants. While the number of plants referenced in the editorial is correct, it should be noted that the use of codes with such techniques were generally not widely used in these plants. While four independent computer codes LSHOCK II, ADLPIPE, WESTDYN (a version of ADLPIPE developed by Westinghouse), DAPS, and PIPDYN II3 did have subroutines (sometimes options) for using the algebraic summation, other versions of these codes as well as many other codes do not use algebraic 4
summation.
t
, , ... _. =.-- -
c - ... - . - . . . . - -
l M r. Fredric C. Olds Reanalysis of piping systems at Beaver Valley, Surry and Fitzpatrick which were originally analyzed with SHOCK II resulted in identification of piping and support overstress. Modifications to correct overstress have been required at these plants.
Reanalysis to date of piping systems at other facilities analyzed with ADLPIPE, WESTDYN and DAPS have not resulted in overstress. However, some supports required modification to correct deficiencies in the "as built" condition at Brunswick 1 and 2 and Pilgrim which were found as a result of DAPS reanalysis. In each instance where the licensee had not already completed reanalysis, the NRC staff met with the licensee to review the method of original analysis to determine the potential for significant '
pipe overstress. For Indian Point 2 and 3 and Brunswick 1 and 2 it was determined that overstress was not likely and these plants were allowed to continue operation during reanalysis. Obviously, the 20 operating plants benefited from the publicity associated with the shutdown orders and in some instances were able to reanalyze the affected systems prior to responding to IE Bulletin 79-07. (Enclosure C). The important point is that reanalysis was performed and 4ndicated that modifications were not required. Whether the original problem was algebraic summation, or some other feature of SHOCK II or improper system modeling, is not certain. What is known, based upon reanalysis, is that the plant piping designs on Surry 1 Beaver Valley and Fitzpatrick were inadequate and that modifications are necessary.
It is also important to clarify your comment that " earlier techniques had been approved by the regulatory authorities (and not disapproved until March 13.1979)." The extent to which computer codes were reviewed and approved varies but in none of these cases was algebraic summation identified during the review or specifically approved by the NRC. The NRC issued Regulatory Guide 1.92 (Enclosure D) in December 1974; the staff generally approved either absolute summation or square root sum of square methods.
The NRC staff has completed a review of the information previously submitted by licensees and has found only one instance (Salem 1) where algebraic summation was proposed by a licensee and apparently overlooked by the .,
staff during its review. However, it can be shown, based upon basic prin'cipals, that algebraic summation is technically not correct because colinear modal response could cancel each other out resulting in no loading when a significant load may actually exist.
A final comment related to system modeling. Clearly no conclusions can i be drawn about the adequacy of the design if the system model is not repre-sentative of the actual plant design. During the reanalysis of these plants t
, .-m 4
-._m _
, . . , - _, . . . .._ , , , _ . , , . . _ _ _ , ,m._e.- ,, . . _ - . _ . - - - - - , - _ _ - . . . - _ _ ~
~.
Mr. Fredric C. Olds significant differences between the "as built" condition in the plant and the original design (system models) have been identified. As a result of this discovery and reports by other licensees of missing and improperly installed supports, the NRC issued IE Bulletin 79-14 on July 2,1979. This bulletin requires inspection to verify that the "as built" conditions conforms to the original design and to correct or show by analysis that,,any deviations are acceptable. -(Enclosure E).
incerely, s .
/
Darrel G. Eisen ut, Acting-Director Division of Operating Reactors Office of Nuclear Reactor Regulation 1
Enclosures:
A. Chronology of Events .
B. Plant Summary C. IE Bulletin 79-07 -
D. Regulatory Guide 1.92 E. IE Bulletin 79-14 O
e
- E9CLOSL'.".E B
SUMMARY
OF PLAN 7S USING ALGEBRAIC SUMMATION ,
I PLANT SHUTDOWN EXTENT OF SYSTEMS REQUIRED ANALYZED USING ALGEBRAIC SUMMATION i TECHNIQUE COMMENTS
. ORDER OTHER Beaver Valley 1 Yes Extensive Complete and Order Terminated 8/8/79 4 ,
Brunswick 1,2* No Voluntary Extensive Reanalysis in progress. Analysis to date indicates modifications will not
, be required.
I Cook 1, 2 No No Main Reactor Coolant Complete
. Loop and some lines in-side containment l'Co:per No No SRV lines Under staff review I
Fitzpatrick Yes Extensive . Order permitting startup issued 8/14/79 i 1
Ginna No No Main Steam and RHR Lines Complete {
Indian Point 2 ' No No 10 Lines Reanalysis in progress. Unit is shutdown for refueling. Operation for 5 weeks to refueling permitted based upon preliminar.
analysis Indian Point 3* No No Extensive Reanalysis in progress. Analysis to date indicates modifications will not La re-quired Maine Yankee Yes 19 lines (Initially Complete and Order terminated 5/24/79 thought to be extensive) 4 Millstone i No No 2 systems (Control Rod Complete Drive Exhaust and CU2 By-pass)
- NRC Staff has reviewed basis for operation during reanalysis, concluded operation was acceptable, and documented the findings.
~
SUMMARY
. OF PLANTS USINr. .GEBRAIC SUMMATION (Continuid) -
~
PLANT SHUTDOWN EXTENT OF SYSTEMS .
REQUIRED ANALYZED USING ALGEBRAIC SUMMATION ORDER OTHER COMMENTS TECHNIQUE Millstone 2 No No 6 systems (Volume Control Complete Tank Changing Bypass, Nitrogen Addition, Charging.
Diesel Generator Exhaust, RCP Top Root Valve Instru-ment SI and Containment lj; Spray Test Line) -
' Nin2 Mlle Pt 1 No No 7 systems (Reactor Recir- Complete
. culation, Shutdown Cooling,
! Pilgrim No Tech Spec Recirculation and Main Steam Complete
'. lines Pt. Beach 1,2 No No 2 CCW and 2SW lines in Complete radwaste system Robinson 2 No No Main Reactor Coolant Loop Complete
~
Salem 1 Extended Extensive Unit shutdown for refueling Re fuel .
Surry 1, 2 Yes Extensive Order permitting operation of Surry 1 issued 8/22/79. Surry 2 shutdown for steam generator repair.
Turkey Pt. 3,4 No No Main Reactor Coolant Loop Complete Zicn 1, 2 No No Main Reactor Coolant Loop Complete i
SUMMARY
OF PLANTS USING ALGEBRAIC StH1ATION (CONTINUED)
PLANT SHUTDOWN EXTENT OF SYSTEMS -
REQUIRED ANALYZED USING (UnderConstruction) ALGEBRAIC SUMMATION ORDER OTHER TECHNIQUE COMMENTS Sales 2 Extensive (Reactor Coolant Reanalyses and implementation of any System excluded) required modifications prior to criticality.
1 i Farked River Containment Spray Reanalyses and implementation of any l
required modifications prior to receipt j
of operating license.
t UNP 1, 4 ASME Code Class 1 Reactor Reanalyses and implementation of any i Coolant System Branch Lines required modifications prior to , receipt Of operating license l '
i i
i .
e