ML20212A407

Responds to 870112 Request for Addl Info Re Util 861212 Response to Insp Repts 50-321/86-28 & 50-366/86-33. Procedures Upgrade Program Intended to Ensure That Procedures Performed to Meet Tech Spec Requirements

ML20212A407
Person / Time
Site:	Hatch
Issue date:	02/23/1987
From:	Gucwa L GEORGIA POWER CO.
To:	NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM)
References
SL-1869, NUDOCS 8703030345
Download: ML20212A407 (8)

Text

Georgia Power Company

, 'e- 333 Pladmont Avenue -

Atlanta, Georgia 30308 Telephone 404 526-6526 Mailing Address:

Post Office Box 4545 Attanta, Georgia 30302 GeorgiaPower t T.Gucwe (tm southern electre system Manager Nuclear Safety and Uconsing SL-1869 1066C February 23, 1987 U. S. Nuclear Regulatory Commission ATTN: Document Control Desk Washington, D. C. 20555 NRC DOCKETS 50-321, 50-366 OPERATING LICENSES DPR-57, NPF-5 EDWIN I. HATCH NUCLEAR PLANT UNITS 1, 2 INSPECTION REPORTS 86-28 AND 86-33 Gentlemen:

By letter dated January 12, 1987, the NRC requested additional information with regard to our response to Inspection Reports 86-28 and 33, dated December 12, 1986. The requested information follows in a question-answer format.

1. "What is the justification for considering the violations to have no potential safety consequences?" ,

Violation A identified a situation where, nile, performing an instrument and control (18C) surveillance che k fort pressure drop across the filters and charcoal adsorber banks,\the fitw rate was not concurrently ensured as being within Technical Specifications. The determination of this flow rate by the I&C surveillance is done by the use of a single point reading and is not used to determine compliance with design or the Technical Specification flow requirement. The determination of the Technical Specificat*,n flow rate, as well as filter train efficiency and leakage, is estublished separately by Engineering. This determination was satisfactorily achieved in July 1986. The Engineering procedure, 42SV-T46-003-1S, utilizes a twenty (20) point flow traverse test procedure and visual inspection is more accurate than the single point determination. The Engineering procedure is scheduled when the following occurs: once every operating cycle or eighteen (18) months; after 720 hours0.00833 days <br />0.2 hours <br />0.00119 weeks <br />2.7396e-4 months <br /> of filter train operation; after charcoal or HEPA filter replacement; i following structural maintenance of the filter train or following j painting; or after a fire or chemical release in a ventilation zone

communicating with the system.

On September 25, 1986, following the I&C surveillance, personnel incorrectly adjusted the flow rate based on the I&C procedure resul ts. We have every reason to believe that the system had been in 8703030345 870223 e i PDR ADOCK 0500032( /, l G PUR q.

Georgia Pcnverb U. S. Nuclear Regulatory Commission February 23, 1987 Page Two full compliance with requirements. The system engineer, in reviewing the flow adjustment made to his system, decided to run the flow traverse test to confirm the accuracy of the new flow setting. In performing the test on September 26, 1986 the engineer found the system had been adjusted below the Technical Specification flow requirement, and he immediately readjusted the volume damper to return flow within the Technical Specification flow requirement. The adjustment to the flow following the I&C surveillance on September 25, 1986, was unique in that flow adjustment is normally performed using the results of the Engineering surveillance procedure. A review was made to determine if adjusting the flow following the I&C surveillance was an isolated event. To the best of our knowledge this is the only occurrence of adjusting flow using the results from the I&C procedure. Since the volume dampers were not adjusted between flow measurements prior to September 25, 1986, and the Engineering surveillance was performed more frequently than the I&C procedure, Georgia Power Company (GPC) can reasonably assume the flow was acceptable during performance of the I&C surveillance up to the previously mentioned date.

Violation B identified the Technical Specification limit of 3.25 times the surveillance interval as being exceeded. The I&C surveillance was performed in October 1982, and was signed off on the computer printout provided to I&C as the scheduling tool for periodic surveillance. The procedural data package showing satisfactory performance of the surveillance could not be retrieved as indicated in the original violation response. This resulted in our inability to demonstrate the performance of the I&C surveillance within the 3.25 surveillance interval. The purpose of preforming the I&C surveillance is to ensure the proper pressure drop across the fil ters. As part of the Engineering procedure the filter media is visually inspected to ensure the media is not damaged and does not exhibit indication of being overloaded. During the time that the I&C surveillance was not documented, the Engineering procedure demonstrated that an overload condition of the filter train did not exist. This conclusion is based on the fact that had such an overload condition occurred during this time frame, the Engineering procedure would have identified this condition by observing decreased flows or through visual inspections that would have required readjustment of the flows and/or replacement of the filter media. A review was made and indicated that no adjustments of this magnitude were made.

We believe that: responses to NRC correspondence should reflect the safety significance of the particular events. These documents are l

1066C l

r=m . .-

Georgia Powerkh U. S. Nuclear Regulatory Commission February 23, 1987 Page Three public and are used in non-technical arenas by non-technical personnel. In this case, the addition of the cited phrase was to convey our view of the actual and potential safety significance of the event relative to other more significant or more probable safety probl ems. The specific past-tense wording of our response, i.e.,

"had no actual or potential safety consequences", [ emphasis added],

was intended to indicate that it was an historic event that did not constitute an actual or potential threat to the public's safety in light of the overall design and capabilities of the plant's systems and operational history.

This event is not safety significant. A small decrease in flow below the Technical Specification range would result in a slight decrease in pressurization of the control room; the control room would have remained at a positive pressure with respect to the turbine building and would not appreciably affect the adsorption capability of the filter trains in removing contaminants from the control room environment.

We believe we met the intent of the Technical Specification requirement by performing the I&C and Engineering surveillances as previously described. Al so , prior to, during and after the relatively short time the system flow was reduced below the Technical Specification requirement, we believe that the system would have performed its intended function had it been called upon to do so.

2. "Was the as-found flowrate on September 26, 1986, outside the TS required flowrate?"

Yes, but it must be realized that when this procedure was performed on September 25, 1986, the flowrate was promptly adjusted from a satisfactory level on the basis of the value from the single point, 1

resulting in a flowrate believed to be 2477 cfm. By utilizing the more accurate 20 point flow traverse test the as-found flow on September 26,1986 was 2196 cfm which is below the flow required by the Technical Specification of 2500 cfm +~ 10% (2250 cfm to 2750

, c fm) . The flow rate was promptly adjusted to a value of 2512 cfm.

The resulting low flow condition was reported by GPC test personnel via Deficiency Report 1-86-lm'). Upon discovery of the flow being below the Technical Specifications value, the actions required by the Technical Specifications were taken, and the associated LC0 was not exceeded.

We are changing our surveillance procedures to reflect the lessons learned from this event. Note that the point of measurement for the two procedures was different. The 18C surveillance procedure is 1066C 700775

~

Georgia Power b U. S. Nuclear Regulatory Commission February 23, 1987 Page Four being incorporated into an Engineering procedure which will no longer allow the one point flow determination, but will require the more accurate 20 point flow traverse test.

3. "Will your Procedures Upgrade Program (PUP) or other programs ensure that procedures can be performed to meet TS requirements and contain adequate detail to prevent occurrences like Violation A and the EDG day tank level problem?"

Yes, it is GPC's intent for the PUP and other programs in place to ensure that procedures are performed to meet TS requirements and contain adequate detail to prevent similar occurrences. When a procedure goes through PUP, it undergoes vigorous review through each state of development, review and approval. The first of these is by the writer, who reviews the TS requirement to determine applicability of the procedure. A comparison is also made to the description of the associated system or components identified in the Final Safety Analysis Report, plant drawings and to actual plant configuration.

In addition, a technical review is performed by knowledgeable individuals to ensure technical adequacy. A review is also performed by the Nuclear Safety and Compliance Department to ensure proper implementation of the identified comitments and requirements. As a final check, each procedure is validated by knowledgeable plant personnel prior to final approval and issuance. This is accomplished by one of the following: actual performance, performance on the Simul ator, walkthrough or talkthrough, or comparison to a similar procedure.

It is the intent of PUP and other existing programs at Plant Hatch to ensure that the procedures meet TS requirements and contain adequate details to prevent similar violations from occurring. GPC does not consider the procedure for determining EDG day tank level to be a problem as discussed in the answer to question #5.

4. "What was the cause of the EDG day tank level problem (i.e., had the gauge been replaced recently, procedure changed, etc.)?"

The EDG day tank level gauges are originally installed instruments, and the associated procedure has not been revised to include a conversion factor. The gauge on Unit 1 was marked in November 1986 to reflect the corresponding value for the Technical Specification requirement to resolve this perceived problem.

5 "How was this procedure being performed in the past, since personnel could not verify the day tank level using the gauge?"

1066C 700775 .)

. Georgia Power b U. S. Nuclear Regulatory Commission February 23, 1987 Page Five During routine performance of the Unit 1 procedure, the conversion factor (between gallons and percentage full ) from the Unit 2 i

procedure was applied to the Unit 1 gauge. Since the procedure required the reading to be in gallons and the gauge registers percent, this conversion would be necessary to satisfactorily complete the procedure. The gauges and day tanks for each unit are j the same; therefore, the practice previously mentioned results in accurate determination of day tank level.

~

I

6. "Have there been and are there any procedures which provide guidance in scheduling surveillance?"

Yes. The " Technical Specifications Surveillance Program" procedure

, (HNP-831 - until May 1, 1985, and now 40AC-REG-001-0S) provides the i

necessary guidance for scheduling Technical Specifications surveillances. . This procedure gave the needed guidance for

, scheduling surveillance, but the refinement to it made in May,1985 should ensure violations such as this one do not recur.

i 7 "What are the details of personnel error involved with Violation A

[ sic]?"

The personnel error causing this violation occurred as a result of

! the oast practice of scheduling the next due date for this i

surveillance based on 18 months from unit start-up following a i refueling outage rather than using the last surveillance performance. This resulted in the due date being greater than 18

! months from the last performance of the surveillance. The error in i past scheduling practice for this surveillance was corrected in May 1985 as described in the original violation response with i

surveillances based on the previous due date of the surveillance.

78. "How does the May 1985 change to the Technical Specification surveillance scheduling program constitute a corrective step for Violation B (i.e., was this violation due to a frequency change)?"

l The May 1985 program change was not intended to serve as corrective action for Violation B. The information on the change to the Technical Specification surveillance scheduling program was provided as an example of how we have improved the program to preclude similar

violations. Based on this previous program change no further changes were deemed to be necessary. The program change was made as a result of another un-related scheduling problem (error in frequency change) #

! that occurred in 1985, but which al so addressed the scheduling

! problem associated with the surveillance described in Violation B.

i i

1066C i

y . , - . - ,.----,_.,m_.,,---e -,__._--.--m_ - - - - - - -

l Georgia Power d U. S. Nuclear Regulatory Commission February 23, 1987 Page Six

9. "Are surveillance due dates scheduled using operating cycle completion dates? If so, what means are in place for not exceeding the TS required frequency?"

The surveillance due dates are based on the previous due date of the surveillance not on operating cycle completion dates. Since the due dates are not based on operating cycle completion dates, no further action is required to preclude exceeding the TS required frequency.

10. "How will you prevent exceeding the TS required frequency using previous due date vice actual completion dates to schedule the next surveillance tests?"

Subsequent due dates are based on previous due dates as explained in this section. If a surveillance is performed "early" (i .e. , before the scheduled due date), the surveillance computer will " flag" this and not accept the performance date. The nominal due date must then be changed to the "early" perforn.ance date. The computer will then accept the perfomance date and schedule the next performance c'ue date based on the actual (early) performance date which was the new due date. If a surveillance is performed on or after the due date, the computer accepts the date and schedules the next due date based on the previous due date, not the performance date.

Example: Assume a surveillance is to be perfonned with a monthly frequency that is due the 5th of each month and a grace period of 7 days (within the 25% allowed by Technical Specifications).

Normally the due and late dates would be as follows:

Month Due Late March 5 12 April 5 12 May 5 12 If the test were performed on March 6, the next due date would remain April 5.

1066C l

-7,, l

GeorgiaPower d U. S. Nuclear Regulatory Commission February 23, 1987 Page Seven If the test were performed on March 14, the Technical Specifications requirements of the 25% " grace period" and 3.25 times the surveillance interval for three consecutive tests would have been exceeded. To prevent subsequent surveillance from exceeding the 3.25 requirement, the next due date would still be April 5.

If the test were performed on March 4, the current due date would have to be changed to March 4, then the subsequent computer generated due date would be April 4. The subsequent due dates would be the 4th of the month unless another test was performed early.

11. "How will you prevent exceeding the 3.25 times the surveillance interval for future surveillance test in general?"

The methodology described in the answer to quest .aii #10 will prevent this from happening.

We appreciate this opportunity to comment on these subjects and we believe that continued technical interchange is beneficial to both NRC and GPC. Based on this interchange, we request that you consider reduction of the previously assigned severity classifications from level IV to level V.

If you have any questions in this regard, please contact this office at any time.

( Sincerely,

c

9t Y~& e_2 l

l L. T. Gucwa MJB/lc l

c: Georgia Power Company U. S. Nuclear Regulatory Commission

Mr. J. P. O'Reilly Dr. J. N. Grace, Regional Administrator l Mr. J. T. Beckham, Jr. Mr. P. Holmes-Ray, Senior Resident Mr. H. C. Nix, Jr. Inspector - Hatch GO-NORMS

{

1066C l

[ toons )