ML20080N187
| ML20080N187 | |
| Person / Time | |
|---|---|
| Site: | Three Mile Island  |
| Issue date: | 09/30/1983 |
| From: | Blake E METROPOLITAN EDISON CO., SHAW, PITTMAN, POTTS & TROWBRIDGE |
| To: | Buck J, Edles G, Kohl C NRC ATOMIC SAFETY & LICENSING APPEAL PANEL (ASLAP) |
| References | |
| NUDOCS 8310040392 | |
| Download: ML20080N187 (219) | |
Text
+
t i
1 SHAW, PITTMAN, PoTTs & TROWBRIDGE
. pA RTNEmSM,p ., .m ,E SSi.N A C.mpenAT .NS ggp 1800 M STREET N. W.
WASHINGTON. D. C. 20036
( } Am G. lie
- AMSav O. aCT's.
- C SMELDQN J. wE8SEL. p C. JEFFE RY L. YASLON E S'E6Amt 6. pit *M AN.
- C. JOMN A. McCUL60uGM.
- C. _ JACM Mc AAY A,.Qatw D. ELuS J. PATm'CM MeCnEv.
GECmGE r ?mCwem*DGE.
- C.
~ J. TMOMAS LENManf.
- C. TELEcopeEm JCMN L CAmm.Je FFLp,,, 0._IMQM wouNASA.E.wM.vECROCnte. Jn
. StE pM E N O #C"S. p C, GEmALD CMANNorr. p C. STEVEN L MELTzEn. p C. pMfum J, MAmVEY
- 'Lup D SC S?wsCK.
- C. DEAN O. AuuCM. p C. 202)822 sce9 & 822 H99 #CSEm? M. Gomo QGg Ag.QsN,sano 4, TIMCTMY M ANLoN. p C. JOMN ENGEL p C. _
sameAmA J. Momo k) SM6TM GEomGt M moGE RS. J a.. p C CMamLES m. TEMaim. p C. soNN.E S. GoTTuEe 3RA ,c,A N S. muTLEoGE FEED A. uTTLE. p C. ' STE* MEN s. MuTTLEn. p C. mApirAx iCo MOwAmO M. SM AFFEmMAN mAy, gm,NE p. CMEER g' JoMN O. mMINELANCEm.
- C. WmTMmop N. snowN. p.C. DESQmAMS.SAUSEE EL GSNCE w. CMueCMsLL P C. JAMES S. MAMuN. 9 C. (202)822-s072 SCOTT A. ANENatso JA,,N6C,g y y y,E,M
,o mE,R-STEIN g, g g, LEft#E A. N CMCLSON, J m.. p.C. MANDALS.RELL.pC. CAMpgELL miLLErte
_ m'CMAm0 M. mmCNTMAL MAm?lN O, m'eA6L p C. mosEm? E. 3AMLEn
- C. Paul M. TMOMAS R*CMANO J. KENOALL. p C. mo ster m.moseiNS.p C. ygLgz SETH M. MOOGASIAN STERMENg_ 3. gMEIMAN,N
, JAY E. SIL3E nG. p C. STEVEN M. LuCAS.
- C. SMEILA McC. MAmVEV E E E " ' ""C** E "
samrAA M. meSSom. p C. oAvio M. musENSTEw. p.C. es 2eos (SMawtAw wSM) oEuSSA A. mioGway PAMELA M. ANOCmSON WICIBO'E V. ALLE N. Jm-. p C. alCMamO E. GALEN .. MENNETM J. MAuTMAN FOES S&,ASNElB.
- C. LYNN WMITTLESEY wtLSON CASLE SMAwbAw OAviO LAwmENCE MILLEm ALERANDEm D. TQMAS2C2um PMium D. pomum ,
R. MSNLY wCSSTE R, p C. MATIAS r. TRAveE SO-OsAZ FREDEmeCN L MLEIN NATMAN.EL p. sorto.Ja p C v Cromia J. mE mmms STEVEN p. miTLER* : M'CMAEL A. SwiGEm Mame AuGENeuCn.
- C. JoMN M. o NEiu Jm. ELLEN SMEmerr miCMAmo J. pammm. o ANeTA J. FmnE LSTEIN EENFST L SLAKE Jm..
- C. JAY A. EpSTIEN " 'e CLLENA.FmCCEL OAIRLETON 5. JONES.
- C. #AND L ALLEN F'
MANNAME.M.uESEmMAN EILEEN M. GLEIMEm TMoM AS A. oAnTEm.
- C. EuSAeETM M. *ENoLEToN couNStu SANoma E. roLSoM oAvio n SAMm JAMES M. SumGER, p C. MAmmV *e.GLAS$psEGEL JUDf 7M A. SANDLEm C. SQwDQiN Team
- esoe acasemo as o c f
j September 30, 1983 waivEn S oimECT oiA6 NUMeEm (202) 822-1084 Administrative Judges Gary J. Edles, Chairman John H. Buck Christine N. Kohl Atomic Safety and Licensing Appeal Board U.S. Nuclear Regulatory Commission Washington, D.C. 20555 IN THE MATTER OF METROPOLITAN EDISON COMPANY (Three Mile Island Nuclear Station, Unit 1)
Docket No. 50-289 (Restart)
Dear Chairman Edles and Judges Buck and Kohl:
Please refer to my letter of September 29, 1983, enclosing a copy of a recent GPU Nuclear Corporation News Release and accompanying letter report concerning the NRC's Office of In-vestigation's report entitled "Three Mile Island. Nuclear Generating Station, Unit 2 Allegations Regarding Modifications, Quality Assurance Procedures and Use of Polar Crane." This letter was inappropriately captioned TMI-2 rather than TMI-l and accordingly this filing is being reserved (without enclosure).
Additionally, because of the Commissioners' concurrent consideration of these subjects, copies of the aforementioned 8310040392 830930 PDR ADOCK 05000289 31D3 C PDR _
.. SHAyv PITTMAN, PoTTs & TROWERIDGE A 94mTNERSwer OF pmOFESSiONAL CompomATIONS Atomic Safety & Licensing Appeal Board, USNRC September 30, 1983 Page Two GPU Nuclear Corporation News Release and accompanying letter report are being provided for their use, as well as the three items served the same date under separate cover with the correct heading, all of which relate to the B&W/GPU litigation record review.
Respectfully submitted,
%/ f. MJ fa.
Ernest L. Blake; Jr.
Counsel for Licensee cc: Attached Service List i
f i
l l
l
UNITED STATES OF AMERICA NUCLEAR REGULATORY COMMISSION Before the Atomic Safety and Licensina Acpeal Board In the Matter of )
)
METROPOLITAN EDISON COMPANY ) Docket No. 50-289 SP
)
(Restart)
(Three Mile Island Nuclear )
Station, Unit No. 1) )
SERVICE LIST Administrative Judge Administrative Judge Gary J. Edles, Chairman Ivan W. Smith, Chairman Atomic Safety & Licensing Atomic Safety & Licensing Board Appeal Board U.S. Nuclear Regulatetty U.S. Nuclear Regulatory Commission Commission Washington, D.C. 20555 Washington, D.C. 20555 Administrative Judge i Administrative Judge Walter H. Jordan John H. Buck Atomic Safety & Licensing Board Atomic Safety & Licensing 881 West Guter Drive Appeal Board Oak Ridge, TN 37830 U.S. Nuclear Regulatory Commission Administrative Judge.
Washington, D.C. 20555 Linda W. Little l Atomic Safety & Licensing Board Administrative Judge 5000 Hermitage Drive Christine N. Kohl Raleigh, NC 27612 Atomic Safety & Licensing Ato,mic Safety & Licensing l Appeal Board U.S. Nuclear Regulatory Board Panel Commission U.S. Nuclear Regulatory
. Washington, D.C. 20555 Commission Washington, D.C. 20555 Jack R. Goldberg, Esquire (4) l Office of the Executive Atomic Safoty & Licensing l Legal Director Appeal Board Panel U.S. Nuclear Regulatory U.S. Nuclear Regulatory Commission Commission Washington, D.C. 20555 Washington, D.C. 20555 Docketing & Service Section (3) Douglas R. Blazey, Esquire Office of the Secretary Chief Counsel U.S. Nuclear Regulatory Department of Environmental Commission Resources l
Washington, D.C. 20555 514 Executive House
_ __ h _@o ESR 2 3 5 7__
l
.O eo _ ,
John A. Levin, Esquire ;is. Gail Phelps Assistant Counsel ANGRY /TMI PIRC Pennsylvania Public Utili 'ty - 1037 Maclay Street Commission -
Harrisburg, PA 17103 Post Office Box 3265 Harrisburg, PA 17120 - Jordan D. Cunningham, Esquire Fox, Farr & Cunningham Mr. Henry D. Hukill 2320 North Second Street Vice President Harrisburg, PA 17110
! GPU Nuclear Corporation Post Office Box 480 Ellyn.R. Weiss, Esquire (1)
Middletown, PA 17057 William S. Jordan, III, Esquire (1 Harmon & Weiss
- Michael F. McBride, Esquire 1725 Eye Street, N.W., Suite 506 LeBouef, Lamb, Leiby & MacRae Washington, D.C. 2.0006 1333 New Hampshire Avenue, N.W.
Suite 1100 Mr. Steven C. Sho11y Washington, D.C. 20036 Union of Concerned Scientists 1346 Connecticut Avenue, N.W.
! Ms. Louise Bradford Dupont Circle Bldg., Suite 1101
! TMI ALERT Washington, D.C. 20036 1011 Green Street Harrisburg, PA 17102 Chauncey Kepford l
Judith H. Johnsrud Mr. Norman Aamodt Environmental Coalition on R. D. 5 Nuclear Power Coatesville, PA 19320 433 Orlando Avenue State College, PA 16801 John Clewett, Esquire The Christic Institute David E. Cole, Esquire 1324 North Capitol Street Smith & Smith, P.C.
Washington, D.C. 20002 2931 Front Street Harrisburg, PA 17110 Michael W. Maupin, Esquire Hunton & Williams Ariministrative Judge 707 East Main Street Gary L. Milhollin Post Office Box 1535 Atomic Safety & Licensing Board Richmond, VA 23212 '1815 Jefferson Street James K. Asselstine, Commissioner U.S. Nuclear Regulatory Commission Nunzio J. Palladino, Chairman Washington, D.C. 02555 U.S. Nuclear Regulatory Commission Washington, D.C. ,20555 Victor Gilinsky, Commission U.S. Nuclear Regulatory Commission Thomas M. Robe s, Commissioner l
Washington, D.C. 20555 U.S. Nuclear Reghlatory Commission l
Washington, D.C. 20555 j
- Frederick Bernthal, Commissioner -
U.S. Nuclear Regulatory Commission Washington, D.C. 20555 I . _
News Releaso Threa Mila Isl:nd Nuclear Station Post Office Box 480 GM Middletown, PA 17057 717 948-8197 Public Information Services For Further Information
Contact:
Douglas Bedell Gordon Tomb For Release: Ime.diately Date: September 23, 1983
- 126-83N GPU NUCLEAR RESPONSE ON TMI-2 CLEANUP ALLEGATIONS Middletown, PA -- The report by the Nuclear Regulatory Commission's Office of Investigations relating to allegations that safety requirements were ignored at Three Mile Island's Unit 2 is " misleading to the public and lacks any perspective as to what the adequacy of controls for the TMI-2 cleanup is all about," GPU Nuclear Corporation President Robert C. Arnold said today.
"We are convinced that the NRC's Office of Investigations report, which deals principally with prncedural issues, fails to identify the extent to which procedural deviations were related to the changes being instituted in the organization so that it could better cope with unprecedented problems.
Further, the report does not make clear that these deviations were, in fact, of no direct safety consequence," Arnold said.
"The investigators took a very narrow approach in their interpretation of the regulatory requirements. They arrived at judgments and conclusions which, we believe, are not supportable by the technical facts or reasonable interpretations of regulatory requirements and guidance," he said.
"From our own investigation and from analysis of the NRC report, we are convinced that: --
- The physical work has been done safely. There have been no allegations to the contrary.
- A small fraction of the refurbishment activities on the polar crane that were carried out under Bechtel administrative procedures
- more -
September 23, 1983
- 126-83N did not conform in some respects with the GPUNC administrative controls. These were controlled adequately to assure safety. Corrective action to prevent recurrence of such deviations has been taken.
T,here was no effort to ' circumvent' quality assurance. Indeed, the quality assurance program was effective in identifying problems that did exist and led to corrective action initiated by the company.
- There was an honest effort to proceed in a safe, efficient and timely fashion with high regard for the need to protect the public health and safety.
- There was no collusion between the Company and the NRC.
"We believe," Arnold stated, "that none of the public allegations about lack of safety at THI-2 or harassment of people who voiced concerns about safety have any merit. Further, while it is not directly addressed in the NRC report, there was no ' mystery man' who secretly turned off high pressure injection pumps the morning of the accident. This issue was raised in the original allegations.
"The manner in which the report overlooks the very deliberate and effective procedural improvements that were achieved in conjunction with an extensive and fundamental restructuring of the organization in 1982 and 1983 detracts from the merit of the report and unfairly mischaracterizes the Company commit-ment to a safe and early cleanup," Arnold said.
"Several of the findings set forth in the report represent differences in judgment between the investigators and the Company as to what the require-ments were," Arnold said. "In those instances where the report asserts the Company did not correct identified shortcomings, the Company's interpretations of the requirements disagree with the investigators' underlying interpretations.
(more)
September 23, 1983
- 126-83N "The TMI-2 cleanup is a totally unprecedented task' undertaken by competent and dedicated individuals who regard their work with a high sense of responsibjlity. As the NRC staff reviews this document, we h6pe they will provide a more balanced interpretation of the NRC requirements and guidelines.
To fail in that regard is to cause further confusion for the public about a report which is already misleading."
Enclosed with this statement is a letter to Arnold from Edwin H.
i Stier, an outside investigator who GPU Nuclear asked to investigate the original allegations last March of safet.y violations in the TMI-2 cleanup by Lawrence P. King, Richard D. Parks and Edwin H. Gischel. Stier, who has soent 17 years as a state and federal prosecutor, most recently was Director of the New Jersey Division of Criminal Justice. He is now in private practice.
Stier's investigation began March 28, 1983. Since then, he and his nine-member staff have received statements from 80 individuals and reviewed more than 1,000 documents. He is preparing a full report that will be submitted to GPU Nuclear and the Nuclear Regulatory Comission.
Arnold said the Company will be describing the refurbishment of the polar crane at an NRC meeting to be held Tuesday, September 27, in Middletown, and at a meeting of the Advisory Panel for the Decontamination of Three Mile Island, Unit 2, to be held Wednesday, September 28, in Harrisburg, and urged the public to attend.
"I ask all who have a concern about these matters to attend one of these meetings to hear the facts firsthand," Arnold said.
l w- -- -, ----,a m. , , - - . - - , . . . , , . - . , , ,w . , , . . --.------,---e-,------ y-.- , . . ,- - - - - - - . - - - . - ~ - - - y,--- ,m - - , - -,,--rvi -yn -7r y
KIRSTEN, FRIEDMAN & CHERIN A PROFESStoNAL CoRPOAATioN *
. COUNSELLORS AT LAW K1CHAAD E. CHERIN* 17 ACADEMY STREET MAnoLo ra CoMAN NEWARK. N. J. 07102 JACM S.KIRSTEN*
PHituP LEwfS PALEY' (201)623-3600 JOSEPH HARRISON (1930 8976)
CowN H. sTrER or counsel oENNis c. uants September 23, 1983 COWARDS. Levy MILTON LoWENSTEIN MARoARET E. ZALESMS or COUNSEL cEnano n. rnEcs-STEVEN PASTERNAK*
JrHN K. ENRIGHT
'utasern N4 & NL SAas
. Mr. Robert C. Arnold, President GPU Nuclear Corporation 100 Interpace Parkway
, Parsippany, New Jersey 07054
Subject:
Allegations by TMI-2 Employees
Dear Mr. Arnold:
You have requested my comments on a. report issued by the NRC, Office of Investigations (OI), entitled "Three Mile Island Nuclear Cenerating Station, Unit 2 Allegations Regarding Safety Related Modifications, quality Assurance Procedures and Use of Polar Crane."
evidence we have gathered in the course of our investigation which includedThe substantially the same subject matter covered by the OI report.
We have reached a stage in our investigation where I have sufficient information to respond to your request for comment.
Our investigation began l
on March 28, 1983 and has continued, full-time, through the present. During i the course of the investigation we will have reviewed in excess of 1,000 documents and have obtained sworn, transcribed, question-and-answer statements from approximately 80 witnesses.
phase of our work and are preparing our final report.We In thathave now report, weconcluded the fac intend to cover the full range of issues raised by Lawrence King, Richard Parks and Edwin Gischel, the THI-2 employees whose public allegations precipitated the investigation by the NRC as well as our own. Although we have not yet completed writing the final report, we have reached conclusions concerning the validity and implications of the allegations that have been made.
These conclusions have been reached by an analytical process independent of CPU Nuclear Corporation (GPUN) management.
than The NRC investigation the investigation that wedescribed in the 01 report iS far narrower in scope have conducted. OI focused heavily on one set of issues relating to compliance with administrative procedures. It is unclear why they did not address issues, such as whether activities at TMI-2 i
Mr. Robert C. Arnsid, Pescidant September 23, 1983 -
Page 2 have in any way endangered public health or safety. They did not evaluate the effectiveness of the safety review system, and the extent of management's concern for the protection of public safety. In my view, it is not possible to assess the seriousness of procedural errors unless these related safety issues are analyzed as well.
i From the limited scope and depth of the 01 report very sweeping conclusions have been drawn by the investigators which were probably not intended to be as categorical as the language of the report suggests. In his cover memorandum to NRC Chairman Palladino, Mr. Hayes, Director of the Office of Investigations, states without qualification "the allegations were not only substantiated, but we found them to be illustrative rather than exhaustive."
Presumably this statement intended to cover only a narrow category of allegations.
It is clear from our investigation (and should have been evident to 01) i that many of the allegations made by Parks, King and Gischel are contradicted by the overwhelming weight of the evidence. For example, all three allege l that TMI-2 management intentionally withheld information from the Site Operations Department about the polar crane load test Safety Evaluation Report (SER) to minimize the time available for them to review and criticize the SER. The testimony and documentation clearly demonstrate that Site Operations had received the necessary information and had the opportunity to comment sufficiently in advance of their review of the final revision of the SER.
Parks suggests that a critical document had not been reviewed or approved by Site Operations or the NRC. In fact the document had been reviewed and approved by both and bears the signatures of Site Operations and NRC personnel. Numerous other examples have been found where witnesses who allegedly possessed evidence supporting the allegations have contradicted them.
The obvious danger in overgeneralizing about the validity of the allegations is the confusion it engenders. As important as administrative procedural compliance may be, it would be unfortunate if one were to infer solely from the verification of an allegation of procedural noncompliance, that a related safety allegation might also be true. We have found that the allegations must be very carefully sorted and separately investigated and analyzed to account for the half-truths and distortions inherent in many of them.
Even within the narrow framework of procedural compliance, the OI report provides no means of assessing the significance of its findings. A procedural violation may be serious or inconsequential depending upon its cause and its public health and safety consequences. In measuring the significance of a XIRSTEN, FRIEDMAN & CHERIN
Mr. Robert C. Arnold, President -
September 23, 1983 Page 3 procedural violation some reasonable criteria must be applied beyond word for word compliance. For example, the OI report does not apply the following criteria:
Has the physical safety of the general public or site personnel been jeopardized to any degree?
Has the work activity been subjected to the scrutiny of the safety I
review groups which constitute the checks and balances system designed by GPUN management to identify and correct potential safety problems?
Has the GPUN Quality Assurance Department (QA) identified and resolved procedural deficiencies in the ordinary course of its work?
Additionally, during che time period covered in the OI report, THI-2 was undergoing a fundamental restructuring of its management. The objectives and die effectiveness of that reorganization must be examined in depth and understood in order to judge whether procedural noncompliance was endemic to TMI-2 or transitory. The OI report makes only passing reference to the fact that the reorganization has been taking place. It does not analyze the actions taken by the Director of TMI-2 to resolve the very problems that the report describes.
Without discussing each issue considered in the OI report, I will offer some consents based upon the above criteria. Our investigation has found no evidence that any work performed on the polar crane created a safety hazard for the public or for site personnel. None of the information contained in the OI report or in any other source leads to a contrary conclusion. As our l investigation report will describe, the engineering judgments made in the course of refurbishing the polar crane, and, in connection with other activities which were the subject of allegations, met reasonable standards and were based upon appropriate consideration for public health and safety. For example, it was alleged that, in addition to procedural violations relating to the propcsed load testing of the polar crane, calculations were not performed to determine the consequences of dropping the test load. We have confirmed that, in fact, such calculations had been done. Similarly, we have investigated many other safety allegations that have been found to be without merit.
The OI report fails to consider the extent to which alleged procedurally deficient activities were nevertheless subjected to the CPUN safety review system. This deficiency may leave an uninformed reader with the mistaken ' '
l l
EIBSTEN, FRIEDMAN & CHERIN
Mr. Robert C. Arnold, President '
' September 23, 1983 Page 4 impression that a system.
the safety review procedure may have been intentionally violated to circumvent polar crane No Load Test. One example is the OI report's discussion of the The OI report concludes that the test violated site test procedures because it was not reviewed by the Test Working Group (TWG).
Seven other specific violations of the test procedures are set forth in the report, including such findings as " failure to include the RBPC (Reactor Building Polar Crane) No Load Test in the Master Test Index."
i 1
What the OI report neglects to mention is that prior to the No Load Test being conducted, it was categorized as "Important to Safety" thereby suh*ecting it to the highest level of safety review. The test procedure was rt Lewed by the Plant Operations Review Committee, QA, NRC and the Site Operations Department, where it was approved by King. The performance of the test was witnessed by representatives of Quality Control. The chairman of TWG was satisfied with the test plan prior to its performance, and reviewed the test results, determining that they were satisfactory. Finally, notwithstanding its prior approval of the test, QA subsequently issued a quality testing deficiency report noting "the administrative program controls for were not followed."
However, QA went on to find, "the test results were technically adequate." Unless the reader of the OI report has all of the information conc.erning the performance o.4 the No Load Test including the details of the review process to which it was subjected, the procedural violation may take on exaggerated proportions. ~
l In addition to procedural violations that had been identified, investigated and resolved by QA, 01 investigators assert several other i
violations based upon their interpretation of GFUN procedures. Varying interpretations of procedures are understandable. Site administrative procedures are highly complex and of ten ambiguious. Their construction must be tempered with logic and recognition of their intended function. Procedural uncertainties procedures should not be resolved simply on a literal interpretation of the themselves.
The 01 report does not identify such ambiguities nor does it fully explain the reasoning process which led to the GPUN interpretation which the report criticizes.
l An illustration of this problem is the discussion in the 01 report of the use of the GPUN maintenance procedure to authorize the refurbishment of the polar crane.
The conclusion of the OI report is that it "was the incorrect procedure to use. . ." This conclusion is significant. It calls into question the procedural validity of all of the refurbishment work on the polar crane.
l l
l
! XICSTEN, FRIEDMAN ar CHERIN
, .=
Mr. Robert C. Arnold, President .
3 September 23, 1983 Page 5 i
Our investigation has carefully traced the process by which the refurbishment work was initiated. By reviewing the maintenance procedure with the individual who wrote its relevant provisions, we have determined the basis for its use for the refurbishment of the polar crane. He identified specific language in the procedure which authorizes its use for such purposes.
However, even if the OI investigators are ultimately correct in their interpretation of the maintenance procedure, the absence of a clear articulation of GPUN's rationale suggests that the procedure might have been used in bad faith. In fact, there was a reasonable, logical basis for the procedural appro'ach taken by GPUN, but the reader of the OI report has no way of knowing that.
It is highly significant that the procedural deficiencies that constitute the primary subject matter of the OI report occurred at a time when TMI-2 was undergoing a major reorganication. Its objectives included the establishment of uniform, practical procedures and measures to assure procedural compliance. Shortly after the reorganization became effective, a new procedural system was initiated. Training has taken place and the effort is ongoing.
When the issue of procedural compliance was raised internally, the new TMI-2 Director took immediate action to~1nvestigate the matter and assure that future activities complied strictly with site procedures. Through his effort, uncertainty about the applicability of Bechtel administrative procedures to recovery work has been resolved.
To evaluate the response of THI-2 management to procedural deficiencies, it is necessary to understand that completion of the reorganization has taken many months. Reorganization has progressed slowly, with apparent
- inconsistencies between the organizational structure and Technical Specifications. The OI report comments upon this. However, it does not '
consider that new Technical Specifications intended to complete the reorganization process have been awaiting NRC approval for 10 months.
In analyzing the issue of misclassification of activities, the OI report makes no reference to the Quality Classification List (QCL). The QCL should identify the proper safety classification for all plant systems. It was correctly alleged by King that misclassification of activities had resulted in large measure from an outdated QCL. Neither this nor any other underlying cause for misclassification was explored by the OI report. The impression created is that the safety classification system was intentionally circumvented.
XIRSTEN, PRIEDMAN & CHERIN l _ _ _ _ - - -_,_ _ - - - - - - - - - - - - - - - - - ---- - -- - - ~-- ~ ~ ~ -
Mr. Robert C. Arnold, President .
September 23, 1983 Page 6 On the basis of numerous interviews, we have identified fundamental differences in engineering judgment which have existed for some time in different parts of the THI-2 organization. Certain groups tend to rely totally on the literal contents of the QCL. If a system is listed as "Important to Safety," any related activity is so classified despite the
! absence of any safety implications. Other groups in determining safety classification emphasize the safety implications of the activity to be performed. These differences in approach have caused disagreements in determining safety classifications. What may appear on the surface to be an intentional misclassification may, in fact, be legitimate difference of opinion by responsible engineers.
' The IMI-2 Director has attemtped to solve this problem by updating the QCL so that obsolete systems are reclassified "Not Important to Safety" and new recovery systems are properly classified according to their current functions. The 01 report makes no assessment of the potential impact of this effort on the safety classification process.
Therefore, the conclusions reached in the OI report concerning the response of management to procedural deficiencies are based upon events which occurred during the most difficult period.of the reorganization. The report fails to take into account either the objectives or accomplishments of the new TMI-2 management.
- t A further finding of the 01 report deserves some comment since it suggests an improper relationship exists between GPUN and NRC representatives at the site. The OI report, however, does not consider the stated objectives of the NRC in establishing a unique regulatory relationship with THI-2 after the accident. Our investigation has found that practices exist whereby NRC personnel attend CPUN meetings, GPUN transmits draft documents to the NRC for information purposes and the NRC staff informally communicates its concerns to GPU N.
However, this consnunication is apparently based upon an expressed NRC policy to " oversee day-to-day licensee activities."
The OI report states that the investigation of allegations of harassment and intimidation of TMI-2 employees and the so-called " mystery man" remain open and will be the subject of future reports. Our investigation has been concluded in those areas and will be documented in our final report.
In previous connaunications to you we have indicated that the sworn statements of witnesses identified by Parks have refuted his " mystery man" allegations.
Additionally, a recent Babcock and Wilcox analysis of system responses indicates that there could have been no " mystery man" at the time of the accident as suggested in the Parks affidavit.
EIRSTEN, FRIEDMAN & CHERIN
(+'.
Nr . Robert C. Arnold, President .
September 23, 1983 Page 7 Our investigation of the harassment and intimidation allegations was l extens ive. However, it did not encompass allegations that Parks had been subjected to harassment by Bechtel which was the subject matter of litigation between Parks and Bechtel. With respect to King, Gischel and Joyce Wenger, King's secretary, we have determined that none of them had been subjected to harassment as they alleged.
l King's employment was terminated based on a conflict of interest. While
( employed as the Site Operations Director, he was an owner of and operated a l consulting firm which recruited GPUN employees. Our investigation traced the I
origin of the information which led to King's termination as well as the internal GPUN investigation of that information. The judgment to terminate King's employment was based upon factors independent of any safety or management concerns raised by him. ,
l Gischel claimed that as a result of expressing concerns about the testing of the polar crane, GPUN pressured him to take a neuro psychological examination. Our investigation has determined that the decision to urge Gischel to take the examination was made independent of and without the l
knowledge of CPUN management by a psychological counselling service under contract to GPUN from which Gischel voluntarily sought help. Approximately a month af ter Gischel had expressed his concerns about testing the polar crane to management, a psychologist from that service requested GPUN to assist it in convincing Gischel to be examined. The" psychologist felt that the test was necessary to diagnose fully the after-effects of a stroke which Gischel had suffered. Failure to take the test would leave a serious question about Gischel's ability to perform in his employment. There was no connection l
between efforts to convince him to be examined and his expressed views on the testing of the polar crane.
j Joyce Wenger's allegations of fabricated evidence to justify her l
termination have been thoroughly examined and found to be without basis. All of the individuals involved in the incidents which led to her termination have been interviewed and refute her claims.
Unfortunately, the allegations which have been made, and the work being performed at TMI-2, are so complex and highly technical that it is difficult to summarize them in a way which is not over generalized. The OI report seems to suffer from a quite natural desire on the part of the investigators to cut through a great deal of detail and reach the heart of the matter. In 1
attempting to do so, however, balance and perspective have been jeopardized.
Sincerely, i
e - mg Edwin H. Stier EHS:lhw KIBSTEN, FRIEDMAN & CHERIN
,* e =
, SHAW, Pr,TMAN, PoTTs,& TROWBRIDGE .
A paartwEpSMap or poorEsseoMAL comeOmations 1800 M STREET. N W wAsmuoTow. o. c. 2cose ,
mAu$av O. aOTTs. s C ' Joe.= a. neecvLLouoM.
- C. (202)8224000 JErrEnvLva.Lom anO=Ew 0. EL: es STEua J. PATmeCK M*CREY.
- C. JACmwemAv msCMamo A. SAMP G'OA.m? E F6 pwTMaN.TSOwe#iOGE.
- C.p C. GEomOE P usCMaELv. Jst, p C. 'Monsas M. tocConeseCm ?McM AS E. CmOCKER. J88 STXpeg1N O 80775. p C. J. TMoesA5 LENMART. 5 c TgLgCoPIER f JoMN L CAN#. =# WENDELipe A. wMITE 2"Ofi"^t *.',h". *C. Un?no'l'*KE "" <=om an..o.. <. ==.. .. ~Ja",,t,:'ag,=,7,0,, g,^;,';=3, ~,;*
"'d::n".Z".'l:n U .* c 2".LE?N.*.. '
. c - rartad.;??a!" an't =a-
'," " M 'P. E .*,O ="f, " ? """"fe" "*-^",," = ~ ~*a,"'"
U"E*#CE*
e S .n . e.C A.
.CM.Lu .
. O,.SO
. e J.c. .c J .E...c .Eu. AL.
..O.
. C.
e X '.ra=. 'a-
,8 G.
,PE,o,,,,.Ae A gem .,mE,,,,
A ,==,,, J.~
-V.S MCMAN KA C Mt.
= T.
u.M.r =. J.
esANTus 4 aftabb 8 e 880 88837E. EAMLES. P c -
SETM M. DeOOGAS4 Ass IteCMamp J. mEmeALL P c 8BOGEfr? S. IBOSSINS. # c STEPD*EN MElb4A88N TELEX S**EILA R40C. MAfEVEY SA8eOIRA E. SituSCA*
Jav E. OLSEmS. 9 C STEVEN 08. LUCAS. 9 C.
- ^:",S ;^.".m"?,",',"."c" PC .! "J""t.". " "" - =au-~ - -.M> 2*E'20-" =;,"" ',= ;"=,, ,
r."A"","E"'J'l2" ca = ~ ~ ~ ~~ Pa' ""l"u*"" i ."" ^u=>-===T=-* = ua .
,T .et.e=,s.T.!M*J c. .c -A J. E... __ =va - Tu - ~~ ===
meeCMAEL A. SwsGEN esAsset #,uGE JOcess M. O fsEfL6 Jft. 8HCMAIBO J. DefMm.meO ELLEN SMEfteFF LAmt. J et.. L 8.AseETT C
THOesA3 A .AKTER.
- C.
L .ssGLeCm. J.Av.A. Ou* C..un-TIMOTMV ..mt E f.BSTIE.N c JO-u. A.e,,,,,,,'OL,.
hecGNIDE ELLEN Maa SAMO88A E. FOLSome A* A. E. F#E. . i. DEE.E.Ma= *=caJ.ra===
EILEEN se. OLEtesER ulBG O COUseGEL A. S Da c EL.s.S.A,GET.M u. to. P.E.N.E.LEToM JUD.IT.M.O e ,ApeO.LER... 00 T-.=
J.1.m>O as18 M J. .E E.R.
- 86. P.c M M. EL EO a Ou .
- c. vtO.It ..E SAMM August 23, 1983
,E. . O.=e? . .u-==
822-1026 Harold R. Denton, Director '.
Office of Nuclear Reactor Regulation U. S. Nuclear Regulatory Commission ,
Washington, D.C. 20555
! Re: Meted (TMI-1), Docket No. 50-289 i
l
Dear Mr. Denton:
./
In your memorandum to your principal Staff members, dated July 11, 1983, you outlined the plans for NRR review l,,
of the GPU v. B&W lawsuit documents. Among other things the memorandum stated that "DL will contact and, if appro-priate, will meet with outside organisations (e.g. Con- '
gressional' Staff and hearing intervenors) to obtain their comments" on those documents. Similarly, by memorandum dated June 30, 1983, Staff counsel advised that counsel was planning to telephone each party t7 the TMI-l restart proceeding in order to obtain identification of any,'of the litigation documents which the parties believe to be per-tinent to the Staff review.
Licensee has apprised its counsel in the litigation, who are familiar in detail with the documents in question, of the NRC's ongoing review of the litigation documents and areas of particular interest as evidenced by pleadings filed in the Restart Proceeding. One such particular araa of in-terest is the so-called " Mystery Man" issue and the' question e ?
9
, _ _ . . . , - . , ._. ,_ _ _ _ _ . , _ . _ _ , _ , . , _ _ . , . . . , - . . , _ . . , . ___.,m _ _ _ _ _ _ , _ _. , _ _ _ . _ _ . _ _ _ _ . _ _ _ _ _ _ _ _ , . _ . _ _ _ _ _ .
i SHAW PITTMAN. PoTTs & TROWCRIDGE
. .r ... o. ore o 6 co o..rio .
Harold R. Denton August 23, 1983 ,
Page Two of actuation of HPI following the TMI-2 accident. For your information I enclose a copy of a memorandum on this subject prepared by Kaye, Scholer, Fierman, Hays and Handler, dated August 16, 1983. Other memoranda, as appropriate, will be forwarded to the Staff for use in its review.
On behalf of Licensee, we request that undersigned counsel be notified of, and have an opportunity to attend, any meeting which your Staff or NRC counsel may have with any "outside organizations," including Congressional Staff and hearing intervenors, for the purpose of identifying or obtaining comments on any of the GPU v. B&W lawsuit docu-ments pertinent to your review.
Sincerely, SH PITT ,P TS & TROWBRIDGE gi eo e F. T owbridge Counsel for Licensee cc w/ encl: Mary E. Wagner, Esq.
t 1
e
MEMORANDUM ON THE 5:41 HPI ACTUATION
" MYSTERY MAN" ISSUE I
l r
l l
l l
l l August 16, 1983 l
l O
l / .
--- .c -_. -
. i
\
?NTRODUCTION The first reference to a so-called " Mystery Man" was in the opening statement in the GPU v. B&W trial by B&W's counsel, who coined the phrase in the context of his argument that (1) there was evidence in prior testimony that high pres-sure injection ("HPI") was actuated at about 5:41 a.m.; (2) if HPI had been maintained at that time, the core would not have been' uncovered; and (3) since no one t.estified to having turned off HPI near that time, there must have been a " Mystery Man" who did so.1 Prior to this assertion by B&W's counsel, none of the studies and investigations by the NRC and other indepen-dent bodies had found or even speculated as to the presence of a so-called " mystery man" in the control room. After the ar-gument was raised at trial, GPU's trial counsel retained EDS Nuclear Inc., a firm of expert consultants, to perform an analysis of this issue. The EDS study conclusively deter-mined, on the basis of hydraulic analyses, that there had been no actuation of high pressure injection during the time period in question. Since the high pressure injection was proven by scientific analysis not to have been actuated at about 5:41 a.m., no " mystery man" turned it off.
4 1 Trial tr. 148-49, 159.
THE EDS ANALYSIS .
EDS Nuclear Inc., an engineering consulting firm with an expertise in thermohydraulic analysis, provides engi-neering, design and analysis services to th'e electric power industry.2 The EDS analysis of the alleged 5:41 full flow ac-tuation of high pressure injection was conducted by Dr. James Holderness, who is the manager of an EDS group specializing in thermohydraulic analysis.3 Dr. Holderness' qualifications to perform the analysis consisted of ten years' past experience ,
in performing hydraulic analyses, including seven years with Combust' ion Engineering, where Dr. Holderness was responsible for performing analyse.s in support of licensing applications for nuclear power plants and for verifying the efficacy of the models used in those analyses.4 Dr. Holderness' analysis of the alleged 5:41 actua-tion of high pressure injection was reviewed within EDS both by another EDS engineer and by the EDS Nuclear Quality Assur-ance Department.5 The objectives of the EDS analysis were to determine whether certain indications of operation of the emergency core cooling system, which were recorded on the day of the acci-2 Holderness', trial tr. 5590-91.
3 Id. at 5591-92.
4 Id. at 5590-91.
5 ld. at 5592.
^% ,,
dent, could be analyzed to identify an actuation of high pres- .
sure injection and, if so, to determine whether an actuation occurred at or about 5:41 a.m.6'.
There are two potential ~ emergency cooling water sup-plies to the high pressure injection pumps: the make-up tank and the borated water storage tank.7 EDS constructed analyti-cal computer models which established that the response of the make-up tank level would. indicate whether there was an actua-tion of high pressure injection.8 EDS confirmed the accuracy of its models, and of the conclusions derived from those mod-els, by comparing those conclusions with data on make-up tank level acquired during undisputed actuations of high pressure injection.9 The EDS analysis found that make-up tank level could be used as an indicator of full flow high pressure injec-tion,10 and further found that "the 5:41 response of the make-up tank level does not exhibit the characteristics of (an ac-tuation of high pressure injection]."ll The EDS analysis 6 " Analysis of Reactor Coolant System Make-Up During the Three Mile Island Unit 2 Event," EDS Nuclear Inca, Decem-ber 29, 1982, p. 1. The analysis was marked at trial as GPU trial exhibit no. 2233.
7 Holderness, trial tr. 5594.
8 Iji. at 5594; 5600. .
9 Id. at 5600-01; 5608-09.
10 GPU trial exhibit no. 2223, p. 2.
11 Id. at p. 3.
3
flatly concluded that the make-up tank level behavior at 5:41 a.m.. established the absence of an actuation of high pressure injection.12 Dr. Holderness, in- testifying at trial, demon-strated that given the behavior of make-up tank level at or about 5:41, it was impossible for there to have been an HPI actuation at that time.13 OTHER INDEPENDENT STUDIES The conclusion of the EDS study that there had not been a 5:41 actuation is consistent with the findings of other independent studies regarding the March 28, 1979 accident.14 Among the technical analyses of the Three Mile Is-land accident containing detailed sequences of events are the NRC Special Inquiry Group (SIG) Reportl5; NUREG-060016; and 12 Id. at p. 3.
! 13 Holderness, trial tr. 5637; 5700.
14 See discussion in the NRC staff's " Report of the Review of the Babcock - d Wilcox - General Public Utilities Law-suit Trial Court Record," March 28, 1983, at pp. 6-9.
15 "Three Mile Island - A Report to the Ccamissioners and the Public," Nuclear Regulatory Commission, Special In-
! quiry Group (NRC-SIG), Volume II, Part 2, "The Accident and Its Analysis" ("SIG Report").
16 "Irvestigation into the March 28, 1979 Three Mile Island Accident by Office of Inspection and Enforcement," Inves-tigative Report No. 50-320/79-10, NUREG-0600, July, 1979
("NUREG-0600").
l l
l
! 4
l NSAC-80-1.17 None of :no scquencos of ovents in thoso studies j include a 5:41 HPI actuation, and none of these studies con-s cluded that there was such an e'v.ent.
A GPU sequence of events, TDR -044,18 ref ers to a 5:41 full flow manual HPI actuation. That entry, however, was based on statements by cperators, which are reviewed below.
Significantly, the TDR limits the possibility of a 5:41-actua-tion to a maximum of five minutes, a conclusion consistent l with the findings of GPU-sponsored studies of borated water storage tank levels that a 5:41 actuation could not have last- i ed for more than five minutes.19 As we will show, this maxi-mum five-minute figure in itself casts doubt on the operators' ability to recall the timing of the full flow HPI manual actu-ation.
Moreover, Dr. Holderness of EDS Nuclear discussed the EDS study with the principal author of TDR-044, Dr. Van
'Witbeck of Energy, Inc. After reviewing the EDS report, Dr.
Van Witbeck concurred that no studies had been done in prepa-
- 17 " Analysis of Three Mile Island - Unit 2 Accident," Nucle-ar Safety Analysis Center, NSAC-80-1 (NSAC-1 Revised),
1 March, 1980 ("NSAC-80 1").
18 GPU trial exhibit no. 2079, GPU Nuclear Technical Data Report, TDR-044, " Annotated Sequence of Events, March 28, 1979," February 6, 1981 ("TDR-044").
19 GPU trial exhibit 2079, TDR-044 at figure 60; Zewe dep.
821-22, 826-28, 830-37.
5
ration of the TDR which contradicted the EDS conclusion as to the absence of any 5:41 actuation.20 OPERATOR STATEMENTS IN INTERVIEWS '
REGARDING A 5:41 ACTUATION In early interviews of the operators, their recol-lection of the approximately 15 hour1.736111e-4 days <br />0.00417 hours <br />2.480159e-5 weeks <br />5.7075e-6 months <br /> accident sequence gener-ally was not assisted by computer data or engineering analyses of the accident.21 Zewe testified at trial that soon after the accident he spoke to and was interviewed by "a great many people"; that he had had no opportunity to review data to con-firm the accuracy of his recollections; and that some of these early interviews were conducted in motel rooms, cars and trailers.22 Moreover, as Zewe noted during the trial, the op-erators' post-accident recollection often tended to " compress" the timing of various actions taken on the day of the acci-dent.23 Zewe explained in an early interview: . . . I was very far off in the times. Times were much longer and I felt that they were much shorter. . . . (Oln a few of the graphs and things that we have seen, we have noticed that the times there were a lot different than we have previously imag-I ined."24 l
20 Holderness trial tr. 5659-60.
21 Zewe trial tr. 3031.
22 _Id.
23 Id. at 3032.
24 TMI Staff interview, 4/6/79, p. 1.
6
_ ~ . . . _ _ _ _ _ , _ . _ _ _ _ _ _ _ - . . - - _ . _ _ - . . . , _ _ . _ _ __ _ . _ _ _ _________-._._,_____.A_
In order to delineate and understand the sequence of events in a nuclear plant transient, large amounts of data'are recorded and then subjected to engineering analysis. As with the " black box" used in airplane accidents, it is more accu-rate to rely on engineering analysis of recorded data than on human recollection of numerous, technically complex events.
No one person could be expected to recall precisely the time and sequence of the thousands of indications and actuations which occurred over a period of some fifteen hours on March 28, 1979.
Unfortunately, alarm printer data was not available for a period of time on March 28, 1979 that included 5:41 a.m.25 The alarm printer data, where available, recorded each actuation and termination of high pressure injection. The alarm printer records one full flow manual actuation of HPI at 7:20. This followed the shutting off of a reactor coolant I
pump at 7:13.26 Faust l
Craig Faust, an operator on shift on the day of the accident, stated in an NRC interview that manual HPI actuation occurred prior to the time of the shutting off of the reactor coolant pumps, at 5:41.27 However, Faust later appeared to 25 GPU trial exhibit no. 2079, TDR-044, p. 42. (Alarm printer data not available from 5:13 to 6:48 a.m.)
26 GPU trial exhibit no. 2079, TDR-044.
27 Faust, NRC I&E interview 4/21/79, pp. 48, 51.
7 l
e
. . - -- -- . .-.n - , - , - . - . - , - , , , . - - --- . - . - . . . , ,- ..-- --- ---- .- . a
realize that his recollection of the manual actuation had been displaced in time, and that he had confused the 7:20 full flow manual actuation with the 5:41 bime period.28 Additionally, in response to what an I&E interviewer characterized as a
" monday morning quarterbacking" question about what he would have done differently, Faust stated that he had felt uneasy about turning off the reactor coolant pumps, and that, had he known at that time what he had learned after the accident, he would have tried high pressure injection rather than turning off the pumps.29 This conclusion appears to be inconsistent with Faust's statement that full flow HPI had been initiated at the time of the 5:41 reactor coolant pump shutdown. Howev-er, the interviewers apparently did not assist Faust in recon-ciling this divergent recollection, either by bringing it to his attention, or by providing relevant data.
Zewe .
Nearly two months after the accident, at an inter-view on May 25, 1979, William Zewe, shift supervisor on the day of the accident, referred to a 5:41 full flow, manual ac-28 Specifically, during an interview on May 29, 1979, Faust, after a lengthy discussion of what was uncuestionably the l
7:20 manual actuation of HPI, said "[t]his [the 7:20 ac-l tuation) would be the manual actuation wnich would be right around when we were stopping the pumps. The one I could think of. So that would be earlier than what it should be." NRC I&E interview, 5/29/79, p. 10 (emphasis added).
29 NRC I&E interview 4/3/79, pp. 39-40, 48.
8
, w- ,n,, - s y 7- - -4 p- -,--- - , , - , , - - , y w y- ,r, ,w- y,,,m- > - - - - - - - - z -n. ,,,,,m-,-,w, - - , - , , - ___,, _ __ -
tuation of HPI when the remaining reactor coolant pumps were secured.30 As of the May 25th interview, Zewe had attended several meetings at which Faust had mentioned a 5:41 actua-
~ tion.31 Zewe's understanding of the time of the actuation ap-parently derived from the earlier statements of Craig Faust.
For example, during the May 25th interview, in response to a question as to whether he was sure full HPI had been on around 5:41, Zewe stated that "I was not as sure as the operator who actuated it was. He is sure, Craig Faust.a32 Notably, Zewe's understanding that Craig Faust had actuated HPI was contradicted by Faust himself, who stated that he thought that high pressure injection had been initiat-ed by Edward Frederick at the time that he, Faust, recalled having secured the reactor coolant pumps.33 Frederick Edward Frederick was the operator assigned to the HPI controls throughout the day of the accident.34 He has never testified to a 5:41 actuation, either during the GPU v.
l B&W trial, during his deposition or in interviews. In an ear-30 Zewe, TMI Staff interview 5/25/79, p. 31.'
31 Zewe trial tr. 2760-62.
l 32 TMI Staff interview, 5/25/79, pp. 5-6; see Zewe dep. 824-26.
33 Faust, NRC I&E interview, 4/21/79, p. 39; Faust, Presi-dent's Commission deposition, p. 158.
34 Frederick trial tr. 3492-96; Zewe trial tr. 2124.
9
. _ . _ . __ .. . _ _ _ _. _ _ _ . _ - - _ _ _ _ _ _ _ . __ _-.1 .
ly interview, Frederick characterized the manual HPI actuation as having occurred during a time when Gary Miller', Station Su-perintendent, was present in th'.e control room.35 Miller ar-rived in the control room at approximately 7:05 a.m.
Frederick also placed the HPI actuation as occurring.
"after the relief valve had been isolated"36 which was at 6:19 a.m.,37 and as occurring "during the time when we had decided that we did not have natural circulation,"38 which was at about 6:54 a.m. when reactor coolant pump 1B was restarted in an attempt to induce forced flow circulation, marking the point at which the operators ultimately were convinced that natural circulation was not going to work.39 These statements establish Frederick's recollection that a manual actuation of full flow HPI occurred, not at ~ 5:41 a.m., but rather after 7:00 a.m. -- conforming to the alarm printer entry at 7:20.
35 Frederigk, NRC I&E interview, 4/23/79, pp. 48-57.
36 Frederick, U.S. Senate Subcommittee interview, 8/22/79,
' pp. 16-17.
l 37 GPU trial exhibit no. 2079, p. 47.
38 Frederick, U.S. Senate Subcommittee interview, 8/22/79, pp. 16-17.
39 Fr'ederick, NRC I&E interview, 4/3/79, p. 36; Faust, NRC I&E interview, 4/21/79, p. 52; GPU exhibit no. 2079, TDR-044, p. 51.
10
"GPU v. B&W" DEPOSITION AND TRIAL TESTIMONT Faust f Faust's deposition testimony further confirms that he did not have a sound basis for his earlier statement re-
/
garding a 5: 41 full flow HPI actuation. He,. testified that
~
"before" they took the pumps of f ,'ETgh pressure injection was initiated; that he did not physically do it, but he heard it was done or was going to be done; that he did not know if it was at full flow; and that he was primarily involved with the secondary side of the plant.40 Zewe During his GPU v. B&W deposition, after he had had access to relevant technical data, Zewe testified that he had recalled that "we did high pressure injection at or about the time we secured the second two pumps. But I cannot void my-self of everything else that has happened and what is fact from other sources," refe: ring to studies.of borated. water storage tank level responses to actuations of HPI.41 Thus, Zewe was unable to distinguish in his mind between what he had previously said based on what he had heard from Faust, and facts that he knew to be scientifically accurate.
Zewe's trial testimony further indicates that his recollection of a 5:41 actuation derived from Craig Faust's 40 Faust dep., pp. 525, 540.
41 Zewe dep., 821-24.
11
. - - . - n --_ ,
. \
e statements. Thus, Zewe testified at trial that Faust had said, referring to a 5:41 actuation, that he, Faust, " felt sure that's when it was" and that Zewe had then said: "if you are sure that's when it was, then that's when it was. . . .
"42 Zewe also testified regarding his own early recollection of the time of the manual HPI actuation: "I for one could not remember exactly when it happened. One of the operators, as I recall, was pretty sure of the exact time."43 During the GPU v. B&W trial, GPU's counsel asked Zewe a series of questions which he had not been asked in any of his prior testimony or interviews by the NRC or GPU person-nel, i.e.: How many HPI actuations had there been? How many of these were automatic? How many of these were manual? Zewe testified that he recalled four full flow actuations, three of which were automatic and one of which was manual.44 Specifi-cally, Zewe recalled a single manual actuation; the initial automatic actuation a few minutes into the accident; an auto-matic actuation "after the declaration of the site and general emergency"; and an automatic actuation in the early afternoon f associated with the hydrogen burn.45 Zewe recalled that the manual actuation was the second actuation of the four.46 42 Zewe trial tr. 2761.
43 Zewe trial tr. 2173, 2761.
44 Zewe trial tr. 2115-18, 2153, 2156.
45 Zewe trial tr. 2115-18; 2153, 2156.
46 Zewe trial tr. 2117, 2121-22.
12 l
(
This recollection is corroborated by recorded plant data. Computer alarm data establish three automatic actua-
~
tions of full flow HPI: at two minutes into the event; at 7:56 a.m., or 30 minutes after the general emergency declara-tion; and at 1:50 p.m., the same time as the hydrogen burn.47 The computer alarm data lists only one manual actuation of full flow HPI, at 7:20 a.m. Thus, the manual actuation was the second HPI actuation, just as Zewe recalled.
Zewe also testified that the manually initiated full flow HPI that he recalled had been maintained "for a consider-able period of time,"48 which he stated meant from ten minutes to half an hour.49 This accords more closely with the record-ed 7:20 a.m. actuation, for seventeen minutes, than with the five-minute possibility at 5:41 a.m. in TDR-044.50 Zewe's ultimate conclusion at trial was that the closest that he could pinpoint his recollection of the single manual high pressure injection actuation at full flow was that it occurred after the reactor coolant pumps were turned off (at 5:41 a.m.) but before the second automatic actuation (at i 7:56 a.m.).51 l
l 47 GPU trial exhibit no. 2084, alarm printer data.
48 Zewe trial tr. 2115.
49 Zewe trial tr. 2173-77.
50 GPU trial exhibit no. 2084, alarm printer data; GPU trial l exhibit no. 2079, TDR-044, figure 60, 51 Zewe trial tr. 2117, 2121-22.
I 13 l e
i
'. \
The operators' earlier confusion may have resulted in part from the fact that some time after the reactor coolant pumps were shut off, the operators began to inject borated wa-ter into the reactor coolant system in order to regulate the boren concentration. This action was necessary to counteract symptoms that the reactor was returning to criticality.52 Furthermore, the 7:20 manual actuation was -- according to the recorded data -- in fact juxtaposed to the tur,ning off of a
! reactor coolant pump,53 an event which Faust and Zewe linked to the manual actuation of HPI. A reactor coolant pump was shut off at 7:13, following an attempt to restart the pumps.54 Frederick Frederick, like Zewe, testified during the GPU v.
B&W trial that he recalled only one manual actuation of HPl.
He put the time of the manual actuation as "around the time of the site emergency."55 Frederick had no recollection of hav-ing actuated HPI manually at full flow at around 5:41 a.m.,
and was quite certain that no " mystery man" could have sneaked by him to actuate or turn off HPI while he was at the con-trols.56 ,
52 Zewe trial tr. 2110-13; Frederick trial tr. 3477-79.
53 Zewe trial tr. 2153 et seq., GPU trial exhibit no. 2084, alarm printer data.
54 GPU trial exhibit no. 2084, alarm printer data.
55 Frederick trial tr. 3493. The site emergency occurred at 6:50 a.m. GPU trial exhibit no. 2079, TDR-044, p. 50.
56 Frederick trial tr. 4046.
14 1
,~ -
At trial, B&W's counsel examined Frederick on the l
issue of a 5:41 actuation pointing out that Frederick had been present in post-accident interviews when Faust and Zewe re-ferred to a 5:41 actuation.57 Frederick consistently denied ever having agreed that there was a 5:41 actuation,58 and B&W produced no statements of Frederick to the contrary. Freder-ick testified:
I just never disagreed with (Zewe] that's
, all. Until we put all the data together, t'here was nothing to disagree about. . . . -
I think that they think I agreed because I didn't disagree."59 Frederick also testified to data concerning the be-havior of make-up tank levels. On cross-examination, B&W sub-I, .
mitted a chart, prepared by its engineers during the prior few days, reflecting the behavior of make-up tank levels at 7:20, which B&W's counsel claimed was similar to the pattern at 5:41 a.m. In order to resolve the issue in a reliable and defini-tive way, GPU's trial counsel decided to call in EDS Nuclear.
As discussed above, Dr. James Holderness testified to the EDS Nuclear study which established that it was not possible for HPI to have been actuated at full flow at or about 5:41 a.m.
in. light of the analysis of make-up tank levels. In particu-lar, Holderness explained that while the 5:41 a.m. and 7:20 57 Frederick trial tr. 3876-77.
58 Frederick trial tr. 3877-79; 3883-84.
59 Id. at 3888.
I 15 l
m
a.m. make-up tank level patterns might appear similar to the naked eye, computer analysis plainly showed critical differ-ences and disproved any 5:41 a.m. actuation at full flow.60 i
CONCLUSION The alarm printer data show an undisputed manual ac-tuation of high pressure injection at 7:20.61 Thus, the alle-gations of a 5:41 manual actuation must hypothesize an addi-tional manual actuation. Yet, uncontradicted trial testimony establishes that both Zewe and Frederick had a clear recollec-tion of there having been only one manual actuation.62 Fred-erick recalls specific circumstances that place the manual ac-tuation near to the 7:20 time recorded on the alarm printer; Zewe accurately places the one manual actuation as being the second of the total of four HPI actuations, as is recorded on the alarm printer at 7:20. The recorded 7:20 actuation fol-lowed a shut-off of a reactor coolant pump, also recorded in plant data, at 7:13. Accordingly, the trial testimony is con-sistent both with the available alarm printer data, and with the conclusion of the EDS technical analysis that there was no l
5:41 actuation.
l 60 Holderness trial tr. 5637, 5700; GPU trial exhibit no.
2223, p. 3.
61 GPU trial exhibit no. 2084, computer alarm data. .
62 Zewe trial tr. 2153; 2156. Frederick trial tr. 3493; 3876.
l i
16* .
_ - - - - - .-. - .i . - . . - . - _.. . .-. -- . -. -.
Earlier statements made by another operator, Craig Faust, and relied upon by his superior, William Zewe, mis-placed the time of the manual actuation back by an hour and forty minutes, to an earlier shutting off of the reactor cool-ant pumps, at 5:41 a.m. In the absence of data, this was an understandable confusion of recollection.
Since the EDS analysis scientifically proves that that there was no full flow manual actuation of high pressure injection at or about 5:41 a.m., the colorful charge that a
" mystery man" turned it off was conclusively rebutted.
i l
)
I 17 i
\
SHAw, PITTMAN, PoTTs & TROWBRIDGE -
A pamTNEmSMip Or enorESSeONAL CompOmATIONS 1800 M STmEET. N. W.
WASHINGTCN. O. C. 20036 m*MSA ,o.pOTTS.p C. SMELOON J. wEiSEL p C. (202) e22.sooo sg,,g,,Ly,,go, g o,,,, ,, ,o,,,, ,
SituAnT 6 mTTMAN. p C. JOHN A. McCutLOuoM. p.C. _ JACn Menar ANonEw o. ELuS
( EOmo~t F. Taowenvoot, a C. J. parmCn MsCnEv.
- C. TMOM AS M. McConMsCm mCMAmo A. SAMp STleMLN o. pOTTS.
- C. J. TMOMAS LENMAmT.
- C. TELECop'Em JOMN 6 Camm. Jm. TMoMAS E. CooCmER Jm.
Of MAL.9 CHANNorr, p C. STEVEN L MELT 2En. p C.
(202) Saa-eOSS & Saa-liSS pMI.up mC En J,. MAmVEY u,gonoon wE,N,DE gy gg U,N u, A.
,,*M,,ITE eM Lu> o. eOSTwsCa.
- C. otAN o. AvuCa.
- C.
SL TIMOTMY MANLON. p C. JCMM ENGEL p C. _ SAMSARA J. MOmOE N gggyg m, SM,TM OtomSE M. mCOEms, Ja..
- C. CHARLES s. TEMNIN.
- C. BONNIE S.GOTTLIES VtmGINGA S. muTLEDGE ret ) A. UTTLE. p.C. STEpMEN B. MUTTLEn.
- C. mAntFAE 100 MOwamo M. SMarFEmMAN RATHEmeNE p. C JOMN s. mMessELANOEn. p.C. weNTM mop N. seOww. p.C. OEsomAM e. eauSEm ga,,ggggnag ,HEER .gygen SmuCE w.CMumCMILL p.C. JAMES 3. MAMLIN
- C. MR) 888 8072 SCOTT A. ANENetmG T= AVIS T. SmOwN.Jm.
LESWE A. N8CMOLSON.Jm p.C. #ANOAL S. RELL p C. _
CAMOGELL RfLLEFEm g,ggggg g,ggggyggg MamTtN O. meALL p C. mosEm? E. 2AMLER. p C. pauk M. TMOMAS STEpMEN S. MEIMANN IhCMAm3 J. MENOALL.
- C. ROSENT 3. mOG SIMS. p C. ygggg SETH M. MOOGASIAN SANOmA E. smuSCA*
JAT f* SILet 26. p C. STEVEN M. LuCAS. p.C. SMEILA McC. ManVEY "o*" E "
sameA=A M. mOSSOTTi. p C. oAvio M. mueENSTEiN. p.C. SS.eeSa (SMAwtAw wSM) otuSSA A, moowAv GEOmeE V. ALLEN.Jm p.C. atCMAmO E.GALEN RENisETM J. MAuTMAN EE E"M' pAMELA ". ANCE mSON Fet 4 DeASNEm. p C. LYNN wMITTLESEY weLSON CAELE OMA*bA*. OAVID LAwmENCE MILLEn ALERANDEm O. TOMAS2C2un PMeupD.PomTER O. KENLY wgSSTEn
- C. MATIAS F. TRAVeESO CIA 2 FmEOtm#CR 6 RLEIN NATMANsELp.SetEO.Jm *C. VICTomeA J. *EmMIN S STEVEN p, pfTLEm* M'OMAEb A* O*'OEE MAmW #1JOEN5uCR. p C. JCMM M. O NEILL.Jm. mCManOJ.PAmmMO ELLEN SMEN'FF EmMEST L BLARE.Jm p C. JAY A. EpSTIEN ELLEN A. rmEDEL 8 ANNA J. FINRELSMN CARLOTON S. JONES, p.C. mANO L ALLEN JCMNF*DEALY. MANNAM E. M. uESEmMAN ElLEEN W. OLEIMEN TMOMAD A. SANTE =.
- C. EUSABETH M. etNOLETON COUNSEL SANDEIA E. FOLSOM DAV8Dm.SAMm JAME3 M. SumGER.
- C. Mammy M. OLASSp*EteEL JuDeThe A. SANOLEN C..~SOwoolM m.M..TRAsN September 15, 1983 wa.TER S o.mte? oiA6 NuMmEm (202) 822-1084 Harold R. Denton, Director Office of Nuclear Reactor Regulation U.S. Nuclear Regulatory Commission Washington, D.C. 20555 Re: Meted (TMI-1), Docket 50-289
Dear Mr. Denton:
l In response to the NRC's ongoing review of the l GPU vs. B&W litigation documents and. areas of particular l
interest, as evidenced by pleadings filed in the Restart Proceeding, counsel for licensee on August 23, 1983 pro-vided your office with a copy of a memorandum dated August 16, 1983 dealing with the " Mystery Man" issue and the question of actuation of HPI following the TMI-2 accident.
This memorandum had been prepared by licensee's litigation counsel, Kaye, Scholer, Fierman, Hays & Handler.
In addition to providing your office with a copy of this memorandum, counsel for licensee indicated its in-tention to provide the NRC staff with other memoranda, as appropriate, for use in its review of the GPU vs. B&W trial record. Accordingly we are enclosing a copy of a report i
I I
~
\
s
. SHAW, PITTMAN, PoTTs & TROWZRIDGE A pamTsegmense OF petCMrtse:ONAL COmpOmAficass e
Harold R. Denton September 15, 1983 Page Two recently prepared by Babcock & Wilcox at the request of licensee entitled " Response to GPUN Questions Concerning HPI Actuation at TMI-2 About 5:41 a.m. on March 28, 1979."
This updated report was received by licensee on September 9, 1983. Licensee is currently reviewing this report.
Sincerely, SHAW, PITTMAN, POTTS & TROWBRIDGE LJ f.' 7%ifr.
Ernest L. Blake, Jr.
Counsel for Licensee cc: Mary E. Wagner, Esq.
(w/ encl.)
t 9
-, - -.- , -- _ - . . -- . _,-v
i I
p^s i
l 3
A 1
D RESPONSE TO GPUN QUESTIONS
(~ CONCERNING HPI ACTUATION AT TMI-2 ABOUT 5:41 A.M. ON MARCH 28, 1979 e
l 6
s
)
Prepared by:
(]: BABC0CK & WILCOX, UPGD j LYNCHBURG VIRGINIA
. . , l l
7 .
TABLE OF CONTENTS Page 1.0 Executive Sumary 1 2.0 Background 2 3.0 Discussion 3 3.1 Background 3 3.0 Scenario Definitions 4 k 3.3 TMI-2 Data /Information Base 5 4.0 Scenario 1: HPI Only 22 4.1 MUT Level Response 22 4.1.1 Calculations 22 4.1.2 Results 24 -
~
4.2 RCS Heat Balance 25 m 4.2.1 Calculations 25 A'" 4.2.2 Results . 25 4.3' Comparison with Plant Data 26 L
4.3.1 BWST Volume 26 4.3.2 RCS Inventory 26 4.3.3 RC Flow 27 p 4.3.4 SRM Signals 28 4.4 Scenario 1: Summary and Conclusions 28 F 5.0 Scen uio 2: EFW Only 37 5.1 Estimated Heat Removal and EFW Flow Requirements -
37 5.2 Comparison of Steam Generator Response at 94 Minutes 38
- and at Eight Minutes After Turbine Trip
?
5.3 Steam Generator Response Analysis using the AUX-II Code 39 5.3.1 Calculations 40 y (m._
v> 5.3.2 Results -
40 y
il i
V . - - -
l l
l
,_,. TABLE OF CONTENTS I (continued) 1 l Page l -
\
l 5.4 Comparison with Plant Data 42 l
5.4.1 MUT Level 42 l l
5.4.2 BWST Volume 43 5.4.3 RCS Inventory 43 5.4.4 RC Flow
- 43 1 5.4.5 SRM Signals 44 5.5 Scenario 2: Sumary and Conclusions 44 6.0 Scenario 3: Conbination of'EFW and HPI 50 N 6.1 Observations from Other Scenarios 50 l ~
f; 6.1.1 Scenario 1: HPI Only 50 A
6.1.2 Scenario 2: EiV Only 51
) 6.2 Scenario 3: Summary and C'onclusions 52
, 7.0 Sumary and Conclusions 54 h 8.0 References 56
~
APPENDICES
- 1. MUT Level Model A-1.0
- 2. Energy Balance Model A-2.0 l
l 3. HPI Cooling Model A-3.0
- 4. GPUN Questions of April & May 1983 A-4.0
- 5. Brief Summary of References Identified in GPUN Questions A-5.0 f of May 20, 1983 P7 m, i
s b
11 b
l i
' f 'i
. LIST OF FIGURES l Page i
3-1 RCS Temperatures, 60 to 120 minutes 9 3-2 RCS Temperatures, 85 to 130 minutes 10 3-3 RCS Pressure, 85 to 130 minutes 11 3-4 Pressurizer Level, 85 to 130 minutes 12 3-5 RC Flow, "A" Loop, 85 to 130 minutes 13 l
1 3-6 Steam Pressure, " A" SG, 85 to 130 minutes 14 3-7 Steam Pressure, "A" and "B" SGs, 85 to 130 minutes 15 I 3-8 "A" SG Startup Level, 85 to 130 minutes 16 3-9 "A" and "B" SGs Startup Level, 85 to 130 minutes 17 i
3-10 MU Tank Level, 85 to 130 minutes 18 .
. 3-11 Source Range Monitor (SRM) Signal 19 3-12 Letdown Cooler Inlet and Outlet Temperatures 20
/
[e(G 3-13 HPI, Makeup and Letdown Systems 21 4-1 Scenario 1: Makeup Tank Level Response, Model vs. Acual 29 ;
4-2 Scenario 1: Assumed Letdown Flow ,
30 4-3 Scenario 1: HPI Flow 31 l l
4-4 Scenario 1: MU Flow 32 ,
4-Sa Scenario 1: MU Tank Level Response and Estimated 33 Letdown, 4 to 6 minutes j l
4-5b Scenario 1: MU Tank Level Response and Estimated 34 l Letdown, 10 to 12 minutes 4-6 Composite: MU Tank Level and Letdown Flow vs. Time 35 4-7 Scenario 1: Comparison of RC Temperature, Energy 36 Balance Model vs. Actual '
- y .
'd 5-3 Scenario 2: AUX-II Model Results, "A" SG Pressure vs. Time 48 i
Scenario 2: AUX-II Model Results, "A" SG Level vs. Time 49
. 5-4 l
- iii b - -.
f
.p 1.0 EXECUTIVE SUMARY This report is B&W's response to questions posed by GPUN concerning HPI ,
flow at TMI-2 on March 28, 1979. The information presented is ba' sed on the technical material developed by B&W during the GPUN/B&W litigation. The analyses described generally cover the time frame from the initiation of the cooling of the reactor coolant system to the point when the "A" loop steam generator level had been increased to -50% on the operate range.
g This 31-minute period (5:34 - 6:05 a.m.) corresponds to the period from 94 to 125 minutes after turbine trip.
{
The two most probable mechanisms that either individually or in codination could account for the RCS behavior during this time frame were high
! pressure injection (HPI) and emergency feedwater (EFW). Based on the 5 analyses described in this report, it has been concluded that:
( o From 94 - 100 minutes after turbine trip, HPI or EFW, either alone oE in combination, could have produced the observed plant cooldown.
V -
o From 100 - 125 minutes after turtiine trip, EFW was the main source of plant cooldown and the HPI pumps were not fully actuated.
s o Based on the evaluations performed, it is not possible to conclude the J. exact cause of the cooldown between 94 and 100 minutes. However, the codination of HPI and EFW appear to be the most likely cause.
o For the codination of HPI and EFW cooling, the flow capacity of at least one, and possibly two, HPI pumps is required to produce the observed plant response.
F l
l :m
/
H 1
l . .
(
2.0 BACKGROUND
In April 1983, GPUN requested information from B&W about the actuation of .
i HPI at TMI-2 at 5:41 a.m. (101 minutes after turbine trip) on March 28,
[ 1979. B&W answered these questions in early May 1983 ;1 however, GPUN l requested more detail and gave B&W an additional list of questions in i mid-May 1983. (The original as well as the additional questions are :
included as Appendix 4.) This report responds to the April and May s questions and summarizes the analyses and data interpretations pertinent to
[ the question of HPI actuation at TMI-2. All time references refer to minutes after turbine trip.
3
. With one exception, the technical material provided in this report was either generated by or available to B&W in late January 1983, when further f analyses were terminated (settlement of the GPUN/B&W litigation). As such, ,
it should be recognized that these analyses are not " final". Further .
refinements of these analyses were beyond the scope of this response.
b# The exception to the January 1983 data base is the limited scope analysis of thq "A" steam generator response in the 94 to 100 minute (05:34 to 05:40 l
a.m.) time period, using the stand-alone 177-fuel assembly (FA), nuclear steam system (NSS) steam generator model from the AUX-I,I code (Section 5.0). The analyses performed by B&W are of a scoping nature - they have not undergone the internal review and Q/A normally provided for B&W's i design basis documentation.
1 This report is an evaluation of three scenarios that could have caused the known reactor coolant system (RCS)/NSS cooling comencing at about 5:34 f
a.m. on March 28, 1979 at TMI-2. Section 3 provides background data used I
in svaluating the postulated cooling scenarios. Scenario 1, cooling by HPI L only, is discussed in Section 4. Section 5 discusses cooling by EFW only, Scenario 2. The final cooling scenario, a combination of EFW and HPI, is discussed in Section 6. A sumary of the results and conclusions are provided in Section 7.
.f 3 V
L 2
P _ . _ . _ _ _ _ . _ _ ._-__ -___
3.0 DISCUSSION
(
).
This section of the report will present background information, outline the .
postulated scenarios, and identify b;. sic informa' tion (data base) use'd in assessing the scenarios.
3.1 Background
i.
For the 94 minutes between 05:14 a.m. and 06:48 a.m. on March 28, 1979, output from the alarm printer at TMI-2 was not available. As a result, data showing component status that could have provided direct information about HPI actuation was also unavailable. Furtner, continuously recorded measurements of other plant parsmeters such as HPI flow, makeup (MU) flow, EFW flow, and letdown flow were also
) unavailable. However, there are numerous other plant parameters, such y as those discussed in Section 3.3, that can be used as clues in ,
examining the question: "What was the HPI status in the time period when the ' A' RC pumps were tripped (about 5:41 a.m. or 101 minutes
[v after turbine trip on March 28,1979)?"
In April 1983, B&W's response to GPUN's original questions included the following information: .
o Cooling of the RCS, as indicated by decreasing RCS pressure, was 7
initiated at about 94 minutes after turbine trip, and continued without significant interruption (except for tripping of RC pumps Al and A2) until 125 minutes when the "A" once-through steam generator (OTSG) operate level had been raised to about 500 m
i o The RCS cooling rate was about 4.4F/ minute between 94 and 100 H minutes.
, o EFW was on and flowing to the "A" steam generator at 100 minutes, as indicated by the "A" steam generator startup level response.
G3 -
3 L -
I From about 100 to 125 minutes, the EFW flow rate suggested by the "A"
~
steam generator lovel response was adequate to provide the observed RCS and steam system responses (RCS pressure and temperature,s, ,
pressurizer level, and "A" steam generator level and pressure).
Analysas performed with the HPI cooling model (Appendix 3) confirm
~
this premise and show that HPI could not have been fully actuated at 101 minutes (5:41 a.m.).
i However, the mechanism for RCS cooling from 94 to,100 minutes is uncertain due to a lack of data indicating component status. If the l
HPI system were actuated, it would have occurred during this time period. Three cooling scenarios have been postulated as the potential f
causes for the RCS cooldown from 94 to 100 minutes - HPI only, EFW only, and a combination of HPI and EFW. Each scenario is discussed in 3
the following section.
I I 3.2 Scenario Definitions t
I -
Scenario 1: HPI On11 This scenario assumes that at about 94 minutes after turbi.1e trip
}
(05:34 a.m.), both HPI pumps (1A and IC) are operating with a total 1 flow less than design capacity. At about 951/2 m'inutes (05:35:30),
l the flow from both HPI pumps is increased to their design capacities and continues at this rate until about 101 minutes (05:41), when it'is i either substantially reduced or terminated. The assumed mechanism for RCS cooling after about 101 minutes is EFW injection into the "A"
{ OTSG. The analyses performed to support this scenario are discussed l in Section 4.0.
Scenario 2: EFW Only L This scenario assumes two conditions: first, there is no HPI flow into the RC system (makeup is not ruled out); second, the RCS cooling yN
! .. effect consencing at about 94 minutes (05:34:00) is due principally to
(> the EFW injection. The analyses performed for this scenario are described in Section 5.0.
]
, 4
e i ,
a C' Scenario 3: Combination of HPI and EFW Scenario 3 assumes that beginning at about 94 minutes after . turbine
- trip (05:34 a.m.), HPI (or makeup) and EFW are actuated at less than full capacity. This combination of HPI and EFW continue (although the HPI to EFW ratio may vary) over the period of interest (94 to 125 minutes or 05:34 to 06:05 a.m.). This scenario is discussed in Section 6.0.
3.3 TMI-2 Data /Information Base The available RCS, steam system, HPI, MU, and letdown data are common to the three scenarios. Figures 3-1 through 3-13 (imediately following) illustrate the data base used in describing and evaluating each scenario. Unless otherwise noted, the time of turbine trip, .
04:00:37 a.m. on March 28, 1979 is taken as time zero. The -
resctimeter data is also indexed to this time (04:00:37 + 3 sec.). ,
/~ Turbine trip on the reactimeter occurs at 45,624 seconds. Control L room strip charts, where possible, use turbine trip, reactor trip, or other known events for indexing to the time of turbine trip. The RCS pressure wide range control room strip chart has been adjusted by a
-75 psi correction to agree wit'h other RCS pressur.e signals (reactimeter, memory trip review, and alarm printer data).
( A brief review of each figure follows. Particular periods of interest are noted.
L Figure 3-1: RCS Temperature: 60 to 120 minutes _
The figure shows the RCS temperature response from 60 to 120 minutes for general trend information.
e Figure 3-2: RCS Temperature: 85 to 130 minutes l At about 94 minutes and 10 seconds, a significant cooling of the RCS J/ ,
occurred. The cooling rate of 4.4F/ minute was sustained for approximately six minutes.
i 5
l i,
Figure 3-3: RCS Pressure 85 to 130 minutes The RCS pressure from 94 to 100 minutes decreased from about 1,100 to ,
900 psig. Pressure then decreased from about 900 psig at 10'0 minutes s to 625 psig at 125 minutes.
1 Fig 9re 3-4: Pressurizer Level: 85 to 130 minutes 1
The pressurizer level decreased about 30 inches in the twa minutes between 94 to 56 minutes, and then slowly increased 10 inches by 101 minutes.
Figure 3-5: RC Flow "A" Loop: 85 to 130 minutes The "A" loop RCS flow signal indicated about a 10% increase (35 million 1b/hr to about 38 million 1b/hr) at 94 minutes, peaked at .
.l j about 95 minutes, and began to decrease at a rate faster than that of the 85-94 minute period until the "A" RC pumps were tripped.
J .
f Figure 3-6: Steam Pressure, "A" SG: 85 to 130 minutes -
L The "A" SG pressure began decreasing at about 89 minutes. At 92 minutes, pressure abruptly increased about 20 psi' Pressure then -
continued to decrease until 94 minutes. At 94 minutes, e pressure increase occurred going from 820 to 865 psig at 98 minutes. From 98 to 124 minutes, a general pressure decrease from 865 to about 600 psig occurred.
Figure 3-7: Steam Pressure, "A" and "B" SG: 85 to 130 minutes This figure presents both the "A" and "B" SG pressures for comparison.
The "B" SG pressure began to decrease at about 72 minutes post trip l
l ("B" loop RC pung trip). This "B" SG pressure decrease continued and was about 350 psig at 94 minutes and 140 psig at 102 minutes.
h{
6
l
. . . I u
Figure 3-8: "A" SG Start-up Level: 85 to 130 minutes The "A" SG level began decreasing at about 85 minutes reaching about 8 inches at 92-1/2 minutes and generally holding 8 to 10 inches until
, about 100 minutes when a positive sustained increase in level began, and continued until a level of about 193 inches was attained at 125 minutes. (One-hundred ninety inches on the startup level corresponds to approximately 50% on the . operate range.) _
Figure 3-9: "A" and "B" SG Start-up Level: 85 to 130 minutes This figure presents the start-up level for the "A" and "B" SG's for comparison. The "B" SG 1evel generally decreased from about 72 minutes to about 92 minutes when it increased from about 67 inches to i 112 inches between 92 minutes and about 94 minutes. Between 94 and 102 minutes, the level decreased to about 92 inches and remained at i that value until after 130 minutes.
U Figure 3-10: Makeup Tank (MUT) Level: 85 to 130 minutes ht about 85 minutes, the MUT level began increasing at a rate of about
- 0.33 in/ min, corresponding to 10.2 gpm. At about 91 minutes, the MUT level rate increased to about 2.2 inches / minute, then began to decrease at about 1.3 inches / minute at about 95-1/2 minutes and continued decreasing until about 113-1/2 minutes, where it began to
< -ise again.
I -
Figure 3-11: Source Range Monitor (SRM) Signal h
The SRM response before about 94 minutes was increasing. At about 94 minutes, the slope of the signal became slightly negative and stayed i
negative until the "A" RC pumps were tripped at about 101 minutes. At j this point, the SRM and intermediate range monitor (IRM) signals abruptly increased and then abruptly decreased.
j, - S.
s k
7 l
l I .
Figure 3-12: Letdown Cooler and RCS Cold Leg Temperatures The letdown cooler outlet temperature is an index of makeup and letdown flow. The letdown cooler outlet te'mperature is taken from a control room multi-point recorder, chart identification SC-0043.
The multi-point recorder also provides the RCS cold leg temperature
. used to infer time. This figure, covering the period from about 05:15 h through 07:30 a.m., shows major temperature inflection points
~
I occurring at about 05:27, 05:33, 05:39, and 06:03 a.m.
t Figure 3-13: HPI, Makeup, and Letdown Systems
[> ~
This figure, a schematic of the RCS, HPI, makeup and letdown systems, is used as reference for interfaces among the systems and gives general information about the piping, pumps, valves, and tanks of the systems. ,
L
-w 1
b l
2\ .J W
8
. . _ _ _ . _ _ _ . , _ . - . , _ _ _ w. _ , _ . . _ . _ _ _ . . . . , _ - _ - . , _ . . _ _ . _ _ . . - . - . _ . - ~ . .
l l
l
,n,
[ FIGURE 3-1 RCS TEMPERATURES L}
/ 60 to 120 Minutes l
l u
2 --
TNI-2 LOSS 0F COOLRNT RCCIDENT 3/28/79 as#CTon coo 6 ANT Loof n louf 65ektol F - asacron soo6a=v Loor o tour sensews:
nsacron coolant Loor a Tiu sa cono-a snonya nsacion coo 6nur Loor a Tin steno onsnsos
'\
- $~ '
- u. .
8 -
w q - ~
- t. y R- .
t s=
M - i .
= Time Period of -Interest ' -
. I W $~ .
a F 6 g
~~
~- 'ag -
W g. N-Ji/~
(- -
s L 3- .
%s
- %s\.,
NN N
s l
, 8 . . .
1, 60 70 00 -
90 -
100 110 120 d - TIME RFTER TURSINE TRIP = S0-320 MINUTES s-
' Source: Figure 7 TDR-044 tt ,
J
.I
.. i 9
i l
FlGURE 3-2 l RCS TEMPERATURES I 85 to 130 Minutes 500 l'
.. .J a30 - [% [
560
..** t 1 - THA >
s H '
540 -
TlB
,/
s . s, ~ / ' .
l 6
m ~
%w
% vfl . 7 b Tu I4
- w Q,hw T
"} CB (
500 -
f; 'w<C !
- . . . - . .m y wn%
0.
P1w .
p 450 l
l 5 460=
440 -
b 420 L' '
Source: Reactimeter
!!. 400 ,' . , , ,
85 90 95 100 10 5 ;7 11 5 12 0 12 5 13 0 l
TIME (MINu fES) 10 s . _ . _ .
FItiURE 3-3 RCSPilESSURE 85 to 130 Minutes
.500 - , ,
! i Sounlce:Reactimeter f' Memory Trip Review -
Widei Range Stric Chart I
. ! l l
. , * ..t.
?
l 1300
>200
{
I. C b
3v 1000 -
- f. 's ';,
C_
y) ~300 0
1
~00 2 -
I N a
700 r
1 H
I 1
o00 -
r> V ll 500 , , , , , , , , ,
55 ao 35 10 0 10 5 11 0 11 5 12 0 12 5 13 0 l2
, TIME (MINUTES) l 11
=
FIGURE 3-4 PRESSURIZER LEVEL p, g 85 to 130 Minutes
>' 400 - -.
ya.>u .s. > ..; ,
l
.\
.. r.
l 5uO - '
. :"ya,., ,.w- 2 <.uw'.,..
l 1. .
%o . r ,w =v. ,,,,,
320, f ,, 2 3 0 -
PW LJ i .I 1OZ 240 -
- v .
LJ s
,*d
,j \ .:. .
d g .~ 200 -
- l QC L.J s N
% 16 0 1 D A
M LJ QC Q.
H 12 0 -
1 a
t 30 H
40 -
ll/ ,,
q
~' / - . .
Source: Reactimeter 0 , , . , , , , , , 9
.5 90 35 '
10 0 105 11 0 11 5 12 0 12 5 13 0 TIME (MINUTES) 12
FIGUkE 3-5 RC FLOW, "A" LOOP 85 to 130 Minutes I' 45 Soudce: Reackimeter 1
43 - Y' -
35 - ,
[ l l C i E T "O l CL j,1 2 '
hi i a
3:
lIl ,
, J o -
.l L
J 25 -
~
l . >-
x l <
1 20 -
4 %
l CL i CL O
O g
J 15 -
4 <C j 10 l
l l ,
p 5
./,
8 h I h I h h h 0 35 90 95 10 0 10 5 11 0 11 5 12 0 12 5 130 TIME (MINUTES) 13 5
FIGURE 3-6 STEAf1 PRESSURE, "A" SG 85 to 130 Minutes
's 1100 i
Sourte: Reac<;imeter o.
300 I. an - k .x te \
k tv M
u 700 x\ .
3'.
( _ . ,~.. . ~ _ -. _ _
t v ,- o00 (A
! W
- l Q::
h1
< 500 C
w
. 400-300 r
200 M. s 10 0 , . 4 i < 6 i * *
~*
35 90 95 10 0 10 5 11 0 11 5 12 0 12 5 13 0 1
TIME (MINUTES) s -
14
e _ _ .
FI'GU,RE 3-7 STEAM PRESSURE, "A" AND "B" SGs 85 to 130 Minutes r_s
[t 110 0 Sour e: Peactkmeter 1000 - . , -
{
L N /
\ '
Lm ,
O ,
~
700 ,
m i .\ ,
- u. CL .
w e
,R; .
- t .
600 - -
_A s~~
f~d "e
g kJ s t' %
- O 1 300 i o e M 5 8
g 400 3
- "B" SG s
e
',i 300 c S 9
t 200
"#4eea,
- ., ,ewuungg ygg,,,,
Lp~ 'g 4*, ,,, swouwsma
'v 10 0 , . , , .. . . 6 35 90 95 10 0 10 5 11 0 11 5 12 0 12 5 13 0 TIME (MINUTES) is 15
FIGl1RE 3-8 "A" SG STARTUP LEVEL 85 to 130 Minutes
'- 250 Source: Reactim ater 225 -
200 17 5 -
m (A
w
<l z O 15 0 j Z .
1O s .J '
< ..~> 12 5 -
w -
l J *
\Dm '
10 0 j.
<C
, O
! M b 75
(
La 50 25 ,
s.( T 4
0 , , , , , , , , , .'
35 90 95 10 0 10 5 11 0 11 5 120 12 5 130 l TIME (MINUTES) y - _ ._ .
16
- n. _
FIGURE 3-9 "A" AND "B" SGs STARTUP LEVEL 35 to 130 Minutes 250 Source: Reactimeter 225 -
1
- 1. er 200 17 5 -
3_
, m L yM I 15 0 / '
o .
s Z
- d w_
SW 12 5 - /
"B" SG
. oM
=
... ' ,5k'
/
,....j' 10 0 ,-
',,*+....
e l . . . . . . . . .. . .. ... .... . . . . . . .
M '
.s
' " 9 av ,
a S 75 - ' *
+
, . f 50 / "A" SG 3
I h h h h 55 90 95 10 0 105 11 0 11 5 12 0 12 5 13 0 TIME (MINUTES) 17
~
FIGURE 3-10 MU TANK LEVEL 85 t'o 130 tiinutes Source : Reactimeter r
.o.
b i
- 32
\ N n 73 -
V)
W b /
L Z /
O 74 -
[. J w
4 N
,, ' . I 70 -
4 M -
I Z H I
. . _ l oo
. D w L_
o M M 4
b2 62 - '
! \/
r l 55 L
54 i ',
ll 50 , , . . . . , , . .
L 35 90 95 10 0 10 5 11 0 11 5 120 12 5 13 C TIME (MINUTES) 18
6
^s .,,4 q ,.c.
.g 4 p _ .
- g. .y
- e .
- 1.0 -5 . '
n T ~ a -
a I j i. M l" . l s
.9 be.
.. g m u, - _ .
,=. .
.'8 . f_ ,, N G .
........,f+n., . . . . .
.. 4...
9 sq g . g._.
....__g
- . _ ; o...
l ..C ge
'~2^
e
,9
, w
_. g.. . . . . . . . . . - .. _. -
. ,, .e . .. .. ..,..
.. g 4....-..._
=
y 3
- % 4 . . . . 8 . -
- e
% %. . , . ... . j. _ _ . 1.e 4 o s
). _ o, ... . %v, g... .. _ ... _ .
g.
.L %- .i p. _ ,
A ....-
. . . . +. _ - . _ .
- c. .,
.,i--
, [.
L Q... . . ,
f... ._. . ......__
.._ _.. g
- g. g g._ i (. ,g.
- h. _
7 *
. _ . . . . , 3. ,
..==.. l
- d. - T ,_
if -
\T p 'f g {',
. . . . _ . ...,.3.. ,' . .. g.
. . . g. ..__. ._. . . . _ . .._ .. r
(.. s ,
j . . .
._. . .T .. ..
i . . . # .
.. _ ,. s u ) (~a p ,
n M i es
'
- 4 i iT._ .
q., - _ . , m. .. qy.
i m s n.
s,, ; - -
ud.r.c d,e i
I : . , .
-_,6 i i_
I.'9
.. .._..._. y. _ g._
.. . _ _1
_. , . , a_ '. .
- g. ._.a,..
I
_) -- *
,. f.. . - ,.. . , , . . , , .
i _
_. -- =
.._..____u.
- V- --*. ;si.+_r .j '.3, _ .
- r. .. _
,. L _
_. .=._ .t . .
4
._-1, * * * *, , 3 p m n.
l . ._ 1
_ ,r. -- .._ .
- u. v 2.a __
- ~
._.. s,, -e n-- As
. . _.. c, ,
._.1 .. l,_
_mi-. . ~. - -
t 8
. 4 ._
i_._-...... . _ !.
. =_ ,
._..._ .i_. . . . .
_. - i,....__
. .._i._._
. ; . . = . w m
i
- _ _ .. - R_,__
FIGURE 3-11 r .l . ._
l F_ SOURCE RANGE MONITOR (SRM) SIGNAL c
t . , . .
. _# + . . n ,_ ,
Source: Recorder Chart
- n. ,.
s _. = [..~
. . . . . . _ . . _ . . _ , _ a..!._. -
. ~
s
.. . . . . . , _.._. .__ . . . . . __ . . _ __ _ . .. ..._. 4. 9 _ . . . . .
i, , ,
m N,, _ . . . .
._i L n..
m.
,. . i.
.. 4
. . I 8
,.._3
. _ I.
_.4 k . .' , , .
c.
, mi
. u' _
(. .
- . * * * * * = * . , ,
, - . . .geP*
1 I
t ..
n 1 - =
..e 3
- 6. . . .
P g y .=. ". . . .
. . . . . ._ :. 1
. . .. - . -_ -j - _
O , ,-
4 . _. -. . .
4
. e. . \ . ._ . . . . - . . . . _ . ,".... .. . _, , . f .; _
4 g ._ _ . _ . ; .. _
- .a 1-:. _;-t .- . . _. ,-+
. . '.p.-- A " - 19 '.'
.g. e.___... ...
n-- . . p_. . _ g. .. .
1..ma e_m .. ,
.. x-;> q q-
_____-;__3-m
I e
1
\n O
FIGURE 3-12 LETOOWN COOLER INLET AND CUTLET TDiPERATURES
- s. CLCCX TIME h %
- a 4
,g n .8 'i < s m c..-www e. ,.
F .. D= n. .,m qq v., - t2.2 r .s = c. v..s=wwrrrms=.n
.4. la << a. i i i,iti, n i i t s t i t i , , i i i t t i i i i mi n.. M, c,r. - , . ,
n,,.,.,.
. . . , ,_ .. . i s t.i v.t . i i t v i i t t t i t i t # 11 1 11 i n s i e t t i e i r i t i t i t t i i ,
g .
. . c ,. ,
. .> ..w. .
Y * ,
., i, - * =
] .. s ,
)
. 5 . . .'..
y
. , 4 ,
o.
. .. .. o . . .
.. g g . . g .
, . . e.
. d.E'W , ,c. ,
s = .
g
. , , . ,a 4 .. 3 .. s.
e .e s 4 .y e ,. .. %'_ , . .
- M *a e 8 ..***l**=*
.ed"***+ * *..
gg m .' . ].,p
.- t .
. . . w. ..
... .-2 , .. 17r w ' .
. ' d , .
. . D' 4 f F. e - g .
. .. . . g
. 4 5
a.
s . .. ... - =.
I i
s m
- m. . '9 .
..s .
, ..~
.I i i W~& m a g ! ..
. - - . ., "-"-""T
'. 5 e i 1v, - . s .L
.w.
,i,.,.,.,.. ..
6
- 3. _7. s ,
.. , ,. ,. .. r-W. .....- J..2 .. ...
, . . . . . , . M.1 . . .
W 6 . . .!= -
..~. .
i
.c. ,.#
j 4 % . .- 2 .r. . .
, .. . . .4 2.... ..c... .. . _. . ,
. . . a . .. s ,. . .... . . . . .w...,.
. - - . . , .;. . .*m. ..
.,. .w.., . . , . , . . .a ..
I_-]. 7 54....
. , +. .
. w . # ., .. . 1,.
- a. . .......
. , e.. .. :. . .r. . .
. s:- . . .
s.
s . . . .
,. < . - .~ .
i .. . .- - . . . .
.s a
E
. c. . ,t 1 -
g s...... . ... ., ,. .
. .. . - 4 )
- w. - , .
. m , m , .,
. .. .... r . .
. . . . . .. . . , ,na w . . . . . .. .
...q -...
1 i
P '--. *
- L ' ' ["" /.1'.~, ,
4 .
- u. . . 4 . . .
1 . . . - .
.. . s . - . . . , .
.. , .. . . . . . d. .. g
. ... > r..; 3. .. , * **
- 4
. . . . .. ~.
. ...... . . , . ..7 ., . . . ' . Source: Multipoint Recorder l )'
1
..m I.D. # SC.0043
({ +
. .. . . w
..4 Ja ..- .. . .. ..
1
.. .- a .
...as.. W 5
.==..b . . .
J 5 "J - L 7 ' . . 'a '. r 5 3 1
.c.. .
. r . .
.*4
, 9 I- 2 E * " r; gi L 1 .t '2 S r.
5 Ki r.
"E'1 I T I i 1 3 1.K . _ , . I
- . '. 'l T T . E i 1. E E E : . EI E Z . . ' I E . . ~ ' * '
e Je . ~ .. .
...~.
.- m .~ m . . . ,, . . .
A ^
m .. .
4 l = g , g . g g g g g g .. g gg ,g 3
. .A
._ a m .,
a _
O .,.
l 3 gu 1 g95 y g 3 } g ,. g g g, g g g g
. . . ' - g,T 1 , F ,5 g 5 ,J g g C O. O ..... O ~O O..O .O2 O O'..O O O. O O O Or-O O O u........... -
O.'..O O O C 'O 'O O O .O O O. ,O O (
a
~
~ . . . . . . . ....:.... ; . i ~.,. . .
~
. ~ . . . . . . . . .-
g . -*.-- - w .:..- -* - .. . ,. --
e, - .
?
g g g OM N . . ... .
' . - . q p. . .6.c .
- 4. 3.,7 g.,. gy.s.,-
- . 20
.g ,. . - %
x=s -e b,_ :,_ .. W ' ..3 . 1_ n m ..g. % n m >--- ~ - - . .
l F ~ FIGURE 3-13 l
l -
HPI,!!AKEUP Af4D LETDOWN SYSTEllS y
b Source: 'NSAC-1
, (July 1979)
F unne voand unne w ano ourmcahon guns.cm.on La oo.n if"***'*' ""d*ng
- noen I tmoo.n oemeneranaere falsers onfece j coWars
~~Ch il I
. ES I H @ l Normal letoown I
uu.ve y x e i s.
{ "t> E RCP i
l 1O coment
- 1 O
Chem cal sead RCP l T system - T
- esa Molouo tanks aamtaan return seel I S JJ.anc oeoorahng coders 0 systern return R' Juminerauters heter ES RCP 1 1
_ g , ." h-
_ . , uamevo l
U RCP NRCP 16A I -
28 2A 3 hnera ES Safety insection ES Safety insection awsT
,)
up tann _
(y ,
-r Seal suppfy
- q g QV5-Ay d-nnere -
g 4c xn h g o as LW v8 hu :i
_ L een,
( g" VT. b3,, A -Js= dip 1iiE2+,,'"'"" --
-, ,, y ~ (Mij-V17A&B) l HPIC 16C I Legend
/ ES Sefesyinsecaon l
@ Motor operaswd waswo ES h 100 I
@ Ast operased wasve j ] Seemyinsectinn l
. Re or,-
TE Temperature seement -j (N Reassornuddwig FE Flow etemerW V ES Opemeo my ESF esonas Note: HPI Train "A" consists 'of HPI pump "A" taking suction from the BWST through VS-A and discharging to RCS Loop "B" through ES valves 16 A&B.
a 3 HPI Train "C" consists of HPI pump "C" taking suction from the BWST through V5-B and discharging to RCS Loop "A" through ES valves 16 C&D, 21
ms 4.0 SCENARIO 1: HPI ONLY Scenario 1 assumes that the RCS cooling that began at 94 minutes after turbine trip was caused only by HPI. Two analyses were performed fo'r this scenario using simple computer codes. These analyses assessed the MUT leve:1 response and the RCS heat balance. The MUT and energy balance models are summarized in Appendices 1 and 2, respectively. The findings of each
~
analysis are compared with the plant data to ascertain whether or not this scenario is consistent with the observed plant response.
4.1 MUT Level Response 6
The purpose of this analysis was to investigate whether or not full activation of HPI, in the period from 94 to 100 minutes, is consistent with the observed MUT level response.
> 4.1.1 Calculations *
! This calculation included several ' assumptions to describe valve operation and letdown flow. When these assumptions were made, the calculational results agreed well with the actual MUT level response (refer to Figure 4-1). These assumptions ~ included:
- 1) The initial MUT pressure at 91 minutes is 20.2 psig at an initial level of 76.2 inches. (The MUT. pressure model is a Boyles' Law model: PV1 i = P22 Y *) -
t 2) The initial makeup flow is set to 40 gpm, a minimum value based on RC pump seal flow.
I' 3) DH-V-5A (Figure 3-13) has been opened before 95-1/2 minutes, allcwing the "A" HPI pump to draw from both the su borated water stora'ge tank (BWST) and the MUT.
\_/
w 22
L L -
- 4) HPI Train B ("C" HRI Pump) is manually actuated at 94 minutes. Little effect of HPI is expected on the MUT level
_ because it was assumed that the HPI Pump "C" is not aligned .
to the MUT.*
?
k.
! 5) "C" HPI pump recirculation is not open to the MUT.
- 6) The "A" HPI pump recirculation is open to the MUT.
- 7) The "A" HPI pump is brought to full flow at 95-1/2 minutes.
(
- 8) Maximum letdow'n occurs at about 95-1/2 minutes. The j maximum letdown flow is estimated to be 220 gpm. The maximum letdown flow continues from 95-1/2 minutes to 97 ,
minutes when it is throttled to about 30 gpm at 101 ,
h minutes. The assumed letdown flow response is shown in
~, Figure 4-2. These variations reflect variations in letdown J cooler response.
\
l 9) HPI flow from Pump "C" is terminated at about 101-3/4 minutes.
P 10) The flow from HPI Pump "A" is terminated at 101-3/4 minutes with makeup flow continuing at about 100 gpm.
L
- 11) DH-V-5A is closed at 101-3/4 minutes.
>=
r
- For this scenario, the "C" HPI pung and HPI valves 16 C and D were actuated at 94 minutes. From about 91 to 95-1/2 minutes, j the MUT level is rising,at 'an average rate of about 70 gpm. If j the HPI flow were from the "A" HPI pump (through valves 16 A&B), the MUT level response would have been quite different and would have resulted in an immediate change in direction and
! slope. It should be noted that the actual time of "C" pump If actuation would not be detected by MUT level response. Rather, v the 94 minute time was chosen based the observed RCS cooldown (refer to Figure 3-2).
I 23
[- The maximum letdown flow used in the analysis (220 gpm, Figure 4-2) is larger than the design value. Examination of the MUT
, level response in two periods earlier in the transient, from -
about four to six minutes, and from 10 to 12 minutes, shows that letdown flows in excess of the design value were observed.
During the first period, the MUT level rate of change corresponds to about 236 gpm and during the second period, r about 190 gpm. These MUT level responses are shown in Figures 4-Sa and 4-5b.
The letdown cooler outlet temperature response was also evaluated to obtain an indication of letdown flow. The results of this evaluation are presented in Figure' 4-6 which also shows 1 for comparison the letdown flow estimated by NSAC3 modified L
to account for two letdown coolers in operation. The MUT level ,
response is also shown. Figure 4-6 also snows the hourly .
- 1etdown flow points obtained from the ' computer log. The B&W 7 analysis estimates a peak letdown flow of 220 gpm versus 190 gpm for the NSAC3 analysis. These results also suggest that
. letdown flows in excess of design values could possibly have occurred.
4.1.2 Results The analysis results that best correlate with the MUT data
- correspond to the following HPI flows:
o The "C" HPI pump flow ranges from about 520 gpm to 540 gpm from 94 to 101-3/4 minutes and the "A" HPI pump flow ranges l from about 460 to 480 gpm from 95-1/2 to 101-3/4 minutes h (Figure 4-3).
P' o The makeup flow (Figure 4-4) initially is assumed to be about l
40 gym, decreasing to about 10 gpm when the "A" HPI flow is L q increased, and then assumed to be about 100 gpm after the "A"
~
HPI is terminated.
l L
24 V _ _ _ _ _ _ . __. . ._ . -_-- - _ -_- - - - .- -- -- -.
M q 4.2 RCS Heat Balance l l
I g The purpose of this analysis was to investigate whether or n.ot the
{ assumed HPI flows would yield the expected RCS heat balance.
4.2.1 Calculations This analysis assumes that no EFW was injected to the "A" steam t
generator between 94 and 100 minutes. The general method of analysis was to assume an HPI flow rate to match RCS temperatures (Figure 3-2). The results of the energy balance
, model analysis indicate that the approximate HPI flow shown in the following table can produce the observed RCS cooling rates.
HPI Flow from Energy Balance Model Analysis -
s HPI Pump Time, Min. HPI Flow, % Canacity_
(3 C 94 - 101 100
[' ,
A 94 1/2 70
. A 95-1/2 - 101 100 l' 4.2.2 Results The HPI flows given above' produce cooling rates that agree with the observed RCS temperatures from 94 to 101 minutes, as shown 3 in Figure 4-7. It should be noted that while good agreement was obtained with the observed plant cooldown over the period, f an inconsistency does exist between the results of the MUT model- and the energy balance model from 94 to 95-1/2 minutes.
During this time, the energy balance model predicts that an HPI flow rate equivalent to approximately 1.7 HPI pump flow would be required to match the observed cooldown rate. However, the a MUT model obtains a flow of approximately 1.1 HPI pumps. It is believed that further refinement of the MUT analysis could gn result in flows consistent with those obtained from the energy
> balance model. However, based on analyses performed to date, s
25 P . . . - _ _ _ _ _ _ _ . _ . - . - _ _ . _ . - _ _ _ _ _ _ _ _ _ _ _
r m
it appears that, at least for this time period, some additional cooling is being provided by the EFW system.
4.3 Comparison with Plant Data This section examines how the HPI only scenario compares with other plant responsas. These comparisons appear in the following order:
BWST volume, RCS inventory, RC flow nesponse, and SRM signals.
4.3.1 BWST Volume The volume injected from the BWST calcul.ated with the models is consistent with that observed during the' accident. Based on I
the pre-accident BWST level and the TMI-2 alarm. printer output b
at 7:30:30, an outflow of approximately 15,000 gallons occurred ,
from the BWST. Further, there are two known periods of HP.I injection starting at about two minutes and another at about
- ] 150 minutes. The two HPI actuations account for about 10,000 3 gallons of the outflow. Thus, an additional 5,000 gallons of
, outflow occurred.
The energy balance calculations indicate ,that about 6,200 gallons would need to be injected to obtain the observed
-] cooldown in the RCS. Based on the MUT calculations, an outflow from the MUT of 1,250 gallons occurred between 94 and 100 y minutes. Thus, the BWST would have to provide 4,950 gallons.
This is consictant with the observed outflow from the BWST.
l 4.3.2 RCS Inventory
r This HPI only scenario results in increased itquid mass in the RC system. Other investigations3,7,10 have not predicted 6 significant HPI addition in the 94-100 minute period. However, the accuracy of these inventory predictions, relative to the
,f) TMI-2 accident, is unknown.
I 26 L
L .
3.; 1 The 5,000 gallons of additional inventory does not appear consistent with the observed increase in reactor outlet temperature at about 112' minutes post turbine trip. However, ,
the energy balance model predicts a ' decrease in RCS void' fraction of only about 5.5% from 94 to 100 minutes. It is possible that the assumed HPI flow resulted in an undetectable redistribution of voids, or if quid, in the RCS. A more detailed simulation of the transient would be necessary to confirm whether or not a redistribution of voids occurred.
4.3.3 RC Flow In the period from 72 to 94 minutes, the RC flow signal from the "A" loop (Figure 3-5) is generally decreasing. At 94 minutes, this flow signal shows e abrupt increase in flow of about 10 percent (from about 34 million 1b/hr to about 38 '
- y. million 1b/hr flow). After this increase, the fridicated flo"w begins to decrease faster than the rate prior to 94 minutes.
l This decrease continues until the RC pumps are tripped at about f,bs 101 minutes.
The initial indicated RC flow r1tsponse at 94 minutes is consistent with this scenario. .The longer term flow response, j however, appears to be more consistent with limited HPI and/or external cooling. If full HPI were actuated from 94 to 101
,. minutes, the increase in inventory would be expected to provide a longer term reduction in the void fraction in the "A" loop, and thereby, promote a relatively constant or increasing RC l flow signal in loop "A."
The RCS flow meter response, however, is not totally incunsistent with the HPI only scenario. The addition of cold s HPI will result in local effects which could have caused the RCS flow meter response to decrease. These local effects include: decreased void fraction in the downcomer and cold leg
~j pump discharge piping due to steam condensation on the cold HPI, flashing at the RC pump inlet leading to further v
27 5
i . - _ . . _ - - - _ _ _ . _ _ _ . _ _ _ _ _ _ _ _._m._ _ _. ._.
- 7 degradation of RC pump performance, and possibly steam flow via the reactor vessel vent valves. A more detailed system simulation would be required to ascertain whether these local ,
effects were the cause of the RCS flow meter response.
4.3.4 SRM Signals
- Starting at about 94 minutes, the SRM signal (Figure 3-11) begins a downward trend. This trend appears to be consistent with the RCS cooling in that colder water and probably fewer voids are being pumped into the reactor vessel downcomer by the "A" RC pumps. The major response of the SRM at 101 minutes is
^
probably that associated with the "A" RC pump trip and
! separation of steam and water (water falling to lower parts of the loop and steam rising to higher parts of the loop).
- s. 4.4 Scenario 1: Sumary and Conclusions
~
!D As discussed above, full actuation of the HPI system during the time period of 95-1/2 to 100 minutes is consistent with the MUT level
'response and the cooldown rate observed during the accident. Between 94 and 95-1/2 minutes, a conbination of both EFW and HPI (Scenario 3) would seem to be required to produce the observed'cooldown. Further, in examining the effect of an HPI only scenario relative to other observed system responses, it appears that this scenario will not
[ directly correlate with either the expected RCS inventory or the measured RCS flow response. However, it is possible that the local .
effects caused by the HPI addition could have produced the responses
[ observed. Analyses of these local effects were beyond the scope of this report. Therefore, based on the analyses which have been performed, this scenario is a possible but not the probable cause of the observed RCS cooldown.
i,/
wm 28
i
(
~
FIGURE 4-1 SCENARIO 1: MAKEUP TANK LEVEL RESPONSE MODEL VS. ACTUAL 90 Actual ilU ilodel Data I
f 56 -
u I
U .
Z 32 -
_J -
W y
W J
t M 73 -
Z l
H -
k P
, 2 74 -
6 70 , . . . . .
30- 9 2' _ 94 96 98 10 0 10 2 TIME (MINUTES) e J~
s s j -
I 29 k
t
h n.
FIGURE 4-2 SCENARIO 1: ASSUMED LETDOWN FLOW 300 ,
250 -
L ^
1 .
c.
O 200 -
- v
,.-w 1
x.~ ' '
O .
d 150 -
i Z 3
o
- O
[ W 10 0 -
l r '
50 -
O
] 30 p 92 94 96 98 10 0 10 2 l T,l ME (MINUTES) l 5
l (a\.
30
+ . -
m.
e FIGURE 4-3 SCENARIO 1: HPI FLOW 300 1, 250 - 16A s* Train y 16s
> 16c a o tenen. tv.) f.++""*"**...........""""""*"a="")
'A' Train m
- 3 200 - .
c.
O v
(/)
s 3 15 0 -
0 J
6 p -
Q. ~
I joo .
L 50 -
0 ' ' 4 '
e' 9O ' 92 94 96 98 10 O 102f TIME (MINUTES) l l
v 31
FIGURE 4-4 SCENARIO 1: MU FLOW 200
, 15 0 -
m JE .s Q
O ~
. v
~
- 3 .
m...- o J 10 0 - -
r L f
- C L D LJ M ,
2 50 -
i i 0. , , ,
[. 90 32 94 96 98 10 0 10 :
TIME (MINUTES) l I
i t
s I
i 4
32
m ..
FIGURE 4-Sa SCENARIO 1: MU TANK LEVEL RESPONSE AND ESTIfMTED LETOOWN
~
4 to 6 flinutes 80 7n -
f-70 -
6 5
L 74 - .
2 .
- u.
Q) m
> 70 -
e J
'.4 og .
I Z h aate in./ min.
66 - i 236 gpm e n ba . ,,.
2 f 1
[ 64 -
62 -
l 60 ,
4 5 6 TIME (M1NUTES) l
\
. .m
~.s .
P
.33
W FIGURE 4-5b SCENARIO 1: MU TANK LEVEL RESPONSE AND ESTIMATED LETDOWN FLOW f
4 10 to 12 Minutes 80 6 -
78 -
I. .
70 - 48 p f %
5 u- -
w
- 3 Z O 73 7Q
=
J -
( b8 -
> r .
, d6 -
e., .
1
- f. 60 '
L 10 33 ,,,
( TIME (MINUTES)
J.
l h(_.:
I 6
34 ll A -.
~ O 9
i l
- g m m e~, Q r - - M- - - - - - - - - - 5 .
i l % ~
B i o L g
- - o l _
> =
1 l m l 0 .
8 L_ 2_x_ _ _ _ _ _ _ _ _ _ _,
, r - - - %- - - - - - - g
- 8W ggw .* / .
gg L_ . ___,
( il 95 .} '
~
i r - -- - -r '-- - - - - - - - - ".
a
$w= i
( ;$9 l Ej I g 3s I . s (p~- )
I ti
3 a2-
- C.3 l O' -
S 3-
"8- I i\ %
^
g l 2 id %
i l [ - -
s I
4 g l i._ _ _ _ _ _ _ _ _ _y_ .
c _ _ _,ec.. ________ _; -
e
[
i k N g
E -
I
- _-.J
- z. ; g- -
a c__s_____________,
I 5
__________.s -
g p .. ..
_. p
- z. E I
I I E s2 E
A R R 3 8 3 g s seuouT 'TeAe-) Musi m a ~gs a ~ HMa s e a 8 s s ea =
g - - - - __ _ _____ __ _ ___ _ _. ___ __ __ __ _ N * @ N GQ@
\.
y c g y==--- c v ,
~- " I
, ~ -
rO.,
(.. 's - -
FIGU.tf 4-7 SCENARIO 1: C0ffARISGsl 0F RC TfiFfRANRES, [?t[RGY BALAftCE ff2EL VS. ACTUAL 565 i -
I
. ...d. .".-
Assumptions: *
~-- ~"
Tcold A -~~ 1.7191 ("C") at 94 minutes 10 seconds *
" ptg;gai;
-,/ l88 __ 2191 ("A&C") at 95 minutes 30 seconds untti 101 minutes 560
, a s a r i hatT A no Elv
- , # g.gg
'i y./
~~ 0
- p ,;; % ,
100 letdoun That 8 Steam release frun PORV
,. .. p-
% N, JF g " s.
555 . m , ,..
sb .
_ . _ sg,. saa m ..#' N k \
w - -
I m q@ ~.
- ~
^ Tcold B (Le N} ,
550 - '.M.:.p - -
m -
- g. -
' +e'3f.
W '
C % ~.
'....N.'..
D %, ' -
p yc- ~ ~ . . .
4 545 4.+ ; .
g.
y M I; .
r- .
i n: -
- s. .
% %% . . . ?' -
y 540 "Model Results d W. j (
8c y 4g, m ,
3
- .w;g.4_'y
's gy,,
i N %i -
'\ 4hm ,
535 '
NJ,tys .
g%. N?g '
N s y,,c .
N
% hm 530 x 9- m '.;ggggen.
w aw
.Vid::..)
s-u-- ' ~ . mu RfACTilETER DATA ieiei...............
O b. M 525 -
s- i a ' e's59.P
. x:
91 92 93 94 95 96 97 98 . 99 100 101 102 103 :sl::3 ty* . s. '
TIME (MINUTES) .:*
.%4., ,.
0
5.0 SCENARIO 2: EFW ONLY This scenario postulates that the cooling of the RC system that began at 94 minutes is due mainly to EFW. The analyses performed to investigata the feasibility'of this scenario included:
- Estimating heat removal and EFW flow requirements,
- - Comparing SG response at 94 minutes and at eight minuter after turbine trip,
- Performing steam generator response analyses with the AUX-II code.
The results of these analyses were then compared with other observed plant j responses to determine the viability of this scer,ario.
l 5.1 Estimated Heat Removal and EFW Flow Requirements The capacity required to remove decay heat and cool the NSS at abodt 4.4F/ minute from 94 to 101 minutes is well within 1.he capability of the EFW system with one EFW pump running. The amount of heat removal required, neglectidg primary heat, is estimated to be:
1 Decay Heat: 1.6 x 106 Btu / min.
1.6 x 106 Btu / min.
RC Water / Steam:
RC Pump Heat: 0.2 x 106 Btu / min. (4 MW assumed)
Total: 3.4 x 106 Btu / min.
From 94 to 100 minutes, the T SG startup level (Figure 3-8) has a near constant level at about eight inches which indicates that only steam exists in the tube bundle between the level taps. The nearly constant level in this period indicates that if EFW is acteted
> and providing the cooling, all the EFW flow is being svaporated without forming a water pool in the generator. If all the water were evaporated, the resulting EFW flow required to remove 3.4 x 106 Btu / min., would be about 360 gpm, assuming heating of 100F EFW to saturated steam. This flow, 360 gpe, is within the capacity of one 6
37 L
. . . ___ _ - - _ _ _ _ - = _ - _ __ _ _ _ _ _ _ _ _ _ -___
motor-driven EFW pump and within the cut-off flow (about 600 gpm) of the EFW cavitating venturi. Thus, there is no known system constraint which would preclude this scenario.
5.2 Comparison of Steam Generator Response at 94 Minutes and at Eight
- Minutes After Turbine Trip Comparisons of the RC temperature, steam pressure, and SG start-up level for the "A" OTSG at about eight minutes (EFW actuation into a l' " dry" pressurized OTSG) and at 94 minutes (also dry and pressurized) are similar. The effect of the EFW actuation at about eight minutes post trip and* the corre>ponding response at 94 minutes post trip are shown in Figures 5-1 and 5-2, respectively. The similarities between j these two actuations are:
i o EFW injection provides significant cooling without an increase in -
the SG 1evel.
3 o Steam pressure increases with EFW injection.
f l 0 The SG 1evel starts to increase only after the difference in e temperature between the RCS cold leg and the saturation temperature corresponding to the steam pressure has decreced to less than 10F.
l e Based on the similar responses in the SG at eight and 94 minutes, it
, can be inferred that EFW may have been delivered at 94 minutes.
However, the "A" SG pressure increase from 94 to 98 minutes is relatively small, especially when compared with the pressure increase that occurred at eight minutes. The steam pressure response in the 94 to 98 minute period indicates that a steam relief path, such as a l partially-opened atmospheric dump valve, apparently is relieving pressure at 94 minutes on the '"A" SG if EFW is being provided. It
[ should be noted that TDR-0442 indicates that during this general l
time frame, the atmospheric dump valve was being used to remove heat.
\ ..
l 38 I
l
.x, . , _ . _ . _ . . . . . _ _ . _ _ . . . . . _ - _ _ . _ . . _ _ _ _ _ . . . . _ . _ . . . _ . _ _ . _ . . . _ _ _ _ _ _ _ . . . - . _ . . _ _ . . _ _ .
l 1
Examination of the "A" SG pressure response from 89 to 94 minutes ,
^-
(Figure 3-6) further supports the position that a steam relief path i could have been open. The "A" SG pressure decreases in a manner -
indicative of a valve opening, closing, holding, and partly ope'ning again, while the RC system is heating and SG "A" is drying out. Thus, l an assuirption that the operator was manipulating the atmospheric dump valve is reasonable. Further examination of the possibility of a steam relief path was performed with the AUX-II code. These results are discussed below.
5.3 Steam Generator Response Analysis Using the AUX-II Code I"
The purpose of the AUX-II analysis'was to gain a better understanding I
of the steam generator response in the period from 94 to 100 minutes.
The analysis performed consisted of sensitivity studies to determine ~
steam flow requirements and the effect of various EFW flow rates on
> steam pressure and level, and to compare the sensitivity results with g the observed "A" steam ' generator pressure and level. The sensitivity studies were performed in ,the following sequence:
- a. Closed steam system - steam pressure and level response to veious EFW flow rates up to 600 gpm. ,
- b. Open steam valve - steam pressure and level response to various.
EFW flow rates with various steam valve flow rates.
6
- c. Benchmark model to the "A" steam generator response in the period around eight minutes and at 20 minutes post trip.
[
l d. Final approximation to response from 94 to 96 minutes.
I The steam generator / steam system model from the AUX-II code was used in a " stand alone" mode to investigate the "A" SG response. The main features of the SG model pertinent to the analyses include:
1 NJ l
l 39
Upper EFW injection spray and penetration, RCS hot leg temperature input, Steam pressure control,
- EFW input. '
1 5.3.1 Calculations
[ The major parameter inputs to the model included the RCS inlet temperature (THOT A) and the RCS "A" loop flow, both of which were obtained from the reactimeter data. The steam .
system model used was the 177-FA NSS version of the AUX-II code for the Davis-Besse i plant. Differences in steam if ne volume between TMI-2 and Davis-Besse are not significant for this purpose.
3 i
5.3.2 Results .
} .
The sensitivity studies indicated that the pressure response in j the closed steam system was characterized by rapidly increasing steam pressure, varying with EFW flow rate. (For example, the model predicts steam pressurization rates of about 60 psi / min.
at 100 gpm EFW flow, and about 120 psi / min. at 500 gpm flow rates). In addition, the model predicts increasing SG 1evel, contrary to observed response, with the higher EFW flow rates in a closed system. These results indicated that a steam valve had to be open for this scenario to be viable.
l The amount of initial steam flow was estimated from the "A" SG depressurization rate (Figure 4-6) from about 91 to 92 minutes and from 93 to 94 minutes. In both periods, the estimated steam flow is about 28 lb/sec. At this point in the analysis, a calibration of the model was performed with the RCS flow and temperature, and estimated EFW flow with the TMI-2 m
f c
u
data at eight minutes and 20 minutes. As a result of the rm calibration, increased heat transfer coefficients were implemented in the model .
The final case in this analysis was based on the following
[ assumptions with respect to the steam valve flow rates and the EFW flow rate: ,
- a. At 94 minutes, the steam valve flow is 28 lb/sec.
l L b. At about 96 minutes, the steam valve is positioned to produce a flow rate of 24 lb/sec.
- c. The EFW flow rate is assumed to be 200 gpm.
i The results of this run are shown in Figures 5-3 and 5-4. The A
SG pressure response (Figure 5-3) indicates reasonable agreement with the observed response. The SG 1evel response
- s. (Figure 5-4) also shows good trend agreement with the observed response, in that the level remains low and does not increase s as the steam pressure is increasing. It should be noted that a F
flow rate of 200 gpm is not sufficient to cause the ob',erved RCS cooldown rate (see Section 5.1). Thus, this analysis suggests that a combination of both HPI and EFW (Scenario 3) was the cause of the plant cooldown.
As a result of the findings with the steam generator model,
_ additional investigation of the TMI-2 steam pressure control valves was conducted. Review of available information indicates that TMI-2 has one steam atmospheric dump valve of about 3.2% steam flow capacity for each steam generator 3 A l steam flow capacity of 3.2% steam flow is about 50 lb/sec E which is more than adequate to relieve 200 gpm of EFW.
s As noted above, the "A" SG pressure and level response are controlled by the steam valve. The actual response of the "A"
( ,
SG pressure and level in the period from 94 to 100 minutes is lV characterized by evaporation of all entering EFW with
- 41
i relatively small pressure' changes. Assuming that atmospheric dump valve capacity is about 50 lb/sec, the EFW flow can be evaporated and produce little level or pressure change. . A steam flow rate of 50 lb/sec corresponds to a EFW flow rate of L 365 gpm. This agrees very closely with the estimated EFW flow (360 gpm - see Section 5-1) required to produce the observed cooling rate of 4.4F/ min from 94 to 100 minutes post trip.
Thus, the analysis performed cannot claarly rule out the viability of this coo 14ng mode.
5.4 Comparison with Plant Data As shown above, if the atmospheric dump valve is fully open from 94 to j 100 minutes, the assumption of an EFW only scenario would be consistent with the "A" SG response and the observed RCS cooldown -
rate. This section examines the impact on other observed plant
v; signals) of an EFW only cooling scenario to attempt to provide further insight on the viability of this cooling mode.
5.4.1 MUT Level f_ In section 4.1, an analysis was performed that indicated that the MUT level response could have been caused by full HPI activation. However, the MUT level response can also be characterized as just a reflection of the mismatch between j makeup flow and letdown - that is, the MUT level rises when U
letdown is less than makeup and falls when letdown is less than I makeup. With this characterization, the EFW only cooling scenario is consistent with the MUT level.
L F
q o
,V l
b 42
~_
1- _. --
5.4.2 BWST Volume 7 s.
Since this scenario does not use any HPI, it does not result in .
any depletion of the BWST volume. Therefore, this cooling scenario implies thac the 5000 gallons of BWST liquid was injected at some unknown time between 4:05 a.m. and 7:30 a.m.
5.4.3 RCS Inventory No calculations of RCS inventory were performed for this
. scenario. However, since the EFW only cooling scenario does not provide any additional mass to the RCS, other than makeup flow, the RCS inventory would be expected to continuously j decrease between 94 and 100 minutes. This decrease in L inventory is consistent with the results of other investigationsp,7,8,9 .
u.
5.4.4 RC Flow V -
The RC flow signal for the "A" loop (Figure 3-5) generaily decreases between 72 and 94 minutes. At 94 minutes, the flow P signal indicates a relatively abrupt increase in flow of about l 10%. After this increase, the indicated fl'ow continues to decrease until the RC pumps are tripped at 101 minutes.
The EFW only cooling scenario is believed to be consistent with .
k
! the RCS flow response. With the initiation of EFW at 94 minutes, steam will condense in the primary side of the steam generator tubes. The additional liquid created by the steam condensation will decrease the void fraction at the pump inlet L and will lead to improved pump performance (i.e. increased pump head and flow). The longer term response of the RC flow is L believed to be a reflection of the decreased RCS inventory that results over this time period.
1.j le lI 1
43
- \
5.4.5 SRM Signals The SRM signal (Figure 3-11) indicates a general downward trend between 94 and 100 minutes. This trend is believed to be consistent with the RCS cooldown in that colder water is being pumped into the reactor vessel downcomer by the "A" pumps. In addition, the EFW cooling is expected to result in a J redistribution of voids in the R".S. This condensation of steam I in the primary side of the steam generator will tend to locally decrease the void fraction in the cold regions of the RCS.
Thus, the SRM signals are consistent with the EFW only cooling scenario.
5.5 Scenario 2: Summary and Conclusions The analyses performed suggest that the EFW cooling only scenario is a -
y viable cooling scenario. There are no major inconsistencies betwedn
- observed and calculated system behavior if the atmospheric dump valve
~ '
is assumed fully opened. It should be noted that other investigations reported in References 6, 7 and 8 have assumed EFW at varying rates as l the cooling mechanism. The analysis described in Reference 6 assumed f 500 gpm EFW and indicated reasonable agreement with the RCS cooling from 94 to 100 minutes. -
l The analyses that have been performed show that the viability of this y approach is sensitive to the assumption of a steam relief path being open on the "A" SG. The analysis described in section 5.3.2 shows that if the steam relief path remained in its previously observed open position, insufficient EFW would have been provided to yield the observed RCS cooldown rate. Thus, a combination of EFW and HPI would be necessary (Scenario 3).
If it is assumed that the operator closed the steam path at 94 l
minutes, the analysis shows that EFW flow would have to be H significantly less than 100 gpm to obtain the observed "A" SG 1
lq
'V P
44 N . _ - - - - - . - . - . - . . - _ . - - - - . - - . - - . - . - . - - - . . . . - - . . - . - . -
pressure response. This would imply an HPI only cooling scenario (m, (Scenario 1). This assumption of a steam valve closure is reasonable in light of the response shown in Figure 3-6. The "A" SG pressure ,
rise at 94 minutes is similar to the rise occurring at 92 minutes that appears to be the result of a valve's closing.
The EFW only scenario is viable only if one assumes that at 94 minutes
- the operator both actuates EFW and fully opens the atmospheric dump valve. While this is possible, it is believed unlikely since no major 3, steam relief path is available between 100 and 125 minutes. Thus, it has been judged that the EFW only cooling scenario, while possible, probably did not occur at TMI-2.
{
t p
q L
b*
l f
l l
?
H 1
, ~7 m
I 45
- m. Figure 5-1 SG RESPONSE 8 TO 25 HINUTES 40 -
600 -
j--TCOLD "A" 590 36 - .
[s(
l 32 - 58 0 -
TCOLD - TSAT < 10F WHEN LEVEL STARTS TO INCREASE: 'l f
r.,o I
/
57 0 S .g 28 I m NOTE: SG "A" PRESSURE I
{ RESPONSE TREND lS ,
o 24 - p 560 . IN DI CATED BY TSAT RESPONSE 1 2.$ $
U
-d P "A" : 1030 PSIG a 20 - 550 -
- % ., , , %-e y c,
/
"h / TSAT " A" 16 - 540 - I .
I I i "A"-STARTUP 530 LEVEL 12 -
f ,%,.-..
[
P // \
8 -
520 yP"A"g700PSIG \\ #
/
4 SOURCE: REACTIMETER L 4 - 510 - TSAT, XENN AN l & XEYES O- 500 I I I i 1 i i i i
- 8 10 12 14 16 18 20 22 24 26 (h
's Time, minutes L
46 l
i
, - , - - - , - - - - - , - , - - - . ...,n_- - . - - - - , , , _ , , , - - - , . n , , _ - , - - - , - - - - . . , - - - . - , - - - - , - . ,
l.
<~,.
L- Figure 5-2 , l SCENARIO 2
[
l
[ SG RESPONSE 90 TO 120 MINUTES L
L SOURCE: REACTlHETER I
560 -
TSAT, KEENAN & KEYES 200 550 -
- ( -
s-j g TCOLD-TSAT < 10F WHEN LEVEL STARTS INCREASE lT \
j j 160 o" 5% g /
0
.a.
5 3 .
\
g P " A" ::: 870 ,
[
g PSIG
_ g D 120 E. 530 -
a 8 \ j N .
e k
-./ \
/
P " A" ~
~ 83 %
80 520 -
p3 q N L '
TSAT "A" N
ks % 510 -
NOTE:
SG "A" PRESSURE RESPONSE
\ N TREND IS INDICATED SY TSAT RESPONSE
" A" STARTUP LEVEL \
6 s - \
1 0
fn
- N.
l l e I e e n )
d
- 92 96 100 104 108 112 116 120 Time After Turnine Trip, minutes 11' ~,
l\ )
.1 47 l
,,,_,,,,m,,-,,---,,,.---r,.,--y_----,,-y,., , , , - , - - - - , - - - -
V Fi~gure 5-3 .
SCENARIO 2:
AUX ll NODEL RESULTS "A" SG PRESSURE VS. TIME A = ACTUAL DATA B = CALCULATED RESPONSE ,
1
.i 880 5
860 /
E.7 g [ [
- i /
U A
! 840 l
820 e
800 0 30 60 90 1 20 1 50 Time,-seconds -
n
, T, = 94 MINUTES POST TRIP ic] .
=
48
l l . .
L Figure 5-4 SCENARIO 2:
AUX li H0 DEL RESULTS "A" SG LEVEL VS TlHE u
A = ACTUAL DATA 8 = CALCULATED RESPONSE
- 2. 50 u..
! 2.00 e i
- m. -
W
-. . - 1.50
., n L
h .
a i
t
= 1.00 m
A 0.50 w.
0.00 0 30 60 N I20 IN Time, seconds T, = 94 MINUTES POST TRIP
- '3 v
i I 49 i
i d
1
( . e e L .
6.0 SCENARIO 3: COMBINATION OF EFW AND HPI r,
This scenario postulates that the observed RCS cooling from 94 to 100 ,
minutes (and beyond) results from a combination of EFW flow and HPI. The studies supporting this scenario principally involve interpretating available data and addressing the inconsistencies found in the other scenarios (Section 4.0 and 5.0). .
~
6.1 Observations from Other Scenarios m
6.1.1 Scenario 1: HPI Only Analyses showed that an HPI only cooling scenario, with some i small supplemental cooling from EFW in the period from
' 94 to 95-1/2 minutes, can explain the observed RCS cooldown.
However, an outstanding question as to why the excess inventory -
- was not detected still exists. As stated previously, a redistribution of voids in the.RCS could have occurred as a
(.J result of the HPI injection. Examining this redistribution would require a more detailed system simulation.
The analyses in Section 4.1 also demonstrate that HPI
~
activation is not inconsistent with the MUT level response observed during the accident. However, an alternate MUT level scenario is also possible. The response of the MUT level in' the time period from about 4:10 a.m. until well after 6:30 a.m.
could be characterized as a "saw tcoth" response. This response suggests that the main cause of variations in the MUT
~
level was caused by imbalances between MUT inflow (letdown) and outflow (makeup or partial HPI). During periods when the HPI was known to be actuated at TMI-2 on March 28, 1979, the MUT level response is detectably different from a "saw-tooth,"
_. e.g., in the period from 04:00:37 through 04:05:00, and again in the period from 07:21:00 through 07:30:00. The analysis l ,T provided in the EDS Report 9 concludes that, based on the MUT
. ./
50
s-. .
level response, HPI was not actuated at 5:41 a.m. The MUT
.m. . response to an "A" train . actuation 1s characterized by either a non-linear response (called " banana shaped") when OH-V-5A is
- opened or a decrease in the MUT level rate which approxiinates the HPI flow ("A" train) when DH-V-5A is closed. The absence I of this type response, without the assumptions defined in Scenario 1, suggest that full HPI actuation did not occur at 94-100 minutes. Based on these arguments, Scenario 1 has been judged possible, but it probably did not cause the RCS cooling i
observed.
6.1.2 Scenario 2: EFW Only
(
[ The analyses discussed in Section 5.0 demonstrate that an EFW only cooling scenario could result in the observed plant cooldown. Although there are similarities in the "A" SG -
L
- response at 94 to 100 minutes to that at about eight minute's,
- the relatively'small increase in steam pressure and the AUX-II h
l analysis suggests' that full EFW flow was not injected to the "A" SG at 94 minutes and beyond. For this scenario to be l viable, it must be concluded that the operator fully opened the S
l atmospheric dump valve and initiated EFW at 94 minutes. Since I
the SG was essentially dry at 94 minutes and pressure control f of the SG was being lost, it has been judged unlikely that the, operator would have fully opened the atmospheric dump valve.'
l Without the assumption of full opening of the atmospheric dump m
valve, the analyses indicate that the EFW flow would have been s
L limited to 200 gpa or less. Since approximately 365 gpm has ll been estimated to be required for the observed plant cooldown, N HPI flow would have been required between 94 and 100 minutes.
Based on these arguments, Scenario 2 has been judged possible 6 but it p"obably did not cause the RCS cooling observed.
w%
lV .
1.
l m
51
, , . - - - ,, , _ _ _ - , - - #-_.- _, _y, , . . , - , . - . _ . , _ . _ _ . - . . . - - _ , , . _ - _ _ _ - - _ - - - - -
6.2 Scenario 3: Sumary and Conclusions _ _ _ __,
j The evaluation of Scenarios 1 and 2 imply that a combination of both -
EFW and HPI led to the plant cooldown between 94 and 100 minutes.
Since no information exists for the positioning of the atmospheric -
dump valve over the period, an exact combination of EFW and HPI cannot be determined. Therefore, no detailed evaluation of the flow rates !
required from both tne EFW and HPI was performed. However, the relative flow rates can be estimated. ,
n
- If the operator had fully closed the atmospheric dump valves, the SG
- response analyses suggest tihat EFW flows of less than 100 gpm would have been provided to the SG (as. indicated in Section 5). Under this -
, condition, Scenario 1 would approximate the required HPI flows needed i d to obtain the plant cooldown. Thus, two HPI pumps would have had to ,
be actuated in this time period. .
However, if the atmospheric dump valves remained in- their apparent ( *-
L(j 50%) open position between 93 and 94 minutes, an EFW flow of .i approximately 200 gpm is suggested (as indicated in Section 5). An energy balance indicates that a total HPI flow of approximately 470 t-gpm would be mquired to obtain the observed plant cooldown. This is t:
the appropriate flow rate of one HPI pump in the pressure range of. 3-interest ( 1000 psig).
- I J The EFW/HPI scenario addresses both of the major questions identifed 1-in-evaluating Scenari_os 1 and 2. The main question with Scenario 1 is ..
the effect of the added HPI on the. system inventory. With one-HPI-
? pung, the minimum required by this scenario, the RCS void. fraction-is - . _
not expected to increase between 94 and 100 minutes. For Scenario 2, 4 1
6 a fully opened atmospheric dung valve is required to obtain the - -
.i.!
observed plant response. Using HPI to provide additional cooling if *,
_i eliminates the need for a full'y opened valve to satisfy the RCS 5e cooldown without an increase in SG pressure.
l.
i .
u 52 y .__y_ _ ,. , ,...r -. _ _ _ . ~ , - . . . . , _ _ , . _ , , - , . . _ _ . ., ,_,_w--,-.,r. - . _ _ . - . . - , _ _ _ , . . - . _ _ _ . _ - - _ . . --
Thus, based on the analyses described in Sections 4 and.5 and the qualitative assessments above, Sienario 3 appears to be the most probable cooling mode between 94 and 100 minutes.
l 4
a i
k l
I
&~
I _
I
')
lJ 53 L
r 7.0
SUMMARY
AND CONCLUSIONS The analyses of HPI and EFW status described in this report cover. the -
31-minute period between 94 and 125 minutes at TMI-2 on March 28, 1979.
During this time, the RCS experienced a significant cooldown and depressurization. The possible causes of this cooldown is the subject of q this report.
k l
It is 6elieved that the cooldown between 100 and 125 minutes was primarily i the result of EFW delivery to the "A" SG. Full HPI actuation did not occur at the time the "A" loop RC pumps were tripped (5:41 a.m.).
Three possible cooldown scenarios have been examined as possible causes of the cooldown between 94 and 100 minutes: HPI only (Scenario 1), EFW only (Scenario 2), a combination of EFW and HPI (Scenario 3). Each of these ,
scenarios could have provided the observed ccoldown. The evaluations of l
- each scenario are provided in Sections 4, 5 and 6 respectively. It,was not
!,'1 possible to conclusively determine which cooling scenario caused the plant y I L cooldown over this period. However, the evaluation suggests that Scenario 3 (bath EFW and HPI) best satisfies the behavior of primary and secondary parameters'.
Table 7-1 summarizes the results of the investigations performed for each cooldown scenario and compares each scenario with the other investigations listed in the Reference Table. As Table 7-1 shows Scenario 1, HPI only, is the least probable cooling model; Scenario 2, more probable; and Scenario l
3, a conbination of EFW and HPI, most probable. A detailed evaluation of the actual HPI flow used for Scenario 3 is not possible. However, the results indicate that a flow capacity of at least one HPI pump, and possibly two, was provided between 94 and 100 minutes.
I 1
a l 54 h . _ _ _ _ _ _ _ _ - - _ _ _ . _ . ______ _. _ _ _ _ __ _ _ ___ _ _ _ _ . _ _ _ . _ _ _ _.
-___,_________7____- __.
O, f- s
( , r. ,
w..
TAliLE 7-1
- COMPARISON OF OBSERVED RESPONSE Al THI-2 WITH THREE INDIVIDUAL SCENARIOS AND COWARISON OF THE SCENARIOS W!itt OTHER INVESTIGATIONL LISTED IN REFERENCE TABLE SCENARIO 3 - - -
DBSIRVED RESPONSE Parameter Agrees ~ Disagrees Agrees Disagrees Agrees Disagrees A LOOP T gg X X X A T cold X 7 --
g,g,,,,,,g ,4,4pj,g,,
X 8 LOOP Thot X X X B LOOP I cold -
X X X
~-
Pressure - decreasing X X X
~~
Pressurizer Level - decreasing X X X
~'
A LDUP llou - Increase, then decrease I partial I X X
-~
Source Range Monitor - decrease, drop X X X
. Leth Cooler Dutlet Temperature /Let : - Flow X partial X X X
, <a agreement ,
MU Tank Level - variable I (with X X assumption)
Let M . Flow - verlable I (wlth I X partial i assumption) agreement RC Inventory Analysis - core dryout p 112 minutes X partial X _. X X
-~ ~ ~
agreement A LDUP 5teae Pressure - slight increase, decrease X X partial X I
~~
agreement A UIMa Startup Level - low - dry 55 X X partial X X
~ ~ ~ ~
Other Analyses: ( W-lZ5 etnute range) '
TRAC - Reference 6 X X Not M - Reference # X Not investigated Investigated RSAC Sequence of Events - Reference 3 X Not investigated NUREG-0600 - Reference 10 hot investigated X agrees w/ lower flow M5AC Reference 4 I X --
X in part Jones Testloony ";;=Im 5 X X ~
X in part Wallis Testloony - Appendix 5 X . X X in part S
9
- o .
p
8.0 REFERENCES
,s ' 1. " Responses to GPUN Questions Concerning HPI Status at TMI-2 on March 28,197 9," May 6,1983.
- 2. TDR-044, "GPUN Sequence of Events, Three Mile Island," March 28, 1979, September 1980, Rev. 3.
- 3. NSAC-80-1, " Analysis of Three Mile Island - Unit 2 Accident," March 1980.
. 4. NSAC/24 "TMI-2 Accident Core Heat Up Analysis," January 1981.
l S. TDR-045, GPUN Service Technical Data Report, " Accident Transient Modeling Analyses, Appendix B: Inventory Change to RCS via Makeup / Letdown System," August 1,1979.
- 6. NUREG/CR-1353; LA-8273-MS, Preliminary Calculations Related to the Accident at Three Mile Island, March 1980.
! 7. NUREG/CR-1219; Analysis of the Three Mile Island Accident and J. Alternative Sequences, January 1980, i
- 8. EDS Document No. 02-0370-112. " Analysis of Reactor Coolant System ,
1 Makeup During the TMI-2 Event," December 29, 1982.
- 9. BAW-1629. TMI-2 RCS Component Evaluation Task 27, May 1980.
(' ,
- 10. NilREG-0600, " Investigation into the March 28,1979 Three Mile Island.
Accident by Office of Inspection and Enforcement, Investigative Report No. 50-J20/7 9-10," August 1979.
u .
e be hw un W.
O me kmi 56
r i..r.
. a t, ,,.M.<K..Fi.
Is p r .i't.M.,. m
?.rs ' v ')i
.i O..cii.;n.
~- i .
, . ..:, . ." q , pf , g g.g. .o -: s. .
I
'I d, n .
b_l M l
l l
l l
t
'l' $
t :. , ,
1 1.v, ., . . w':,.
zw.n s , . .
1
, ~ .
- < , . 6. ( 0 . t. , , I
'. ' w , . . .' ,1
' 5.ispi)%.y *
's C -
3, : . ' . ,
- a , ,
9.g - . , ,,2 -
/
t>
' y 'l.f;, .. ..ts: i, !
~
-( -
ca, < ;J.;ut,1,.c,,
,.t.. . . i .. m
, 3. . , ,
- ~
3:,. . i7,a. , . . . , ,
g , h.
- i. . Pli f
+Gla!dDif>lf, ..,
..,. <w.- ,
.s
...l...,<,...,.,,- .
. , . : ,. r. v ,
- . N + t . I e O *'
g J. t ,,, 6 s , tz : *3 -
e1..,.
',y_.,6s,' >?,.,
. c. ,
-' :t,{ ; il i*f~F \. .
- e
., y.-if,. .'1)8'1.*r$,'*,'
l , '{.
- o 2
,j ' : 4t
..ti2'.
t
, , ' . ('s,
, e e,
p'. .Q1
?
! '. ?.' tic N' 1. ,,Q'
- p. *.
1-a ,.d ,
- ' .;..,. s y . .v. y.., '* ; , >e 7 o ; ..
,. ..v
. 'w.m.
R:
) 1'. .1. y! .g
. g. o. . rp
.m' ,e
. i n .. .../6 . ' . ; . . . .f.
'_ ; ' :, .,<- 3., ; ' <q.; .
- ,, .. i'
,,. 4-! E Ij*:t.j ..
i, 1' ')l,'f t .s. ? n
, ' ' , . ,i.%,
. . , . . .% s.,...; ' .. . . c, ..
i
. . + , .
, n.. ) p !.
e ng'neering Services l
l 1
l v',
l Babcock &Wilcox a McDermott company l
.. . _ . . . . . . ... ~ ..._..
e 0 0 e
e A
1
.j APPENDIX 1 L
MUT LEVEL MODEL .
g_. .
O 1
4 J -
?
l
- l( >
\ -
P N - '.
s ,
A-1.0 I
Y -
APPENDIX 1 ,
MUT Level Model l
>- Introduction ^ '
l l
The computer model of the TMI-2 MU/HPI system was' developed to analyze MUT response during HPI actuation. Make-up tank level response was thought to be a key indicator of HPI flow. Hence, substantial effort was expended to develop a stable computer model that could be used with confidence. ,
The final program configuration includes many of the important features of the TMI-2 system (see Figure A.1.1). The BWST and MUT are modeled,' as are both HPI 3 trains. Any normal configuration of the three HPI pumps is allowed. Pump recirculation flow paths, letdown, and RC pump seal return are included. The program also has isolation valves for the MUT and BWST, along with MU control f
valve MUV-17 and four HPI control valves'(16A,168,16C,160). The program runs .
interactively on the B&W CDC-855 computer. Written in BASIC, the program ailows any or all valves to be stroked while the program is running. Data on RC pressure and. letdown flow vs. time is entered before execution begins. Starting L, MUT pressure and level initialize the model to match transient data. Valve stroke times are usually ir.ept constant from run to run unless an aby'ious correction is required. .
The flow network transients response is solved by a quasi-steady state approximation. At the beginning of each timestep, boundary conditions for the flow network are updated. (Valve positions, tank levels, and RCS pressure are 6
examples of boundary conditions that influence MU/HPI system response.) A new cteady-state solution of the flow network is achieved for each timestep. A L disadvantage of this solution technique is that it does not account for i localized transient conditions within the system, i.e., pressure waves due to y punp starts and/or rapid valve closures. However, for moderate to slow changes ,
in the system configuration, the code has proved to be a reliable and fairly accurate tool for scoping work (see Figure A.1.2, Benchmark Against Known HPI f
Actuation at 7:20 a.m.).
The following pages provide a brief analytical basis for the code, along with a program itsting. -
l s
i A-1.1 u
s Fi gu re A. l .1 GENERAL QUTLINE OF' HU-HPI H0 DEL w
i L
LETDO#M W(7)
(RECIRC) 8 RECIRC FLOW W(6)
Ill ll 13 DOUBLED IF SOTH i L6(FT) P6
'A' AND 'B' PUNPS U' ARE RUNNING K(6) (RECIRC VALVE)
( CATER LEVEL a
LEVEL U
V5(GALS: 51, U2,V2(GAL) a Li (INCHES) R(6)
-MU TAN (. 9 HUP A
, ][ W(2) } [ "
K(5) l 1 V12 K(3) R(8) O R(4),
aa' ELEY. HEAD W(o k
l l O
,;;;; R ;;; z '
y P1 Q y. R(1)
R(2) p W(5) -
VI6A
P3 K(4)
'( " MUP 8
" R(7) R(3) a'#'
g .'."" = , ,
W 4) Vl68 W(4)+W(9) b((9)K )
W(14) db M yg7
\ MU LINE X(2)
C PUNP RECIRC VALVE 3 E(7)
ASSUMED h.R(6) g{g) g g(g)
_ WATER SOURCE 4 i sai, ua m s W(13) Vl6C W(ISC)
~~ R(2) 3 > P5 i f PI-K(1 )
- W(13 )2 -- P(7) r- P(g)
K(9)
HPI PUNP C R(3) R(4)
'YA v ?NS 2 y Vl60 W(16)
A-1.2
F O
- 3 3
' (DVERN I NG STEADY- STATE EQUATIONS l.0 BWST Vo l ume in upper head = V5 ga l lons Level in upper head ='L6 feet L In the spherical r e gi on f rom 51 tu 56 feet of actual wat~er level, the i ni t ial BWST level is approximately 55 feet. The BWST level as not expected to fall below 51 feet for the duration of the transient. Therefore, the f ollowi ng equa t ions appl y:
(1.1) L5 = ini t ial BWST level, ft. (initial input) k (1.2) L6 = L5 - 51 ft. (line 4700) 2 Vol = [ TT D H - 4 TT H3](7.4805)
- 3 F .
l Vol = ( 7. 4 805 TT ) ( 3 7. 5 ) 2g, (4).tr(7.4805)H 3 .
4 3 L
l (1.3) V5 = 8261.981H - 31.334H 3 (.!ine 4800) p l To o)tain BWST level af ter a volume change has occur red, equatton (3) must be solved for L6 (or H). Being a cubic, the analytical m solution is somewha t complex. Therefore, the program solves for L6 us i ng a Newton-Raphson i terat ion scheme of, the f orm:
L6 -
f (L6 )
' 1+1 .= L6 I f'(L6 )
where f(L6;) is equation (1.3) with V5 known, and 5
f ' (L6 ; ) is the derivative of the equation with respect to L6.
L Thus, i
l (1.4) L6 i g = L6 -
([(31.334)(L6;)3 -
(8261.981)(L6 g)+
V5 ]/[ ( 94.002 )(L6 ; )2 - 8261.981]}
The static head from the base of the BWST to the tsolation valve I, DH-V-5A is approximately 22 feet. This gives an ini tial value of it 33.37 psig f or P1 when the BWST level is at 55 ft.
g '
v P! = (55 + 22) ft. HO 2 L A-1.3
L .
W t'
- b. P! = (77 ft)(62.4 lbn/ft 3)( g/ ge )( f t /144 in2)
P1 = 3 3. 37 ps i g To update t he BWST vo l ume a t the end of each timestep, I
L (1.5) V5 = V5 - At[W1 + W13](Gl/60) (line 21500) where W1 and W13 are the ' flows from the BFST to the two parallel' HPI trains.
s_
2.0 MU TANK User input can'sists of MU tank initial level in inches (L 1 ) ,
and initial pressure P6 ( ps i g) .
L_ Ul = MU tank vapor volume ( ga l l on s )
(2.1) Ul = 3925 - (L1)(30.75)
(line 8300) .
This f o rmul a wa s de r i ve d f rom th e f ol l owi n g da t a o n t he AUT leve l/ ga s vol ume relat ionshi ps MUT level,
-- Vapor volume -
0.0 inches 3925 gallons
,, 50.0 inches 2375 gallons 100.0 inches 850 gallons The general form of the equation is VOL = A + (B)(Lmu)
Using the above data, A = 3925 and B = -30.75 L is an expression of the water volume in the NU tank.
mu Knowledge.pf the MU tank volume a!!ows one to wri te the expression f or the MU tank ligiud volume U2 f rom equation (2.1):
(2.2) U2 = (L1 + 14.63)/(.0325) (line 8400)
-- From the ideal ga s r elat i onshi p, PV = nRT. Hydrogen is normally used as the cover ga s f o r t he NU t ank , so the ideal ga s a s sump t i on is appropriate.
a
.e.
A-1.4
l Assume an isothermal expansion or contraction for the tank cover gas as the MU tank level changes. At state (1),
,_ P gVg = ngRTg At state (2), the same relation applies:
i b P V 22 "2RT2 If the temperature and mass of the cover ga s remains
- b. constant during the level change, then Pg y, = p2V2 or, (2.3) P2 = PgVg/V2 (line 21700)
H! represents the static head contributed by the MU tank ,
when measured at t he T-connec t i on wi th the BWST outlet. The el eva t ion f rom the T-connec t ion to the AU Tank out let is
. , estimated to be 24.8 feet.
i '
'- (2.4) HI (psig) =
(.036)(Lmu) + 10.746 (1ine 9700)
,, The MU tank conservation of mass equation can be expressed as a change i n MU t ank l iq ui d vo l ume a s f ol l ows : ,
(2.5) AV = ( A t )[W7 + W8 + W14 + 0'6 - W2 ] (GI /6 0 )
(line 21100)
Where:
AV = MU tank volume change during timestep. ( ga l s )
At = timestep (sec)
W7 =' Letdown flow (Ib/sec)
W8 = RCP seal return flow (!b/sec)
W6 = Recire flow from HPI pumps A and/or B (Ib/sec)
W14'= Recirc flow from HPI pump C (l b/ sec)
- m. W2 ='MU tank outlet flow through V12 (ib/sec)
G1 = conversion factor from Ib/see to gpm Up
. A-1.5
l . . ,
e,x L 3.0 MU TANK - BWST FLOW SPLIT CALCULATION Two equations can be written for the pressure at the L T- conn ec t i on , P(2):
(3.1) PI -
P(2) = (Kche_x + K5A)EWI3 (3.2) P(2) - P6 = Hi -
(R1 +K 12+Kcheck)EW3 - WIl Where:
P(2) = Pressure at T- co nn ec t i on P6 = hU tank cover gas pressure 6- H1 = MU tank static head (does not include P6)
R1 = MU tank outlet line resistance K5A = Equivalent K- f a c t o r for DH-V-5A K12 = Equi va lent K-factor for M'U-V-12 -
Kcheck = Equivalent K-factor for check valves -
WI = Flow contribution f rom BWST (I b/ sec)
') W2 = Flow contribution from MU tank (Ib/sec)
W3 = Total t r ai n A MU/ HPI pump f low (I b/ sec)
,_ Equations (3.1) and (3.2) can be added to yields (3.3) P I - P6 = HI + (Kcheck + K5A)[Wl] -
(R l' +Kcheck +
K12)[W3 - Wl]
.. Thi s equa t ion i s a standa rd quadratic. Let WI be the variable to be determined:
.. Let A=Kcheck + K5A - R1 -K12-Kcheck Let B = (R1 + Kg2+Mcheck)(2)[W3]
Let C = -(R1 +K 12+Kcheck)[W3 ] + HI - P1 + P6 ,
Let R= [B 2 - 4AC]I (3.4) W1 = (-B + R)/2A .... upon inspection, R must be positive to yeild a legi t imat e va l ue f o r W1. Fl ow i n t he ne gat i v e j di rection will be prevented by the check valves.
.L s _,. (3.5) W2 = W3 - W1 (line 18000) ,
A-1.6
r.
There are two conditions which a e outside the bounds'of. '
this flow split calculation. If the darection of flow through DH-V- 5A o r MU-V- 12 r eve r se s , the check valves will b close. Program l ogi c is created to simulate the action of the check valves ba sed on pressure di f f erences. !ines 16700 16900 simulate the MU tank outlet check valve, while lines.
17000 - 17100 simulate the BWST check valve.
The logic is as follows:
- _ (1) If the pressure from the BWST less line losses with total pump flow is greater than the NU tank static head plus tank pressure, then the MU line check valve closes.
The program sets MU tank outlet flow to zero and assumes all flow comes from the BWST.
j (2) If the MU tank pressure and static head less line losses
- 2. with total pump flow is greater than t h e BWS T p r e s
- a r e ,
then the BWST flow is zeroed and all flow for the "
timestep is drawn from the MU tank. ,
i 4.0 HPI PUMP ADDEL (4.1) P(3) = P(2) - R2 [W3 2)
WHERE:
P(3) = Pressure at MU/HPI pump suction P(2) = pressure at T-connection -
W3 = pump flow R2 '= line resistance f rom T-connection to HPI pump.
The MU/HPI pump H/Q curve can be represented by:
(4.2) P(4) = P.(3) + 2860 -(Cl)[W32) (psi g)
P(4) is the pump outlet pressure.
Cl is a coefficient which determines whether the H/Q
__ expression is for one HPI pump or two pumps operati ng in parallel. If one pump is running, C1 = 0.29071 If two pumps are operat i ng, C1 = 0.14535. ,The pump flow doubles when both pumps are runni ng. Cl is specified by the user in a value vs. time table so that the pump can be " ramped" up or down as necessary.
W-g A-1.7
5.0 HPI/MU/RECIRC VALVES AND FLOW SPLITS .
DERIVATION OF PLCW SPL IT THROUGH HPI AND MU CONTROL' VALVES:
K6 RG PG ;>< "::,:
l (Mut- PRESK) p wg gg g, pa; S
..... mm mu
,l 4 . .. . p
,R1 ws k ,No - (ges Mess) s MuP(s) . " g,.y "= '
s p W 9-+
(5.1) P4 - P5 = (W9)2(R5 + KS) + (W9 + W4)2(R3)
(5.2) P4 - P5 = (W4)2(R7 + K4) + (W9 + W4)2(R3.) ,
1 *
> Subtract (5.2) from (5.1):
m (W9)2(R5 + KS) = (W4)2(R7. + K4)
(..
g .-
) -
(5.3) W4 = [(W9)2(R5 + K8)/(R7 + K4)]Y L (5.4) F1 = {(R5 + K8)/(R7 + K4))
Substitute (5.3) and (5.4) into (5.1): -
P4 - P5 = (W9)2(R5 + K8) + {W9 + [(FI)(W9)23 f(g3))
l (5.5) W9 = { [ P4 - P5 ]/ [R5 + K8 + R3 + ( 2 ) (R3 ) (F I ) I +
(R3 ) (F1 ) ])I (line 19500) l From (5.3) and (5.4),
(5.5) W4 = (W9)(F1)Y (1ine 19600) b By inspectson of the itowpaths, (5.6) W5 = ((P4 - P5)/(R8 +,K5 + R4)}Y (line 19700)
L" (5.7) W6 = ((P4 - P6)/(R6 + K6))I (1ine 19000)
If two pumps are runni n g, the reci rc f low (W6) is doubled.
y3 -
A-1.8 I l
l 1
5.
5 s-6.0 DETERMINATION OF THE TOTAL FLOW NETWORK SOLUTION (6.1) Wil = W4 + W5 + W6 + W9 (L INE 19800 )
Wil is the sum of the calculated outlet valve flows based on A P's . W3 is the total pump flow, and is chosen at the beginning of the iterat' ion. All " exterior" parameters are fixed for each timestep: i.e., P!, P5, P6, MU tank level, BWST level, and valve positions, if the calculated Wil is greater than W3, then the pump outlet pressure is t oo hi gh , resulting in outlet valve flow which is greater than actual pump flow. The next pump flow choice should be larger in order to reduce the calculated pump outlet pressure and outlet valve flows.
l On the other hand, if Wil is less than W3, then pump flow is
& reduced for the next iteration. A smaller pump flow selection will i nc.r ea s e the calculated pump outlet pressure, -
thereby increasing the outlet valve flowrates. When the ,
difference between Wil and W3 becomes less than the preset P error criteria, then the flow' network has been solved and
, the program advances another timestep.
w 7.0 SOLUTION 'OF FLOW NEMORK FOR PUMP RUNOUT CONDITION
" When RCS pressure falls and the HPI valves (V16s) are open, the pump flow will increase to a maximum capacity. This limit was established at 550 gpn/ pump for the T41-2 model.
If the flow limit is reached, pump out put pressure will l continue to drop without a si gni fi cant increase in flow.
Therefore, it becomes necessa ry ~ to calcul at e pump di scharge 6
pressure when RCS pressure, MU tank pressure, and the total flowrate are gi ven. A subrout i ne (beginni ng at line 22300) was written to solve for this condition.
m The subroutine uses the same principles established earlier in sections 5.0 and 6.0. The equations for the valve flows t are identical, as is the technique of summing the calculated L valve flows f or compari son agai ns t the total pump flow.
However, in this case the changi ng va riable is the pump l outlet pressure, P4, rather than the pump flow which is fixed at the maximum value. P4 can be va ri ed in the search f
for a solution be' tween RCS pressure (P5) and the pump shutof f head of 2860 psi g. When a value of pump out let pressure has been found for which the calculated valve flows (l
A/
equal the pump flow, (line 23100), the subroutine returns control to the main program.
l A-1.9
~
l 1
- 8.0 CALCULATIONS FOR THE C HPI TRAIN The C train solution is identical in principle to the A/B
, HPI train solution, but is less complex. The C train is solved independently of the A/B train. At the begi nni ng of the iteration, P!, P5, BWST level, and the C train valve positions are inxed.
P -
l The calculation begi ns at line 25200. -
All line resistances are assumed to be identical to corresponding resistances in the A/B train. The suction valve to the BWST is also assumed to be open when the C train is activated (otherwise a loss of NPSH condition would result for the C pump). Therefore, DH-V-5B is not modeled for the C train.
The allowable flow range for the pump i s 0 - 5 50 gpm.
Subrouttne 28100 calculates the solution for the C train .
L network in the event a runout condition occurs. The f technique is identical.to the A/B train solution in section
{'}
i 7.0. The pump discharge, pressure is allowed to vary between
_s 2 860 p s i g a nd RCS .p r'e s s u r e (P5). Upon f indi ng the proper
> value for the C pump discharge pressure (P8), control returns to the C train subroutine.
9.0 VALVE MODEL ,
L The program allows the user to interact with the program while it is running to initiate valve movement. Any or all valves may be moved at any time. Each valve has a stroke time which i s programmed into the model, but can be changed L easily if required.
I Valve movements, tank level changes, and pressure changes l
occur once each timestep, but are held constant during the flow network solution. Tabular valve K-f actors as a function i
of position are programmed into the model. As a valve st rokes open o r cl osed, i t s K- f acto r chan ge s incrementally k f or each t imes t ep. Check va l ves a re no t modeled ex pl ici t l y, but are accounted for by appropriate logic in the flow network solution (s ee s ect i on' 3.0 ) .
-O A-1.10 k
n PROGRAM LISTING
^
00100 DIV T2(100,100),W(50),Th(50,2),T5(50,2) 00200 FILE # 1 =" TAPE l",#2=" TAPE 2" 00100 .? CA O G I,T(3),El L 00400 DATA 7.1923,.5,.01 00500 C9:0 00600 Ai = 1E6 00700 FOR 3 = 1 TO 9 00800 READ N(3) -
00900 NEXT 3 01000 D ATA 6,3,2,6,6,3,6,4,6 5 01100 FOR I = 1 TO 9 01200 FOR 3 = 1 TO 10 01300 FOR K = 1 TO 2 .
g 01400 TI(I,3,K) = Al 01500 NEXT K 01600 NEXT 3 01700 NEXT I 01300 FOR I = 1 TO 9 01900 FOR 3 = 1 TO N(I) .
} 02000 FOR K = 1 TO 2 .
L 02100 READ TI(I,3,K) 02200 NEXT K 002300 NEXT 3
'02400 NEXT I 02500 REM VALVE TABLES K VS VALVE POSITION 02600 DATA 0,1,5,62-6,10,3E-6,25,1.2E-6,50,6E-7,100,3E-7 02700 DATA 0,lE6,95,1.M38,100,lE-7 02300 DATA 0,lE6,100,l E-7 02900 D ATA 0,l E6,6.25,1.856,12.5,.923,25,.M4,50,.232,100,.116 03000 D ATA 0,1 E6,6.25,1.856,12.5,.928,25,.M4,50,.232,100,. I 16 03100 DATA 0,lE6,95,1.M38,100,l E-7 03200 DATA 0,l E6,6.25,1.856,12.5,.928,25,.%4,50,.232,100,.!!6 -
03300 DATA 0,l E6,10,66.76,50,19.15,100,.668 03400 DATA 0,l E6,6.25,1.856,12.5,.923,25,.M4,50,.232,100,116 f- 03500 FOR 3 = 1 TO 9 03600 READ S(3) 03700 NEXT 3 L 03800 REM VALVE STROKE TIMES (SECONDS) 03900 DATA 14,4,4,10,10,4,10,10,10 04000 READ W(8) 04100 REM W(8) IS THE SEAL RETURN FLOW (GPM) ;
04200 DATA 5 04300 W(8) = W(8)/G1 - i 04400 READ L5,P6 i
- 04500 REM L5=BWST LEVEL (FEET)
- P6 = INITIAL MU TANK PRESS (PSIG) 04600 DATA 54.6,16.5 04700 L6=L5-51 04800 V5=826I.98i *L6-31.334*L6^3 b (,104900 FOR I = 1 TO S 05000 READ R(I)
L-A-1.11
-+ . _ -
r 7
05100 NEXT I 35200 REO AlUT SUCT --HPl N'OZZL ES-- MULINE RECIRC HP! TRAIN Eh VALV -
0 5 300 D \ T.\ 4.00 E- 3,2.! F. 4,,106 2?,311,.10623311,1,037 36,14.49,1.9517E-2,l.9317E-2 05400 READ N2 05500 FOR I = 1 TO N2 6 05600 R EAD T2(I,1),T2(1,2) 05700 NEXT I 05300 REM TABLE 2:. TABLE OF TIME,RC PRESSURE (IST ENTRY = NO. OF DATA PAIRS)
, 06010 DATA 20,4320,1054,4380,1052,4440,1048,4500,1046,4560,1058,4620,1081 06020 DATA 4680,1100,4740,l108,4800,1108,4860,1096,4920,1077,4980,1069 06030 DATA 5040,1073,5100,1085,5160,1100,5220,1112,5340,1135,5400,1146 06040 DATA 5460,1154,5520,1163
' 06300 READ N3 06400 FOR I = 1 TO N3 06500 READ T3(I,1),T3(1,2) 06600 NEXT I 06700 REM TABLE 3: DETERMINES HOW MANY HPI PUMPS ARE RUNNING- FORMAT = TIME, COEFF
, TIM E, CO EF F, ETC.
j 06800 REM .2907 = l PUMP, .14535= 2 PUMPS. (IST ENTRY = NO. OF DATA PAIRS) 06900 D A TA 6,0,. 2907,41,. 2907,46,.14535,180,,14535,4200,. 2907, l E6,. 2907 07000 READ NS .
- 07100 FOR !=1 TO N5 ~
1 07200 READ T5(1,1),T5(I,2) 07300 T5(I,2)=T5(1, /GI
~,07400 NEXT I
. 07500 REM TABLE 5: LETDOWN TI'ME, FLOW DATA (IST ENTRY = NO. OF DATA PAIRS)
' 07600 DATA 8,4320,118,4388,118,4389,50,4818,50,4322,105,5109,105,5110,65 07650 DATA 5460 Of 000 READ Li *,65 f
- 08iOO REM L1 IS THE INITIAL MU TANK LEVEL QNCHES) 08200 DATA 63.514 08300 Ul = 3925-Ll*30.75 '
08400 U2 = (Ll+14.63)/.0325 08500 FOR I = 1 TO 9 08600 READ X(1) 08700 NEXT I
] 08800 REM TABLE OF INITIAL VALVE POSITIONS (VALVES 1-8) 08900 DATA 0,0,100,0,0,100,0,100,0 09000 FOR 3 = 1 TO 9 3 09100 D(3) = X(3) 09200 Y(3) = X(3) 09300 NEXT 3 09700 HI = .036*Ll+10.74 2 09800 PRINT "!NITIAL TIME FOR THIS RUN (SECONDS)";
09900 INPUT T(1) 10000 PRINT ,
g 10100 PRINT " TIME FOR VALVE ACTUATION, SEC (<0 IF NONE)";
10200 INPUT T(2) 10 40 0 I F T(l )>T(2) TH E N I 070 0........................................................
10600
[gm==COTO L5100 ==============m============
b A-1.12 C
L -
~
r 10700 PRINT "TIVE IS"; T(1)-T(3); "SECOND5"
_ 10S00 T(3)= 5 -
10900 A$ =""
11000 PRINT " ENTER THE NUMBER COR RESPONDING TO THE V ARI.\i3LE YQt; W!iH TO CHANG. :"
11100 PRINT 11200 PRINT "<l> "; "VSA - SUCTION TO BWST"; TA8(40); "<2> "; "MUP-lC RECIRC CTRL
~
' VALVE" i 11300 PRINT i !!400 PRINT "<3) "; "V12 - SUCTION TO MU TANK"; TA B(40); "<4> "; "Vl6B - LOOP B MU/HPI" 11500 PRINT
"~
11600 PRINT "<5> "; "Vl6 A - LOOP B HPl"; TAB (40); "<6> "; "MUP-I A, MUP-IB RECIRC" 11700 PRINT
!!S00 PRINT "<7) "; "V!6C - LOOP A HPl"; TAB (40); " "; "V17 - MU CONTROL" l1900 PRINT
- 12000 PRINT "<9) "; "V160 - LOOP A HPI"; TAB (40); "<10> "; " CHANGE MU TANK PRESSURE" 12100 PRINT 13000 PRINT "(VALUE <0 TO END PROGR A V)";
I3!00 INPUT S I 315 0 i F S = R O F (S ) A N D S < l i TH E N I 3 20 0............................................... ~
EL5EI3000...................................................................
- l 3 2 0 0 I F S < 0 TH E N 2 210 0... . . . . . . . .. . .. .. . . . . .. . . . . .. . . . .. .. . . .. .. .. . . .. . . . . . . . . . . . . . .
13300 PRINT
- l 3305 IF S=10 THEN I3310...........................................................
j E L S E I 3 4 0 0.. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
_ ~ 13310 PRINT " CURRENT MU TANK PRESSURE IS"; P6; "PSIG" 13 320 PRINT "NEW PR ESSURE IS ..........";
13330 INPUT P6-13340 GOTO 13700
=============== ======================================= ==================
13400 PRINT " CURRENT POSITION (%) FOR"; S; "1S"; X(S) 13500 PRINT "NEW POSITION DEMAND IS";
_ 13600 INPUT D(S) 13700 PRINT 13800 PRINT "ANYMORE TO CHANGE (Y,N, OR OPTION NUMBER)";
13900 INPUT Al$
! 4 0 0 0 IF A 1 $ =" Y" TH EN I I 0 0 0 ........................................................
14100 IF Al$="1" OR Al$="2" OR Al$="3" OR Al$="4" OR Al$="5" OR Al$="6" OR Al$="7" OR A l$ ="3" OR A l$ ="9" OR A l$ ="10" TH EN S=VA L(A l$) ELSE 14300 ....................
.. 14200 GOTO L3200
==================================== ====================================
14300 PRINT 14400 PRINT " TIME FOR NEXT ACTION (SECONDS)";
14500 INPUT T(2) 15100 FOR I = 1 TO 9 ,
15200 YO) = T(3)*l00/S(I)*SGN(D(1)-X0))+X0) 15300 IF XO)<D(I) AND YO)>D(1) THEN YG)=0(1) 15400 IF XQ)>DU) AND YG)<D(1) THEN YU)=DU) 15500 IF YO)>100 THEN Y(I)=100
~
.I k.j L . . . _ _ _ _ _ _ _ _
A-1.13
(
/m
'15600 IF YUM0 THEN YU)=0 I 5 7 C 0 3 0 5 U 6 2 3 4 C 0 .. . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . .. .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . -
1530G NEX T I ,
I 5 9 0 0 G 0 5 U S 2 4 0 0 0.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . .. . . . . . . . . . .. . . . . . .
16 0 0 0 G C S U B 2 4 6 0 0 .. . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . .
16 G 5 0 G C S U B 3 3 6 0 0 ... . .. .. . . . . .. . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . .
16 1 0 0 C 0 5 U S 2 5 2 0 0 . .. . . . . . .. . . . . .. . . . . . . . . . . . .. . . . . . . . .. . . . . . . .. . . . . . . . .. . . . . . . . . . . .
16 2 0 0 G C S U B 2 9 2 0 0 . .. . .. . .. . . . . .. .. . . . . . . . ... .. .. .. . .. . . .. . . .. . .. . . . .. . . . .. ... . . . . . .
, 16400 IF Cl=.2907 THEN M2=550/G1 ELSE M2=(.14535/Cl)*(!!00/GI) 16500 Ml=0 -
16600 W(3)=(M 1+M2)/2 16700 IF P i - (K (1 ))
- W(3)^ 2> P6+ H I TH EN 16 80 0 .......................................
- E L S E I 7 0 0 0 . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
16800 W(1) = W(3) 16900 GOTO 18000
======================== == ===== ============= === ==== ===============
170 00 IF P6+ H i-(R (1 )+K (3))*W(3)* 2> P I TH EN 17100......................................
g E L S E I 7 5 0 0 . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
17100 W(1)=0
( 17200 P(2)=P6+HI-(R(thK(3))*W(3)*2 ~
17300 W(2)=W(3) 17400 GOTO 18200
=================== ===================== =========-=====================
17500 A=K(1)-R(1)-K(3)
(DD17600 l7700 C=-(R(t B=2*W(3)*(R(1)+K(3))
hK(3))**(3)*2+P6+Hi-PI ,
17800 R=SQR(B"2-4*A *C) 17900 W(1)=(-B+R)/(2*A) 18000 W(2)=W(3)-W(1)
> 13I00 P(2)=PI-(K(1))*W(1)'2 18200 P(3)=P(2)-R(2)*W(3)*2 .
18300 P(4)=P(3)+2860-Cl*W(3)*2
. I 8 40 0 I F P(4) <P 5 TH EN I 8 5 0 0..........................................................
l EL S E 1 90 0 0 . .. .. .. . . . . . . . . . . . . . . . . . . . . . . . . . . ... . . . . . . .. . . . . .. . . . . . . . . . . . . . . . . .
l 18500 M2=W(3) -
i L
13600 GOTO 16600 7 =====================================================================================
19000 W(6)=SQR((P(4)-P6)/(K(6hR(6)))
19100 IF Cl<.2907 THEN W(6)=(.14535/CI)*(2*W(6))
19400 Fl=(K(8)+R(5))/(K(4)+R(7))
19500 W(9)=SQR((P(4)-P5)/(K(8h.R(5)+R(3)+2*R(3)*SQR(FI)+R(3)*FI))
19600 W(4)=W(9)*SQR(FI) 19700 W(5)=SQR((P(4)-P5)/(K(5hR(3)+R(4)))
P 19800 W(I I)= W(4)+ W(5)+W(6)+ W(9) l 19 90 0 IF A BS(W (11 )-W(3)) <E I TH EN 20 40 0 .............................................
l 20000 IF (.14535/C 1)*1100/G 1-W(3)<El THEN GoSUB 22300................................
p E L S E 2 0 2 0 0 . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . .. . . ... . . . . . . . . . . . . . . . . . . .. . . .
l 20100 GoTo 20400
=========================================================================
h m20200
~ 20300 coro IF 16600 W(ll)>W(3) THEN Ml=W(3) ELSE M2=W(3)
==================-=============================================================
l r
A-1.14 1
a
i p
~
/~'s 2040 3 IF(T(11/3)= R O F(T(I )/3) THE N GO SUB 29800 .......................................
20SOG FC.7 I= 1 TO 9 '
20500 \'(I):Y(I) 20700 IF X(i)=D(I) THEN E2:E2+1
, 20300 NEXT I 20900 JF E2=9 AND T(l)=ROF(T(l)) THEN T(3)=1 21000 T(1)=T(1)+T(3) 21 100 V2= U2+ T(3)*(W (7)+ W(I 4)+ W(3)+ W(6)-W(2)) *G I/60 21200 L1=.0325*V2-14.63 F 21300 Hl=.036 *Ll+10.74 21400 VI=3925-30.75*Li .
3 21500 V5=V5-T(3)*(W(1)+W(13))*G1/60 21700 P6=(P6+14.7)*Ul/V1 21750 P6=P6-14.7 21300 Ul=VI 21900 U2=V2 22000 GOTO 10400
= = = = = : : u = = = = : : : : : : : : = = = = = = = = = = = = = = : : = = = = = = = = :- = : ::: : = = = = = = = = : : : :: :: : : = = : : :-
L 22100 PRINT TAB (35); "300 FINISHED" 22200 STOP ,
b 22300 REM SUBROUTINE. TO CALCULATE CONDITIONS FOR PUMP RUNOUT 22400 H(1)=PS
's22500 H(2)=2860 '
d22600 P(4)=(H2+HI)/2 . .
22700 W(9)= S Q R ((P(4)-P 5)/(K (8)+ R (5)+ R (3)+ 2 *R (3)*SQR(F I ) R(3)*F 1))
22300 W(4)= %(9)*SQR(?!)
22900 W(5)=SQR(lP(4)-P5)/(K(5)+R(2)+R(4)))
, 23000 W(6)=SQR((P(4)-P6)/(K(6)+R(6)))
23100 IF A BS(W(9)+ W(4)+ W(5)+W(6)-W(3)) <E I TH EN R ETUR N................................
23200 IF W(9)+W(6)+W(5)+W(4)>W(3)THEN H2=P(4) ELSE Hl::P(4) 23300 GOTO 22600 b- ============u=================== ============ ======================================
j 23400 REM SUBROUTINE TO CALCULATE VALVE K'S VS POSITION L 23500 FOR L=1 TO N(I)-1 l 23600 IF Y0)<TI(i,L,1) OR YG)>T1(I,L+ 1,I) THEN 23700..............................
EL S E 2 3 8 0 0 . . .. . ... . . . . . . . .... . . .. . . .. . . .. . . . . . . .. . . . . . . .. . . .. . . . . . . . . . . . . . . . ..
23700 NEXT L h 23300 KO)=TI(I,L,2)+(YO) - TI(i,L,1))*(T10,L+1,2) - TI(I,L,2)) / (TI(1,L+1,1) -
Ti(I,L,1))
23900 RETURN I 24000 REM SUBROUTINE TO CALCULATE RC PRESS VS TIME ,
l 24100 FOR I = 1 TO N2-1 -
L 24200 IF T(1 )<T2(I, !) OR T(1)>T2(!+ 1, !) TH EN 24300 .................................
P E L S E 2 4 4 0 0 . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . .
l 24300 NEXT I w.
d i
k A-1.15 -
s s'
24400 P5 = T2(I,2)+(T(1)-T2(I,0)-(T2(T I,2)-T2(I,2))/(T2(1+ 1,1)-T2(1,1))
N 0 TETURN .
10600 RE\1 SUBROUTINE TO CALCULATE COEFFICIENT FOR H/Q CURVE PROM TABLE 24700 FOR != 1 TO N3-1 6 2650 0 IF T(l )<T 3(I,1) OR T(1)> T3(I+ 1,1 ) TH EN 24 900...................................
' E L 5 E 2 5 0 0 0 .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
24900 NEXT I g 25000 C1 = T3(I,2)+(T(1)-T3(1,1)) *(T3(I+1,2)-T3(I,2))/(T3(I+ 1,1)-T3(I,1))
25100 RETURN 25200 REM SUBROUTINE TO CALCULATE FLOWS IN C HPI TRAIN 25 3 00 IF Y(2)= 0 A N D Y(7)= 0 AN D Y(9)= 0 TH EN 2540 0.....................................
E L S E 2 5 8 0 0 .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
25400 W(14) = 0 I
25500 W(15) = 0 25600 W(16) = 0 j 25650 W(16) = 0 g 25700 RETURN
=======================================: ==== =================.:========
25300 M3=0 -
25900 M4=550/Gl -
S 26000 W(13)=(M3+M4)/2 26100 P(7) = P I-(3E-7+R(2))*W(13)'2
~ /.)26200 P(3) = P(7)+2360 .2907*W(13)^2
,# 26 3 0 0 I F P(8) < P 5 TH E N 2 6 4 0 0..... ...................... .. .. ..... ...... ...... .. ... .....
E L S E 2 6 9 0 0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . .
26400 M4=W(13)-
26500 GOTO 26000 3
======================= ============================= === ================
26900 W(14)=SQR((P(8)-P6)/(K(2)+R(6))) .
27200 W(15)=SQR((P(8)-P5)/(K(7)+R(3)+R(4)))
. 27300 W(16)=SQR((P(S)-P5)/(K(9)+R(3)+R(4)))
l 27400 W(21)= W(14)+ W(15)+W(16) 27 5 00 IF A BS(W (2 I )-W(! 3)) <E I TH EN 230 00 ............................................
[ 27600 IF 550/G i-W(13)<E l TH EN GOSUB 28100 ..........................................
P EL S E 2 7 8 0 0 . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
27700 GOTO 28000
======= =================================================================
27300 IF W(21)>W(13) THEN M3=W(13) ELSE M4=W(13) 27900 COTO 26000
=========================================================================
28000 RETURN 23100 REM SUBROUTINE TO CALCULATE C PUMP RUNOUT CONDITIONS '
28200 H(3)=P5 28300 H(4)=2860 28400 P(S)=(H(3)+H(4))/2 23500 W(14)=SQR((P(8)-P6)/(K(2)+R(6)))
g ,28600 W(15)=SQR((P(8)-P5)/(K(7)+R(8)+R(4)))
s.> .
w.
A-1.16 m-
28700 \V(i6)=%;R((?(3)-P S)/(K(9)A (3)M(4)))
2em ir MSPV(i * )* T(l 5). 3 (l6)-T(13))<EI TH E N 2 91 0 0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
23900 IF &(14)d(15)+T(16)>?s(13) THEN H('4)=P(8) ELSE H(3)=Pf 3) 290 % GOTO 2.%00 u2-.:223:::: : . :::::=::::::::::=::==== ::::::::.:::::::::::=:::::=====:;::::: :::::::
29100 RETURN
[' 29200 REM SUBROUTINE TO CALCULATE LETDOWN FLOW FROM TABLE 5 29100 FOR I= 1 TO NS-1 29400 IF T(l )<T5(I,1) OR T(i)>T5(I+ 1,1) TH EN 29500...................................
E L S E 2 9 6 0 0 .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
- 29900 NEXT I 29600 W(7) = T5(I,2)+(T(1)-T5(1,1))*(T5(I+1,2)-T5(I,2))/(TS(l+1,1)-T5(I,1))
29700 RETURN
~
29800 REM PRINTING SUBROUTINE 29900 MARGIN #1,80 l 30000 MARCIN J2,140 30200 PRINT #2," "; " ELAPSED TIME IS"; T(1)" SECONDS"; TAB (60); "MU TANK LEVEL ='; Ll; "InCHE5" .
30300 PRINT #2," "; RPT$(" ",135)
- _ 30400 PRINT #2," "; TAB (20); " FLOW RATE"; TAB (40); " VALVE"; TAB (60); " VALVE";
TAB (30); " FRICTION"; TAB (100); " PRESSURE" 1
(m 0500
- PRINT #2," "; TAB
" FACTOR"; TAB(100); (20); "PSIG" "GPM"; TAB (40); " POSITION"; TAB (60); " DEMAND"; TAB (80);
" 30600 PRINT #2," "; TAB (20); RPT$(" ",10); TAB (40); RPT$(" ",10); TAB (60); '
3PT$(" ",10); TAB (80); RPT$(" ",10); TAB (100); RPT$(" ",10) 30700 PRINT #2',"0"; "HPI VALVE 16A"; TAB (20); W(5)*Gl; TAB (40); Y(5); TAB (60); D(5);
, TAB (80); K(5)+R(8)+R(4); TAB (100); P5; TAB (120); "RCS PRESSURE" 30800 PRINT #2,"0"; "HPI VALVE 16B"; TAB (20); W(4)*Gl; TAB (40); Y(4); TAB (60); D(4);
TAB (80); K(4)+R(7)+R(3); TAB (100); P(4); TAB (120); "MU PUMP OUTLET" 30900 PRINT #2,"0"; "MU VALVE 17"; TAB (20); W(9)*Gl; TAB (40); Y(8); TAB (60); D(8);
TAB (80); K(3)+R(5)+R(3); TAB (100); P(2); TAB (120); "MUT-BWST TEE" 31000 PRINT #2,"0"; "MU PUMP RECIRC"; TAB (20); W(6)*Gl; TAB (40); Y(6); TAB (60); D(6);
TAB (80); K(6)+R(6); TAB (100); P(3); TAB (120)"MU PUMP SUCTION" .
31100 PRINT #2,"0"; "MU VALVE 12"; TAB (20); W(2)*Gl; TAB (40); Y(3); TAB (60); D(3);
TAB (80); K(3)+R(1); TAB (100); P6; TAB (120); " MAKEUP TANK (TOP)"
31200 PRINT #2,"0"; "BWST VALVE 5A"; TAB (20); W(1)*Gl; TAB (40); Y(1); TAB (60); D(1);
TAB (80); K(1); TAB (100); Pl; TAB (120); "BWST STATIC HEAD" 31210 PRINT #2,"0"; "HPI VALVE 16C"; TAB (20); W(15)*Gl; TAB (40); Y(7); TAB (60); D(7);
TAB (80); K(7)+R(4)+R(8); TAB (100); P(7); TAB (120); "C PUMP INLET" 31220 PRINT #2,"0"; "HPI VALVE 160"; TAB (20); W(16)*Gl; TAB (40); Y(9); TAB (60); D(9);
TAB (80); K(9)+R(4)+R(8); TAB (100); P(8); TAB (120); "C PUMP OUTLET" 31300 PRINT #2,"0"; " LETDOWN"; TAB (20); W(7)*Gl; TAB (40); " SEAL RETURN"; TAB (60);
W(8)*Gli TAB (80); "C PUMP RECIRC"; TAB (100); W(14)*GI
.. 31400 PRINT #2,"0";"BWST LEVEL (FEET)"; TAB (20);L6+51 31600 C9=C9+1 31700 IF C9=2 THEN PRINT #2,"1" ELSE PRINT #2," "
31800 IF C9=2 THEN C9=0 c 131350 PRINT #1," ELAPSED TIME IS"; T(1); "SECONOS","MU TANK LEVEL ="; Ll; " INCHES" V
A-1.17
>+
9 31900 CR INT # 1,R PT$(" 'r,75)
"FRICTICN"; TA8(64); " PRESSURE" 12100 PR!iT # 1,TA 3(12); " CPM"; TA 8(25); " POSITION"; TA 3(3Sh "DEM.\ND"; T V(SI):
6 "FA CTOR"; TAB (64); "PSIG" 32200. PRINT #I,TA3(12); RPT$(" ",10); TAB (25); RPT$(" ",10); TA B(33); R P T$(" ",10);
TAB (5I); R PT$(" ",!O); TA B(64); RPT$(" ",10) 32300 PRINT #1,"HPI V16A"; TAB (12); W(5)*Gl; TAB (25); Y(5); TAB (38); D(5); TAB (49);
K(5)+R(3)+R(4); TAB (62); P(3); TA 8(72); "MUP IN" 32400 PRINT #1,"HPI V16B"; TAB (12); W(4)*Gl; TAB (25); Y(4); TAB (38); D(4); TAB (49);
j K(4)+R(7)+R(3); TAB (62); P(4); TAB (72); "MUP OUT" .
L 32425 PRINT #1,"HPI Vl6C"; TAB (12); W(15)*Gl; TAB (25); Y(7); TAB (18); 0(7); TAB (49);
' K(7)+R(4)+R(8); TAB (62); P(7); TAB (72); "C INLET"
}
32450 PRINT #1,"HPI Vl6D"; TAB (12); W(16)*Gl; TAB (25); Y(9); TAB (33); D(9); TA B(49);
( K(9)+R(4)+R(8); TAB (62); P(bh TAB (72); "C OUT" 32500 PR:NT #1,"VU Vl7"; TAB (12;; W(9)*Gl; TAB (25); Y(8); TAB (38); D(3); TAB (49);
K(3)+R(5)+R(3); TAB (62); P(2); TAB (72); " TEE" 32600 PRINT #1,"RECIRC"; TAB (12); W(6)*Gl; TAB (25); Y(6); TAB (33); D(~); TAB (49);
b K(6)+R(6); TAB (62); P(6); TAB (72)"MU TANK" 32700 PRINT #i,"\iUT V12"; TAB (12); W(2)*Gl; TAB (25); Y(3); TAa(3S); D(3); TA B(49); ,
K(3)+R(1); TAB (62); Pi; TAB (72); "BWSI" , ,
y 32800 PRINT #1,"BWST V5A"; TAB (12); W(I)*Gl; TAB (25); Y(1); TAB (38); D(1); TAB (49);
K(I); TA B(62); P5; TAB (72); "RCS"
, ] 32900 PRINT #1, " LETDOWN"; TAB (12); W(7)*Gl; . TAB (25); "BWST LEVEL (FEET)"; TAB (45);
V L6+51 ,
33000 PRINT #1, " SEAL RET"; TAB (12); W(8)*GI; TAB (25); "C PUMP RECIRC"; TAB (45);
- W(I 4)*G I ,
33200 PRINT #1 3 33300 PRINT # 1,RPT$("*",75) 33400 PRINT #1
- 33500 RETURN 33600 REM SUBROUTINE TO CALCULATE NEW BWST LEVEL
, 33700 L6(1) = L6 ,
l 33800 L6(2)=L6(1)-(11.334*L6(1)*3-8261.981 *L6(1)+V5)/(94.002*L6(1)'2 - 8261.981) 33900 IF A BS(L 6(2)-L6(1)) < l E-8 TH EN 34200 ..........................................
EL S E 3 40 0 0 . . . . . . .. . . . . . . . .. . . . . . .. . .. . . . . .. . . . . . . .. . .. . .. . .. . . . . . . . . . .. . . . . . .
34000 L6(1) = L6(2) 34100 GOTO 33800
=========================================================================
34200 L6=L6(2) 34300 Pi=(L6+73)*62.4/144 6 34400 RETURN 34900 END ,
w 4
tu ne A-1.18
5 L
VARIABLE LIST L
A Dummy va r i a bl e us ed to simplify the calculation of MUT-BWST f l ow sp l i t s . Corresponds to a coefficient in the s t anda r d quadra t i c eq ua t i on.
L Al Dummy va ri abl e. Used to load K- f actor tables with an ini t ial value of IE6 A$,Al$ Dummy st ri ng va ri abl e f or obtaini ng user interactive
,_. response.
B Dummy variable used to simpl i f y the calculation of MUT-BWST flow splits. Corresponds to a coef fici ent in the standard quadra tic equa tion.
C Dummy va riabl e used to simpl i f y the calculation of MUT-BWST f l ow sp l i t s . Co r r e s po nds to a coef fici ent in the standard quadra tic equa tion.
. Cl Coe f f i ci ent in H/Q expression for MU pump (s). Cl is int erpolated f rom table 3 :as a function of time.
C9 Counter to hel p de t ermine ca r ria ge cont rol for output file.
D(1) Current position demand (%) for va l ve numbe r I.
~
El Error margin cri teria for flow iterations: typical l y = -
.01 .
E2 Counter which determines if all val ve s have reached te demanded positions.
, F1 Funct i on used to simplify flow split calculation
,_ through A/B t rain HP! va l ve s an d MU va l ve . ,
'G1 Conversion factor gpm/ l b / sec (co ns t ant )
H1 ,
MU tank static head (psi)m, neasure.d at the tee connection with the BWST line. (H1 is a function of
" L1, the MU tank level .)
H(1),H(2) Range of minimmn and maximum pressures for A/B MU pump '
discharge (R uno u t condi tions onl y) .
, , . H(3),H(4) Range of minimum and maximum pressures for C HPI pump di s cha r ge (Runout condi t ion onl y) .
1,3,K,L DO LOOP co un t e r s .
K(1) Current K- f ac t o r f or va l ve numbe r I.
- Ll MU tank water level (inches) (user initial input).
L5 BWST level (f eet ) (user ini tial input).
L6 BWST level (f eet ) above the st rictly cylindrical
.. portion of the tank.
L6(1) Calculated values of SFST level as a f unction of vo l ume . Used in L6(2) Newton-Raphson i tera t ion whi ch sol ve s f or SFST leve l .
MI,M2 Ran ge of mi nimum and max imum f l ow f o r A/ B MU/ HPI train (Ib/sec).
NO,M4 Range of minimum and maximum flow for C HPI train (Ib/sec).
N2 Numbe r of data pairs in table 2, ( t ime, RC pressure).
N3 Number of data pairs in table 3, (time, H/Q w,.. .
coef ficient ) .
A-1 19
\ .
L N5 Number of da t a pairs in table 5, (time, letdown flow).
N(1) Number of va l ve po si t i on/ k- f ac t o r da ta pa i r s for va l ve
. number I. ,
Pl Pressure (psig) f rom BWST s t a t i c head just up s t ream of BWST i sol at t on va l ve 5 A.
[' P5 RCS pr e s s u r e (psi g) at HPI nozzles, int erpola t ed from table 2.
P6 Cal cula t ed MU tank gas pressure (psi g) . (user initial input) i _. P(2) Pressure at T connection between BWST and MU tank (psig).
P(3) MU pump suction pressure (p s i g) for A/B train. .
P(4) MU pump di schar ge pressure (psi g) for A/B train.
~
'P(7) HPI pump C suc t ion pressure (psi g) .
P(3) HPI pump C di scharge pressure (psi g) .
R Dummy variable used to simpl i f y the calculation of
-- MUT-BWST f l ow spl i t s . Co r r esponds to the ra di ca l term in the standa rd quadra ti c f ormula.
l R(1) Flow resistance of Ith re gi on , ps i d/ (I b/ sec) 2
,1 5 Option choice (user interactive input). (i nt e ge r , 1-10)
S(l ) Stroke time in seconds for va l ve number I.
Ti(1,3,K) Array of valve po s i t i on/K- f ac to r entries where: -
!= Valve number. .
3= Data pair location [1 to N(1 )]. Posi t i on J= 1 woul d no rmal l y co rrespond to 0% open, and posi tion 3=N(I )
would normally be 100% open. ,
i . K= Either I or 2. If K=1, then Ti(I,3,K) i s a va l ve position entry fo'r the I th va l ve . If K=2, TI(1,3,K) is the corresponding R-factor.
T2(1,3) Array of time in seconds (3 = 1 ) and RC pressure in psig (3=2).
- T3 ( 1, 3 ) Array of time in seconds (3= 1 ) and pump H/Q coefficient (3=2). (Determines whethe'r the H/Q correlation is for
_. one or two MU/HPI pumps. A & B pumps only.)
T5(1,3) Array of time in seconds (3= 1) and l et down f l ow i n gpm (3=2).
- T(l) I ni tial starting time for transient (seconds).
T(2) Time for next valve actuation (s eco nds ) .
T(3) Timestep (seconds).
U1 MU tank ga s vo l ume ( ga l l on s ) .
- U2 MU tank l iq ui d vo l ume ( ga l l on s ) .
VI MU t ank ga s vo l ume a t end of times t ep ( ga l l o n s ) .
V2 MU tank l iq ui d vo l ume at end of time s t ep ( ga l l on s ) .
V5 Vol une of wat er i n th e BWST ( ga l l ons ) .
~'
W(!) F lowra t e f rom BWST to A/B MU/HP1 t rai n th rough VSA, valve #1 (Ib/sec).
W(2) Flowrat e f rom MU tank to, A/B iMU/HPI .t rain through V12, valve #3 (!b/sec).
W(3) Flowrate through A/B MU pump (s) (Ib/sec).
W(4) Flowrate through HPI co nt rol valve 16B, valve #4, (Ib/sec).
U i
1 A-1.20 s.
en
~
'N W(5) F l ow ra t e through HPI co n t ro l va l ve 16A, valve #5, (Ib/sec).
W(6) MU/ HPI pump (s) reci rc f l ow th ro ugh va l ve #6, train A/B (ib/sec).
W(7) Letdown flowrate (lb/sec). (I nt erpolated f rom table 5, user input.)
{ W(3) Constant seat return flowrate (Ib/sec). (use r ini tial input is in gpm . )
W(9) MU flowrate through Vl7, th e no rmal MU co nt ro l valve u (valve #9) (Ib/sec).
W(ll) Sum of W(4) + W(5) + W(6) + W(9) (!b/sec).
W(13) Flowrate through C HPI pump (Ib/sec).
W(14) C HPI pump reci re flowra t e th rough val ve #2 (ib/sec).
- W(15) Flovrate through HPl co nt ro l va l ve 16C, valve #7, (Ib/sec).
W(16) Flowrate through HPI control valve 16D, valve #9, t_ (ib/sec).
W(21) Sum of W(14) + W(15) + W(16) (lb/sec).
X(1) Val ve po si t i on (0-100%) f or val ve numbe r I at be gi nni n g d of timestep.
b Y(I) Valve po s i t i on (0-100%) for va l ve numbe r I at end of timestep. .
1 l
l b
? ' %.
N._.) ,
c>
A-1. 21 6
F I GURE A.l.2 MU/ HP I PROGRAM BENCHMARK
~).
73 - -
Actual u
HU/HPi Model 75 ,
u ,o E
74 -
L_
w I
~
O 72 -
Z '
s.
s, a
g
> 70 -
\
w 5, J
M ',
L z 2, r
68 -
D 2
66 -
64 -
w', 62 , ,
/ W8 200 2$2 2$4 . 2d6 20S TIME (MINUTES)
V e
O h
e w
L APPENDIX 2 -
ENERGY BALANCE MODEL 40 .
O
?
1 j
J 6
/
9 A-2.0
\
l l
L .
- 1. Introduction A review of the reactimeter data and,other available data showed that a cooling of the RCS occurred at approximately 94 minutes after turbine trip
- at TMI-2 on March 28, 1979. A simplistic energy balance model was developed to examine this cooldown of the RCS. This appendix describes the
, energy balance inodel and the analyses performed, l
M 2. Calculational Technique and Assumptions l
L The energy balance model is an automated, hand calculational technique to
. examine the overall RCS temperature response to HPI injection and/or EFW I
delivery to the "A" steam generator. The model was used to examine system re.sponse starting at 94 minutes (5:34 a.m.) into the transient. The RCS
{
g was assumed to be 60% voided based on the CRAFT 2 prediction of the accident. Since the RCS experienced a continual depressurization between ,
94 and 101 minutes, the steam in the RCS was assumed to remain saturated-(
and control the RCS pressure, i "},
l The calculation was performed using one second time steps. The calculational procedure used is described below.
N
- a. Guess RCS pressure for the next time step. .
L
- b. Calculate new RCS steam and liquid masses accounting for
- HPI/MU addition,
- PORY leak flow,
- Steam condensation by HPI and/or EFW,
- Steam production by core decay heat addition, pressurizer heaters, s and RC pump heat,
- Liquid loss via letdown,
- Steam production via flashirig.
e ll lL ' -
O
~_
A-2.1
, and the calculated steam and liquid masses.
I Some of the basic assumptions for these analyses include:
j b. The RCS inventory in the "B" loop was 2200 ft.3 based on the CRAFT 2 L prediction. The liquid is assumed to be subcooled based on the cold leg temperatures recorded on the reactimeter. Until the RCS pressure -
decreased below 1082 psia, this liquid was not allowed to flash.
' c. The PORY flow was calculated by using the Moody Critical Flow Table for P saturated steam conditions.
- component behavior.
s j 3. Results b
l>
The sensitivity studies described below were performed using this model.
3 The results are sumarized in Table A2.1.
l
- l b
l w 3 a
[~ A-2.2
- 1) HPI Only
(~ .
This scenario is described in detail in Section 4.0. In this study, it was assumed that the RCS response resulted from HPI flow into the system. The HPI flows were varied to match the cooldown rate shown in
-- the TMI-2 data. The case that matched the cooldown rate is shown in Table A-2.1. Letdown flow was not modeled in this study. ,
i L.
- 2) EFW Only I
L This scenario is described in Section 5.0. In this study, it was assumed that the RCS cooldown response resulted from EFW flow to the
-- "A" steam generator without any HPI flow into the RCS. The study indicated that varying EFW with a flow of >300 gpm would be required to
- 4. . match the TMI-2 data. It was assumed that a constant flow of letdown (15.4 lbm/sec) occurred throughout the transient period (94 to 100 .
o minutes).
) 3) Combination of HPI and EFW This scenario is described in Section 6.0. In this study, it was assumed that the RCS cooling response was caused by a combination of HPI and EFW flow to the "A" steam generator. Results of this case indicate that 100 gpa EFW to the "A" steam generator with one HPI train actuated at 94 minutes and the second HPI train initiated at 95.5 minutes produce RCS cooling rates that exceed *those shown by the data.
'~
A constant 1etdown flow of 15.4 lbs/sec was assumed in this case.
b_
6.
. _- 'N
/
A-2.3 l
_ Table A-2.1. - Summary of Results Analyzing Scenarios Using th's Energy Balance Model Scenario
~
' 1 1.7994.0 minutes None Matches the data well 2 0 95.5 minutes 2 None 300 gpm Overpredicts the system to "A" pressure and temperature.
SG Undercooled the system.
Indicates EFW flows in excess of 300 gpm would be necessary to match the .
-- system response. -
j,,- ,'
3 1 HPI @ 94.0 minutes 100 gpm Overcooled the system 2 HPI 9 95.5 minutes ,
E mm h.=
b me em I
1 4
i
, A-2.4
w.
L.
I s
L L
i L
APPENDIX 3
(=
m HPI COOLING MODEL
( Ud .
l l
>[
d, l
e W
l t>
i A-3.0
3 L
- 1. Introduction s
Ai part of the TMI-2 litigation support, a task was undertaken to ,
6 investigate the effect of continued full HPI operation after 100 minutes.
Automated hand calculations were used to estimate the RCS pressure and a liquid inventory with HPI injection following the "A" loop RC pump trip at 100 minutes into the transient. The calculations were begun' at 100 minutes into the transient assuming the system was at saturated conditions with initial liquid inventories supplied by the CRAFT 2 predictions of the I
- accident. Assumptions included RCS pressure being 920 psia and the core being completely covered.
i 2. Calculational Technique and Assumptions The pressure calculation estimates were performed using a simplistic approach. There are two stages of the calculation: (a) the first 39
' system pressure. During the second. stage, repressurization of the RCS .
occurs as a result of the HPI injection. The calculations depend on
,_ detennining both the mass and volume of the steam in the RCS, which are functions of steam loss through the PORY plus RCS boiling, condensation, c.
and flashing rates.
Assumptions for Stage 1 (Depressurization)
- 1. The CRAFT 2-predicted best-estimate liquid inventory and distribution
-- values are used.
- 2. The "A" steam generator heat demand versus time is based on the actual TMI-2 transient. The heat transfer is calculated between 100-135 minutes assuming the heat transfer ceases once the primary and secondary liquid levels coincide (QSg_A = 12500 Btu /sec over 35
. minutes).
y A-3.1
t
mcond = mHPI Ceff (hf - hgpg)/hy g ,
The condensation efficiency for injection under the liquid level is L taken to be zero. The effect of the condensation efficiency was studied in a sensitivity study described below.
- 4. The HPI liquid created a subcooled liquid volume that was tracked into the core. The fraction of core heat added to the subcooled volume is
' ratioed based on the fraction of the core covered by the subcooled liquid.
u
, 5. The PORY flow is obtained using the homogeneous equilibrium model for i , saturated steam conditions at the existing pressure.
- 6. Core flood tanks were not used, even though the pressure ultimatelf fell below the actuation point.
i ;
Assumptions for Stage' 2 (Repressurization)
'- Once the PORY closes, the calculations simulate a simple repressurization from the HPI injection. As a conservative approach, a simple isentropic m compression of the steam was used without any wall Or interface heat transfer with the steam. During this phase, the core is subcooled and the steam ganerators will not be removing energy. The mass of steam remains constant and the volume of steam is calculated from the specific volume of the steam undergoing the isentropic compression.
- 3. Results The HPI cooling model was used to examine four cases. A sumary is provided in Table A-3.1 and pressure vs. time history for each case is shown in Figure A.3.1. The results indicate that HPI could not have been en after i 100 minutes because the HPI flow would result in rapid depressurization of the RCS. Thus, the HPI pumps were not actuated at the time of the "A" loop RC pump trip. -
l
.L
. A-3.2 l
A
.. .. - _ - - _ _ _ - _ - - _ - - \
V--- '~
U=- t," \F=- (F=== (F=- W tr - \; \;= - ; & r m y-( .
~
)
Table A-3.1. - Susunary of Results from the HPI Cooling Model Conditions at 139 Minutes When PORV is Closed Case Number of Qssg_A from 100-135 Condensation Pressure Steam Mass Total Steam Volume in the RCS Number HPI Pumps Minutes (Btu /sec) Efficiency (psia) (lba) Including the Pressurizer (ft 3) 1 1 12500 0.75 322 3339 4799 2 2 12500 0.75 ' 240 1051 2037
~
3 - 2 12500 0.5 2 95 1399 2188 4 2 6250 0.75 372 1632 2050 O
y -
N e
r*
Fi gu re A. 3.1
' COMPARISON OF THI-2 PRESSURE DATA .
t WITH VARIOUS HPl COOLING ANALYSES CASES 1800 -
- 3
- 2 l'600 - %
1 6 CASE I: 1 HPl 1400 - CASE 2: 2 HPI 1 CASE 3:' 2 HPI(CASE 2 WITH 50% CONDENSATION EFFIClENCY) '
CASE 4: 2 HPl (CASE 2 WITH 1/2 SG HEAT 1200 -
REMOVAL)
=
l ( e '
$ 1000 -
e .
2 TNI-2 DATA h
l -
1 800 -
600 -
Hi 400 -
200 -
p I I I I i i 6~
w' Off 100 120 140 160 180 200 i
Time (min)
A-3.4 Y . _ _ _ _ _
9 .
3 L
L L -
L l
APPENDIX 4 4
g GPUN QUESTIONS OF APRIL AND MAY 1983
% y
~
1 G d
u r
,t s:>,
A-4.0 Y:
p .
3 APPENDIX 4 i
w -
April 1983 Questions t
- 1. Provide the r.esults of each analysis performed by or at the request of
- BW of makeup tank /BWST response which indicates any change in the status of HPI (1nitiated, throttled or terminated) between 5:40 and l 5:50 a.m. In BW's opinion, do these results confirm whether HPI was or was not initiated on or about this time?
L 2a. Provide the results of each analysis performed by or at the request of BW (e.g. detailed computer analysis or energy balance calculations)
L that would indicate whether HPI was or was not initiated, throttled or terminated between 5:40 a.m. and 5:50 a.m. Explain how these results L are related to the operation of plant equipment (i.e., HPI and EFW)" .
Could the RCS and OTSG respnse at this time have resulted from operatin ,
h of the 1C HPI pug alone?' Please state your detailed basis for these conclusions. .
2b. Are there any sensitivity studies (e.g. as a result of your Task 27 or any other analysis) that indicate the effects of variation of HPI flow from 100 minutes on. What are the results of those studies?
6 f
P l v 3
i A-4.1
l . .
u
,_ May 20,1983 Questions
- 1. For each of three scenarios postulated by B3W for the 05:34 to 05:41 time frame provide:
- a. Assumed EFW flowrate vs. time and operator actions needed with the
- system pumps and valves to achieve these flows;
- b. Comparison of the predicted "A" SG secondary size level / pressure response using the actual plant data;
,, e. Assumed operator actions and time's with respect to the operation of'the HPI pumps, the HPI pump suction valves to the MUT/BWST, the recirculation valves to the MUT, the RCP seal injection and return valves, the HPI valves MUV-16A/B/C/0, and the normal makeup valve MUV-17;
- f. Assumed letdown flowrate vs. time, and assumed operator actions and
- times to change letdown flow and/or destination;
,, 9 Comparison of the predicted vs. actual RCS pressure /temperaute (TH/T); C
- h. Comparison of the predicted vs. actual MUT level;
'- 1. Comparison of the expected or predicted and the actual source and intermediate range detector response vs. time; w.
- j. Comparison of the expected or predicted and the actual RCP flow vs.
time; m
A-4.2 m
5
- k. Tha total additional quantity of water withdrawn from the BWST between m 05:34 and 05.40 above the figure in Appendix B to GPUN TDR-045 that was used as the 05:40 base ineeni.ory in the BW Task 27 study; I
- 1. An estimate or prediction of the delay in core uncovery and hot leg 6 superheat beyond the 05:46 point in time (predicted by the BW Task 27 results or any other available analyses) due to the additional quantity of water injection from item (k) above; I m. Analytical results in the form of graphs, plots or taoles which were used in supporting the conclusions of both the May 6 report and the responses to these additional questions; a n. BW's technical judgement on whether this scenario occurred given the actual RCS response.
- 2. With respect to the July 1980 Task 327 results mentioned on page 6 of the BW response, the HPI flow vs. time assumed in that work.
I i f 3. With respect to the analyses showing. sensitivity to HPI *flowrates referred to ati the bottom of page 4 of the BW response, .the EFW flowrstes vs. time assumed and a comparison of the "A" SG secondary side level / pressure
- response to the actual plant data. -
- 4. Consider, as a minimum, the following references in preparing responses 'to l these questions:
H
- a. GPUN Trial Exhibit 2222; L
- b. BW Task 27 Analysis Results, November 16, 1972; e
- c. Jones TMI-1 Restart Testimony at transcript 4626/8; b
- d. Jones TMI-1 Appeal Board written testimony of February 16,1983 at i 17/18; -
s V
l A-4.3
~ '
t.
_ e. Jones oral testimony at transcript 526/9; C f. B&W expert witness Wallis, GDUN vs. B&W trial testimony at transcript 6615/8 and 6635/17; *
- g. " Preliminary Calculations Related to the Accident at Three Mile Island."
g NUREG/CR1353, LA 8273MS, March 1980; L
- h. "Analysdis of Three Mile Island Accident ~and Attendant Sequences,"
Batte11it.Colunbus Ldoratories, NUREG/CR1219, January 1980;
{ i. "TMI-2 Accident Core Heatup Analysis." NSAC-24, January 1981.
50 State explicitly whether, in B&W's technical judgement, there was either a full or partial HPI initiation at 5:41 a.m.
a 3 -
~
3 1
g I
6 A-4.4
9 4 t>
, ~.
3 D
D L
APPENDIX 5 ,
I .
BRIEF SUWARY OF REFERENCES IDENTIFIED l {l IN GPUN QUESTION OF MAY 20, 1983 D
~
H k
1
$(?,
i c
b A-5.0
u APPEhDIX 5 I
.. 1 L .
- 1. Introduction i This appendix is a brief summary of the references identified in GPUN's questions of May 20, 1983. Each summary describes the part of the reference applicable to the HPI status at 5:41. Where possible, the cooling scenarios supported by these references are identified.
u
- 2. Summary of Reference s
2.1 GPUN Trial Exhibit 2222 l
This reference could not be specifically located. It is our understanding that this exhibit is the BE Task 27 Results described, 5 below.
- 2. 2 BE Task 27 Results, November 16, 1972 BW's Task 27, TMI-2 RCS Component Evaluation, made the following f) assumptions with regard to MU/HPI flows:
Makeup Flow Rate: 50 gpm (22 gpm through the HPI/MU nozzle and 28_ gpm through the pump seals.
a HPI Flow Rate: 101 - 165 minutes = 70 gpm 4 165 - 200 minutes = 0 gpm 200 - 210 minutes = 760 gpm The enphasis of the Task 27 analysis was to examine RCS component respnse after all the RC pumps had been tripped (after 101 minutes),
a Thus, the analysis results in the Task 27 report provides no insight to the cooling mechanism between 94 and 101 minutes. After 101 minutes, HPI flow rates were minimal and the overall plant response (RCS v,
A-5.1
L.
m pressure) was governed by the EFW. cooling effect. The conclusions m drawn in the main body of this report also support EFW as the main j cooling source after 101 minutes. .
I 2.3 Jones' Testimony s
2.3(a) Jones' TMI-1 Restart Testimony at Transcript 4626/8 This testimony relates to the observed system response at TMI-2 following the tripping of the RC pumps at 101 minutes. The oral statements basically imply that following 101 minutes, the RCS cooldown was being controlled by EFW. This is consistent
?] with the conclusions drawn in the main body of this report.
The statements in the testimony do not address the observed cooldown between 94 and 101 minutes.
2.3(b) Jones's TMI-1 Appeal Board Written Testimony of February 16, o 1983 at 17/18
,b As in the TMI-1 restart testimony discussed above", this testimony only addresses system behavior after all RC pumps had been tripped. The appeal board testimony also addresses the
- effect of actuation and maintenance full HPI after the RC pumps 1
had been tripped and concluded that no core damage would have occurred. Even if one assumes that full HPI was activated.,at
[
l 101 minutes, the BWST level response for the accident suggests that HPI could only have been maintained for approximately six minutes.
b 2.3(c) Jones' Oral Testimony at Transcript 526/9 P
l As in the testimony references above, this discussion mainly L centers on the system response after all RC pumps had been tripped. However, on transcript page 528, there is a s discussion of the cooling mode between 94 and 101 minutes which b
l A-5.2 i
N
t claims EFW was being delivered to the "A" SG. Thus, the L. '~
transcript supports Scenario 2.
It should be recognized that the SG response analysis presented i in Section 5 of this report was not available until after the b TMI-1 ASt.AB proceedings. The SG response analysis concludes that while EFW was probably being delivered between 94 and 101 minutes, its rate was probably insufficient to cause the observed plant cooldown. With this additional work, it still appears that the Jones' testimony is fundamentally corfect, i.e., the SG was removing energy during the period. However, supplemental cooling via HPI (Scenario 3) is required.
s 2.4 B&W Expert Witness Wallis, GPUN vs. B&W Trial Testimony at Transcript s 6615/8 and 6635/17 ,
Dr. Wallis' testimony at these two reference points appear to deal
~') with:
(a) A "what if" event in which HPI is actuated at about 4:45, a.m., 'and (b)' . At 6635/17, an additional "what'if" assumed HPI at 6:00 a.m.
'- Sumary .
improve. Translating to 5:34, the point here is that HPI, if actuated could be expected to improve RC pump performance. This then is an argument which does not support Scenario 1, HPI only, as the cooling mode between 94 and 101 minutes. At 6635/17 and following, Dr. Wallis is responding to another "what if" HPI is actuated question. It can be
-. implied from his response, that HPI will enhance the cooling in progress (from 5:34 to 5:41 a.m.) by the EFW. This is consistent with Scenario 3. - '
A-5.3 l - ____ _ _ _ _ _ _ _ _
L. .
- u. 2.5 " Preliminary Calculations Related to the Accident at Three Mile Island," NUREG/CR 1353, LA 8273MS, March 1980 L. .
This reference is a Los Alamos Scientific Laboratories (LASL) report on the Computer Code TRAC analysis of the TMI-2 Accident. The TRAC analysis assumes the cooling of the RCS that began at 94 minutes is by 500 gpm of EFW. The TRAC analysis provides reasonable agreement with
- the observed cooling response. Thus, the TRAC analysis is similar to Scenario 2, EFW only.
2.6 " Analysis of Three Mile Island Accident and Attendant Sequences,"
Battelle Columbus Laboratories, NUREG/CR 1219, January 1980 s
j This reference is the analysis of the TMI-2 accident by Battelle
' Columbus using the MARCH code. The Battelle Base case analysis assumes that:
6.
,. Net MU/HPI flow in the period of interest = 25 gpm jj (165 inje'ction,140 letdown from Table 2-3)
' Cooling at 94 minutes post trip and beyond is by EFW (Section 4.2, page 4.7 of NUREG/CR 1219)
The MARCH Code analysis using EFW in the period of interest agrees well l
with the observed response, based on Figure 4.1 of NUREG/CR 1219.
P The Battelle base case is also similar to Scenario 2, EFW only.
L 2. 7 "TMI-2 Accident Core Heat Up Analysis", NSAC-24, January 1981 The pertinent points in NSAC-24 related to HPI status at 5:41 a.m. on 3/28/79 include:
k Makeup and Letdown Operations Calculated Makeup and Letdown Flows Iv -
l w
A-5.4 e
L BWST Inv:ntory Ba' lance RCS Inventory
.m Activation of the "C" Makeup Pump w
A summary of NSAC-24 points follows.
- Figure 5 of the report presents calculated MU and letdown flows from 100 to 205 minutes post trip. During this period, the makeup flow is o-estimated to be about 127 gpm and the letdown flow is estimated to be about 110 gpm suggesting a net makeup of about 17 gpm to the, nozzles plus 36 +4 gpm to the pump seals. Total makeup is about 57 gpm.
, - BWST inventory unaccounted for is about 4800 gal, assumed to be injected in the period 73-166 minutes.
- Makeup /HPI Pump C is estimated to be actuated at about 112 minutes at .
100 gpm until about 160 minutes post trip. .
w The NSAC-24 analysis is generally consistent with Scenario 2, EFW only
- (e by inference. It is specifically not consistent with Scenario 1, HPI only, beca.use the unaccounted for BWST in'ventory is assumed to be by
,. the "C" MU/HPI pump, operated at throttled flow, in the period from 112 to 160 minutes post trip. .
- 2. 8 Additional
Reference:
NUREG-0600, " Investigation into the March 28, 1979 Three Mile Island Accident by Office of Inspection and
~
Enforcement, Investigative Report No. 50-320/79-10," August 1979 On page 1A-35, item 192, the NRC sequence of events indicates at about 93 minutes post trip that " Operators report increasing HPI flow, RCS pressure per.ks and drops rapidly, SRM and IRM indication spike upward, and injection flow was maintained at 150-200 gpm/ loop."
This entry, apparently based on interviews, tends to support Scenario 3, Combination of EFW and MU/HPI.
M -
N.
A-5.5
tV ,
SHAw, PITTMAN, POTTs & TROWBRIDGE A JeARTNEmSwep OF pmOFE SSsON AL CompOmATiONS 4800 M STRE ET. N. w.
WASMf NGTON. D. C. 20036 (202) e22-ioco meMSAv O pOTTS.
- C. SMELOON J. wEiSEL.
- C. sg,,,,,g,y,,go, g ow.,o o, ,oy,o, m CTJu1RT L piTTMAN. p C. JOMN A. McCULLouGM.
- C. _ JACM MenAv ANonEw0.ELuS C% nGE F. TooweniOGE p C. J. pArmeCu Mecntv.
- C. TMCMAS
- McConM*Cm m*CM Amo A. SAMp CTG PMEN O. PCTTS.
- C. J. TMCMAS LENMamT.
- C. TELECoptER JCMM L CAmm. JR. TMOM AS E. CROCp Em J m.
GEmALOCManNOFF *C. STEVEN L MELTZEm. # C. DM eup J. M A RvEv wENOguN A, wMiTE
- MeLup O. SOSTwsCK. p C. DEAN O. AuuCM.
- C. (202) 822-8099 & S22-H99 mosENT M. GOmOON STANLEY M sARG O. TIMOTMv MANLON.
- C. JOMN ENGEL
- C. _ SAMSARA J. MONGEN LF SUE R. SMtTM ff Omot M. noOEmS, Jp.. p C CHARLES s. TEMRIN.
- C. SONNIE S. GOTTUE S FmE A A. uTTLE, p C. STEPMEN s. MuTTLER p C. mapirAm ggo MowamO M. SMAFFERMAN y,po,iNsA gay g ,,g S. muTLEOGE
,,g ,ggg JOM .6 O. RMINELANDEn.
- C. WINTHRCP N. SmCwN. p.C. OEBONAMs.sAuSER JANiCE %EMmEm STEIN smuC]w.CMupCM*LL*C. JAM E S e. MAMuN. p C. (2021822 sO7E SCOTT A. ANENSEmG TRAVIS 7. SnowN. JR.
LESua A. NiCxOLSON, JR.. p C. mANDAL 5. RELL. A C. _
CAM DS EU. Mf LLE FE R meCMARO M. Mm MAmTo D. aRALL.
- C. mostNTE.2AMLER.* C. Dau6M.TMOMAS ,7g ,,g , ,, ,,C,,NTMAL m'CMAmOJ.mENOALL pC. NCSE RT s. ace ssNS.
- C. TELEX SETM M. MOOGASIAN JAv R. SILBERG
- C. STEVEN M. LuCAS. p C. SMEILA MCC. MAmVEY SANOmA E. SRUSCA*
sa meAmA M. noS SOTTi. p C. DavtO M. mumE NSTEiN. p c. a S.2e es (S MawtAw wSM) OEuSaA A. miOGwar E'LEEN u mmOwNELL f ' OmO E V. ALLE N. J m.
- C. R'CHARO E. GALEN RENNETH J. MauTMAN PAM EJ M. ANDERSON Fat # OftaSNE R.
- C. LvNN wMITTLESEv wtLSON CABLE.,SMAwLAW.. OAvs0 LawmENCE MeLLER ALEXANDE R O. 70M ASZC2um FREDEmeCK L MLEIN *Mtup D. PORTER O. K NLY wt BSTE R. p C. M ATIAS F. TmavetSO-OsA2 ~
NAYM.' NoE L p e nE E D. J a.. p C. vtCTomeA J. PEmmens STEVEN p.StTLEm' M8CM AEL A SW'OI R Mima K uGENSUCK.
- C. JCMN M. O NEILL JR. alCMamO J. PAmmi ELLEN SMEntFF E RN 7.T L SLARE Jm.. p C. JAv A. EpSTIE N
"" " e ELLEN A. FRECEL.NO ANe?A J, FINRELSTEIN CALL 3 TON 3. JONES p C. mANDL. ALLEN MANNAM E. M. UESE mM AN EtLE EN M. GLEIMER TMOMAS A. SasTER.
- C. EUSABETM M. PENOLETON COWNSEL SANOmA E. FOLSOM OAveOR.SAMR J1M '.O M SupG E R. P C. mammy M. GLASSp*EGEL JuCITH A SANDLER C. BOwOOsm TRAIN
- e . .mi m c.
September 22, 1983 we,,Em S O.mECT O.Am ou.E m (202) 822-1084 Harold R. Denton, Director Office of Nuclear Reactor Regulation U.S. Nuclear Regulatory Commission Washington, D.C. 20555 Re: Meted (TMI-1), Docket 50-289
Dear Mr. Denton:
I enclose for your use in the ongoing review of the GPU v. B&W litigation documents, a copy of GPU Response
~
to "TMIA Interim Comments on B&W Trial Record." This memorandum was prepared by GPU's litigation counsel Kaye, Scholer, Fierman, Hays and Handler in response to TMIA's Interim Comments, which were filed on July 1, 1983.
Sincerely, SHAW, PITTMAN, POTTS & TROWBRIDGE hwl l. /NAbjk Ernest L. Blake, Jr.
Counsel for Licensee cc: Mary E. Wagner, Esq.
i
l
- 1
= &
UNITED STATES OF AMERICA NUCLEAR REGULATORY COMMISSION -
BEFORE THE COMMISSION In the Matter of )
)
METROPOLITAN EDISON COMPANY ) Docket No. 50-289 g q, (Three Mile Island Nuclear ) '
a '
't Station, Unit No. 1) ) g
[7I
- e. i r8
- SEPg ggg, E*
'La-0 S p. P. & T.
/sb"
% $<9 ktl s GPU RESPONSE TO "TMIA INTERIM COMMENTS ON B&W TRIAL RECORD" i
September 21, 1983
. L TABLE OF CONTENTS Introduction........................................... 1 April 23, 1978 overcooling Event....................... 2 PORV Problems.......................................... 11 PORV Maintenance.................................. 11 PORV Indicator Light.............................. 14 Tailpipe Temperatures Analysis.................... 16 Alleged Leak Rate Falsification................... 21 Hartman Cross-Examination.........,................ 22 Condensate Polishers................................... 24 Water in Air Line................................. 26 The Keaten Task Force.................................. 28 Maintenance............................................ 39 Maintenance Budget and Staffing................... 40 Maintenance Staffing......................... 41 Budget Cuts..................... ............ 42 Repair Parts...................................... 43 Paperwork......................................... 45 Overtime.......................................... 45 Daniel Shovlin.................................... 46 Training............................................... 54 Richard Zechman................................... 57 The Book Letter................................... 59 Management Structure.............................. 62 Managers.......................................... 62
. a .
LICENSEE RESPONSE TO "TMIA INTERIM COMMENTS ON B&W TRIAL RECORD" Introduction TMI ALERT filed on July 1, 1983 its "TMIA Interim Comments on B&W Trial Record" (TMIA Comments). As demonstrat- ,
ed in detail below, the TMIA Comments distort the trial record and repeatedly mischaracterize both the oral testimony and documentary exhibits.
The TMIA Comments are divided into two main sections in which TMIA purports to show that (1) "the B&W record sup-ports the argument that the Licensee is inherently incapable of learning from past mistakes or correcting recognized prob- ,
lems"; and (2) "the B&W record, read in light of other new in-formation which has come to light reflecting on Licensee's management, fundamentally undermines the credibility of the ASLB decision, and the PIDs cannot be used to lawfully justify restart of TMI-1." The TMIA Comments fail to establish either of these propositions. When analyzed, the Comments prove to be primarily a collection of unsupported and unsupportable as-sertions or misstatements about the facts regarding past GPU Service Corporation and Met-Ed (TMI 1 and 2 Licensee at the time) actions which fail to sustain TMIA's arguments that Unit 1 should not be restarted and that the ASLB decision should be overturned.
. t . .
April 23, 1978 Overcoolinq Event In the first section of the TMIA Comments (after the l introduction), TMIA argues that some of the B&W trial evidence showed that Licensee was " incapable of learning f rom past mis-takes." (Comments, p. 4.) Its first purported example is the April 23, 1978 overcooling event at TMI-2. This is an inci-dent which occurred when the main steam safety valves on the secondary side of Unit 2 failed to re'close following a reactor trip. The resulting loss of energy from the secondary side caused the temperature and pressure to drop in the reactor coolant system. This was a fundamentally different transient l from the Unit 2 accident where the pilot-operated relief valve (PORV) on the primary side pressurizer failed to close result-ing in a loss of coolant accident. In contrast to the March 1979 accident, there was no steam or water release from the primary side of the unit during the April 23, 1978 event.
,The TMI A Comments' position is that management per-sonnel currently with GPU Nuclear and who were with Met-Ed or ,
i GPUSC in. April 1978, failed to adequately analyze and respond to that plant event. They endeavor to support that assertion with discussions on three primary points:
- 1. The event showed pressurizer level could in-crease while pressure was decreasing and "because Licens-i ee presumably failed to attach significance to th'is as-pect of the transient (see, e.g., Keaton dep. at 225), it did not bother to modify emergency procedures or training 2
_ -- - - - - , - - , - - - - . _ _ . - . - - - ?-
i <
- l
. s to instruct operators what to do in the event the system experienced the condition again." (Comments, p. 7);
- 2. Neither Robert Arnold (President of GPU Nuclear Corporation, and then Vice President-Generation, GPU Ser-vice Corporation) nor John Herbein (then Met-Ed Vice President-Generation ~ but no longer with GPU Nuclear)
" learned tha most fundamental lesson from the event."
(~ Comments, p. 7); and
- 3. " Operators were extremely confused by the number of alarms which went off at the time of the event" and the response by Met-Ed and GPUSC was unsatisfactory.
(Comments, pp. 7-8.)
Assessing each of these specific issues requires an understanding of the response that management did make to the April 23 event. As indicated by the TMIA Comments (p. 7),
management at the time " considered the event very signifi-cant." A special task force was set up and chartered to con-duct an extensive review and analysis of the event. (See Ap-pendix A, attached, Memorandum from R.C. Arnold to R.W. Kea-ten, Re: TMI 2 Cooldown Transient of April 23, 1978, May 2, 1978.) As can be seen from Appendix A, the charter was very comprehen.sive and called for a thorough and objective ef fort by appropriate technical personnel. Because of the work al-ready completed or underway, the task force was able by May 9, 1978 to make twenty-three (23) specific and substantive recom-mendations to reflect lessons learned from the event. The 3
recommendations were given priorities according to categories:
prior to restart, near term and long term. All were included in the final task force report, as was an additional recommen-dation for more extensive post-trip monitoring capability.
(GPU Service Technical Data Report TDR No. 001, B&W Ex. 186.)
(The latter recommendation in fact resulted in the addition of equipment which provided much of the data utilized for analy-sis of the TMI-2 accident.) A complete review of the record shows that management's response to the April 23, 1978 event was thorough and professional. That is not to say that one cannot go back to that event and find the footprints of other lessons since obtained from subsequent events. However, the management competence issue is whether a responsible and tech-nically competent effort was made to learn from the April 23 event and to utilize those lessons to improve safety, not whether 20-10 hindsight can identify traces of other insights that might have been gained if the organization had recognized the significance of relatively obscure peripheral details.
Regarding the first point, both before and after the April 23 event, B&W instructed users that the pressurizer's water level was the appropriate and accurate indicator of wa-ter inventory in the reactor coolant system (Lind trial tr.
5227-28; Frederick trial tr. 4022-4023; GPU Ex.'2347, pp. 61,
. 74; Marzec trial tr. 6398-99). Throughout the April 23, 1978 event, the water level in the pressurizer in fact continued to ,
be a reasonable -- though possibly not perfect -- indicator of 4
-. x. . -
the volume of water in the reactor coolant system and in fact dictated correct operator action. Scrutiny of this overcool-ing event by not only Met-Ed and GPUSC but also by the NRC and B&W, failed to produce any suggestion that the pressurizer should be abandoned as a guide to water inventory in the reac-I tor coolant system. -
During the April 23 event, the operators monitored pressurizer level and regulated high pressure injection in ac-i cordance with their training and plant procedures with no del-eterious effects. As pressure and pressurizer level fell, an operator promptly (43 seconds after the reactor trip) started l
1 a second makeup pump and increased make-up flow through the high pressure injection flow path. At one (1) minute, eleven (11) seconds after the trip, high pressure injection was auto-j matically initiated and full high pressure injection flow would have been established if it had not already been accom-
] plished as a result of prompt operator action. Pressurizer level had by then dropped below the range of level indication.
At two (2) minutes and thirty (30) seconds into the event, about twenty (20) seconds after pressurizer level was back in the indicating range and thirty (30) seconds before pressure l started to increase, the operator began to reduce the flow of i high pressure injection pursuant to his training and proce-dures. For this event, the operator's manipulation of high pressure injection was accepted as appropriate by all who re-5
viewed the event including B&W. Thus, the existing procedures and training had directed a correct operator response.
I B&W analyzed the April 23 event extensively, specif- l ically commented favorably on the operator *. imndling of the high pressure injection system (B&W Ex. 186, TDR 001, p. A4-3) and attached a graph showing the points where the high pres-
~
sure injection was initiated and throttled on April 23 in the same manner as would be done eleven (11) months later, on March 28, 1979. Thus, the April 23, 1978 event, far from be-ing an event that should have alerted GPU to improper proce-dures or operator action, was instead an event that appeared ;
i to confirm the appropriateness of existing training and proce-dures.
Subsequent to April 1978, both the NRC and B&W ana-4 lyzed a number of different overcooling events at various i plants, including the April 23 event at TMI-2. On February 14, 1979, the NRC convened a meeting at B&W's Lynchburg head-quarters specifically to address the issue of pressurizer lev-el indication. In attendance were several top technical peo-ple from B&W, including Bert Dunn (author of the B&W internal memorandum which stated that the Davis-Besse event showed that training and procedures for LOCA's were incorrect) and also representatives of Met-Ed and other owners of B&W plants. Mr.
J. E. Foster of the NRC, who conducted the conference, " stated that as far as he was concerned loss of pressurizer level in-
- i. dication was merely an operational inconvenience and the loss l 6 l
I._ - .. . . - , - - - .
1 . .
of pressurizer level was not a safety concern. He was recom-mending that this issue be closed." (GPU Trial Ex. 84, inter-nal B&W memorandum, March 9, 1979.) When asked about his j failure to express any safety concerns at this meeting, Bert Dunn stated that the issue of loss of pressurizer level in j l
these overcooling events was a separate issue from his concern
- that operators would secure HPI based on misleadingly high j pressurizer level indications during a loss of coolant acci-dent (Dunn NRC Special Inquiry Group Deposition, 10/4/79, pp.
63-64).
Thus, both the NRC and B&W concluded that overcool-ing events such as the April 23, 1978 event did not produce I any significant revelations and therefore the existing proce-dures did not need to be revised. In iight of these analyses performed by engineers from several different organizations, TMIA's assertion that Met-Ed and GPU Service Corp. should have 1 recognized the need to revise its pr,ocedures for loss of cool-ant accidents after the April 23 overcooling event is simply i not supportable.
The TMIA Comments state at pages 6-7 that during the April 23 event, the pressurizer level increased while pressure dropped, a " condition . . . plainly not contemplated by . . .
procedures or training." However, as pointed out above, no one -- including the NRC or B&W -- deduced anything of signif-icance from the very brief period, of no more than a minute's duration, when pressurizer level turned upward slightly in ad-7
J vance of pressure also turning upward. A review of the graph attached to the B&W analysis of the April 23 event (B&W 186, TDR 001, p. A4-4) reveals that pressure and pressurizer level i
did trend in the same direction during that event, as expect-ed. The high pressure injection flow added water and caused pressurizer level to increase. The fact that an increase in pressure followed less than a. minute later rather than simul-taneously did not appear to the NRC, B&W or GPU to undermine the then current understanding that pressure and pressurizer level would trend together during a LOCA. Indeed, the minimal delay in the recovery of pressure during the April 23 over-cooling event -- a delay so slight that it would not even have been discernible on the control room recorder chart and which resulted in no incorrect action on the part of the operators
-- is an entirely different phenomenon than a sustained low pressure condition in conjunction with a high pressurizer lev-el such as occurs during a LOCA at the top of the pressurizer.
TMIA's own comments show responsible action by man-agement in responding to the April 23 event. As TMIA points out, GPUSC considered the April 23 event significant and formed a task force to study the event and to supplement a Met-Ed report on the transient (Comments, p. 7). These ef-forts by GPUSC and Met-Ed management to assess the overcooling event demonstrate the importance which management placed on learning from prior transients and illustrate the steps which it took to respond to significant plant events.
8
TMIA's assertion that GPU Service Corp. did not re-port technical specification violations occurring on April 23 ,
to the NRC (Comments, p. 10) is misleading. The Metropolitan Edison Company had reported April 23 technical specification violations to the NRC (L.E.R. #78-33, May 8, 1978; L.E.R. #78-34, May 8, 1978).
The citations from the B&W record by the TMIA Com-ments do not support its second point -- that neither Arnold nor Herbein to this day have learned "the most fundamental lesson from the event." TMIA states:
"And Arnold still maintains in testimony that '(a)ny time we were identifying a loss of coolant accident. . . . the pres-surizer level and pressura would both ,
trend together' . . . . Thus, not only did the President of the company fail to learn the most important lesson from the April 23 transient . . . he failed to learn the very same lesson from the acci-dent itself. This perhaps says more about the Licensee's fundamental inability to learn from the past . . . than any other single incident in this trial record."
(Comments, pp. 7-8, emphasis in original.) TMIA has misread Arnold's testimony, attempting with the words "still. maintains in testimony" to make it appear that Arnold "still" believes that pressure and pressurizer level trend together during all loss of coolant accidents. The entire context of the testimo-ny quoted by TMIA indicates that Arnold was referring only to his pre-a'ccident understanding -- which was consistent with the understanding of the industry generally. (See, e.g., GPU Ex. 132, B&W internal memorandum.)
9 j
o *
- The third point associated with the April 23 event that is identified in the TMIA Comments concerns management action regarding control room alarms. TMIA states that during the April 23 event and during the Three Mile Island accident, operators were confused by the number of alarms (Comments,
- p. 8). TMIA alleges (p. 9) that an alarm window correction program, recommended after the April 23 event, was never put into effect.
TMIA correctly asserts that an alarm window correc-tion program had been recommended; but it incorrectly asserts that the alarm window correction program was " explicitly dis-approved by Met-Ed corporation management, B&W 767, and thus never put into-effect." (p. 9.) The only cited reference for this assertion is B&W Ex. 767, which' concerns the PORV indica-tor light and has nothing to do with the alarm upgrade pro-gram. Indeed, testimony established that the alarm upgrade program was put into effect and that it remained in effect, as an ongoing project, up to the day of the accident (Zewe dep.
845; Zewe trial tr. 2180-82; Frederick trial tr. 3567-68; GPU Ex. 2067).
TMIA does not support its assertion (p. 8) that op-erators, during the 1979 accident, were " extremely confused" by the number of alarms that went off during the accident.
Mr. Zewe did testify that the sheer number of alarms present on the day of the accident may have interfered with his abili-ty to prioritize available information (Rogovin dep. 177-78, 10 l,
1
- _ - - - - - - - - - - - - - - - - - - - - - . - - _ - - _ - - - . - _ _ - - - _ - - - - - -j
, 9/11/79). However, the presence of a large number of control room alarms during a major transient was the norm rather than the exception in power plant control rooms prior to the acci- ,
I dent, and operators received training at the B&W simulator as to how to systematically analyze large numbers of alarms.
Thus, Mr. Frederick specifically testified that he had been trained by B&W to analyze large groups of alarms and so had not been confused by the alarms present on the day of the ac-cident (Frederick trial tr. 3278-79, 3568-69, 3224-25, 3271-l 72).
I l PORV Problems In its section entitled "PORV Problems" (p. 11),
TMIA jumbles together a number of loosely related assertions. -
The gist of TMIA's allegations appears to be that the PORV was poorly maintained, that the PORV was leaking before the acci-dent, that management purposely ignored this alleged leakage, and that operators falsified leak rate data in order to keep the plant on line. As described below, TMIA's citations to the trial record fail to support these propositions and, in-deed, ignore the substantial body of contrary evidence.
PORV Maintenance TMIA states at page 11 of its Comments that the PORV had a history of problems. Thus, TMIA points out that the PORV experienced some leakage in 1977. However, as the Com-ments also note, the valve was repaired and the leakage cor-rected (Comments, p. 12). Indeed, Sieglitz specifically tes-11
tified that the valve was redoved, sent to the TMI machine .
shop where mechanical " lapping" was performed, tested, and confirmed to be satisfactory (Sieglitz trial tr. 5764-65). As is discussed further below (pages 16-21), although there was some mention of PORV leakage in post-accident interviews -- at a time when attention was very much focused on the PORV --
there is no scientific evidence of PORV leakage after the 1977 repair of the PORV and before the day of the accident. To the contrary, the evidence reveals that a code safety valve rather than the PORV was leaking.
The TMIA Comments also note that the Unit 2 PORV had at one time been installed at Unit 1 (Comments, p. 11). TMIA fails to mention Met-Ed's action in sending the valve back to -
its manufacturer, Dresser, to be refurbished and retested af-ter this initial use at Unit 1 (Technical Data Report No. 160
("TDR 160"), PORV Investigation by Q. Billingsley and J. Cor-rea, 7/8/80, p. 14, B&W Ex. 456). This action by Met-Ed and the valve manufacturer indicates, as does the 1977 repair of PORV leakage, continued responsible activity by Met-Ed with regard to PORV maintenance.
TMIA further states, at page 11, that the difference in voltage at Unit 1 and Unit 2 may have damaged the PORV.
TMIA asserts, as if it were an established fact, the possibil-ity raised in a GPU technical data report (TDR 160 at p. 10) that the PORV may have been damaged by operation at a higher voltage than it was designed for during its temporary instal-12
lation in Unit 1. It is speculative as to whether a higher voltage supply to the Unit 2 PORV harmed the PORV. Moreover, as is discussed above, after its use at Unit 1 and prior to its reinstallation at Unit 2, the PORV was completely over-hauled and tested for any problems by its manufacturer, Dress-er Industries. It was returned to the site in "like new" con-dition and in conformance with the technical and quality as-surance requirements imposed on the original purchase order for the valve (TDR 160, Appendices 5 & 6, B&W Ex. 456 at pp.
11353, 11355). Thus, even if there had been any damage as a result of the temporary installation of the valve on Unit 1,
~
l
- the return of the complete valve, including the solenoid, to Dresser IndustrieG should have resolved any such problem. -
The TMIA Comments also assert (p. 11) that the al-leged difference in voltage was "never believed significant enough" to warrant discussion at the plan of the day meetings.
! However, in the testimony of Mr. Sieglitz cited in s'upport of this proposition, Sieglitz merely agreed with the questioner that the issue of different voltages had not arisen in his presence at plan of the day meetings.
TMIA also contends (pp. 12-13) that the frequent cy-cling of the PORV could have had an adverse effect on PORV re-liability. The cited Keaten testimony does not support that proposition. Moreover, even if in theory the frequent cycling of a valve can damage a valve, there is no evidence that the cycling of the Unit 2 PORV during the April 23 event in fact 13
caused damage of any kind. The PORV functioned normally dur-ing several subsequent Unit 2 avents prior to the accident and, as is discussed below (pages 16-20), there is no convinc-ing evidence that there was leakage through the PORV following 1 .
that event.
PORV Indicator Light .
The TMIA Comments allege (p. 12) that a "better" I PORV indicator light was requested in 1978 and that this re-quest was denied by Met-Ed corporate engineering. TMIA tries to create the impression that Sieglitz had recommended a change in the PORV light and that " corporate engineering" had rejected the request out of hand, shortly before the accident.
This assertion is substantially misleading. -
The PORV position indication instrumentation which was added to the Unit 2 control room following a March, 1978 event had been installed only after review by several engi-neers within B&W's Lynchburg headquarters (GPU Ex. 368; Rogers trial tr. 5462-72) and was wired to the PORV solenoid. A mod-ification of the position indication mechanism to employ a limit switch was proposed within Met-Ed in December, 1978 as a result of PORV testing (referred to as an " incident" in the ,
TMIA comments, p. 12). As Sieglitz testified, an engineering decision was made to reject the proposed modification because it wss not a better alternative (Sieglitz trial tr. 5804).
Richard Noll, an engineer in the Met-Ed Generation Engineering Department, carefully reviewed the proposed modification, 14
spoke with other engineers regarding the proposal, and con-cluded that the modification was not an improvement over the existing PORV position indication mechanism (Seiglitz trial tr. 5804; Noll cep. 92-93, 96; B&W Ex. 767). Thus, the change was not effected.
TMIA further misrepresents the nature of the deci-sion regarding the proposed change in the PORV indicator light
~
by implying that Sieglitz made a request for operator training on the proposed modification and that that request was turned down. In fact, the change request submitted to Sieglitz sim-ply indicated that if a change to the PORV light were made, the operators should be trained on the modification (Sieglitz tr'ial tr. 5799-5803; B&W Ex. 767, p. 4). As Gary Miller tes-tified, operators were trained on virtually all plant modifi-cations (Miller dep. 298).
Since the accident, and as a result of the Lessons Learned from the accident, the TMI-l control room has been up-graded by installing devices which measure flow past the PORV and the pressurizer code safety valves -- an approach prefera-ble to either the. prior indication or the December, 1978 pro-posed indication.
g 15
Tailpipe Temperatures Analysis In a somewhat garbled discussion of the tailpipe temperatures (Comments, pp. 12-18), TMIA appears to conclude that simply because the PORV tailpipe temperatures were ap-proximately 180 prior to the accident, (a) the PORV must have been leaking and (b) Met-Ed, and S'ieglitz in particular, were negligent for not investigating and repairing this alleged leakage.
A brief summary of the trial record regarding leak-age from valves at the top of the pressurizer indicates that Met-Ed engineers reasonably determined before the accident that a code safety valve and not the PORV was leaking before the accident. This was confirmed after the accident by GPUSC engineers. A further elaboration of this information may be found at pages 34-37 of the Licensee's response to the NRC no-tice of violation, B&W Ex. 707.
The testimony of Sieglitz and others establishes that leak rates and tailpipe temperaturas were routinely moni-tored prior to the accident (Sieglitz trial tr. 5719-20, 5746; Zewe trial tr. 2847-48; GPU Exs. 2095 and 2096). For months prior to January 1979, temperatures on the PORV tailpipe had been consistently running in the 180 range (Sieglitz trial tr. 5746-47; GPU Exs. 2095 and 2096). During this period, there'was minimal leakage from the reactor coolant system (Sieglitz trial tr. 5746-47; Ex. 2095).
16
l' .
The fact that only minimal leakage existed for months while the PORV tailpipe temperature was in the 180 range confirmed that a 180 temperature was normal for the PORV discharge pipe on Unit 2. The 180 reading was also con-sistent with the configuration of the PORV thermocouples in relation to the pressurizer and other hot equipment in the same area of the containment building (Sieglitz trial tr.
5734-38, 5747). Thus, the persistence of a 180 temperature while there was no significant identified leakage meant that 180* was not an indication of PORV leakage (Sieglitz trial tr.
5809).1 Following a two-week outage at TMI-2 in January, 1979, there was simultaneously a marked increase in the iden-tified leakage and in the temperature of the code safety valve discharge pipes. As the TMIA Comments indicate (p. 14), after that outage, the code safety valve discharge pipe temperatures jumped to 200 from a previous range of 110 to 120 . There was a prompt discussion of this temperature rise at plan of the day meetings (Sieglitz trial tr. 5713, 5719-23). In light of the sudden appearance of significant identified leakage and 1 The leak rate involved here is the " identified" leak rate and not the " unidentified" leak rate which was the main focus of allegations by Harold Hartman. See p. 21, in-fra. All leakage going past either the PORV or pressur-izer code safety valve seats was collected in the Reactor Coolant Drain Tank and measured. Thus, such valve leak-age has no direct relationship to the "unide,ntified" leak rate.
17
elevated temperatures on the code safety valves, it was deter-mined that the higher temperatures and leakage were caused by a leaking code safety valve (Sieglitz trial tr. 5716, 5719-23; Exs. 2230, 2231, 2232, 2233).
As a result of the determination that a code safety valve was leaking, a repair ticket was written for a code safety valve (Sieglitz 5739; Ex. 2072) and repair of a code safety valve was placed on the list of work to be completed during the next outage (Sieglitz trial tr. 5471; Ex. 2086).
The valve leakage through the code safety valve, as measured at the time and confirmed after the accident, never reached the point where the plant was required to shut down pursuant to the technical specification limit on identified leakage.
GPU management was certainly ready and willing to repair any pressurizer valve if necessary. This is illustrat-ed by the fact that other valves on the pressurizer (isola-tion, or " root" valves for the level transmitters) were, in fact, repaired during the January, 1979 outage (Sieglitz trial 2
tr. 5747-49; Zewe trial tr. 2832-34; Ex. 2092) and that, when a code safety valve started to leak'after the January outage, that valve was promptly scheduled for replacement.
In summary, as was concluded by post-accident stud-ies of the issue, the evidence available to and appropriately assessed by Met-Ed personnel indicated that a code safety valve, not the PORV, began to leak after the January 1979 out-age (TDR 160, pp. 1, 17, 20; TDR 126, Investigation of TMI-2 18
Pressurizer PORV Discharge Pipe Temperatures, 2/28/80, pp. lb, 12). TMIA's apparent arguments to the contrary are based on little more than a characterization of the 180 PORV tailpipe temperature as an " elevated temperature," and on various mis-characterizations of testimony, discussed below. Furthermore, contrary to TMIA's assertions (Comments, p. 13), emergency procedures did not require closure of the PORV block valve prior to the accident to check for leakage as a result of the high PORV tailpipe temperatures. The procedure pertaining to PORV leakage to which TMIA refers never became applicable be-cause it had been appropriately determined that the PORV was not in fact leaking. A further discussion of this issue may be found at pages 3A-36 of the Licensee's response to the NRC notice of violation, B&W Ex. 707.
The section of the TMIA Comments on valva leakage contains the type of mischaracterizations of testimony that are found th oughout the Comments. For example, TMIA states, at p. 13, in italics, that "no leak rate was being monitored,"
citing as support Sieglitz trial tr. 5811. , However, at the page cited by TMIA, Mr. Sieglitz testified that "no laak rate was being monitored back to that area" meaning, as Mr. Sieg-litz explained during his testimony, that it had been deter-mined that there was no leakage from the PORV ("that area")
(Sieglitz trial tr. 5896). Mr. Sieglitz did not testify that leak rates were not monitored. Indeed, he expressly described 19
the review of leak rate reports at the plan of the day meet-ings (Sieglitz trial tr. 5715-21).
TMIh also states (p. 13) that Sieglitz was'"so igno-rant of the situation that he did not know that 180 was even an elevated temperature, claiming that it was not his iob to know such details," citing Sieglitz trial tr. 5809-10. In fact, Sieglitz testified that when he looked at the leak rate and at the 180* temperatures and saw that there was no signif-icant identified leakage, he and others concluded that the 180* temperature was normal (Sieglitz trial tr. 5809, 5816).
Mr. Sieglitz did testify that his job responsibilities, which pertained to the maintenance of equipment and not to opera-tions, did not require familiarity with the details of a plant operating procedure which referenced PORV tailpipe tempera-tures (Sieglitz trial tr. 5809-10).
TMIA goes on to state (p. 13), citing Sieglitz trial tr. 5815, 5817, 5844, that because Sieglitz "never thought to ask" if 180* could signify leakage, "these high temperatures were thus never discussed at ' plan of the day meetings.'"
However, contrary to the implication of TMIA, Sieglitz did not give this testimony regarding plan of the day meetings in re-sponse to a question about whether he "never thought to ask if 180* . . . could signify leakage" (Comments, p. 13). In fact, Sieglitz testified that the 180 PORV temperature was not dis-cussed prior to the January, 1979 outage because it was known, 20
based on leak rate data, that there was no significant identi-fied leakage (Sieglitz trial tr. 5815-16).
Alleged Leak Rate Falsification Repeatedly in its discussion of alleged PORV leak-age, TMIA refers to Hartman's allegations that tests for un-identified leak rate were falsified (Comments, pp. 11, 14, and
. 16-17). TMI_A's citations fail to establish either that leak rates were falsified or that GPU management knew of any pur-ported falsification.
The allegations that leak rate test results were falsified or " fudged" originated with Harold Hartman, who was a control room operator at Three Mile Island prior to the ac-cident.2 Hartman, during cross-examination under oath at his
- GPU v. B&W deposition, substantially limited his allegations regarding leak rate testing falsification. As a result, the implications of his earlier accusations were seriously, if not totally, undermined. It is therefore remarkable that TMIA (p. 14) relies on Hartman for the proposition that every shift supervisor and shift foreman knew that leak rates were being falsified, citing only Hartman's speculation ("I thought that it was just the fact that everyone knew that these leak rates were hard to get . . . "). Hartman's cross-examination re-vealed that Hartman could not identify with certainty a single 2 Hartman was discharged by Met-Ed shortly after the acci-dent. Only after his discharge did he make the allega-tions regarding leak rate falsification.
21
o operator other than himself who had attempted to falsify leak rate tests -- a marked change from his earlier allegations. .
The results of the Hartman cross-examination are summarized below.
Hartman Cross-Examination Operators routinely add both hydrogen and water to the makeup tank for legitimate reasons (Hartman dep. 263, 261). However, if water is added during a leak rate test, the addition must be " logged," so that the test results will take that water into account. Hartman, at his sworn deposition, admitted that he had seen only one operator add water during a leak rate test (id. at 260-61). Hartman also admitted that although it was his belief that the operator had added the wa-ter in order to affect test results, he had not looked to see whether the operator in question had logged the addition of water (id. at 261-62). Hartman testified that he personally had never added water to the makeup tank in order to affect test results (id, at 260).
Hartman believed that adding hydrogen to the makeup tank during a leak rate test would result in inaccurate test results. In his deposition, Hartman repeated his assertion that he personally had added hydrogen in an attempt to influ-ence leak rate test results. However, on cross-examination he admitted that he had no recollection of seeing anyone else add hydrogen in order to affect a leak rate test (id. at 274). He also admitted that operators were not required to record hy-22
l- - -
drogen additions made during leak rate tests and that he did not know whether hydrogen additions ~had ever in fact affected .
leak rate test results (id. at 264, 272-74).
Thus, Hartman's cross-examination revealed that the ,
only attempted manipulation of a leak rate test result to which he could testify with certainty was his own occasional attempt to affect results by the addition of hydrogen.
Other than TMIA's misleading citations to Hartman's testimony, the only support cited for the contention that op-erators were " fudging" leak rates is the Faegre & Benson re-port. That review of the Hartman allegations was prepared by outside consultants at the request of Metropolitan Edison.
~
TMIA cites page 17 [vol. 1] of the Faegre & Benson report for the proposition that operators were "' fudging [ leak rates] '
continuously" in order to keep the plant on line (Comments,
- p. 16). However, at the cited page, Faegre & Benson recog-nized only that in the three months preceding the accident the rate of leakage from one of the pressurizer relief valves was increasing. Faegre & Benson did conclude that that increase in identified leakage may have made it more difficult to ob-tain an accurate unidentified leak rate figure, because of the way the leak rate test computer program was designed (Faegre &
Benson, vol. 1, p. 26). .
However, the Faegre & Benson report nowhere endorses, on page 17 or elsewhere, the proposition for s which it is cited by TMIA.
23
Condensate Polishers The third topic raised by the Comments (pp. 19-20) concerns the Unit 2 condensate polishers and the failure of GPU to install an automatic bypass valve on these polishers.
TMIA cites no technical authority whatever to establish the advisability of such a bypass. Moreover, the Comments ignore the fact that management gave serious consideration to the suggestion to add a bypass and rejected the idea for sound en-gineering reasons (Toole dep. 446-47).
Mr. Zewe's suggestion, in the handwritten memorandum referred to by TMIA (Comments, p. 19), that an automatic by-pass be installed on Unit 2 was the result of familiarity with a bypass at Unit 1 (Zewe trial tr. 2202; Toole dep. 446; Ross dep. 113-14). However, the condensate polisher system on Unit 1 is significantly different from the type of system on Unit
- 2. The Unit 1 system consists of filters embedded with a com-po.und called Powdex which causes impurities in the secondary-side water to coagulate and stick to the filters; the Unit 2 deep-bed demineralizer system used ionic resins to remove im-purities.
Unit 1 had to have an automatic bypass valve in or-der to protect the Powdex filters from damage (Toole dep.
396). As shown by the testimony of Ronald Toole, a GPUSC startup and test engineer, protecting the Powdex filters from high, differential pressure was the sole design use and func-tion of the automatic bypass valve on Unit 1 (Toole dep. 400).
24
The Unit 2 deep-bed demineralizers did not require protection from high differential pressure; thus, they did not need an automatic bypass valve (Toole dep. 447).
Any bypass of the condensate polishing system on Unit 2 would have degraded secondary-side water purity, and was, therefore, viewed by GPU, B&W, and Burns & Roe (the ar-chitect engineer) as inadvisable (Arnold trial tr. 1502; Toole dep. 447). In fact, B&W specifically advised against bypass-ing the condensate polishing system. During a May, 1978 B&W User's Group meeting, B&W reviewed Unit l's condensate polish-ing system operating experience and stressed the importance of maintaining secondary water purity to avoid damage to the steam generators. As summarized in the meeting minutes, B&W
- concluded with the imperative that the condensate polishers should "never" be bypassed (GPU Ex. 2031, p. 3, emphas_is in original; GPU Ex. 2028, p. 5; O'Hanlon trial tr. 1067; Toole dep. 320-21; 452; Ross dep. 95-96).
Moreover, an automatic bypass would probably not have functioned rapidly enough to prevent an isolation of the condensate polishers from inducing a feedwater trip (Miller dep. 783; Zewe trial tr. 2202; 2209-10; Toole dep. 448). As discussed above, the bypass on the Unit 1 Powdex system had not been designed to prevent a feedwater trip, nor had it been tested to determine its ability to do so.
25
Water in Air Line Not only are the Comments misleading with respect to the purported need for an automatic bypass on Unit 2, but TMIA goes on to assert, incorrectly, that nothing was done to rec-tify the problem of water getting into the instrument air lines of the condensate polisher system (Comments, p. 19). In
. response to a 1977 startup and test event when' water got into .
the instrument air lines, Michael Ross, a shift supervisor, and John Brummer, an instrument and control engineer, wrote an extensive memorandum outlining steps to be taken to reduce the possibility of a similar occurrence (Ross dep. 121; B&W Ex.
165). Every one of the nine recommendations in this memoran-dum was carried out (Ross dep. 129-35; Zewe dep. 415-16). Ad- -
ditionally, in or about April, 1978, chemistry technicians were assigned to monitor closely the operation of the Unit 2 condensate polishers (Toole dep. 478).
The Comments state (p. 20) that, as of the date of '
the accident, there were " fully thirteen" work requests out-standing "with respect to" the Unit 2 condensate polisher sys-tem. The Comments fail to state either the nature of these work requests or how long they had been outstanding. More-I over, thirteen other work requests involving the condensate polishers were carried out within the three months preceding the accident (Shovlin dep. 165). Such activity with regard to the polishers supports Mr. Arnold's testimony that a number of decisions were made and actions taken with regard to the pol-26
l* .
ishers following the condensate polisher event in May of 1978 (Arnold trial tr. 1497-98, 1644, 1649).
Finally, the section of the TMIA Comments concerning the condensate polishers contains significant mischaracteriza-tions of testimony. For example, at page 1642 of Arnold's trial testimony, the following question and answer appear:
"Q: Did you become aware of an incident in the spring of 1978 in which water again l got into the instrument air lines on the l
discharge valves for the condensate
] polishers?
A: I think I was aware of that incident at that time contemporaneously with when it was happening, when it had happened."
(Arnold trial tr. 1642.) Incredibly, TMIA cites this testimo-ny for its contention (p. 20) that " Bob Arnold claim [s] to be entirely unaware . . . of the 1978 [ polishers] incident pro-voking Zewe's memo," citing Arnold trial transcript at pages i 1497, 1642. As indicated, the actual Arnold testimony, at l page 1642, is contrary to the TMIA assertion. At the other page cited by TMIA, Arnold testified to discussions which oc-curred at an October, 1978 meeting regarding progress in the resolution of problems with water in the instrument air lines.
The Comments assert (p. 20) that "perhaps most im-portant for present purposes, Arnold still does not consider this problem as representing any threat to the safe operation of the plant." (Id. at 1498, emphasis in original.) Once again, the cited Arnold testimony (at p. 1498) does not sup-port TMIA's assertion. The cited testimony says nothing re-27 4
-, . ~ . . _ , - , ~ . - . . _ . _ . , . , - _ , . _ _ . _ _ _ . . . - _ . . _ _ _ . - . - _ _ _ _ - - _ . . _ - . _ , _ . . . _ , , . . . . _ . - _ _ . - , , - , _ . - - . , . . ,
garding Arnold's present understanding. Rather, Arnold, after testifying that activities had been underway in 1978 with re-gard to the condensate polishers, explained that at that time he "did not consider the problem with water in the instrument air line as representing any threat at all to safe operation of the plant from a nuclear safety standpoint. . . .
"3 In sum, the focus placed by the Comments on the con-densate polisher system at Unit 2 mischaracterizes testimony and seriously misrepresents the efforts taken by GPU to deal with the pre-accident issues regarding design and operation of the polishers. As shown, the GPU/ Met-Ed management did, in fact, take responsible action to evaluate and deal with con-densate polisher matters prior to the accident.
The Keaten Task Force The next section of the TMIA Comments - " Licensee's response to the accid (nt" (pp. 20-37) -- concerns the Keaten task force. This was an internal task force investigation un-dertaken on the initiative of GPU management to explore the causes of the accident. The Comments attempt to establish 3 of course, the secondary side of the plant, where the condensate polishers are located, does not contain any nuclear fuel. Although a secondary side upset, such as a loss of feedwater, might take the plant off line, causing operational inconvenience, such upsets were expected and considered an anticipated operational occurrence for which the plant was designed and licensed (Womack trial tr. 4645; 10 C.F.R. 50, Appendix A: Definitions and Ex-planations, " Anticipated Operational Occurrences"; Crite-rion 29).
28
that "[a]fter each draft (of the task force report] was sub-mitted to management for comment, the task force rescinded significant findings of culpability which had been previously reached, unanimously, by members of the task force." (Com-ments, pp. 21-22.) This insinuation is not consistent with the evidence in the B&W trial record and is'at odds with the actual development of the task force report. Keaton testified at great length regarding the work of the task force and man-agement's overriding directive that the task force produce a j
report which contained only what was " factual and clear and
, accurate" and " consistent with the best belief of the task force." (Keaton dep. 60, 641.) Keaten described in his tes-
~
timony how the task force was still conducting its investiga-tion as the various drafts of the report were being prepared.
He also noted that some of task force's investigative work re-mained incomplete at the time of the final report and that, in some areas, work continued subsequent to the completion of the final report (Keaten dep. 679).
TMIA's Comments attribute to management a role in the development of the task force which is not supported by the B&W record. For example, according to TMIA, management reviewed the first draft of the report (dated September 28, 1979) and directed the removal of conclusions unfavorable to GPU. In fact, there was no management review of the report until after the October 29, 1979 draft, which was intended as .
the task force's interim summary report, and there is no evi-29 -
, _ _ . . . _ . - - . _ - , . ...,.-_..-..--.x._--.-_.. . . - _ - . _ - _ . - . - _ _ -
dence in the record of any prior management review.4 More-over, Keaten's testimony makes clear that.the September 28, 1979 first draft was intended not as a definitive report for management's consideration but rather as a talking piece to
" trigger discussion" among the task force members (Keaten dep.
732).
Keaten testified that he dictated this first draft into a hand-held recorder while driving across Pennsylvania (Keaten dep. 460), which is corroborated by the cover memo transmitting the draft to the task force members, in which Keaten invited them to " Feel free to attack without mercy."
(B&W Exhibit 347.) This memo was sent to Mr. Arnold without the draft report attached (id.).
4 Keaten elaborated on his rationale for preparing this draft in his testimony by stating that his purpose was:
. . . to dictate something myself for the entire report, not worrying about if what I dictated was right or whether I even agreed with what I dictated, but just sim-ply to get down on paper so people can shoot out [ sic], so this is what I did.
And if you look at this initial draft, you l will find errors of fact as well as shaky l conclusions or conclusions that were later
! . completely reversed, so I am perfectly willing)to but [to aanswer questions lot of them on this the answer is draft, going to be that's just something I put down to i get something down to get people to start l working on."
4 4 Indeed, the statement by Keaten in his memo transmitting the October 17, 1979 draft to the Task Force that "R.C.
Arnold has directed me to issue the interim report with-out his review" indicate's that up to that point, Arnold had not yet seen the report. (See B&W Exhibit 350.)
30
(Keaten dep. 461.)
Keaten went on to testify that "he elected to start preparation of the report recognizing that the investigation was not complete, but . . . feeling that we had learned enough in some areas to make it profitable to start the process of getting it into a report." (Keaten dep. 502.) As a result, Keaten testified, the first draft was dictated "without any attempt to be very careful or very precise in the wording since [he] recognized that there would be a great deal of work that would be required in order to convert this draft into one which was really clear and accurate." (Keaten dep. 566.)
Thus, the statements contained in the first draft did not rep-resent task force conclusions at all and were not even conclu-sions Keaten had reached in his own mind (Keaten dep. 461-62).
Above all, there is no evidence of any management review or revision of the draft report at this early stage or any other attempt to influence the content of the report.
The B&W record reveals that the first management re-view of the report came after issuance of the Interim Summary Report of the task force on October 29, 1979 (B&W Exhibit 351). The cover memo accompanying transmittal of this report to Mr. Arnold and Herman Dieckamp, president of GPU, notes that "although the investigations are not yet complete, the task force has been able to draw conclusions regarding many of the relevant issues." (Id., emphasis added.) Following issu-ance of this report, Keaten testified, he and Dieckamp dis-31
cussed it. During this discussion, Keaten and Dieckamp large-ly concentrated on Dieckamp's understanding of the report as written, with a view towards clarifying ambiguities in pas-sages in which Dieckamp felt that the task force had not ade-quately communicated what it had intended to say (Keaten dep.
638-40). Keate.n further testified that he and Dieckamp also discussed the bases for informat. ion contained in the report and possible areas of additional work for the task force (Kea-ten dep. 640-41).
Keaten testified that subsequent to this discussion, revisions to the draft report were made in an attempt to clar-ify areas previously seen as ambiguous, consistent with the direction given Keaten by Dieckamp that nothing should be add- -
ed, removed or revised unless it was " consistent with the best belief of the task force." (Keaten dep. 641-43.)
As a result of this effort, the task force produced a revised Interim Summary Report on November 28, 1979 (B&W Ex-hibit 352). Keaten testified that Dieckamp and Arnold were the intended audience for this report and that since he had ,
previously discussed task force activities with them, an in-terim report would be "a useful opportunity for management to give the task force any revised or new direction regarding things that we should be doing." (Keaten dep. 655, 657.)
Keaten further testified that there were no substantial changes in direction, but that Dieckamp and Arnold did suggest 32
some further investigation by the task force (Keaten dep.
659).
A comparison of the October 29, 1979 and November 28, 1979 versions of the Interim Summary Report shows that most of the revisions reflect Dieckamp's expressed concern over possible ambiguities in the report. Minor editorial changes can be found throughout the report.Section II.A of the November 28th version also contains the task force's new conclusion that water in the instrument air system probably caused the sudden closure of the condensate polisher outlet valves.Section II.B, dealing with the rationale for the con-trol room and staff personnel response, was extensively re-vised. Dieckamp had told Keaten that he did not believe the interim report adequately conveyed what he believed to be the task force's conclusion that the operators were misled on the
. day of the accident by the location of the break in the pres-
! surizer vapor sp' ace which resulted in parameters which did not resemble the LOCA response which the operators hsd been trained to expect (Keaten dep. 638). Accordingly,Section II.B was substantially restructured and expanded, in particu- ,
lar, to explain the effect of the break location on operator response, but also generally to convey more clearly the task force's conclusions (Keaten dep. 642-43). Certainly, no fair reading of the revised Section IT.B in comparison with the Oc-tober 29th version can yield anything but the conclusion that the authors were attempting to express themselves more clearly L
33
and fully in explaining the operator's response. No material which could be considered critical of'GPU was removed from this section.
The other significant change in the November 28th report is in Section II.F dealing with the closure of the emergency feedwater block valves. The October 29th version notes that "The surveillance procedure [for checking the oper-ability of certain valves in the emergency feedwater system]
clearly violates this technical specification," referring to a technical specification which, the report states, requires that three independent emergency feedwater pumps and associat-ed flow paths shall be operable as a limiting condition for operation. The November 28th version states that the emergen-cy feedwater system is defined in the technical specification as three independent pumps and flow paths, which must be oper-able, even though one " system" may be inoperable for 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br />.
The report points out that there is no statement as to pumps or flow paths out of service and the actual requirement is thus unclear. The report concludes that "while the surveillance procedure may not have violated the literal req luirement (of the technical specifications , the Task Force believes that it was contrary to the intent of the specifications. In con-trast, the TMI-l surveillance procedures 1 indicates that at no time may two emergen-cy feed trains simultaneously be out of service."
Also, the November 28th version does not make mention of vio-lation of this technical specification by a second surveil-34 wy ---p-9--w-- +m---- r-e= - - + , - - - -wwq- ----gy-,- -- p w-- em=->.pa-- p-wg v~w-,- ,-e-- ye www--y .,- - e_w
lance procedure (although that procedure and its requirements are discussed). -
I Reference to the " violation" of the technical speci- !
l fications was omitted from the " Conclusions" section of the l November 28th draft. At his deposition, Keaten addressed this revision of a " conclusion," which first appeared in the rough draft of September 28th. Keaten testified that:
"At the time that this was written, I think I believed based upon work done by others, not by me personally, that the way in which that surveillance testing was done was contradictory to the requirements of the technical specifications. I should add that a later and more careful investi-qation of that contradicts what is said here."
(Keaten dep. 490-91, emphasis added.) The changes made in the -
November 28th version thus do not reflect a management " white-wash" with respect to this surveillance procedure, as TMIA im-plies at pp. 32-33 of its Comments. Rather, they reflect the results of a more careful investigation of what the technical specification actually says and what it actually requires, and a finding that the specification may not have been violated.
Further, any allegations of a " whitewash" should be stilled by the finding that although the literal requirements of the specification may not have been violated, the procedures as implemented were nonetheless contrary to the intent of the specifications.
In general, the November 28th report, which incorpo-rated, among other things, Dieckamp's comments, continued to 35
- - - . . . - . . - ~ . - - , -.._y ,-_~.,_.,,,,y_..,,,,,,,,y,,_,mm,.m .,..,,--,,,..,,_,...-,,.,_.-.,.m.,_y,
,,__r ,y
=.
be what he directed, an accurate summary by the task force of its findings. The November 28th version cites several defi-ciencies within GPU and the nuclear industry in general which contributed to the accident, and strongly recommends correc-tive action. TMIA's mischaracterization of the report as a management cover-up ignores these self-criticisms which are evident throughout the substance of all versions of the re-port.
The November 28th Interim Summary Report was sent to both Dieckamp and Arnold (Keaten dep. 653-54). The next re-port produced by the task force was a draft Final Summary Re-port dated March 24, 1980 (B&W Exhibit 354). This draft con-tained various minor changes in the section on operator re-sponse (II.B) and assessment of core damage (II.G). The " Con-clusions" section of the November 28th report noted that while the task force had not performed a thorough review of the rea-son for the widespread existence of problems in the TMI opera-tion, there appeared to be a lack of management awareness of problems, insufficiently stringent standards to evaluate oper-ations, and a management tolerance of this situation, at least in the short run. This conclusion does not appear in the March 24, 1980 report, which notes that the task force did not thoroughly review the role of TMI management " relative to the identified problems," because of the significant changes made .
since the accident,in management structure.
36
--wi-- -----
$ h
$' d'#4 k> IMAGE EVALUATION
((//p%j* f4 #4, 4 pyx!$> TEST TARGET (MT-3) pzz[ %z, 1.0 N 034 l5l8EE u [5 EE if 1.25 1.4 i 1.6 I
4 150mm >
6" >
- ?
/? < :l
.e; %
P 499 h
o
$big, y*k gp %f f/
IMAGE EVALUATION TEST TARGET (MT-3) ////p% 4
\'{pf %/,[,4 #
l.0 lf Ea Un y @ Est i,i [' HM
'8 kl 1.25 1.4 1.6 4 150mm >
6" >
4
' #kl;%
b
,,? ,..
TMIA has seized on this section as further "evi-dence" of a management cover-up (Comments, p. 34).
In fact, the role of management is not listed among the seven areas the task force was directed to investigate. (See Table I at the beginning of each version of the report.) In complaining about the removal of " pejorative" statements describing man-agement's " culpability," TMIA has ignored the purpose of the _
task force, which, Keaten testified, was not to fix blame for the accident, but rather "to understand where were areas where improvement or change were either necessary or desirable in order to reduce, as far as possible, the probability of this type of event occurring again." (Keaten dep. 583.) TMIA to-tally ignores the fact that the March 24, 1980 report fulfills this mandate by making several recommendations not contained in the November 28th report which involve changes by plant management. These include recommendations 3 (implementation of an integrated system for procedure development and review),
, 13 (formal system for assuring action on employees' sugges-tions for improved plant operation), 14 (assuring that plant improvements are not hindered by difficulties in obtaining au-
! thorization and establishing an annual resource fund to that end), and 17 (careful evaluation of plant operation under de-graded conditions). Each of t hese recommendations appears in the final version of the Final Summary Report, dated December 15, 1980 (B&W Exhibit 356), along with an added recommendation (number 14) that a formal system to document degraded plant 37 s
-v,- - - - - , . , - - - , . - - ...r ,----,,-.,,,.,..,-.,,.-_,..-.-,-.,~,-.---,,,n. -
equipment conditions and ensure corrective action be imple-mented. Far from being a cover-up, the March 24, 1980 report instead represents GPU's continuing attempt to learn from the accident and to improve its nuclear operations. TMIA's empha-sis on the removal of certain " pejorative" language should not be allowed to obscure the larger point that GPU continued to engage in penetrating self-criticism and to take corrective action.
A subsequent draft of the task force final summary report was produced on May 12, 1980 (B&W Exhibit 255), and, Keaten testified, was later distributed widely among GPU's up-per management, including William Kuhns, Philip Clark and Richard Wilson, along with Dieckamp and Arnold (Keaten dep. -
683; see also, B&W Exhibit 357). The final version of the Fi-nal Summary Report was released on December 15, 1980; there was no Keaten testimony about management comments on the May 12th draft. Aside from some editorial changes, and a supple-mentation of the section on " Pressurizer Relief Valve Failure Modes" (II.D), the December 15th report is substantially simi-l lar to the May 12th draft. Indeed, TMIA admits as much in its comments (see TMIA Comments at pp. 31-32). Thus, the draft of j the report most widely distributed to GPU management resulted l
l in virtually no significant changes.
I l A careful review of the B&W record refutes TMIA's attempt to impeach the integrity of the work done by the Kea-ten task force. Plainly, the earliest drafts reflected tenta-38 4
--vrwe-- - -. s --.m~,se,-s,- --m-- --,-,-.-m-,,--r-o- -e-e-,-e-,--e , s -,e w-,,w,-, .--,-,,w ,m w e--- - - w--- ------ - ,,,
4 tive hypotheses, sometimes exaggerated for use as talking e
points, based on incomplete investigative work. The later drafts, which were the only ones reviewed by management, re-flect not a " whitewash," but a continual development of under-standing and a refinement and clarification of the task force's conclusions. Indeed, all of the drafts were main-tair.ed, and produced in the B&W litigation. Above all, the tr,sk force effort represents a thc.ough self-examination, en-couraged by GPU's management, and the final report contains numerous criticisms of both GPU and the nuclear industry, ac- .
companied by strong recommendations for reform. It is upon this process of self-criticism and its follow-through that the NRC staff should concentrate.
Maintenance In its section on Maintenance, pages 37-50, TMIA makes a number of derogatory assertions regarding the quality of maintenance at TMI both before and after the accident. In fact, however, TMI-l's maintenance record had built a reputa-tion in the industry for excellence. B&W's manager of Operat-ing Plant Services, James Phinney, recognized that TMI-l's good operating record and history of reliability was a result of the high quality of the outage planning and of the mainte-nance performed during outages (Phinney dep. 6-10; GPU Ex. 394, Metropolitan Edison minutes of BTW Operating Seminar, at p. 4).
39
TMI's argument that a favorable assessment of cur-rent maintenance conditions cannot be accurate if there were unfavorable maintenance conditions in the past is dubious at best. In any event, the examples given of allegedly poor maintenance, drawn almost exclusively from the GPU v. B&W tri-al and depositions, do not demonstrate safety-related prob-lems, either in the past or the present.
Maintenance Budget and Staffing Seeking substantiation from the B&W trial racord, TMIA's first criticism of maintenance (pp. 39-42) focuses on what it labels " budget cuts and understaffing." Purportedly relying on a tape-recorded interview (B&W Ex. 360) with former station superintendent Gary Miller, the Comments seek to un-dermine certain ASLB findings. Among the ASLB findings favor-able to GPU which TMIA attacks are: (1) that the " Licensee's records under the old system were auditable" (Comments, p. 42; PID 1 314); (2) that the Licensee has, under the past and pres-ent systems, adhered to company standards in performing main-tenance in a timely fashion (Comments, p. 42; PID 1 289); (3) that even in ti.mes of financial stress, GPU management has shifted available resources to meet its clear obligations (Comments, p. 42); and '(4) that there is no improper deferral of maintenance work "of significance" (Comments, p. 42; PID
! 1 206). As developed below, TMIA fails to marshall any evi-l t
9 40
dence from the B&W trial record or elsewhere to undermine these findings.5 Maintenance Staffinq The Comments argue (pp. 40-41) that prior to the ac-cident there were about 800 open maintenance items and that the maintenance staff and budget should not, therefore, have been reduced. TMIA has neither cited nor quoted any testimony to show that 800 open maintenance items was an unreasonably i large number for a 2-unit nuclear station, which has hundreds of thousands of pumps, motors, pipes, lines, valves, switches, filters, relays and other instruments and machines. Nor has TMIA cited any evidence to establish that those 800 items were left "open" for unreasonable lengths of time. TMIA has also failed to indicate how many of the 800 items were major or
! whether any of those items were safety related. In fact, Miller, in his recorded interview, indicated that " priority one" jobs, i.e., the important jobs, were always done (B&W Ex.
260 at p. 22, Comments, p. 41). Nothing in the Miller state-ments indicates improper deferral of maintenance work or that significant work was deferred.
Miller's comments indicate that, prior to the acci-i dent, maintenance staffing for the two units was approximately l
- 5 For example, with respect to whether the TMI maintenance records were "auditable" (Comments, p. 42), nothing in the Miller statements quoted by TMIA in any way pertains to the Board finding.
41 i
a
- 135 persons, plus outside contractors, whose staffing was be-ing gradually reduced (Arnold trial tr. 1661-65). Arnold tes-tified that, in or about 1978, the maintenance staff increased by about 50%, to approximately 150 people (Arnold trial tr.
1656-58). Before the accident, GPU studied the adequacy of its maintenance staffing. It found that-the level of staffing .
at TMI compared favorably with other utilities and that GPU was on the high side of the range of expenditures for all nu-1 clear activities when compared to other utilities (GPU Ex.
2056; Arnold dep. 395, trial tr. 1774-75, 1782-84).
- Budget Cuts As TMIA concedes (Comments, p. 41), the ASLB has al-ready rejected TMIA's contention that a proposed 1979 budget
- cut lacked due regard for safety. TMIA's attempt to find evi-dence to the contrary in the GPU v. B&W trial record has i failed to provide any evidence to further support its conten-i tion. The fact that Miller would have preferred not to reduce the maintenance budget does not undermine the Board's conclu-sion. Miller presented his views on the budget to GPU manage-i ment (Arnold trial tr. 1664; Miller dep. 1038-39) and the Com-ments make no allegation that Miller's views were not fairly considered. TMIA states as a criticism (Comments, p. 41) that budget cuts were also considered in 1977 and 1978. In fact, the decision of management not to implement budget cuts in those years, if anything, seems to bear out the ASLB finding 42
l that GPU shifted available resources to meet nuclear obliga-tions, even in times of financial stress (Comments,.p. 42).
Recair Parts In a further part of the TMIA Comments on mainte-nance, TMIA attempts to criticize the pre-accident repair parts documentation at TMI (Comments, pp. 43-44). As TMIA recognizes, the issue of repair parts documentation was iden-tified by Met-Ed management in its 1978 management " audit,"
B&W Ex. 843. The thrust of the portion of the audit quoted by TMIA was that, in 1978, it took a long time to obtain repair parts. However, the Comments (p. 43) quote from only a por-tion of the 1978 audit which found that there was " difficulty in locating repair parts known to be in the warehouse." The Comments fail to indicate that 'the audit went on to explain what it was referring to: "There were complaints of long lines at the service window, and that often the service window is not manned for long periods of time." (1978 Audit at
- p. W45228.) The other audit finding quoted by TMIA (p. 43) is that it took "an inordinate amount of research to identify the Met-Ed stock numbers required for requisitioning material" (id. at p. W45228). These relatively minor inc~nveniences that existed five years ago do not even begin to be valid rea-sons for delaying the restart of Unit 1.
The references.to the "Glickman Audit" of late 1979 in the TMIA Comments (p. 43) misrepresents both the effort and the contents of the document. Mr. Glickman's effort was di-43 l
rected at interviewing a number of management personnel in or-der to be~sure that any concerns they may have had would be met with a sympathetic ear in the new GPU Nuclear organization being set up. Thus, to the extent that problems in fact ex-isted, they would be addressed effectively. The report pro-vided as a result of that effort drew no conclusions on these issues and Mr. Glickman in writing the report was acting as a reporter of the informally received opinions of the interview-ees and not as a confirmatory editor (Glickman dep. 476-77; Arnold trial tr. 1771-72).
TMIA's example of " concrete evidence" from the trial record of an alleged repair parts problem is that "B&W discov-ered that a spare PORV was ordered in 1975, and delivered to the Island until 1978 [ sic]. Yet in early 1979, the Supervi-sor of Maintenance [Sieglitz] was still unaware that it had been delivered" (p. 43, citing Sieglitz trial tr. 5766-67).
As Mr. Sieglitz pointed out at the citad transcript page: "I did not keep track of each and every component. There are thousands and thousands of line items in the warehouse. I was not specifically looking for any one component when it arrived unless I had a component that was broken down. . . . I had no reason to look for [a spare PORV) unless I felt I had a main-tenance problem." (Sieglitz trial tr. 5766-67.) As Mr. Sieg-litz understandably pointed out, he relied.on a computer printout, not his memory, to keep track of spare parts, and
! that printout was approximately 18 inches thick (Sieglitz tri-l l 44 l
l
._. _ _ _w_,.
F-i . ; -
al tr. 5792-94). Clearly, Sieglitz should not have memorized l
a list of the equipment components contained in the warehouse; records were available from which that information could be determined.
In short, the purported " concrete evidence" from the B&W trial record fails to demonstrate any safety problem in
~
the location or documentation of spare parts. .
Paperwork The third part of the Comments' section on mainte-nance (pp. 44-45) addresses the time spent by the maintenance department on paperwork. As Arnold has testified, a paperwork burden is endemic to the nuclear industry, which requires an enormous amount of paperwork (Arnold dep. 380). As is also
~
established in Arnold's testimony, GPU has taken many steps to
. reduce this burden. The steps taken prior to the accident in-cluded additional new office machine equipment, additional staffing (e.g., use of a " switching and tagging" operator),
and evaluations by the Met-Ed operations analysis section of ways to reduce paperwork and streamline administrative contro1 ,
procedures (Arnold dep. 379-86).
a overtime TMIA addresses the issue of overtime on page 46 of the. Comments. As TMIA admits, the ASLB has dismissed the pu-tative " overtime issue," finding no abuses (Comments, page 46); TMIA, in its Comments, provides no evidence relevant to the Board's finding on this point. In its discussion of ha,1f 45
a page (Comments, p. 46), TMIA cites only so-called " audit" findings that are four and five years old, regarding working hours on site.
The need five years ago for personnel to put in long hours arose from a unique and temporary situation existing at that time, i.e., the intensive efforts needed for startup and test of Unit 2 (Arnold dep. 348-49, 367). The Comments do not assert that an overtime problem presently exists or indicate i why concerns with overtime in 1978-79 are relevant to restart.
Daniel Shovlin The final portion of the TMIA Comments on Mainte-nance (pp. 46-50) attempts to utilize the deposition testimony i of Daniel Shovlin, the current Unit 1 Maintenance Supervisor, I
! to attack his competence and credibility. TMIA is unsuccess-ful on both counts. As demonstrated below, its unfounded at-tacks on Mr. Shovlin rely either on statements taken out of context or on his understandable inability to recollect the details of events which occurred between three and a half and seven years prior to the date of his deposition.
The TMIA Comments on Mr. Shovlin begin by citing his deposition for the proposition that when he first came to TMI as supervisor of maintenance, "he felt that there was nothing he needed to do to familiarize himself with TMI-1." (Com-ments, p. 47, citing Shovlin dep. 30.) The actual testimony given by Mr. Shovlin reads quite differently:
46 l
l I
"Q. And in July 1973 you found yourself as a supervisor of maintenance for a nu-clear power station (i.e., TMI). Did you do anything to familiarize yourself with how that station was constructed or de-signed, or how it worked?
"A. Yes." (Shovlin dep. 30.)
After giving this clear testimony directly contrary to what TMIA asserts, Mr. Shovlin observed correctly that, with the exception of "the radiological hazards" (id.), which he noted as a " major difference," a fossil plant and a nuclear plant have many similarities. It was only in that context that Mr. Shovlin gave the testimony under cross-examination which TMIA cites (id.).
TMIA's assertion that Shovlin has never "seen a ,
maintenance log" (Comments p. 47) is similarly misleading.
Shovlin has seen and was familiar with logs kept in the Unit 2 maintenance department (e.g., Shovlin dep. 88), including the maintenance log itself. The context of the testimony cited by TMIA, which concerned solely start-up log entries, makes clear that B&W's counsel was incorrectly using the terms "mainte-nance log" and " start-up log" interchangeably. Shovlin did testify that he had not reviewed GPUSC start-up logs. Those logs were kept by GPUSC engineers and pertained not to mainte-nance, but to start-up and ' - 'ctivities within the respon-sibilities of the GPUSC engineering . -tment.
Another example of TMIA's loose tu. *nt of testi-mony is its assertion that "in 1975, [ Gary] Miller ... ' Shov-lin, among others, responsible for a ' low level of interest 47 1
1
and seriousness'" with which the Unit 2 procedure writing ef-fort was purmied (Comments, p. 47). In fact, however, Miller did not attribute such an attitude to Shovlin or to any par-ticular person. The document containing the language quoted by TMIA, B&W Exhibit 763, is a memorandum from Gary Miller which is merely addressed to Shovlin, among other persons, and which also contains a lengthy distribution list, including "each engineer" involved in the procedure-writing effort.
Shovlin was not even questioned about this memorandum at his deposition. As for its author, Mr. Miller, he testified as i
l follows regarding his memorandum:
"The reason for the statement . . . is that . . . I wanted to have more resources
- which means more individuals devoted to writing the chemistry, HP, and maintenance procedures, and that is why the addresses are Mr. Sawyer, Shovlin, Dubiel, and Ro-manski, and I don't think it is fair to say this characterized a low level of in-terest in Unit 2 procedures, but it did characterize the fact that the Maintenance
- and Health Physics Departments needed to recognize Unit 2 was to get some more pri-ority." (Miller dep. 135-36.)
Thus, TMIA's attack on Shovlin, based on the Miller memoran-dum, falls wide of its mark.
In addition to taking statements out of context, TMIA's attack on Shovlin seems to assume that as of July 1982, l
l when his deposition was taken, Mr. Shovlin should have re-l called the details of incidents which' took place as many as seven years before. Thus, for example, TMIA challenges the fact that he couldn't recall whether or not he had seen a let-48 l
l l
l l ._ ._, . _ . . . _ . _ _ _ _ . _ _ . , _ _ _ _. . _ . _ _ _ _ _ _ _ _ _ _ _ , _ _ _ . . _ _ _ . _
i ter which was sent from Lee Rogers to John Herbein dated July 30, 1975. Similarly, TMIA attacks Shovlin for not recalling the details of various incidents which took place in 1977 and 1978. For example, TMIA criticizes Shovlin for not recalling that in September of 1977 the Unit.2 PORV was removed for in-house repairs. There is no reason why that standard repair, properly performed, should have stood out five years later as a memorable event. It is simply unfair to criticize someone for being candid enough to state that he is unable to recall such events when they took place so many years before.
TMIA's attack on Shovlin for failing to recall such events is particularly unfair in light of his responsibilities during much of the period in question. Beginning in late 1977, Shovlin was superintendent of the TMI maintenance de-partment with overall responsibility for the planning, organi-zation and direction of maintenance for both Units 1 and 2.
At that time, Shovlin also assumed the additional position of supervisor of maintenance on Unit 1, a position he had held from 1973 through 1976 and which he had held throughout most of 1978 (Shovlin dep. 4-12). Thus, with regard to detailed maintenance issues, as opposed to the overall administration u
! of maintenance, Shovlin focused principally on Unit 1 during that period. For the details of Unit 2 maintenance matters, l Shovlin relied on Richard Sieglitz, the Unit 2 supervisor of maintenance, who Shovlin testified "ran the maintenance activ-ities at Unit 2." (Shovlin dep. 161-62). As both superinten-49
.. l dent of maintenance on the Island and supervisor for Unit 1, Shovlin appropriately relied on the Unit 2 supervisor to han-die the. detailed maintenance oversight of Unit 2 while dealing with such matters himself on Unit 1 (Shovlin dep. 195-98).
Accordingly, TMIA is simply off base in criticizing Shovlin for not recalling, years later, minute and detailed facts con-cerning events at Unit 2. There is no need to commit to memo- .
ry and retain in one's mind for as long as seven years de-tailed information on numerous maintenance efforts which are all accurately recorded by the company and retrievable by any managerial or supervisory' personnel when they need to know about it.
In light of this background, many of TMIA's specific criticisms of Mr. Shovlin must be rejected out of hand. For example, TMIA criticizes Shovlin for not knowing that there were 13 work requests outstanding on the Unit 2 condensate polishers at the time of the accident and that 13 more were carried out from January to March of 1979. Since Unit 2 main-tenance was performed under Sieglitz's direct supervision, not Shovlin's, Sieglitz certainly did not need to inform Shovlin of each of these work requests, nor would Shovlin be expected to recall these statistics years later even if he once knew them.6 6 Ironically, in a subsequent section of its Comments deal-ing with " Management Structure," TMIA criticizes manage-ment for allegedly failing to delegate responsibility ef-(Footnote continued) 50
-m m
r- v-- g_v , ---v-- w -- - , ,e a -- ,-r-w, , , - e a, -cem ---- e .wm v-----e -
1 Shovlin's personal involvement in the details of Unit 1 maintenance is clearly reflected in his deposition.
Thus, while TMIA criticizes him for not recalling leakage from valves at the top of the pressurizer in Unit 2, it fails to mention that he did recall learning of such leakage in Unit 1.
i (Shovlin dep. 125-56.) In view of his line responsibility for Unit 1, dating back to 1973, it only stands to reason that his knowledge and recollection with respect to that Unit, which is also the sole subject of his present responsibilities, would be greater.
! TMIA's criticisms of Mr. Shovlin also ignore the ba-sic division of responsibility between GPUSC and Met-Ed, par-ticularly as it relates to Shovlin's area. Most of the inci-dents at Unit 2 with respect to which TMIA criticizes Mr.
Shovlin took place during the start-up and test phase of the Unit in 1977 and 1978. As TMIA is well aware, GPUSC, not Met-Ed, was responsible for the start-up and test program. In-deed, a number of the specific incidents cited by TMIA, in-cluding the PORV failure in August 1977 and the condensate
, system problem in the fall of 1977, were addressed by GPUSC engineers in start-up problem reports. As Shovlin testified j (pp. 161-62), when Unit 2 went commercial in December 1978 it l
l l
(Footnote 6 continued from previous page) fectively (Comments, pp. 57-58), while in its section on Shovlin TMIA inconsistently criticizes management person-l nel for not being involved in the minute details of de-
. partment activity.
51
was the responsibility of Mr. Sieglitz to handle maintenance issues remaining from the start-up and test phase; nnly those problems which were considered at that time to be highly sig-nificant would be referred by Sieglitz to Shovlin. Thus, the details of at least some of the events which TMIA criticizes shovlin for not recalling would not even be expected to have been brought to his attention in the course of his duties.
Finally, in criticizing Mr. Shovlin for failing to recall the details of a particular incident, TMIA often fails to note, or discounts the fact, that Shovlin did have a gener-al recollection of the event and his participation in it --
which is all that could reasonably be expected so many years later. Thus, for example, TMIA states that Shovlin is "en-tirely unaware of the April 23 (1978] transient or of doing anything in connection with it." (Comments, p. 47.) At best, this is a half-truth. While Mr. Shovlin could not, four years later, specifically recall the events of April 23, 1978, he
! " remember [ed] very vividly" the fact, which was mest pertinent from a maintenance point of view, that the main steam relief valves had lifted and stuck open (Shovlin dep. 58-59). Since Shovlin was in charge of maintenance and not engineering or operations, it stands to reason that the maintenance aspects of an operating event would remain with him longer than the engineering or operational details.
l 52 -
l l
l
T Similarly, TMIA criticizes Shovlin for not recalling specific instances where problems were encountered with the condensate polishers, resulting in the loss of feedwater, but fails to note that he did recall participating in discussions as to whether an automatic bypass system would address those problems. Specifically, Shovlin testified that he had pursued the matter first with one of his lead foremen, Doug Weaver, and then with GPUSC's lead start-up and test engineer, Ron Toole, who co'ncluded that such a system was not necessary since the plant was designed to accommodate safely any loss of feedwater (Shovlin dep. 163-64). Far from serving as a basis for criticism, Mr. Shovlin's recollection of this incident,
~
omitted by TMIA, illustrates the fact that he attended to his duties in a conscientious and appropriate manner.
In sum, Mr. Shovlin's deposition testimony does not raise any serious issue with respect either to his credibility or to his competence to serve as supervisor of maintenance at Unit 1.
Training The Comments attack the current training at TMI by mining the B&W trial record for alleged indications of "long-l ,
standing and well-recognized" training problems which "the company did absolutely nothing about until forced to do so af-ter the accident." (p. 50.) A fairer reading of the trial record, however, reveals that Met-Ed was addressing training concerns before the accident.
53
- The. focus in the Comments on "non-attendance" at re-
! qualification training (pp. 52-53) is seriously misleading.
In the first place, all of the operators enrolled in the re-qualification program were already licensed, had passed the NRC licensing exam and were qualified to operate a plant. In fact, operators from TMI performed better than the average on the NRC licensing exams (B&W Ex. _ 707, p. 1, memorandum from Arnold re: Response to Notice of Violation; GPU Exs. 2305, i
. 2306). The material presented in the requalification program (to which the TMIA Comments refer) was presented for the pur-pose of maintaining operator qualification in accordance with the guidelines established by the NRC in the mid-1970's. (See Frederick trial tr. 3210; GPU Ex. 2320, 10 CFR 55, Appendix A.) Both the NRC, in 1977 and 1978, and the GPU Quality As-surance Department, in 1978, audited the training department programs (GPU Exs. 509, 510, 511). The results of all of those audits were favorable.
Moreover, before one can draw any conclusion from I
classroom attendance figures, it is crucial to note that the number of scheduled classroom hours at TMI was far greater than the hours requir, either by the NRC or by Met-Ed's own internal procedures (Arnold trial tr. 1760-63; Zechman dep.
596). For example, a minimum of only six training lectures is recommended in the NRC-endorsed "American National Standard for Selection and Training of Nuclear Power Plant Personnel" (GPU Ex. 2256, ANSI /ANS-3.1-1978, p. 9; GPU Ex. 2258, NRC Pro-l 54 i
^L posed Revision 2 to Regulatory Guide 1.8, February 1979, pp.
2-3; Arnold trial tr. 1760-61). Yet, Met-Ed committed itself to offering 60 hours6.944444e-4 days <br />0.0167 hours <br />9.920635e-5 weeks <br />2.283e-5 months <br /> of training (GPU Ex. 2155, Unit 2 Final Safety Analysis Report ("FSAR") S 13.2.2.1; Frederick trial tr. 3328), and typically offered many more hours of training lectures during each annual requalification cycle (Zechman dep. 900 - more than 200 hours0.00231 days <br />0.0556 hours <br />3.306878e-4 weeks <br />7.61e-5 months <br /> of scheduled training offered).
That operators in fact attended classroom hours in excess of requirements is illustrated by GPU Exhibits 3030 and 3031 (Frederick trial tr. 3327-38; GPU Exhibit 2154) which graphi-cally depict the excellent 1978 attendance records of the four operators on shift on the day of the accident.
In the event that operators were not able to attend i
- a classroom training session, they studied the material pre-
< sented through make-up packages, a practice provided for in the NRC-approved FSAR (Frederick trial tr. 3327, 3333; Arnold trial tr. 1759; GPU Ex. 2155, Unit 2 FSAR, S 13.2.2.1). This study, which was done on company time, was tracked by the
, training department (Arnold trial tr. 1760; Zechman dep. 191).
The level of attendance at training was in part the result of the five-shift schedule in effect prior to the acci-dent. (See Herbein dep. 367-68; Miller dep. 501-03; 1978 Au-dit, p. W45229.) The five-shift schedule in fact contributed to valuable operator training in that the operators had in-creased exposure, on a five-shift schedule, to important tests and evolutions performed only during the start-up and test pe-55
riod (Arnold trial tr. 1755-56, 1958; Herbein dep. 366-67, 371). Met-Ed management planned to return to six shifts after the start-up and test program and did so shortly before the accident (Herbein dep. 367, 371-72; Arnold dep. 366-67; Arnold ,
trial tr. 1755-57; Miller dep. 503; Herbein dep. 503). The six- and five-shift schedules at TMI compared favorably to the schedules of most nuclear plants, where four- and five-shift schedules were generally found (Arnold trial tr. 1756).
l The memoranda cited by TMIA at page 51 do not sup-port the allegations regarding training; i f anything, they in-dicate that management gave training serious and appropriate attention. Thus, f or e:: ample, the Comments quote from a memo written by Miller in 1977 stating that greater progress was l
- necessary in on-the-job training if a' deadline (four months away) was to be met. (The on-the-job training referred to is a process whereby operators witness and participate in specif-ic aspects of plant operations, so-called plant " evolutions.")
Miller's tracking of the on-the-job training program, and the fact that there were frequent training department audits of this program, speak well for the organization of the training ,
program. TMIA presents no evidence that operators did not complete on-the-job training in 1977, or that operator train-ing re,quirements were not met.
~
i TMIA notes a concern with the relevance of classroom training to plant operations (Comments, pp. 51, 54-55). Inev-
- - itably, a certain amount of training will be devoted to sub-56 1
- - , , -,v- .,---,., , , , ,, nw--. -.w-.,,,.ww._,v. -..a,_p,,eg-, . , ,
,,,,,-,,,,--m, , , , - _ _ , , , , , , , - , . ,-w,,~m.w_,,,,- ,
F jects on which operators are to be tested on the NRC licensing examination and on the requalification examinations given by the utility and audited by the NRC. Some'of the operators might consider some of the topics in these exams not immedi-ately relevant to plant operating needs. However, that does
- not change the fact that the NRC training requirements needed to be met and are met by the TMI training department.
Richard Zechman The Comments (p. 53) state that the trial record contains the " revelation" that Richard Zechman, supervisor of training, did not have an operator license. In fact, there was neither a requirement nor a necessity for Zechman to be a licensed operator.
Zechman had an extensive' background in nuclear'oper-ator training and had held both a RO and a SRO license at the Pennsylvania State University research reactor prior to join-ing the TMI training department in 1974 (B&W Ex. 554, Zechman i
resume). His position as supervisor of the Met-Ed training department was primarily administrative, and his teaching re-sponsibilities'were in theoretical areas such as reactor the-ory (Zechman dep. 737, 744, 751-52, 760). Thus, his position did not require a knowledge of plant-specific information, such as that possessed by a licensed operator.
It was not a job requirement that the head of the Met-Ed training department hold an operating license; nor was it standard in the industry for the supervisor of training to 57
~
hold a license. For example, Richard Marzec, GPU's expert witness on training at the GPU v. B&W trial, and one of the nation's leading experts in operator training, did not himself hold an operating license on a B&W plant during the time that he was the head of operator training programs at Duke FTwer Company, which had three B&W reactors.
Met-Ed and GPU management knew of Zechman's decision to take the challenging senior reactor operator license exami-
, nation-(rather than.the simpler control room operator's exami-nation) (Herbein dep. 278; Arnold trial tr. 1705). During the time that Zechman was studying full time for this examination, administration of the training department was properly handled by others within the department (Beers dep. 110; Herbein dep.
274-75, 281-83).
The Book Letter TMIA quotes at length (pp. 51-52) from a letter by T. L. Book which expressed his view on certain training is-sues. The contentions made in the Book letter have already been investigated by the NRC's Office of Investigations (Memo-l randum for Chairman Palladino, et al., from Darrell G. Eisen-l hut, Director, Division of Licensing, NRR, re: Management Competence and Integrity - Board Notification 83-71A, June 27, I
1983, attaching a memorandum from Ben B. Hayes, Director, Of-fice of Investigations, re: TMI-l Allegations of Falsified i
j Training Records, with enclosures - sworn statement by Book,
. and report of the OI interview of James O'Hanlon).
58
The Office of Investigations interviewed both Theo-dore Book and James O'Hanlon, to whom the Book letter had been addressed. O'Hanlon was at the time TMI Unit 1 Superinten-dent; he is presently Manager Operations Department, Evalua-tion and Assistance Division, at the Institute of Nuclear Pow-er Operations (INPO). The Office of Investigations concluded, as a result of its inquiry, that no further investigative ef-fort is necessary.
In his sworn statement, Book stated that he knew of no falsification of training records or of any other impropri-
, , eties at TMI. He explained in his statement that the language in his letter regarding documentation of more hours than were 4 actually used for training meant that when the training de-l partment sent out material that was to be covered in one hour, the shift did not always expend a full hour on the topic.
This resulted from interruptions to on-shift training caused i
by plant events, which would result in the shift's losing i
track of the time. Although Book therefore believed it was <
possible for some hours to be documented that were not used, he did not believe that it was the norm or that it was done blatantly or. irresponsibly.
James O'Hanlon, during his interview, stated that he had spoken with Book after receiving the letter and had learned that Book's concern was indeed that, although required training material was covered, it did not always utilize the l
time allotted it. O'Hanlon also stated that he has absolutely 59 1
l l
no knowledge of any record falsification at TMI and that he had monitored and was involved in training activities until he left the TMI site in December, 1978.
Regarding the other contentions in Book's letter, O'Hanlon stated that there was a real possibility that Book had not had any training since his requalification exam in -
! February, as was stated in the letter, because they were in a refueling outage at that time. (See GPU Ex. 2256, ANSI /ANS-3.1-1978, "American National Standard for Selection and Train-ing of Nuclear Power Plant Personnel" (" ANSI /ANS-3.1"),
5 5.5.1.1.1, page 9, indicating that lecture schedules should
! be spaced throughout the year, taking into consideration heavy vacation periods and infrequent operations, such as refueling periods.) As is discussed supra, pages 54-55, TMI operators did attend a substa:itial number of hours of regularly sched-uled classroom training.
O'Hanlon also stated in his interview thit Book's concern that management or training was not responsive to sug-gestions was unfounded. He explained that Book had been un-aware of plans to upgrade training facilities, and to include a sixth shift (see discussion supra, at pages 55-56), and had been unaware of the addition of an operator [as instructor] to the training department. O'Hanlon also said during'his inter-
- view that he had told Book that having the operations person-nel read emergency procedures on backshifts and during "down time" was a worthwhile training method. (See GPU Ex. 2256, i
60 1
i J
+ ---.,-iv. - ---. ,~, ,
. , , - - ~ . - ...--_,..e,,.--.--:.__,,,-..---..,wy ,..%..
,,,-mm--,._.,,----,-,.w- . , - - - - ~ ~ , - - - . . - - - - - - , . -
l
- * =
ANSI /ANS-3.1, 5 5.5.1.2.4, page 10, which indicates that on-shift review and self-study of emergency procedures was in
, conformance with applicable guidelines.)
- Theodore Book, the author of the " Book" letter, has clearly stated that he was not referring in that letter to any kind of impropriety regarding training records or otherwise. .
The other concerns raised in the letter, all expressed in equally hyperbolic form, are equally unfounded, as is indicat- -
ed by the interview with O'Hanlon. Thus, the Book letter clearly does not necessitate further inquiry nor raise any is-
~
sues relevant to the restart of Unit-1.
Management Structure
. The inconsistencies in the Comments' discussion of alleged problems in management structure (Comments, pp. 55-59) .
have already been discussed, supra, pages 50-51 n.6. The crux of this section of the Comments is TMIA's allegation that "Li-censee must be fundamentally incapable of taking responsibil-ity for its own wrongdoing." (Comments, p. 59.) Nothing in the Comments supports this conclusion. TMIA failed to show l
any " wrongdoing" by Licensee management. Moreover, the action l ,
items that were consistently developed as a result of self-initiated reviews evidence an important capacity for self-evaluation and improvement on the part of General Public Util-
~
ities and Metropolitan Edison. (See, e.g., B&W Exs. 883, 884 (follow-up minutes of 1978 Audit); B&W Ex. 186, TDR 001, pp.
I 61 l
1
56-58 (action items arising out of 4/23/78 event); Arnold dep.
349-50 (previous management audits acted upon).) -
Managers In the last section of its Comments (pp. 59-61),
TMIA purports to set forth "the more revealing statements and criticisms of Bob Arnold" Irom the record in the B&W case. In fact, the excerpts cited by TMIA do not reflect poorly on Mr.
Arnold, nor do they form the basis for any valid criticism of the exercise of his management responsibilities at GPU. To the contrary, as the following point-by point refutation makes clear, quite the opposite is true.
Several of TMIA's comments purport to rely on the so-called "1078 audit findings" found in 3&W Exhibit 843.
That exhibit contains the results of interviews of fifty su-pervisory personnel conducted over a period of five days by three selected employees. They were assigned to this task by l Mr. J. G. Herbein, Vice President of the Generation Division i
of Metropolitan Edison, to ascertain the concerns of the TMI supervisory staff within Met-Ed. Mr. Arnold testified that as Vice President of Generation for GPUSC, he did not recall hav-ing received the report of these interviews, which was ad-dressed to Mr. Herbein. Herbein did have some discussions with Arnold on issues raised by the effort as, based on their past relationship when they were both at Met-Ed, Arnold en-couraged Herbein to continue to use him as a sounding board (Arnold dep. 345-353). Arnold also explained that he "would 62
-e - - - -
yn -,,w--, -r-- ,-,n- - - - - - - --,y,,-
y
, .,-w,.-wm,.,,., , ,,, , ,-ew ,,,,,m ,-, - _, --- - ,, - , , , , , , , , , , _ ,_.,,,,-,.,_,wne
not have chosen the nomenclature of an audit for this type of effort" which was in fact one of a series of " attempts to un-derstand the concerns of people within the organization."
(Arnold dep. 345.) Accordingly, the TMIA comments regarding the " audit" must be viewed in the true context of the report.
-- TMIA criticizes Mr.' Arnold first for not recall-ing and then for not agreeing with concerns expressed ic the 3
interviews of TMI supervisory personnel that " upper manage-ment" did not have trust in their abilities and permitted too much "buckpassing" by not holding supervisors accountable in 4 their areas of responsibility. As to the purported lack of recollection, the 1982 deposition testimony to which TMIA cites makes clear that while Mr. Arnold four years later did ,
not recall these specific findings in a Met-Ed memorandum that may not have been sent to him, he did testify that "these types of concerns on the part of the supervisory levels within the organization were generally known to me and I think that I would not agree with many of those assessments." (Arnold dep.~
I 388.) In regard to the merits of Mr. Arnold's disagreement, as a member of " upper management" he was certainly in a posi- -
\
' tion to express his view as t'o the real source of the concerns l expressed. Thus, Mr. Arnold explained that, in hi3 vi'ew, the l
l interview responses to some extent grew out of the require-l l ments in the nuclear industry which differ from those in the areas (such as fossil plants) that many of the employees were accustomed to, and that upon entering the nuclear industry 63 i
I
,. _ _ , _ . . . _ . . _ _ . , _ . _ . _ _ . _ _ . _ _ . _ _ _ _ . . _ _ . _ . , _ _ . , . . , _ . _ _ _ _ . _ . _ . . . . . , , _ _ - . _ _ ~ _ _ . _ . - _ _ _ .
some reorientation had to take place (Arnold dep. 388). Thus, what a lower level employee might characterize as "buckpass-ing" would merely reflect a disgruntlement with the fact that Met-Ed and GPU supervision and management were more involved in his day-to-day activities than he was accustomed to in pri-or, non-nuclear positions.
-- TMIA criticizes Mr. Arnold for not recalling or agreeing with a post-accident statement by Gary Miller about the difficulty of firing someone at TMI. However, TMIA cites nothing which suggests that Mr. Arnold was not justified in disagreeing with Miller that GPU had any particular problem in this regard. In view of federal regulations on equal opportu-nity for various groups, unfair labor practices, union in-volvement and the like, the extensive procedural requirements to support the firing of employees is a universal constraint on American industry.
-- TMIA. implies incorrectly that there is some in-consistency between Mr. Arnold's testimony that he disagreed that " members of the operating staff" had too great a work load at TMI-2 in 1977-78, and his testimony that there was a "need for very heavy commitments on the part of senior members of the staff." The fact that relatively heavy time commit-ments are required by senior staff members to complete a major project, such as the startup and test of a nuclear plant, does not mean that such commitments are "too great" in the context of such a project.
64
~
-- TMIA implies that Mr. Arnold admitted he was re-sponsible for solving the problems identified in the 1978 au-dit and other audits inasmuch as he was the " focal point" for addressing these issues. TMIA mischaracterizes the testimony.
The term " focal point" was used by Mr. Arnold in response to a question in his deposition as to why Mr. Herbeln, a Met-Ed em-ployee, was discussing some of the problems raised by the 1978 interviews with Arnold, who was head of the GPUSC Generation Division. Arnold simply explained that some of the issues, which had been raised by Met-Ed's employees, were already be-ing addressed by him on a system-wide basis, so that he "was sort of the focal point for assuring that the systemwide dis-cussion of these issues was undertaken [and] that we were able to get resolution of some of the systemwide policies and prac-tices necessary to address these kinds of problems." (Arnold l dep. 352-54.) This testimony plainly reflects a responsible approach by management.
TMIA also asserts that Mr. Arnold purportedly "ad-l mits" that the same problems discussed in the 1978 interviews were found in a 1975 " audit" that Mr. Arnold had initiated.
Mr. Arnold explained that in 1975 there had been an analogous _
although less structured effort to ascertain the nature of managerial problems and to obtain recommendations to address them (Arnold dep. 347-48). however, he explicitly refuted the suggestion by the cross-examiner that the same problems had continued, and instead explained that circumstances and spe-l l
65 l
. . - - - - . - . . . - . . - , - _ - - . . - . . _ -,~
cifics had changed. The reasons that the areas examined for problems in 1975 and 1978 were the same is, as Mr. Arnold ex-plained, because they were a reflection of the kinds of activ-ities in which the company was engaged during those time peri-ods (Arnold dep. 350). TMIA is plainly off base for criticiz-ing Met-Ed management for conducting interviews which focused on subjects similar to those covered three years earlier.
Rather, management acted responsibly in making continual ef-
~
forts via the interview or " audit" procedure to revisit poten-tial problem areas, identify current concerns and take posi-tive steps to make improvements.
The 1978 interviews referred to by TMIA indicated that, as with prior analogous efforts, the focus of the 1978 effort was on conc. erns over the division and assignment of management responsibilities within the company (B&W Ex. 843).
Testimony in the B&W trial record establishes that steps were being taken to deal with just those concerns. For example, Mr. Miller testified that the company was moving towards the mane tement organization that had been suggested by an indepen-
! dent audit conducted in 1977 by an outside consulting firm i
(Miller dep. 74). Those changes included putting Mr. Miller in a more direct reporting relationship with Mr. Herbein, and improving the organization of the training department (Miller
! dep. 74, 541-43). The responsiveness of Met-Ed and GPUSC man-agement to the issues raised in company reviews is further ev-idenced by the minutes of meetings held after the 1978 " audit" 66
was performed, in which action items were established for each area of concern discussed in the audit (B&W Exs. 884, 883; Herbein dep. 362). In view of these and other actions being taken to respond to identified problems, Mr. Arnold was acting appropriately and responsibly as a member of GPUSC's senior management.
-- TMIA cites Mr. Arnold's testimony that he under-stood Gary Miller's statement in 1979 that a " dollar crisis" existed at TMI to refer not to a shortage of " total dollars" available at TMI, but rather to the dollars available for
" compensation packages." (Comments, p. 60.) There is no ba-sis in fact for disputing Mr. Arnold's interpretation of Mr.
Miller's statement, which appears in a memorandum under the heading " Personnel Retention & Hiring." Mr. Arnold further testified that in the 1978-79 time frame, he and others were examining the competitiveness of the salary levels and grades l of the'TMI plant staffs; that he talked with human resources people regarding data on industry practice; and that in early 1979, fairly substantial modifications had been made to staff salaries (Arnold dep. 404-05).
l
-- TMIA's effort to criticize Mr. Arnold with re-spect to Met-Ed's selection of Richard Zechman as head of its training department at TMI-2 is misguided for the reasons stated above in the discussion of Mr. Zechman's qualifications l
(pp. 45-46). Zechman's position was,primarily that of an ad-ministrator and a teacher of theory, for which he was well-67 l
l l
l
_ __ _ _ =__ _ . -
4 qualified and which did not require an operator's license.
Furthermore, Mr. Arnold's testimony, like that of Mr. Herbein and others, indicates that there was appropriate direction of the training program during the period Zechman was studying for his license (Arnold trial tr. 1705).
-- TMIA asserts that Mr. Arnold admits that Met-Ed failed to instill in its operators a sense of respect for its post-accident training program. The statement to which TMIA refers was made to the NRC by an attorney for GPU Nuclear, who was referring to incidents of operator cheating after the ac-cident. Mr. Arnold was asked about that statement during the B&W trial in the context of an evidentiary issue. The fact that Mr. Arnold supported the company's candid recognition, expressed through counsel, and had been involved in developing i
that assessment of the problem indicated by the incidents of
- operator cheating, reflects well upon Mr. Arnold, as do the corrective actions subsequently initiated by the company to further ensure appropriate operator attitudes towards training and testing.
During.the course of Mr. Arnold's direct testimony, counsel for B&W objected from time to time that Mr. Arnold, as head of' Generation of GPUSC in Parsippany, had had insuffi-
, cient direct contact with certain day-to-day affairs at TMI to l
l provide admissible testimony on those matters. (See Arnold trial tr. 1504, 1506.) Therefore, on cross-examination, al-though testimony regarding post-accident events generally was 68 :
o ,-- .
excluded from evidence under Rule 407, B&W's counsel was al-lowed to ask about Mr. Arnold's awareness of the July, 1979 VV ,
incident in order " arguably" -- as the Court put it -- to de-velop contentions as to Mr. Arnold's knowledge of Met-Ed oper-ations in general and particularly training. TMIA misleading-ly suggests that the Court's ruling went to an issue of fact rather than being, as was the case, an evidentiary ruling as to the areas on which Mr. Arnold could be examined. The Court merely ruled that evidence concerning Mr. Arnold's knowledge I of and reaction to the VV incident was admissible. Mr. Arnold then went on to testify that when he learned of the VV inci-dent he took quick and decisive action to apply appropriate sanctions for the circumstances as he then knew them, and to
- remove Mr. VV permanently from his supervisory position (Ar-nold trial tr. 1749-50).
-- TMIA criticizes Mr. Arnold because he allegedly "still does not recognize that Mr. VV cheated in 1979." For this erroneous proposition, T!!IA cites Arnold's deposition taken in July 1982 -- a time when all of the facts relating to VV's conduct were not yet known to Mr. Arnold. Even at that time, Mr. Arnold had imposed serious disciplinary action upon VV by having him permanently demoted. Moreover, after reading the ASLB Partial Initial Decision (Reopened Proceeding - TMI-l Restart), dated July 27, 1982, on operator cheating issues, Mr. Arnold, in August of 1982, initiated the Speaker investi-gation, an independent inquiry into, among other things, the 69
circumstances surrounding VV's actions on the make-up examina-tion in the 1979 training program. After obtaining all of the facts resulting from these investigations, Mr. Arnold recog-nized that his initial understanding of the circumstances of the incident may have been faulty and led to an erroneous con-clusion as to whether VV had cheated (Oral Presentations on -
TMI-l Restart, Nov. 9, 1982, Presentation of Mr. Arnold, pages 31-32).
-- As is discussed above (pp. 7-8), TMIA blatantly miscites Mr. Arnold's testimony for the proposition that "Ar-nold still insists that pressurizer level and pressure trend together." In fact, the testimony cited by TMIA deals not with Mr. Arnold's present understanding, but explicitly and exclusively with Mr. Arnold's Navy training, received over ten years before the TMI accident.
4
-- TMIA next misconstrues Mr. Arnold's testimony with respect to the response of RCS pressure to a rise in pressurizer water level. During normal operations, there is a steam space at the top of the pressurizer. Dr. Richard Lahey, expert trial witness for GPU, testified that an increase in pressurizer water level would result in some condensation of that steam and therefore that one would not experience the same increase in pressure as one would when dealing with a perfect non-condensable gas, as in Boyle's law (Lahey trial tr. 330-31). The general context of the Arnold testimony cited by TMIA, as well as the question posed to Mr. Arnold by 70
, -B&W's counsel, make it clear that Mr. Arnold was testifying to the same effect as Dr. Lahey (Arnold trial tr. 1605).
Moreover, TMIA mischaracterizes the hot functional test of September 1977. It is wrong that "it took days to re-move the steam which had formed in the reactor coolant system." The test logs indicate that the day after the possi-bility of a steam bubble was suspected (its existence was nev-er proven), the engineers applied nitrogen to the pressurizer which apparently collapsed or purged from the system any bub-ble that may have developed in the RCS (B&W Ex. 175 at p.
WO6073). TMIA's reference to trial tr. 1473 -- a colloquy among counsel on an evidentiary point -- has no bearing on the proposition for which it is cited. Nor is there any other ba- -
sis for TMIA's allegations that "one of the most impcitant lessons" of the September 1977 test event, was that "a steam bubble cannot be compressed by a mere increase in pressure."
This conclusion by TMIA is wrong -- in the circumstances of the September 1977 event, any possible steam bubble in the legs of the reactor coolant system would have been quickly compressed by an increase in pressure. (See B&W Ex. 877, TDR
! 286, 10/7/81.) In any event, Mr. Arnold did not testify about i
this proposition one way or the other. He merely said in re-i gard to the steam space at the top of the pressurizer, that l
l steam is not a perfect gas which is subject to Boyle's law, l and, that since there is some condensation of steam, it dif-a fers in lehavior from an ideal gas. That proposition is an l .
l
,' 71 I
i I
l
inarguable scientific principle, supported in the trial by Dr.
Lahey as well as a B&W document (GPU Ex. 2347, p. 55).
-- TMIA's final Comment, with respect to Mr.
Arnold's testimony on the problem of water in the instrument air lines, is also wrong. As set forth above (pp. 27-28), the testimony it cites deals exclusively with Mr. Arnold's under-standing of the problem in 1978, not with his perception of the problem after the accident.
.TNIA has failed to present any evidence from the GPU 4
- v. B&W trial record, or any other source, that detracts from Mr. Arnold's exemplary record throughout his tenure with GPU Nuclear, GPUSC and Metropolitan Edison.
4 l
[
t-72
e a e 8 e S e
0 0
e T
APPENDIX A
.r.,.,,
D co May 2, 1978 Su:wel TMI 2 Cooldown Transient of April 23, 1978
[W Serm,;,ee To Mr. R. W. Xcaten Locabon Mountain Lakes We have been requested by Met-Ed Genera, tion (J. C. Herbein) to take the lead respon.wibility for conducting a thorough investigation and analyses of the subjoet incident. You are assigned as Chairman of an AdHoc Committee -
consisting of yourself and Hessrs. T. G. Broughton, R. C. Cutler, Jr., J. L.
Seelinger, R. M. Toole und E. C. Wallace - to conduct the investigation and prepare appropria:e interim and final reports. The investigation should include:
- 1. A detailed chronology of the physical events from the time of initiation of the incident until establishment of normal plant condicions. ,
- 2. A review of the automatic response of the plant and the manual actions taken by the operator, the consistency of both automatic and manual response with the design philosophy for the unit, and the appropriateness of the design philosophy in light of the circumstances of the incident.
- 3. The cause for the failure of the steam generator safety valve discharge pipe and expansion joint 1.iners and the adequacy of the corrective action.
! 4. A preliminary review of any concerns about the adequacy of l
Pl ant instrumentation and control raised by Start-Up and Test or plant staff personnel.
! 5. Any 1tems which come to your attention which you believe warrant ,
further investigation. -
Your committee should provide a preliminary oral report in a meeting scheduled with Met-Ed Generation at its facilities at Wyomissing 'for 1:00 p.m.
on Honday, May 8,21978. The primary purpose of that meeting will be to discuss l the information developed as of that tima and identify any open items which l need to be resolved prior to a re-start of the unit. Also, at that time please be prepared to provide me with a target date for a preliminary written report.
A group headed by Jim Seelinger of Met-Ed has been collecting and analyzing information on the incident and I request 'you work with Ron Toole and Jim to insure we take full advantage of the efforts to date in this regard. You have the freedom and anchority to utilize resources within the Division and at Burns and" Roe as appropriate. Please keep Hessrs. Heward, Mirst and Wilson apprised of your efforts.
h R.C.prn'old att
.- . - - . . - _ . . _ - - -____ A x. M _ _ _ - - _ -
V. RECOMMENDATIONS A. Prior to Restart
- 1. The liners in all horizontal bellows should be replaced by thicker liners designed for the anticipated service conditions. (Complete)
- 2. Analysis by Burns and Roe indicates that the liners in the vertical bellows may be acceptable for at least the next fuel cycle. This analysis should be carefully reviewed for acceptability. (Complete)
- 3. Tests should be performed to determine (as precisely -
l as possible) the blowdown characteristics of each safety valve. This can be accomplished by heating the system with the reactor coolant pumps to achieve a steam pressure of 950 - 1000 psi, using a hydraulic essist to individually pop the safety valves, and observing at what system pressure valve reseats. (New valves being installed)
- 4. The C1 contamination caused by the safety injection l should be cleaned up. (Complete) ~
- 5. Defective components in Nuclear Instrumentation Channels NI-5 and NI-7 should be replaced. (Complete)
- 6. Recorders should be installed to monitor selected positions in the circuitry for NI-8, to attempt to determine the maint of origin and cause of the spurious spikes which have been observed on this channel. (Complete)
- 7. Set points should be adjusted in the feedwater control system to eliminate the instability which occurred in the transfer from the start up control valve to the main control valve. To help select the appropriate transfer point, the
~
shut off leakage in the main feedwater control valve should be measured, if necessary. (Complete) l 8. The control system for the pressurizer pressure and level should be tuned up to reduce the cycling which was observed.;during power escalation prior to the trip. (Complete)
- 9. Checks should be made on a representative sample of pressurizer heaters, to verify that the low level in the pressurizer did not affect the integrity of the.. heaters.
(It is recognized that the best test of the heater integrity may be obtained during system heatup and operation).
(Complete) '
9 e
b
-- ---.rw., e-. g .w --.,---vv- - - mw-w ww*- e--- ew w ---*-'w--**-* * - - " - ' ' " " - - ' - - " " * ' ' - ~ ~ ^ ~
\ ., . : . % , . l 10.
The ifeedwater that latch system should be checked to vor.ify designt dwill P, fully close the FW Block Valve at (or near) and that it will not of the emergency feedwater system. interfere with operation (Complete) 11.
The B&W assessment of the impact of the transient on the RCS should be reviewed for concurrence with the assumptions (Complete) as to the nature of the transient.
B. As Soon As Practical i
1.
The Task Force understands that a source of sodium Hydroxide 0.1% has beenwith located.
chlorine contamination less than approxLmately An adequate supply of this low chlorine sodium hydroxide should be purchased and used to replace the existing sodium hydroxide. (Complete) 2.
The logic for sodium hydroxide injection should be modified so that aand system pressure combination low level inofthe lowhorated reactorwater coolant storage tank is required before injection occurs. This
. requires as well asa wiring modification changes. to the tech ical specifications, .
(Complete) 3.
Some method of monitoring when safety valves pop and (if practical) when they reaeat should be installed. (Complete) 4.
Open/ closed indicators for the turbine bypass valve and in thethe main feedwater control block valve should be installed room. Action -
- GPUSC 5.
There are some indications that the "A" turbine bypass valve responded sluggishly to the transient overpressure.
controller should The be res3onse checked. of this valve and its Action - CPUSC 6.
There were indications that the #3 turbine stop valve l did not close completely. The valve and the indicator should be checked and if necessary repaired.
Action - GPUSC 4
, , _ _ , ,,_,----,,e = - ' " '"""* "~ "
- , ' . .'w s
, , C. Longer Term
- 1. An alternate configuration for the relief valve
. piping should be explored in an attempt to eliminate the need for be.llows. Action - GPUSC
~
- 3. The following aspects of the plant control system should be reviewed to determine if modifications are desirable. (GPUSC)
- a. The transfer point from the startup feedwater -
valve to the main feedwater control valve, and the valve trim.
- b. The set points for the turbine bypass valve,
- c. The actuation logic for the atmospheric dump valve. '
- 4. An in-house capability for performing transient analysis-should be developed. This should permit a prediction ofchow the system would respond under various upset and emergency transients, as well as after-the-fact analysis of: unanticipated transients. In addition, this capabiliQ can be 'used to optimize the control system. (In Progrest
Action - GPUSC
- 6. To. permit accurate post event analysis, the possibility of obtaining up to five minutes of post--trip, five resolution data, such as that available from the l reactimeter, should be investigated. Action - GPUSC 7
O e
9