ML20056B518

From kanterella
Jump to navigation Jump to search
Responds to NRC Re Order Imposing Civil Monetary Penalty & Violations Noted in Insp Repts 50-327/90-01 & 50-328/90-01.Corrective Actions:Organizational Capabilities Reviewed.Payment of Civil Penalty Wired to NRC
ML20056B518
Person / Time
Site: Sequoyah  Tennessee Valley Authority icon.png
Issue date: 08/20/1990
From: Medford M
TENNESSEE VALLEY AUTHORITY
To:
NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION II)
References
IEB-88-004, IEB-88-4, NUDOCS 9008290097
Download: ML20056B518 (17)


Text

I- -

,s ' 24

,1 l p; aTENNESSEE VALLEY AUTHORITY CH ATTANOOGA. TENNESSEE 374ot -

6N 38A Lookout Place AUS 201990 I

L lU.S'.'NuclearIRegulatory Commission Regional Administrator,' Region II 101'Marietta Street, NW,' Suite 2900 Atlanta, Georgia;i30323 Gentlemen:

In the. Matter of-

)

Docket Nos. 50-327 H

Tennessee Valley Authority

)

50-328

[

SEQUOYAH NUCLEAR PLANT (SQN) - NRC INSPECT' ION REPORT NOS. 50-327/90-01 AND p

50-328/90 ORDER IMPOSING A CIVIL MONETARY PENALTY ~ $75,000 Enclosed is TVA's: response to ~J. Lieberman's letter to 0. D. Kingsley, Jr.,

3 dated July 20,'l1990, which transmitted an Orde,r Imposing a Civil Monetary Penalty.as.a. result' of violations at SQN involving the potential for residual-

~

heat removal pumpdeadheading.

' Enclosure '1.toi this fletter concains the additional information requested. in'

'Mr.;Lieberman'siletter. to this letter contains TVA's revised violation ' response incorporating the response to Item 1 of Mr. Lieberman's letter as discussed with J. B. Brady of your staff. The revised response L

. clarifies the reason and corrective action for Violation A, Example 1.

.For NRC'.s convenience, changes from TVA's original response are designated'by l

l revision bars.

Payment of the civil penalty in the amount of $75,000.is being l'

wired to:the Director, Office of Enforcement.

I

_If.youlhave any questions concerning this submittal, please' telephone I

m-M.-A'.. Cooper at,(615) 843-6422.

3 Y

O Very truly yours, o

TENNESSEE VALLEY AUTHORITY

[

WM Mark 0. Medford, Vice President Nuclear Technology and Licensing

' Enclosures cc: 'See page 2 m.

nnne, U~CL1 l

/

9008290097 900820 j

i PDR ADOCK 0S000327 PNV An Equal Opportunity Emetoge g.

--M

,l'l

.s cmg,

3 _-

e U.S. Nuclear-Regulatory Comission Mh jQ'g y,w cc (Enclosures):

Ms. S. C. Black, Deputy Director Project Directorate II-4 U.S. - Nuclear' Regulatory Comission One White Flint, North

.g"'

. (1555 Rockville Pike Rockville, Maryland20852 Mr.-J.;N. Donohew, Project Manager.

U.S.. Nuclear Regulatory Comission One White Flint, North 11555 Ro kv111e-Pike Rockville, Maryland 20852 NRC Resident Inspector

^

' Sequoyah Nuclear Plant 2600 Igou Ferry Road.

- Soddy Daisy, Tennessee 37379 U.S. Nuclear Regulatory.Comission~

ATTN:

Document Control Desk Washington, D.C.

20555 Mr. B. A. Wilson, Project .hief U.S. Nuclear Regulatory. Comission Region II 101'Marietta Street, NW, Suite 2900 Atlanta, Georgia 30323 g

O i

e

~

'V fVy

)

en

+

4.

'# iMi ENCLOSURE'1 m

h,g N

ADDITIONAL INFORMATION-REQUESTED IN..

i MR. J. LIEBERMAN'S LETTER TO 0. D.'KINGSLEY,-JR.

DATED JULY 20,-1990 1

0-

1. :Your. response _of May~9, 1990,_did not clearly address the reason and corrective action for Violation A, Example 1.

Some>information on1this violation was_provided.in'your enforcement. conference < follow-upfletter

- 1 dated March 5, 1990. Please-address.why no' action was taken-to preclude damage to a: residual heat removal (RHR) pump because of deadheading'when, information; indicated.that the SQN design was susceptible to;deadheading'.

Specifically, your. response did not_ address why 20 minutes was used.during:

your emergency procedure review as an acceptable time for-RHR punp
deadheading when TVA's calculations indicated that damage would occur s

_after approximately 11 minutes. That information is considered important.

H because had 11 minutes been~used, it appears that.these violations might.

not have occurred.

In addition, please provid: your corrective actions to i

4 assure ;that when potential degradation of. plant dystems 'is indicated, appropriate measures will be taken to prevent the consequences.

,l

, Response In addition to the following discussion, TVA's response pursuant to 10 CFR 2.201 has been revised to provide clarification of the reason and 2

corrective action for Violat'.on A, Example 1. - The revised violation response is included as Enc'.osure 2 to this letter.

! Actions to preclude the RHR pump deadheading issue prior to

~ December 5,1989, were taken but those 'hetions were not adequate.. The reason further actions were not taken was the belief that the corrective l

action plan.put in place in response to Bu11etin 88-04 was adequate to L

ensure that pump deadheading would not occur. As is discussed in more detail in the revised violation response in Ei.alosure 2, it was believed J

that an imbalance between the pumps that could result in deadheading did not exist and that mecsures were in place to monitor for degradation in '

' 7 l

-advancelof development of such a condition.

However, because the 1988 evaluation performed for Bulletin 88-04 was_ inadequate, this violation l

occurred.

In hindsightc we agree that use of 11 minutes in establishing j

administrative controls would have eliminated some of the vulnerability of 1

L3 the chosen corrective action plan to implement and evaluate errors.

Weaknesses. identified in TVA's evaluation of and response to the subject j

-bulletin are not considered to be indicative or generic inadequacies in j;

addressing conditions,with potential for degradation of plant systems. -

' Corrective actions taken to address weaknesses in handling of bulletins and more generally in responding to conditions adverse to quality are included in TVA's violation response.

Future conditions indicating 14 potential degradation of plant systems will be addressed in accordance with TVA's corrective actica program.

b I

. x 6

W

?

m [

2.

Please clarify your position on the utilization of the dif ferential pressure data in your review of the deadheading issue bstw en November 28 and December 5, 1989s In your letter of March 5, 1990, s a indicated that TVA was not satisfied with the perfwrmance of the system eng!neering group in the evaluation of the data. This concern is not reflected in your i

i response to the violation.

Response

. The performance of the Technical Support organization during the discovery and evaluation of the Rl!R pump deadheading issue between November 28 and December 5, 1989, could have been improved, The operobility assessment of the pumps could have been taore thorough, and thn evaluation of the pump data could have been more expeditious. As indicated in the May 9, 199J, response, after notifying Wuclear Engineering ZE) of *.he expected problem in meeting the eight pounds per square inch differential (psid) acceptance criterion on November 27, 1989, Technical Support continued to evaluate the RHR pump data in parallel with NE. 'As previously discu m d, considerable scatter was observed in the data that introduced uncertainty into the evaluation of the pump-to-pump differential pressure.

N subject data were not direct readings of pump-to-pump differential head, but rather were differential pressure readings across each pump (suction to discharge) taken at different times on each pump during the operating cycle.

Thus Technical Support did not reach a firm conclusion in early December 1989 because of the data scatter and doubts about the averaging 4

methods to be used.

The RHR pumps were still believed to be operable at that time based on the 1988 NE evaluation performed for Bulletin 88-04 and the inclusion of the 20-minute limit on parallel pump operation in the

. operating and emergency procedures. A more detailed and expeditious review of the basis for the 1988 evaluation at.d data available since that time could l

resulted in earlier identification and corrective. action.

TVA had previously recognized weaknesses in the systems englueering l

program and had initiated actions to upgrade this area. At discussed in the March 5,1990, letter, trajor reorganization and management changes had been made within Technical Support during the June to Oe'.aber 1989 i

timefreme.

A new Technical Support manager with experleace in managing a successful system engineering program was hired from outside TVA. TVA vas also actively (in March 1990) recruiting experienced system engineers and had established a lead engineer concept to best utilize existing l

strengths.

A reevaluation of existing supervisors and engineers was being conducted and the key elements of system engineering (ownership of problems, leadership in problem resolution, sensitivity of regulatory a~nd operational aspects and focus or. operability and problem solving) were i

being reinforced through direct involvement of Technical Support's management.

The TVA letter to NRC dated March 12, 1990, describing the status and results of performance improvements at SQN related to the 1989 Systematic Assessment of Licensee Perforrrance (SALP) responce discusses some of the improvements to the system engineering program initiated since tLs sumer of-1989.

In ts.e July 16, 1990, response to the 1990 SQN SALP report, 'VA discussed additional improvements being made to enhance plant support including the formalized involvement of system enBineers in the preventive maintenance program, increased involvement in the modification program, and development of a formalized system engineer trending

?-

,3, Y

program. Continued management emphasis on increased system engineer knowledge through the walkdown program and establishir.tr.t a' a formal system training program including classroom, plant walkdowns, and simulator training are underway.

In summary. TVA had previously recognized weaknesses in the Technical Support area and has initiated a number of efforts that are believed to have significantly improved the organization's performance. The perf,prmance of Technical Support with regard to resolution of the RHR pump deadheading issue did not meet expected standards of performance in addressing such issues. Hevever, Technical Support has improved over the past year and, overall, is performing effectively. TVA will continue to focus close managerrmat attention on this area to ensure continued progress.

3.

TVA addressed whether the response to NRC Bulletin 88-04 contained complete and accurate information in relation to the RHR pumps. However, the revised bulletin response dated March 15, 1990, also indice ed that the portion pertaining to the auxiliary feedwater pumps had been revised.

Please address the errors or omission of information in the entire Bulletin 88-04 responae at.d corrective actions to preclude submittal of incomplete or inaccurate data in the future.

Res pec.ste The original Bulleti- *,8-04 response dated August 2,1988, described the corrective action thut had been implemented for the motor-driven auxiliary feedwater pumps (MDAWPs).

That corrective action was a precaution in the system operating instruction to ensure 'the MDAWPs would be operated 'n an intermittent flow mode during the start-up to five percent power mode of operation.

The response also listed three hardware modifications that were being evaluated, and stated that the bulletin response would be upcated if any system modifications were implemented.

One of the potential hardware modifications listed in the response was to add an additional bypass line to increase MDAWP recirculation capacity during low steam generator flow conditions.

The revised bulletin..oronse dated March 15, 1990, added three items to the MDAWP portion:

A.

A clarified description of intermittent flow operation, B.

A brief description of the full-flow MDAWP test program designed to detect and trend any pump degradation and s

C.

As a result of the hardware modification evaluation, a commitment to add a bypass line in parallel with the miniflow recirculation piping to allow the pumps to operate at a higher flow rato during low steam generator flow conditions.

The information added to the MDA NP portion of the revised bulletin response was of a clarifying and supplemental nature.

1

,~..

I

_4 P

t Programmatic changes and improvements that had been instituted prior to the discovery of the RHR deadheading event to ensure completeness and accuracy in evaluating licensing issues and developing responses for j

submittal to NkC are described in TVA's May 9, 1990, violation response.

4.

Please clarify the statement that "TVA believes that this event resulted from past programmatic weaknesses that had been previously recognized and for which extensive corrective actions had already been implemented or ini,tiated."- Although NRC acknowledged that TVA had taken extensive L

programmatic corrective actions for some of the problem areas, it is ' nt clear that all areas of concern were previously identified or corrected.

Examples it.clude the deficiencies in the emergency operating procedure revision procesa and the evaluation and promptness of corrective action in regard to the pump differential pressure data on Unit 1.

Response

A number of corrective actions and impro'vements were initiated both before and after confirmation of the pump deadheading problem.

It was not TVA's intention to convey that all areas of concern had been identified or corrected prior to discovery of the deadheading problem. The corrective actions described in the May 9, 1990, response were characterized as either reactive or proactivet reactive if they were initiated as a result of the deadheading issue; proactive if they were initiated prior to discovery of the problem to address underlying programmatic weaknesses that TVA had already recognized.

Proactive corrective actions are-especially significant because they demonstrate a licensee's willingness and ability to aggressively identify add correct problems.

A few of the proactive corrective actions discussed in the May 9, 199f, response are:

A.

Programmatic changes affecting how licensing issuet are managed using the licensing project management concept.

B.

Functional responsibility changes made with the NE trganization.

l C.

Organizational and management changes made to the Tectnical Support organization.

D.

Changes regarding how the Nuclear Experience Review program is staf fed and nmnaged.

E.

Management's actions taken to encourage safety consciousness with regard to problem identification and reporting.

TVA's May 9, 1990, response identified the cause of the violation as an inadequate evaluation for and response to NRC'sBulletin 88-04 in 1988.

The first three' corrective actions listed above involve the upgrading of organizational capabilities and focusing of responsibilities in the specific areas that resulted in the inadequate bulletin response in 1988.

l

C, f,

<a

.m,

.s'7f, y 3'l Y

t a

)

pl s

1 e

.(

.s a

[

+,, ' F a

J s:, -

i w

These corrective actions were implemented af ter the response to l

'Bulletin 88-01., but before discovery of the problem in December 1989. Had

.these actions been taken prior to the original: bulletin response, the U-violations would likely not have pecurred.. Thus TVA considers the initiating i

Levent (inadequate bulletin response in 1988) to have resulted from past programmatic weaknesses that had been previously recognized and.for which extensive corrective actions (including the first..three listed above) had-

'already been implemented'or initiated. Additional weaknesses, such as the-emergency oper

.t+r, 'rocedure revision process and corrective actions'

}

identified as a resi; of this event, are detailed i:t our May 9,1990, r

response.

s p -r

' i

[

e i

f r

~ >

a Se il i

i c,

i 1,

l, 1, -

l3, y

l a

4 9

.i

I'

(

~ -.

ENCLOSURE 2 REWISED RESPONSE TO NRC INSPECTION REPORT NOS. 50-327/90-01 AND 50-328/90-01 D. M. OkUTCHFIELD'S LETTER TO 0. D. KINCSLEY, JR.,

DATED APRIL 12, 1990 Violation _50-327, 328/90-01-03 "A.

10 CFR Part 50, Appendix B Criterion XVI, corrective Action, requires in past, that measures shall be established to assure that conditions adverse to quality, such as failures, deviations and nonconformances are promptly identified and corrected.

NRC Bulletin 88-04 Potential Safety Related Pump Loss, issued May 5, 1988, alerted licensees to a significant condition adverse to

-quality that involved the potential for the deadheading of one or more pumps in safety-related systems that have a miniflow line common to two or more pumps or other piping configurations that do not preclude pump-to-pump interaction during miniflow operation.

Licensee engineering calculation DNE SQN-74-D053, dated July 22, 1988, determined that RHR pump damage would occur for a pump that was run deadheaded for greater than 11 minutes.

10 CFR 50.9 requires, in part, that information provided to the Commission by a licensee, be complete and accurate in all matcaial respects.

Licensee letter to the NRC in response" to NRC Bulletin 88-04, dated August 2, 1988, stated that the potential existed for deadheading a safety-related RHR pump due to pump-to-pump interaction under miniflow conditions when the head differential between the pumps exceeded 11 pounds per square inch (psi). The letter also stated that recent surveillance test data demonstrated that the head differential between l

i the two RHR pumps was less than 11 psi, ensuring a minimum flow of 100 gallons per minute to allow pump operation for up to 20 minutes without requiring operator intervention.

Contrary to the above, as of December 5, 1989, the licensee failed to l

adequately identify and correct a significant condition adverse to quality regarding the potential for safety-related RHR pump damage from deadheading due to pump-to-pump interaction during miniflow conditions in that:

l 1.

No action had been taken to preclude damage to a RHR pump should deadheading develop due to pump-to-pump interaction under miniflow j

conditions, until a special test demonstrated that the Unit 1 RHR pumps deadheaded under those conditions on December 5, 1989.

L 2.

The licensee's evaluation of Unit 1 RHR pump surveillance test data, referenced in their August 2, 1988 letter to the NRC, was inadequate to identify that an RHR pump was likely to deadhead due to pump-to-pump interaction, as the majority of the test data from July 1977 through August 1988 indicated that the head differential pressure between the pumps exceeded 11 psi. As a result inaccurate information was provided to the Commission on August 2, 1988."

I

.g 8

l 4

Admission or Denial of the Alleged Violation (Violations A.1 and A.2)

TVA admits the violation.

Reason for the Violation (Violations A.1 and A.2)

The cause of both examples of this violati9n was an inadequate evaluation of and response to NRC Bulletin 88-04 in 198'4.

Violation A.1 was a direct result of Vio(ption A.2.

A brief review of the event chronology will illustrate this cause.

During TVA's evaluation of the bulletin, the potential for RHR pump-to-pump l

interaction was recognized, and an analysis was performed by Nuclear Engineering (NE) personnel in Knoxville,-Tennessee. This analysis showed the following:

1.

Eleven minutes of residual heat removal (RHR) pump operation would be needed to exceed design temperature at deadhead (no flow) conditions.

2.

Flows as low as 50-100 gallons per minute (gpm) would prevent an excessive temperature rise in the pump.

3.

A relationship exists between weak pump flow and pump-to-pump differential head (e.g., 0-gpm flow at greater than 12.6 pounds per square inch differential [psid]

100-gpm flow at 11.1 psid, and 200-gpm flow at 8.9 psid).

The pump vendor was also contacted regardi6g operation with lower recirculation flows. The vendor established that operation at 100-gpm flow was acceptable for no more than 20 minutes and that continuous operation required 500-gpm flow. The most recent RHR pump data at that time (second quarter 1988) showed pump-to-pump differential head as 10.5 psid, which was interpreted as approaching a low-flow condition (not a no-flow condition).

Other historical pump test data was not considered during the NE evaluation, i

This oversight was the result of inadequate coordination and communication i

between Knoxville NE and knowledgeable site organizations.

Having established in the NE evaluation that pump deadheading was not occurring, the corrective action plan was to monitor future pump performance to be alert for the onset of deadheading conditions. As a result of the analysis and vendor information, 11.1 paid and 20 minutes were selected as the limits to be placed on RHR pump operation with a lower differential pressure threshold to initiate further investigation.

These limits would. guard against a potentially low-flow condition (100 gpm), and thus, were considered adequate to protect against potential future pump deadheading. The emergency procedures were reviewed and judged to be adequate to ensure that parallel RHR pump operation on miniflow would be terminated within 20 minutes.

Thus, the use of the 20-minute limit during review of the emergency procedures was consistent with the belief that a pump deadheading condition did not exist and would not develop undetected. Stopping the pumps within 11 minutes was not considered necessary; while it would further prevent deadheading from a.

r occurring, it would unnecessarily (as then be'ieved) require stopping an automatic safety system early in the accident scenario.

Benefits were perceived with the alternative approach iseng utilized.

Similarly, the system operating instructions were revised to limit parallel pump operation to 20 minutes. To provide for continued monitoring of the RHR pumps, a surveillance test program was set up to be performed during each refueling outtgo.

The suryeillance test program averaged RHR pump performance data collected during routine testing over the operating cycle and specified an 8-psid acceptance criterion as the threshold value above which an evaluation of the data by NE would be required. The routine test data were not direct readings of pump-to-pump differential head, but rather was differential pressure readings across each pump (suction to discharge) taken at different times on each pump during the operating cycle.

The tests were performed on each RHR pump individually and were not performed on both pumps on the same day.

Consequently, the RHR surveillance test program did not contain a ready mechanism to signal developing pump-to-pump Interaction problems between refueling outages. Thus, the corrective action plan was inadequate in that existing data that could have indicated a deadheading condition was not considered, and the monitoring program was inadequate to detect change before the deadheading condition developed.

The first performances of the surveillance test program were scheduled for the~ Unit 2 Cycle 3 and Unit 1 Cycle 4 outages.

In preparation for the Unit 1 Cycle 4 outage, the EHR system engineer reviewed the Unit 1 surveillance test procedure during the week of November 20, 1989.

The system engineer determined that the 8-psid acceptance criterion would not be met based on existing Unit 1 test data.

He notified his supervisor of the situation on November 22, 1989. The system engineer also notified NE of the situation by correspondence dated November 27, 1989. At this point, the system engineer was not familiar with the problems identified in Bulletin 88-04 and did not recognize the full significance of not meeting the 8-psid acceptance criterion.

l In parallel with but unrelated to the above-described system engineer's actions, routine quarterly American Society of Mechanical Engineers (ASME) pump testing was conducted on Unit 2 on November 28, 1989. During this l

testing, the 2A-A RER pump exceeded the developed head acceptance criteria specified in the test procedure. Additional testing was performed, but the l

pump still exceeded the pump suction-to-discharge diffr ential pressure requirements. The pump was determined to be acceptable a 1 operable in accordance with ASME,Section XI, pump testing requirements on November 29, 1989. However, because NE was concerned with the potential of a pump-to-pump interaction from the increased head on the 2A-A pump, an engineering evaluation was performed. The results of that evaluation demonstrated that the differential pressure between A and b train pumps did not exceed the 11-psid limit established for the NRC bulletin response. As a result of the disposition of the Unit 2 pump problem, the NRC resident inspector questioned the system engineering supervisor on November 30, 1989, about the Unit 1 RHR pump-to-pump differential head data. The NRC inspector was informed that NE was evaluating the Unit 1 data based on the November 27 notification by the system engineer.

In parallel. Technical Support was

k

. further evaluating the Unit 1 RHR data in which considerable scatter had been obs e rved. The data scatter, while not significant for evaluating differential pressure across the pump, did cause the system engineer to question the validity of the data for use in determining pump-to-pump differential pressures.

Review and evaluation of the Unit 1 RHR pump data from quarterly testing continued on Deceniber 1,1989. Two basic problems were noted with the data.

First, when. data on a single pump was compared from one run to the next, there was significant data scatter. RHR Pump 1A-A showed up to 14 psid between performances with an average of approxiraately 7 paid. RHR Pump 1B-B showed up to 11 psid with an average of 4 psid. Second, when comparing single-point data over the time period following the Unit 1 Cycle 3 refueling outage, seven instances were observed when the pump.to-pump data would pass the 11-psid criteria and 10 other instances where it would not. When the data points since the last refueling were averaged, the computed value was between 12 and 13 paid as indicated in the memorandum to NE f rom the system engineer dated November 27, 1989.

By the end of the day of' December 1, Technical Support was not able to reach a firm conclusion because of doubts about both the accuracy of the data and the averaging methods used to evaluate deadheading. To resolve these doubts, a special surveillance test was proposed to test both RER pumps at the same time to obtain a direct reading of pump-to-pump differential pressure.

Operations management reviewed the system operating instruction for RHR and determined that both pumps could be run in parallel in accordance with that procedure. Operation of both pumps in parallel was conducted on the evening of December 5, 1989, and the Unit 1 RHR deadheading problem was confirmed.

In summary, the cause of this violation was an inadequate program review of the pump data in preparing the response to NRC Bulletin 88-04.

This bulletin response contained a conclusion that was reasonably drawn f rom the data used l

as its basis, but.that was not predicated on a sufficient data base, i.e.,

only the nest recent data pair was utilized rather than all previous data.

The corrective action plan itself, even properly implemented, was inadequate to ensure a deadheading condition would not occur.

Corrective Steps That Have Been Taken and Results Achieved (Violations A.1 and A.2)

Immediate corrective actions included placing the 1B-B RHR pump handswitch in the pull-to-lock position to ensure one train of emergency core cooling system (ECCS) was operable.

Subsequent corrective actions included consulting with Westinghouse Electric Corporation and then revising Emergency Instruction E-0,

" Reactor Trip or Safety injection," to ascertain if one RHR pump should be stopped. This revision resulted in the following change to the procedure:

If the reactor coolant system (RCS) pressure is greater than 180 pounds per square inch gauge and both RHR pumps are running, then one RHR pump is stopped and placed in standby.

Because this action ensures RHR pump operability, it elimi,nated the need for future evaluation of the pump data for indications of deadheading. Consequently, that surveillance test program was cancelled.

In addition, the RHR system operating instruction has been revised to ensure that the RHR pumps are not run in parallel on miniflow for greater than 10 minutes.

I

- As a result of confirmation of the RHR pump deadheading condition, TVA also reevaluated its response to NRC Bulletin 88-04.

A revised response was submitted to NRC on March 15, 1990, describing the revised short-term corrective actions (as described above) and long-term corrective actions.

The long-term corrective action pertaining to the RHR pumps is to install check valves in the discharge piping downstream of the miniflow line branch. This measure will preclude the possibility of pump-to-pump interaction. As committed in the revised bulletin response, this installation will be comple.te before ptart-up from the Cycle 5 refueling outage for each unit.

Installation of the check valves will alleviate the need for one of the RHR pumps to be stopped, as currently required by Emergency Instruction E-0.

Following identification of the inadequate Bulletin 88-04 response, TVA conducted a review of the 1987-1989 SQN bulletin responses and records. The conclusion reached was that the responses are appropriate and valid with this single exception,Bulletin 88-04. Thus, the cause of this violation (inadequate bulletin response) was an isolated occurrence.

The review also established that condition adverse to quality reports (CAQRs) were written when deficiencies were identified, with the exception of a single prerestart item that was tracked as a specific nuclear performance plan restart item, rather than a CAQR.

TVA also reviewed the ASME,Section XI, program following this event to determine the cause for recent testing problems, including'the RHR pump test data scatter mentioned previously.

TVA concluded that the program is in overall compliance with requirements; however, several recommendations were made regarding (1) the use of snubbers on test instruments to reduce data scatter (2) the use of dedicated test equfpment for ASME,Section XI, testing, and (3) the consideration to upgrade installed instrumentation as an alternative to test equipment. The NRC senior resident inspector reviewed this report during routine monthly activities in January 1990.

As reported in the associated exit meeting and subsequent inspection report, no specific safety or technical issues were identified.

Prior to discovery of this event, TVA had instituted major programmatic changes that improved the methods used to manage significant licensing issues, such as responses to NRC bulletins. The Licensing project management system was instituted in January 1989.

In particular, Site Licensing now has the lead responsibility for evaluation of and responses to NRC bulletins. A licensing project manager (LPM) is assigned to each issue and is responsiblic for developing detailed action plans defining tasks, scope, schedules, and responsibilities. Additionally, the LPM ensures that the appropriate individuals, disciplines, and organizations are involved and are. assigned clearly defined responsibilities regarding plan development and issue resolution.

Information provided by the various organizations is handled through the formal licensing infora tion request process that has the necessary controls to ensure comphceness and accuracy. The LPM initiates and drives bulletin investigations and evaluations.

Specific CAQRs are written when deficiencies are identified.

In summary, this change, initiated af ter the B'ulletin 88-04 response but before this enforcement action, ensures that the appropriate parties are involved in evaluating licensing issues and developing responses. This new process addresses the root cause of the violation.

i b

L 4

Several other programmatic changes that are having a positive impact on the resolution of licensing issues had been made at SQN prior to discovery of this event. For example, the SQN NE Project Engineer now has the sole responsibility for and control of engineering design activities at SQN.

In addition, 90 percent of all engineering is now donc at the SQN site.

Technical Support is now the focal point at the plant for system-related problems.

rhe close proximity and direct involvement of these organizations are contributing to a more systematic and thorough evaluation of plant pro;1 ems.

In addition, major organization and management changes were made to the Technical Support group during the June to October 1989 timef rame to improve performance. A new Technical Support manager with experience in managing a successful system engineering program was hired from outside TVA, and the Institute of Nuclear Power Operations was requested to perform a peer evaluation of the Technical Support area during the November plant evaluation. A complete assessment of Technical Support was recently completed, and needed improvements are being'impicmented.

TVA is also actively recruiting experienced system engineers and has established a lead engineer concept to best utilize existing strengths. An evaluation of existing supervisors and engineers is also being conducted.

In addition, the key elements of system engineering (ownership of problems, leadership in problem resolution, sensitivity to regulatory and operational aspects, and focus on operability and problem solving) will continue to be reinforced through direct involvement of senior plant and Technical Support management. These changes are intended to further strengthen the system engineering capabilities at SQN.

In addition, in February 1990. TVA conducted a review of two previous nuclecr experience review (NER) items (NRC Information Notice S7-59 and a Westinghouse letter issued on October and November 1987).

These items discussed two potential problems with the RHR systemt deadheading of pumps with common miniflow lines and adequate tr'niflow capacity for single pump operation.

Later, Westinghouse correspondence (November 1987) concluded that the earlier information was not applicable to SQN because of separate RHR miniflow lines and the hydraulic isolation of other ECCS pumps with common miniflow lines through the use of separate flow restricting orifices for each pump.

TVA's evaluation of these items led to the same conclusion and also confirmed the adequacy of miniflow capacity for single RHR pump operation. The pump-to-pump interaction discussed in Bulletin 88-04 was not sufficiently defined until receipt of another Westinghouse letter in late May 1988, which described the interaction and noted that previous conclusions regarding pump deadheading potential were no longer correct. As a result, the RHR pumps were included in the evaluation for Bulletin 88-04 af ter receipt of the Westinghouse letter.

While the NER program was not directly involved in this problem, several changes to improve the NER program were made in June 1989. A dedicated and expan_ded staff is now onsite at TVA plants.

The experience and qualification of personnel in the NER program have been upgraded. Ucekly reviews are conducted by conference calls between the sites and corporate NER groups to identify significant safety issues. The new organization and staffing provide

t

= '

improved capability to recognize and act on CAQRs and potential significant safety issues.

In particular, items that are identified as potentially safety significant are designated as immediate attention iteios and hand carried to the appropriate principal managers for evaluation. The line managers complete the evaluation and make operability and inanediate reportability determinations.

Items that meet the criteria are written as CAQRs and are reviewed by the Management Review Committee.

Recent management actions have been taken to change personnel attitudes about problem reporting and to encourage problem identification to supervisors and management. Two site dispatches e-a issued from the Vice President, Nuclear Power Production, stating these expectativct.

This topic was also ernphasized in the recent Site Director's quarterly n ietings held with over 1300 SQN employees (16 meetings). The CAQR Managernent Review Committee bas been restructured to include senior site management to assess the extent to which the message regarding problem reporting is being understood, to ensure prompt corrective actions to identified problems, and to ensure quick resolution of potential problems. A multisite task force was formed in early November 1989 i

to evaluate the implementation and structure of the problem identification process and recommend improvements. The Vice President, Nuclear Power Production, was briefed on the team recommendations in January 1990. The planned changes to this program include the utilization of a single problem reporting document to ensure identification of potential problems, a lower threshold for incident investigations, and the establishment of criteria to ensure the appropriate resolution of potent.ial problems.

The change to utilize a single problem reporting document was implemented in June 1990.

Further improvements to the corrective action program are currently scheduled for the fall of 1990.

In summary, it is clear from the above discussion that the programmatic problems that led to these violations had been previously identified by TVA.

Correcttve actions to address these broad problem areas had been put in place or were ongoing at the time the deadheading issue was identified. Specific efforts were underway and continuing to improve engineering evaluations, to emphasize timely identification and resolution of problems, to upgr.de the Technical Support organization, and to improve licensing submittals (including bulletin responses).

Corrective Steps That Will Be Taken to Avoid Further Violations (Viointions A.1 and A.2)

No further actions are required.

Date When Full Compliance Will Be Achieved (Violations A.1 and A.2)

+

SQN is in full compliance.

it e

\\..

1

?g

, Violation 50-327, 328/90-01-01

,"B. Technical Specification 6.8.1, requires in part, that written procedures be established, implemented and maintained covering the applicable procedures recommended in Appendix 'A' of Regulatory Guide 1.33, Revision 2 February 1978.

Appendix 'A' of Regulatory Guide 1.33 Revision 2, requires procedures for combating emergencies and other significant events.

Technical Specifications [ sic] 6.8.2, requires in part, that changes to procedures be reviewed and approved prior to implementation as set forth in Specification 6.5.1A.

Technical Specification _6.5.1.A, requires in part, that each review determ. se whether or not an unreviewed safety question is involved pursuant to 10 CFR 50.59.

Contrary to the above, on December 6, 1989 the licensee performed an inadequate review of Emergency Instruction E-0, Reactor Trip and Safety Injection, Revision 7, required by Regulatory Guide 1.33 to combat emergency events. The procedure change would terminate RHR operation prior to the procedure steps requiring operator examination of certain parameters to diagnose whether a LOCA was occurring. The reviaw failed to ensure that the procedure change did not involve an unreviewed safety question pursuant to 10 CFR 50.59.

Violations A.1, A.2, and B are a Severity Level III Problem (Supplement I).

Civil Penalty - $75,000 (assesed equally among the violations)"

Admission or Denial of the Alleged Violation (Violation B)

TVA admits the violation.

Reason for the Violation (Violation B)

Revision 7 to E-0, which required stopping both RHR pumps, was technically deficient. The technical evaluation performed for Revision 7 was done solely on the basis of the accident analyses presented in the Final Safety Analysis Report (FSAR). During this evaluation, the reviewers did not adequately review the potential impact of the change in that they only considered the '

specific break sizes addressed in the FSAR.

Chapter 15 of the FSAR addresses a specific set of bounding 10 CFR 50, Appendix K, breaks. The reviewers considered these breaks to be bounding for all cases. There were however other break sizes than those in the FSAR that could,be impacted by the procedure change. The FSAR did not explicitly define the key assumptions regarding RHR operation for small break loss of coolant accidents (SBLOCAs); therefore, the reviewers assumed that no credit was taken for RHR injection for these breaks.

The Operations personnel responsible for i

9 implementation of the emergency procedure program had recently assumed this responsibility during realignment of Operations staff duties.

Special L

emphasis had not been provided during indoctrination in this function, and program requirements were not fully understood by the personnel involved with the subject revision. Although knowledgeable personnel previously responsible for changes to emergency operating procedures were consulted for this revision, the program did not require the appropriate checks and balances in that Westinghouse was not consulted regarding the change.

Review by the Plant Operations Review Committee (PORC) was not required, and the emergency procedure change evaluation was not well documented.

During the preparation of the safety assessments for both Revisions 7 and 8 to t

E-0, the individuals involved made an incorrect determination that because the change had no consequences (based on their review of the accidet'. analyses discussed in the FSAR), it had no effect on RHR system operation or information provided in the FSAR. Accordingly, the safety assessment questions were checked "no," indicating that a safety evaluation was not required.

In both Revisions 7 and 8, it was' clear that the reviewers were concerned with whether E-0 itself was described in the FSAR as well as whether the RHR system operation, as controlled by E-0, was described in the FSAR.

A safety evaluation was performed for Revision 8 at the direction of plant management.

Corrective Steps That Have Been Taken and Results Achieved (Violation B)

TVA has taken prompt and comprehensive steps to address this aspect of the violation. First, the administrative controls have been strengthened for emergency operating procedure changes. PORC review and Plant Manager approval are now required for all emergency operating procedure changes.

Administrative Instruction (AI) 2. " Guidelines for Preparing Verifying and Validating Operating Instructions," has been revised to require Westinghouse concurrence with any change to the emergency operating procedures that deviates from the Westinghouse Owner's Group / Emergency Response Guidelines.

Verification and validation requirements were also strengthened with

}

particular emphasis on simulator validation whenever possible.

In addition, the step deviation document process has been et.hanced by detailing specific evaluation criteria in AI-2.

+

i Second, training on the above changes has been conducted for the Operations personnel responsible for management of the emergency procedure change process. Westinghouse had previously confirmed the technical adequacy of the emergency procedures in September 1989. Revisions 7 and 8 to E-0 were the only emergency procedure changes made since the Westinghouse evaluation.

Third TVA reviewed the safety assessment problem and identified two areas for correction. A clarification to the FSAR regarding RHR operation for SBLOCAs was submitted in the April 1990 update. A training letter was sent to Level I and Level II 50.59 reviewers describing the RER event, the subsequent procedure changes, safety evaluations, and the lessons learned.

Particular emphasis was placed on the fact that a change without ultimate adverse consequence could still result in having an effect on the system and therefore l

L require a safety evaluation.

This training is considered an enhancement to the major 50.59 program changes made in November 1989.

i

"N# :; A; j

.' i '

1.

a o

o-

.[ *.

-Ni

, wr.

l, r,

..j,

p.qs

,,r.-

g g o c

-10,.

s c

?l.h".% Q

-f.x m..i The inadequate safety evaluation,is. the single implementation problem

>$'if'-

. identified under the revised 50.59 process-as noted in NRC Inspection

' Report 90-01.

,w k

h Corrective Steps-That Will Be Taken to Avoid Further Violations (Violation B).

i

~

j r>

- No further actions are required.

. i

,j.'

i i

- Date When Full Compliance _Will'Be Achieved 9-SQN is'in full compliance.-

l n

}.,

[7 iV -

t l{$ :

.4

!,.' 4 4

y s'

\\

t.

)

r

+

},-

a t

k 1

L p

3

,;c

-4 y.

t 1

s k

L i

k i

s f

't ll' t

3 4

e 0

t i

'i l '

l t

t

,. -;.[h

b i'

.