IR 05000281/1986036

Insp Rept 50-281/86-36 on 861113-20 & 870209.Violations Noted:Failure to Properly Valve Penetrations for Type a Test in Accordance w/10CFR50,App J.Four Unresolved Items Identified

ML20207T588
Person / Time
Site:	Surry
Issue date:	02/23/1987
From:	Arndt G, Jape F, Maura F, Wright G NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION II), NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION III)
To:
Shared Package
ML20207T580	List: IR 05000281/1986036 ML20207T579 ML20207T582
References
50-281-86-36, NUDOCS 8703240159
Download: ML20207T588 (12)

Text

.

e n .neou UNITE 3 STATES

/ 'o NUCLEAR REGULATORY COMMISSION

/ , REGION 11 g ,j 101 MARIETTA STREET. *

• ATLANTA. GEORGI A 30323

%

l

ReportTo. 50-281/86-36 Licensee: Virginia Electric and Power Company Richmond, VA 23261 Docket No. 50-281 License No. OPR-37 Facility Name: Surry Power Station, Unit 2 Inspection Conducted: November 13 - 20, 1986 and February 9, 1987 Inspectors:

5)% r. a. naura 3//i /f7

'

Date'

hwL G. Arnat / whoIt r Date W//wh &Wo 1/ 2 3/F7 F. Jape (p20/87) Tate '

&A O Approved By: G. C. iirishi, Chici V' 3 Test Programs Section N8

~

Region III

'

Inspection Summary Areas Inspected: Routine announced inspection by a Region III inspector, accom-panied by a member of the Office of Nuclear Regulatory Research, of the Contain-ment Integrated Leak Rate Test (CILRT) procedure, CILRT performance witnessing, and CILRT results. NRC inspection modules utilized during this inspection included 70307, 70313, and 7032 Results: Of the areas inspected, four unresolved items were identified and one violation was identified (failure to properly valve penetrations for a Type A test in accordance with 10 CFR 50 Appendix J and Technical Specification require-ments in paragraph 2.a(3). This violation has the potential to cause the r ontainment to exceed the maximum allowable leakage rate stated in the Technical Specification DESIGNATED ORIGINAL Certified By //M6 d h/Mb

.

0703240159 PDR 870302ADOCK PDR 05000281 O

www - - -

Q

,

.

t

DETAILS 1.' Persons Contacted

~

• D.lBenson, Assistant Manager - O and M

.

• R.1Blount,~ Supervisor Surveillance and Test

-

-

.* W.-Craft,' Licensing. Coordinator

_

'* W. Grady,' Supervisor Quality . .

• E. Grecheck,. Super _intendent_ Technical Services-

'** H. Miller, Assistant Manager.- NS and L R. Nicholls~,- Outage Coordinator R. Saunders, Station Manager'

E. Turko,' Test Director

• Denotes person attending the preliminary exit meeting of November 20, 1986.'

• • Denotes persons participating in the telephone. exit interview'at the conclu-

_

'sion of the inspection on February 9, 198 The in_spectors also contacted other licensee personnel including members of-the technical and operating staf .- Containment Integrated Leak Rate Test Procedure Review (Unit 2) Procedure Review The inspector' reviewed periodic test procedure 2-PT-16.3, " Reactor Containment Building Integrated Leak Rate Test (Type ' A')," dated October 9,1986,: for conformance against 10 CFR 50, Appendix J;- the FSAR, and the Technical Specifications. The inspector's comments were discussed with the licensee prior to test performance. The following

' problems were identified:

(1) The procedure called for the liner weld leak chase channel. plugs to be installed throughout the test. Since the weld leak-chase channels may not be designed to the same criteria as the'contain-ment shell or liner each licensee must provide justification to NRR for not venting the channels during the Type A test. 'With the leak chase channel plugs installed, any leakage paths existing in the containment weld seams may not be detected. The licensee ~ felt that the exemption granted by the NRC to Beaver any Valley Nuclear Plant (Duquesne Light Co.) applied to Surry also since - both have Stone and Webster Engineers designed containment In conversations with NRR (J. Pulsifer) the inspector was informed that the Beaver Valley exemption was specific and did not apply to other facility, therefore, Surry (VEPCo) would have to request their own exemptio This informa-tion was given to the license This is an Unresolved Item (281/86-36-01) pending licensee submittal of a justification for not venting liner weld leak chase channels and NRR acceptance of the justificatio r;5 ev <

-

,

a

'

' ~

s .

o >c-. '

.

-

. 12'

- >

c (2) .1The data 1 sets ~ utilized for' the . leak rate calculations were not .

'

. individually _ measured data points. - The licensee's process comput--

,

er; scans the sensors as often as every two seconds and averages the_ readings to: generate a_ ten minute data. point. The;1icensee was informed that statistically the method is not correct because -

the." time" assigned ~to.the data point is not the true time. The:

-licensee stated that their method had been reviewed and, accepted ~

-by previous inspectors. The inspector: stated that neither. ANSI N45.4-1972 nor BN-TOP-1, Revision-1 permitted data averaging and actually required individually. measured data points. The licensee

.amendedithe computer program prior .to. the CILRT'to perform the leak rate _ calculations using individually measured data points-taken everycten minute The resolution satisfied ANSI N45.4 an BN-TOP- (3) During,the partial review of the valve: lineup, the inspector noted r that certain penetrations' listed in Table 3.8-2 of the Technical

' Specification as requiring Type C testing were neither being s properly vented.and drained for the. CILRT, nor was their leakage -

rate contribution being applied as a penalty to the CILRT result The penetrations in? question were:

Penetrations System 46 Loop fill 63 and 64 Containment spray pump discharge

-

66 through 69 Recirc spray suction 70 and 71 Recirc spray ~ pump discharge

-

The licensee stated the reason these penetrations were not vente Land drained was because they are, and remain, water filled during

~the accident. A review of the licensee's Technical Specifications

,. showed that Table 3.8-2 requires that these penetrations be Type C tested. Table 3.8-2 is very explicit as to which containment

'

isolation valves are exempt from Type C testing.10 CFR 50, Appendix J, requires that Type C test be performed on containment isolation valves, by pressurizing to Pa with air or nitrogen, in order to measure their leakage rates and that the combined leakage rate for all penetrations and valves (Type B and C tests) be 1 0.60 L '

The only valves exempted by Appendix J from the Type C air or nitrogen test, and from the 10.60 La acceptance criteria, are those valves which are pressurized with a seal system (normally water), provided they are tested at 1.1 Pa, the seal water inven-tory is sufficient for at least 30 days at a pressure of 1.1 Pa, and their leakage rate does not exceed the limits specified in the Technical Specifications. Since these penetrations were not

,

g =

'

.

exempted by the NRC 'from Type C testing in Table 3.8-2 of.the -

Technical Specifications (meaning that no credit was given to their water seal during the accident) they must be vented, and drained as required, for the CILRT. Failure to properly valve the penetrations for the CILRT, or to take a penalty equal to the local leak rate test (Type C) results for such penetrations which deviate from the proper configuration, is a violation (281/86-36-02) of 10 CFR, Appendix b. Clarifications of Appendix J Requirements To. ensure the licensee's understanding of. Appendix J requirements, the inspector conducted numerous discussions with licensee personnel during the course of the inspectio The following is a summary of the clarifications discussed with the license (1) The only methods of data reduction acceptable to the NRC are total time or point-to point as described in ANSI N45.4-1972 including a statistically calculated instrument error analysi The following options are available to the licensee and are suggested in the following order:

(a) Total time (<24 hour duration test) in accordance with Bechtel Corp. Topical Report BN-TOP-1, Revision 1.- Whenever this. method is used BN-TOP-1 must be followed in its entirety except for any section which conflicts with Appendix J requirement (b) Total time (>24 hour duration test) using single sided 95%

UC (c) Proposed Regulatory Guide MS 021-5, Regulatory Position 1 If this method is utilized the licensee must submit an exemption request to NRC and receive approval for its use prior to the expiration of the Type A test frequency require-ments stated in the Technical Specification (2) Periodic Type A, B, and C tests must include as-found results as well as as-lef If Type B and C tests are conducted prior to a Type A the as-found condition of the containment must be calculat-ed by adding any improvements in leakage rates, which are the results of repairs and adjustments (RA), to the Type A test results using the " minimum pathway leakage" methodolog This method requires that:

(a) In the case where individual leak rates are assigned to two valves in series (both before and after the RA), the penetra-tion through leakage would simply be the smaller of the two valves' leak rate r

~

~ - .

- ' - '

y .

=

'") a 4'

l (b) In 'the case where a leak rate-is obtained by. pressurizing

~

between twoLisolation valves and the individual valve's leak-rate is not quantified,:the as-found and as-left penetration through-leakage for each ' valve. would be _50 percent of the:

measured leak rate if both valves are repaire (c)' In L the case where a . leak rate is,obtained by pressurizing

'

between two isolation valves and only one valve is repaired,-

<- the as-found penetration leak' rate would conservatively be -

the final measured leak rate, and the _as-left penetratio <

through leak rate would be zero (this assumes the repaired valve leaks zero).

(3) Penetrations which are required to be Type C tested, as describe in:the FSAR and SER, must be vented inside and outside the contain-ment during the CILRT. All. vented penetrations ~must be drained of water-inside the containment.and between.the penetration valves to assure exposure of the containment isolation valves to containment air test pressure. The degree of draining of vented penetrations  ;

outside of containment is controlled by the requirement that the valves be subjected'to the post-accident differential. pressure, or proof-that .the system was built to stringent quality assurance standards comparable to those required for a seismic syste j (4) Whenever penetration configurations during a CILRT deviate from

' the ideal, the results of LLRTs for such penetrations must be added as a penalty to the CILRT results at the 95% confidence .

level. This penetration leakage penalty is determined using the

" minimum pathway leakage" methodology. This methodology is '

defined as ; the minimwn leakage value that can be quantified through a penetration leakage path (e.g. , the smallest leakage ,

through two valves- in series). This assumes no single active failure of. redundant leakage barriers. Additionally, any increase in containment sump, fuel pool, reactor water, or suppression pool level during the course of the CILRT must be taken as a penalty to the CILRT results. If penalties exist, they must be added (sub- -

'

i traction is never permitted) to the upper confidence level of the~

+ CILRT result (5) The start of a CILRT must be noted in the test log at the time the i licensee determines that'the containment stabilization has been satisfactorily completed. Reinitializing a test in progress must

be " forward looking," that is, the new start time must be the time

'

at which.the decision to restart is made. This also implies that

the licensee has determined that the test has failed, and has

!. enough data to quantify the leakage rat Any deviation from i= these positions should be discussed, and documented, with the NRC

inspector as they occur to avoid later invalidations of the test

results. Examples of acceptable deviations of reinitializing the

.

b

.

, _ - . . - . - - - --.- . - _ _ . - - - - - . - . - - , , _ _ _ . . _ _ - - - . .

- _ . ___ . _ _ _ _ _ _ _ _ _ _ - _ _ - _ - _ _ - _ _ _ _ - _ _ _

~

.

i I

start time of.the-test in the past are: time at which a leaking I penetration which has an obvious effect on the test data was secured, accidental opening and later closing of a valve which has an obvious effect on the test data, the time at which an airlock outer door was closed and the inner door was ope (6) The supplemental or verification test should start within one hour after the completion of the CILRT. If problems are er. countered in the start of the supplemental test,- data recording must continue and be considered part of the CILRT until the problems are cor-rected and the supplemental test can begi (7) For the supplemental test, the size of the superimposed leak rate must be between 0.75 and 1.25 time the maximum allowable leak rate La. The higher the value, the better. The supplemental test must be of sufficient duration to demonstrate the accuracy of the tes The NRC looks for the results to stabilize within the acceptance criteria, rather than the results being within the acceptance criteri Whenever the BN-TOP-1 methodology is oeing used, the

- length of the supplemental test cannot be less than approximately one-half the length of the CILR (8) During a CILRT, it may become necessary to reject or delete specific sensors or data points due to drifting or erroneous sensors, or data outlier Data rejection criteria should be developed and used so that there is a consistent, technical basis for data rejection. One example of an acceptable method for data outliers is described in an appendix to ANSI /ANS 56.8-198 Sensor data rejection criteria should be plant specific and based upon a sensor's trend relative to the average scatter, slope, and/or absolute output of the senso (9) The water level in the steam generators during the CILRT must be low enough to ensure it does not enter the main steam lines unless flooding of the main steam lines is called for in the loss of coolant emergency procedur (10) An acceptable method for determining if the sum of Type B and C tests exceeds the 0.60 La Appendix J limit is to utilize the

" maximum pathway leakage" metho This methodology is defined as the maximum leakage value that can be quantified through a pene-tration leakage path (e.g., the larger, not total, leakage of two valves in series). This assumes a single active failure to the better of two leakage barriers in series when performing Type B or C test (11) Test connections must be administratively controlled to ensure their leak tightness or otherwise be subject to Type C testin One way to ensure their leak rightness is to cap, with a good seal, the test connection after its use. Proper administrative controls should ensure valve closure and cap re-installation

,

, ,--n--

n ., _

, . , - . , , , , - ,._,- , .r--.--- , . - -

n ,

'

"

,

L i

, .

-

within the local leak rate. testing procedure, and w'ith a chehklist

prior to unit. restart.

-(12) Whenever a valve is replaced, ; repaired,'. orirepacked during' an

'

outage for which Type 'A, B, and/or 'C: surveillance testing was scheduled,-local leak rate. testing for the as-found as well as the as-left condition must be performed on'that penetration. In the case .of a replaced valve, the as-found test.can be waived if no other containment-isolation valve of similar design exists at the sit . ' Containment Integrated Leak Rate' Test Witnessing

- Instrumentation The following instrumentation was used throughout-the tes Type Quantity RTDs 22 Dewcells 5'

Pressure Gauges 2 Flowmeter 1 The inspectors reviewed the calibration data for the pressure, humidity-and-flowmeter instrumentation used during the test and determined.that it was satisfactory. With respect to the twenty-two RTDs the licensee was unable to supply the. inspector with the calibration data by the last date of the site inspection. .During the exit interview the licensee stated that the only calibration data available was the original calibration performed at time of purchase :and the in-situ check performed prior to this test. Subsequent to the-site-exit'

-

interview the inspector received a copy of the RTDs calibration. It consisted of the original . vendor calibration, dated October 1979, a computer input calibration performed November 11,'1986, and a one point in-situ check performed on November 14, 198 The inspector reviewed the data submitted by the licensee and deter-mined that:

'

(1) the calibrations performed and the results obtained appear to be sufficient to ensure the accuracy and repeatability of the temper-ature data obtained during the test, however, (2) the fact-that the sensor was not included in any of the calibra-tion performed prior to this Type A test violates previous inter-pretations of the requirements of 10 CFR 50, Appendices B and G

~:

.

-

,

_

y -

' Similar questions have risen at other nuclear plants regarding

~

whether the temperature sensor must be included in the calibration of temperature channels used in. safety related activities. -As a-result NRR has been requested to address this issue. Pending receipt of.a response from NRR the calibration of the RTDs for the

. November 1986 Type A~ test iis considered an Unresolved Item

'(281/86-36-03). Temperature Survey-The. licensee was unable to suppi / the inspector the results of the temperature survey utilized to justify the location of the ~ sensors used for-the CILRT. .During the exit. interview the-licensee stated that due to a manpower shortage they ~ had been unable to obtain a copy of the survey from.the vault. Subsequent.to the site v.isit'the licensee sent-to the inspector the results of a temperature survey performed in 1980 with containment' recirculation fans runnin It included the: data taken with 69 thermocouples and.the-22 RTDs. It also contained dew point . readings taken at 13 locations with two psychrometers and that recorded by the five chilled mirror sensors then in use. The packag did not include any drawings or sketches showing the physical location of where all the readings were taken and since the temperature varia-tions recorded are considered significant by the inspector, interpreta-tion of the data is not possibl This is an Unresolved Item (281/86-36-04) pending review of the temper-ature survey results by a Region II inspecto Witness of Test-The inspector witnessed portions of.the CILRT on November 13 through the :20, 1986 and noted that test prerequisites were met and that the appropriate revisions to the test procedure were being followed by test-personne Valve lineups for the following systems were verified to-ensure that no fluid could enter the containment atmosphere and that adequate venting and draining was provided:

System Penetration (s)

Service Air 42 Air Monitor Sample 43, 44 Instrument Air 47, 58 Primary Vent Header 48 N2 to Pressure Relief Tank 53 Primary Vent Pot Vent 54 Pressurizer Steam Sample 57A Post Accident Sample Return 578 PRT Vapor Space 570 Air Ejector Discharge to Containment 89 Vacuum Ejector 94 Containment Instrument Air Compressor 112

r

~

...

No violations or' deviations were identified; however, two issues requiring further information from the licensee and/or review and evaluation by the NRC were identified and will be followed as unresolved item . Test Results Evaluation CILRT Data Evaluation A 32-hour CILRT was performed during November 19-20,-1986 at 61.2 psia, following satisfactory completion of the required temperature stabili-zation period. Data was collected every ten minutes. The inspector independently monitored and evaluated leak rate data using total time (ANSI N45.4-1972) formulas to verify the licensee's calculations of the leak rate and instrument performance for the first 26 hours3.009259e-4 days <br />0.00722 hours <br />4.298942e-5 weeks <br />9.893e-6 months <br /> of the tes Following completion of the test the inspector reviewed the rest of the data received from the licensee on February 2,1987, at the Regional Office. The inspector used all the data taken (every 10 minutes) while the licensee only reduced the data taken every 20 minute There was acceptable agreement between the inspectors and licensee's results as indicated by the following summary (units are in weight percent-per day).

Measurement Licensee Inspector Leak rate measured during CILRT (Lam) 0.029 0.026 Lam at upper 95%

confidence level 0.063 0.036 Appendix J acceptance criteria at 95% UCL: <0.75La = <0.075 weight percent per da Supplemental Test Data Evaluation After the satisfactory completion of the CILRT a known leakage rate of 5.03 scfm, equivalent to 0.100 weight percent per day as induced at 1330 hr on November 20, 1986. Data were collected and analyzed by the licensee every ten minutes. The inspector has reviewed and evaluated the data received from the licensee using start time of 1340 hr. The licensee did not declare the start of the supplemental test until 1530 hr. At 2140 hr on November 20, the test was terminated by the licensee

.

c;

. . . -

1 with satisfact'ory results as indicated by the following summary:

Medsurement . L!cenave Inspector:

- (Approximately Approximately 6 hrs of data)

'

_

(8 hrs,of data)

Wt. % per day Wt. % per day Measured leakage rate, Lc, 0.123 0.131 during supplemental test

,

Induced leakage-rate, Lo 0.100 0.100 Lc-(Lo+ Lam) -0.006 0.005 Appendix J. acceptance criteria: -0.0251[Lc - (Lo+ Lam)]s +0.025 Raw Data Retrieval The inspector was not able to verify the containment air mass calcula- i tions being used by the licensee to generate the leak rateLresults by using the raw data being printed by the Unit 2 plant computer. The reason was that two sets of instantaneous raw data were being collected during the-test a few minutes apart as follows:

-(1) one set of instantaneous sensor data were collected by the plant conputer at ten minutes intervals and then processed by the ILRTLPRT 2 program to generate weighted average temperature, vapor pressure, and containment pressure. The results were printed and then entered by the licensee into a PC to complete the ILRT-calculations. The raw data was not saved by the plant compute ,

(2) a second set of instantaneous sensor data was collected by a Trending Program in the plant computer at five minute interval This raw data was printed, but not processed. The two sets'of-sensor data were offset by a few minute .

This system of data collection and reduction makes -verification of test results by the inspector very difficult. The inspector request-ed that the licensee revalidate the calculated leak rate of the

containment using the raw data obtained by the Trending Program since

that is the only historical raw data available and that for future tests the computer programs be modified so that the raw data used for the leakage rate calculations are also available for review. This is an Unresolved Item (281/86-36-05) pending satisfactory revalidation of the test results and modification of the program for future tests, CILRT Valve Lineup Penalties 7 Due to valve configurations which deviated from the ideal penetration '

i (excludes those discussed in Paragraph 2.a.(3)) valve lineup require- i ments for the CILRT, the results of local leak rate tests for such t F

t i

..,,,.-y, . ~ , - - -

e... . - - , . _ . . - , . . . _ _ _ . _ , . .m - - - . _ . _ , . . _ , . , , - _ . . _ . , . _ , , . . . , - - . . 7-.~~-

,

,

.

penetrations must be added as a penalty to Lam at the 95% UCL. The following penalties must be added using the minimum pathway leakage method:

Local Leak Rate Test Value Penetration (Units are in scf/hr)

24 - RHR 0.225 50 - N2 relief 0.150 92 - Containment vacuum 0.220 56A -Liquid sample, hot leg 0.430 1.025 scf/hr or 0.0003 wt%/ day After taking these local penalties into account the upper confidence value for containment leakage remains 0.036 wt%/ day, which is within the acceptable value of 0.07 e. As-Found Condition of CILRT The as-found condition is the condition of the containment at the beginning of the outage prior to any repairs or adjustments to the containment boundar The inspector reviewed the licensee's summary of the containment penetrations local leak rate test (Type B and C)

performed prior to the CILRT in order to determine the amount of leakage rate improvement due to repairs and adjustment Based on the results reviewed it was determined that the containment as found leakage rate was <0.247 weight %/ da The Appendix J acceptance criteria is 0.075 weight %/ day; therefore, the containment failed its as-found condition. Two penetrations were the main contributors: the sump pump discharge (No. 38) with a leakage rate <284 scf/hr and the loop fill line (No. 46) with a leakage rate <300 scf/h As a result of the repeated problems experienced with the sump pump discharge isolation valves the licensee performed a modification during this outage in an effort to correct the problem. Both two inch stain-less steel, air operated, globe valves were replaced. The erosive nature of the fluid being handled was causing erosion of the plug and seat. The new valves are two inch stainless steel, double trunnion, air operated ball valves with a spring loaded spool / seat section which provides a wiping action as the valve rotates from open to close, thus reducing any damage from debris in the fluid. In order to eliminate cycling of the valves every time the pump operated the control circuit was modified to keep the valves continuously ope The valves still close upon receipt of a safety-injection signal, loss of power or air, and control room operator action. Once closed delib-erate operator action (reset push button) is required to open the valves. While it appears that the licensee has taken adequate correc-tive action to prevent recurrence of the problems experienced with the sump pump discharge penetration, only future Type C tests results can determine if the modification was successfu , _ _ _ _

,

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - _ _ _ _ _ _ _ _ _ _ _ _ _ _ - _

.-

o

With respect to the loop fill line penetration no extra measures to prevent recurrence were taken. However, the licensee's position that the penetration remains water filled at 1.1 Pa for at least 30 days after the accident (Refer to paragraph 2.a.(3)) should be sufficient if the Technical Specifications are revised to reflect their positio As indicated earlier the' as-found condition exceeded the allowable Appendix J limit. Presently the unit is on accelerated Type A testing (each refueling outage) due to two consecutive as-found failure Although the June 1985 CILRT was successfully completed, the regulation requires two, consecutive, successful CILRTs before the licensee can return to the normal Type A test frequency. Having failed to meet that criteria a Type A test shall be performed at the next Unit 2 refueling outage unless exempted by the NR No violations or deviations were identified; however, one issue requiring further work from the licensee and review and evaluation by the NRC will be followed as an unresolved item 281/86-36-05, Revalidation of Test Results using Recorded Dat . Unresolved Items Unresolved items are matters about which more information is required in order to ascertain whether they are acceptable items, violations or devia-tions. Unresolved items disclosed during the inspection are discussed in paragraphs 2.a.(1), 3.a, 3.b, and . Exit Interview The inspector met with licensee representatives (denoted in Paragraph 1) on November 20, 1986 at the conclusion of the inspection, and by telephone on February 9,1987. The inspector summarized the scope and findings of the inspection. The licensee acknowledged the information ano did not indicate that any of the information disclosed during the inspection could be consid-ered proprietary in natur The following is a list of new items identified during this inspection:

UNR 281/86-36-01, Status of Liner Weld Leak Chase Channel Plugs, paragraph 2.a.(1)

VIO 281/86-36-02, Failure to Properly Valve Penetration for the CILRT, paragraph 2.a.(3)

UNR 281/86-36-03, RTD Calibration for the November 1986, Type A Test, paragraph UNR 281/86-36-04, Review of Tceporature Survey to Justify Sensor Locations, paragraph UNR 281/86-36-05, Revalidation of Test Results using Recorded Data, para-graph 4.c.(2)

-