IR 05000295/1987032

Insp Repts 50-295/87-32 & 50-304/87-33 on 870921-1026. Violations Noted.Major Areas Inspected:Rcs Pressure Isolation Check Valves,Lers & Related Licensing History & Associated Licensee Practices

ML20236F912
Person / Time
Site:	Zion  File:ZionSolutions icon.png
Issue date:	10/27/1987
From:	Hinds J NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION III)
To:
Shared Package
ML20236F862	List: IR 05000295/1987032 ML20236F861
References
50-295-87-32, 50-304-87-33, NUDOCS 8711020338
Download: ML20236F912 (21)

Text

_- . __

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

k g* . *

.,

-

, ,3

-

L U.S. NUCLEAR REGULATORY COMMISSION ,

REGION III

L Reports No. 50-295/87032(DRP); 50-304/87033(DRP)  ;

-

!

Docket Nos.- 50-295;.50-304 Licenses No. DPR-39; DPR-48 Licensee: : Commonwealth Edison Company Post Office' Box 767 Chicago, IL60690 Facility Name: ' Zion Nuclear Power Station, Units 1 and 2

, Inspection At: ' Zion, Illinois l

Inspection Conducted: September 21 through October 26,.1987

Inspectors: M. M. Holzmer  !

P. L. En .

.q, ..

Approved B M. Hinds, I o 2 1 827 Reactor Projects Section 1A Date Inspection' Summary Inspection on September 21 through October 26, 1987 (Reports No. 50-295/87032(DRP);

'

50-304/87033(DRP))

Areas Inspected: Special, unannounced safety. inspection of testing associated with reactor coolant system (RCS) pressure isolation check valves (PIVs),

Licensee Event Reports (LERs); and related licensing history and associated licensee practice Results: Of the two areas inspected, potential violations of quality assurance j program requirements as applied to testing of PIVs, a potential violation of j Technical Specification requirements to conduct testing in accordance with '

approved test procedures, and a potential violation of an NRC Order were identified. Together, these violations indicate a potential breakdown of the licensee's management control system for PIV testin gDR ADOCK 0 % 295

- - _ _ . _ - _ __ . _ _ _ - _ _ _ _ _ - - _ - _ . _ _ _ _ _ _ _ - _ _ _ ___-_ A

j

'

. *

,

DETAILS l1. Persons Contacted

1. • G. Plim1, Station Manager

• E. Fuerst, Superintendent, Production j

• T. Rieck, Superintendent, Services W. Kurth, Assistant Station Superintendent, Operations

• R. Budowle, Assistant Station Superintendent, Technical Services

• L. Pruett, Unit 1 Operating Engineer-R. Cascarano, Technical Staff Supervisor C. Schultz, Regulatory Assurance Administrator

1. L. Holden, Regulatory Assurance Engineer

• J. Ballard, Quality Control Supervisor

• W. Stone, Quality Assurance Supervisor

1. • T. Printz, Assistant Technical Staff Supervisor

• J. Tiemann, Primary Group Leader, Technical Staff

1. • M. Pigon, Engineer, Primary Group W. Reecher, Engineer, Primary Group

• M. Madigan, In-Service Testing Coordinator, Technical Staff

• B. Soares, Engineer, In-Service Testing Group

1. P. LeBlond, Licensing Staff

1. • F. Lentine, PWR Licensing Supervisor

• J. Johnson, Westinghouse Site Representative

1. J. Yost, Quality Control

1. A. Padleckas, Technical Staff Engineer

1. R. J. Neeley, Quality Assurance Inspector

1. J. Reiss, PWRE Field Engineering Supervisor

• Indicates persons present at exit interview conducted on October 9, 198 # Indicates persons present at exit interview conducted on October 26, 198 . Pressure Isolation Valve (PIV) Testing (93702)

Backgrocnd The Reactor Safety Study (RSS), WASH-1400, identified, in a pressurized water reactor (PWR), an intersystem loss of coolant accident (LOCA) which is a significant contributor to public risk from core melt accident The Event V scenario described in WASH-1400 consists of a LOCA caused by the failure of two check valves in series; the escaping coolant then bypasses containment by subjecting a low pressure system to full RCS pressure. Event V therefore results in both a LOCA with potential for core melt and a simultaneous breach of containment. The design examined in the RSS consisted of two check valves in series as PIVs, although other Event V configurations exis Zion station has two Westinghouse 4-loop PWRs, the emergency core cooling systems (ECCS) for which include both a safety injection (SI) system and a residual heat removal (RHR) system. The design pressures of these systems are 1700 psig and 600 psig, respectively. Overpressurization l

, i l l 2 j l i

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

,. . ,

of these systems is prevented by the use of two series check valves as PIVs, as well as by relief valves, the setpoints of which are designed to prevent pressure in the low pressure system from exceeding the design pressure of the piping. There is shared piping between the SI and RHR systems for the four cold leg injection lines and for two of the four hot leg injection line The PIVs and their descriptions are provided below (set Attachment 1): I Valve N Description Event V Valve 251-9001 8" combined SI/RHR cold leg Yes A-D injection, closest valve to RCS 2SI-9002 8" RHR cold leg injection, second No ;

A-D valve frce RCS -. -'

2SI-9012 2" SI cold leg injection, second Yes A-D valve from RCS 2RH-8949 8" combined SI/RHR hot leg Yes A and B injection, closest valve to RCS 2SI-8949 2" SI hot leg injection, closest No C and D valve to RCS  ;

i 251-8905 2" SI hot leg injection, second No 1 A and B valve from RCS 2RH-8736 8" RHR hot leg injection, Yes )

A and B second valve from RCS 2SI-9004 2" SI hot leg injection, No i C and D second valve from RCS Event Chronology February 23, 1980 As a result of WASH-1400 studies, the NRC issued a letter pursuant to 10 CFR 50.54(f) to all light water reactor (LWR) licensees requesting information regarding measures taken to test Event V check valves for pressure isolation capabilit The letter delineated acceptable test methods for testing

"each check valve" and stated that each licensee's response would aid in the determination of whether its facility license should be modifie The letter specifically requested:

• a description of the valve configurations at each facility,

___

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

u y h

L

• -a 'ddetermination of whether. Event V valve

configurations exist,.the type _of. testing-conducted on and -the past. test history of'

these valves, and

• a statement:of whether' plant procedures or-i modifications are necessary'to adequately test the valve February 29; 1980 The NRC issued Confirmatory Orders to Zion and

' Indian Point which imposed short and long term l; requirements to both licensees; these plants-

,

were considered.to pose a disproportionate risk

, to the public because'.the plants are situated

! . *

in' densely populated regions. The' Confirmatory Orde included Itam A.5,. requiring the. licensee to conduct

. testing.to assure that the LPI.'(Low Pressur Injection)/RHR check valves are installed. correctly, and functioning as pressure isolation barriers prior to plant startup and periodically thereafte March 14, 1980 Th'e licensee responded to the February 23, 1980 letter, stating that'some, valves in-the SI and RHR systems meet ,

the Event V configuration, that. station procedures'have 1 been revised'to perform leak-testing on the'first in

.each set-of two check valves once each refueling outage

, to verify its integrity, and that the' valves!will be tested every time the RCS is brought to within 100 psig

'

of the maximum possible pressure on the low pressure .

side of the valve March 28, 1980 The licensee responded to' Item A.5.of the February 29, 1980 Confirmatory Order and stated that changes to station procedures were being prepared to require.an LPI/RHR check valve test whenever RCS pressure has decreased to within 100 psig of the RHR system design pressur April 20, 1981 Confirmatory Orders were issued to various LWRs with specific instructions for testing PIVs and acceptance criteria to be applied for such testin These Orders  !

were much more specific with respect to Event V valves 'l than was the February 29, 1980 Zion Confirmatory Orde : June 22, 1984 As part of questions generated by review of the licensee's In-Service Testing (IST) program for pumps

,

and valves, the NRC requested the licensee to identify l

those valves which are leak rate tested because they

[ perform a pressure boundary isolation (PI) functio The licensee was also asked if a failure of L

L- _ _ -- _ _ _ _ _ _ _ - ._ _ _ _ .l

f o. . ,

.  !

L Valve MOV-SI 8802 (cold leg SI injection header isolation valve).in the closed position during quarterly exercising would render the analyzed l flow path for the SI system inoperabl January 2, 1985 The NRC granted interim approval for the licensee's )

IST program as submitted, effective until December 31, i 198 l January 25, 1985 The licensee responded to the NRC questions on its IST program, identified valves that serve a PI . function, and stated that the IST program will be revised to reflect PI testing on a refueling outage basi i The response also stated that failure of MOV-SI 8802 l in the closed position would render the SI cold leg i injection path inoperable and that the licensee will )

revise the IST program to stroke the valve on a cold 1 shutdown frequenc ]

February 1985 The licensee's implementation of the IST program l was inspecte The licensee was cited for lack of l administrative procedures governing'IST and lack j of evaluation and trending of test result The j licensee was also cited for one example of using {"

noncalibrated, nontraceable test equipment (stopwatches) to determine component operabilit June 1986 During performance of PT-2A, " Safety Injection Pump Test," the licensee discovered that the SI pump performance was not consistent with past results. Further investigation by the licensee revealed that IL SI pump discharge header was pressurized. Informal, undocumented temperature surveys of the cold leg injection piping revealed that valves 2SI-9012 A and D and 2SI-9001 A and D appeared to be leaking. Work requests to repair or replace the subject valves were writte A temporary procedure change to PT-2A was written to provide for !

depressurizing the SI pump discharge header for pump i test purposes by opening valve 2SI-8961, a normally locked closed containment isolation valve (CIV).

Some external leakage, as well as seat leakage,.from SI relief valves 2SI-9030 and 2SI-9031 was observe i These valves are located on the discharge piping coming 1 from the "A" and "B" SI pumps, respectively. The relief setpoint for each valve is approximately 1750 psig. RCS unidentified leak rates did not exceed 1 gp i

l l

.. . . . _ _ _ _ - - _ - _ _ _ _ _ -

_

- _ .

.- l October 15, 1986 As a result of significant check valve failures at ;

several plants, INP0 SOER 86-03, " Check Valve Failure l or Degradation" was issued. The SOER Identified RHR and LPI system check valves as potential problem March 13, 1987- Generic Letter 87-06, " Periodic Verification of Leak Tight Integrity of Pressure Isolation Valves (PIVs),"

was issued, repesting submittal of a list of all PIVs and a description of testing performed "for each valve" to assure valve integrity as an independent Larrier, including pertinent acceptance criteri March 28, 1987 The licensee performed PT-2P, " Safety Injection System' Backup Check Valve Leak Check - RHR Cold Leg Injection Line," prior to starting the scheduled ,

Unit 2 refueling outage. PT-2P tests as many as

'

16 valves simultaneously in series and parallel combinations. Valve leak rates met the procedure ,

acceptance criteria of PT-2P. Consequently, PT-2M l and PT-2N, which would have more nearly determined the leak rates of individual valves, were not performe PT-2M and PT-2N are normally required to be performed only if PT-2P indicates gross leakage, 1 but may be performed at an engineer's discretio The test data met the PT-2P acceptance criteria; however, as later discussions will show, the test methodology did not adequately test the valve March 1987 The scheduled Unit 2 refueling outage began. Two check valves in the boron injection lines to the cold legs were replaced (these are not Event V valves, since the boron injection lines are not low pressure lines). Parts for repair or replacement of valves 2SI-9012 A and D and 2SI-9001 A and D were not available because they were not ordered in time to accommodate the extensive lead time needed for procuremen April 6, 1987 The NRC issued Temporary Instruction (TI) 2515/84,

" Verification of Compliance with Order for Modification of License: Primary Coolant System Pressure Isolation (Event V) Valves." Appendix I to this TI identified that Zion had received an order prior to April 20, 1981, and was subject to inspection as prescribed by the T June 11, 1987 The licensee responded to GL 87-0 The Zion response identified PIVs which were leak tested and delineated

the acceptance criteria as 5 gpm. The response did not indicate that PIVs were series tested in groups

,

- _ _ .

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

e- .'

as opposed to individually. The response stated that all potentially leaking valves were " repaired to prevent overpressurization of the lower pressure system thereby lifting the relief valves."

July'28-29, 1987 A meeting at Zion station took place to resolve differences between NRC reviewers and the licensee about the licensee's second ten year IST progra The licensee agreed to withdraw relief request VR-25, which would have exempted SI PIVs from leak. testin Although the valves were included in.the IST program, the licensee stated that PIV leak test results were not subject to the trending evaluation of.Section XI, because the licensee had not been specifically directed to impose all Section XI requirements to components which they stated were added to the IST program "for tracking purposes only."

July 29, 1987 During performance of PT-2P, approximately 3500 gallons of water were injected into the RCS, as reported in LER 304/87006. Initial conditions of the test required the RCS pressure to be above 1800 psig. The actual

'

RCS pressure was 900 psig. Valves 2SI-9012 A through D and 2SI-9001 A through 0, which were tested in the section just prior to the injection, were lifted off their fully closed seats due to the injectio July 30, 1987 The licensee performed PT-2P following the injection on July 29, 1987, but omitted Section II of the procedure because RCS temperature exceeded the 450F initial condition for Section I Section II tests the SI-9001 A through D valves. PT-2M and PT-2N were not performed because the leak rates obtained during PT-2P met the acceptance criteri August 26, 1987 The licensee was cited for inadequate corrective action taken in response to the 1985 violation for using noncalibrated, nontraceable measuring '

and test equipment for determining component !

operabilit ;

September 14-18, An NRC Vendor Branch team inspection regarding the licensee's response to INP0 SOER 86-03 was '

conducted at Zion, as well as at one plant in ;

each of the other four NRC regions. The team i leader expressed concern related to Zion's series l testing of PIV !

5eptember 18, 1987 The licensee and the NRC resident inspectors I held discussions regarding plant operations with Manual CIV 251-8961 ope The licensee felt that i

!

l 7 l

_ ______ )

__

1. . - ,

-

leaving 2SI-8961 open would prevent the continuous )

SI discharge header pressurization. The intent of l the discussions was to determine whether leaving this j valve open would cause a regulatory or safety concern I with respect to containment integrit !

l'

September 21~, 1987 The NRC resident inspectors, in coordination with NRC headquarters reviewers, began a detailed review of the licensee's PIV Test Procedures PT-2P and PT-2 The review identified several significant test deficiencies. 'During the week of September 21, l the inspectors held discussions with technical staff l personnel in which the inspectors expressed concern !

that several Unit 2 valves did not have valid test i '

results to assure that each valve was performing its PI function and to verify compliance with the ;

February 29, 1980 Confirmatory Orde i September 24, 1987 The licensee radiographer Valves 2SI-9012 A through j D to verify that the valve disks and springs were 1 installed. Informal temperature surveys of cold leg !

injection piping using a contact pyrometer indicated that 2SI-9012C and 25I-9001B (two series PIVs in the 1 same SI cold leg discharge line) are leakin l l

September 25, 1987 The licensee was informed of the NRC's concerns I associated with PIV testing by a conference call I between Region III, the resident inspectors, Zion l Station and CECO corporate office Issues discussed included PIV testing deficiencies which resulted in untested PIVs, operations with CIV 2SI-8961 open, and test procedures in which 2 MOV-SI 8802 is close The licensee stated that it believed that its test program met regulatory requirements, and that PIV i testing was intended only to detect gross leakag l l

September 25, 1987 The licensee administrative 1y took 2SI-8961 and 151-8961 out of service, pending determination of l whether manipulation of SI-8961 is allowed by the .

Zion Technical Specifications (TS).  !

September 28, 1987 The NRC informed the licensee that, based on the October 10, 1972 Safety Evaluation Report of the licensee's Final Safety Analysis Report (FSAR),

valve 2SI-8961 should remain shut, but may be opened periodically for short durations for plant operations or testin This acceptance was based on missile p protection for the SI test header piping as well as a low usage factor.

!

l-l

_ _ _ . ..- _

.

p ,

October 2, 1987 The licensee conducted TSSP-74-87 to check leakage past 2SI-9001 A through D. The test used 2FI-928, a rotometer, which had never been calibrated by the licensee. The resident inspectors informed Region III and the licensee that test results obtained when using a non-calibrated test instrument are not acceptabl October 2, 1987 The licensee radiographer valves 2SI-8905 A and B !

and 2SI-9004 C and Unit 2 PIV Testing Concerns PIV Testing and Test Results (1) PIV testing procedure PT-2M, "RHR/SI Cold Leg, SI Hot Leg Injection Check Valves Leak Check" will not produce the intended results if performed as writte PT-2M was last performed on March 28, 1987, prior to the last Unit 2 outag The following discrepancies were noted:

• In Sections II and III, check valve leakage flow is estimated by opening a pressure gauge vent valve and visually determining the flow rat * In Section IV, the procedure fails to instruct operators to valve in pressure gauge PI-933 prior to reading the gauge. In the March 28, 1987 test, a pressure of 0 psig was recorded for the upstream side of the Loop B check valve Since the procedure did not direct operators to valve in the gauge, it cannot be demonstrated whether the actual pressure was zero, or the gauge was isolated. In the test, this deficiency would produce non-conservative result * The procedure does not provide for bleeding off pressure on the upstream side of tested PIVs between successive sections of the test. Consequently, high pressures may persist into the next section, yielding inconclusive test result * In Sections II, III, IV and V, operators are required to measure the flow rate after throttling pressure to 1750 psig; however, the applied test pressure is only 650 psi As a result, leak rates are not quantifie :

(2) PIV testing procedure PT-2P, " Safety Injection System Backup '

Check Valve Leak Check - RHR Cold Leg injection Line," will not produce the intended results if performed as writte PT-2P was performed on July 29 and 30, 1987, while the licensee was preparing to place Unit 2 on line following the last ]

refueling outag The following discrepancies were noted:

a l

i 9 l l

l

m -

,

-

.'. .

g

,  ??

"'

. .

1f

• In Section I.of the procedure, the f'llowing o PIVs are not adequately tested due to a test lineup which equalizes 1 pressure,across~the. valves: 2SI-9012 A'

through D, 2SI-9004 C;and:D, and 2SI-8905 A an :*' lIn-Section II of the procedure, the following.PIVs are not adequately tested due to a test lineup which-equalizes pressure across.the valves: 2SI-9002 A through'D and 2RH-8736 A and B.-

,

C Because of these inadequacies in PT-2P,'and because later sections of PT-2P do not repeat tests of these j valves; there were no valid test results for severa '

- PIVS (2SI-9012 A through D, 2SI-9004 C.and D and 2SI-8905 A and B).

a

-* PT-2P tests several PIVs in series. This method-does not adequately test each of'the two valves, but determines the effectiveness of the~one valve in the pair which is leaking'the least, lhe ability of the other valve to-perform its PI function is indeterminat * Because the July 29, 1987, PT-2P initial test conditions werenotmet~,the2Asafetyinjectionpum) injected 350 gallons the RCS,of--refueling as documented water storag/8700 in LER 304 tank (RWST) water into Thisinjection invalidated the leak test results-for several PIVs by stroking 2RH-8949 A and B, 251-9012 A through 0,

'2SI-9001 A through D, and 2SI-8905 A and (3) As'a consequence of these deficiencies, and the fact that only PT-2P was performed at the conclusion of the last Unit 2 outage, the following valves are considered untested: l Valve N Reason Not Tested Note

'

251-9012 PT-2P fails to adequately test 1,2,4 l A through D these valves because the test lineup equalizes pressure across thevalves(July 30, 1987 test).

2SI-9004 PT-2P fails to adequately test C and D these valves because the test lineup equalizes pressure across thevalves(July 30, 1987 test). 1

)

.

251-9001 Thesafetyinjectionwhichoccurred 2,4 . 4 L A through D on July 29, 1987, during Section III i of PT-2P, lifted these valves off their t seats, invalidating the test results, ( and Section II of PT-2P, which adequately tests these valves, was not subsequently i

performed (July 30, 1987 test).

t

,

10 .

\

L .._

__

.

.'

2SI-8905 The safety injection which occurred 3 A and B during Section III of PT-2P on July 29, 1987, lifted these valves off their seats, invalidating the test results, and PT-2M, which adequately tests these valves, was not subsequently performe Notes: These valves were radiographer on September 24, 1987 (2SI-9012 A through D), and October 2, 1987 (2SI-8905 A and B and 2SI-9004 C and D), and the disks and springs appeared to be intac . Violates the NRC Confirmatory Order dated February 29, 1980, which required PIVs to be tested'each time plant pressure is reduced to within 100 psig of the RHR system !

design pressure. (Plant pressure was reduced to L atmospheric pressure during the Unit 2 refueling.) , Is a potential deviation from a March 28, 1980, response i to a 10 CFR 50.54f request dated February 23, 198 . These valves were leak checked using a non-calibrated flow instrument. The licensee has expanded the list of untested Unit 2 PIVs following the identification of the non-calibrated flow instrument. In addition, the licensee reviewed Unit 1 PIV test results and concluded that, because the Unit 1 flow instrument was also not calibrated, Unit 1 PIV test results obtained using that instrument were also not vali (4) PIV testing requirements imposed on Zion appear to be much less stringent than those imposed on all other licensees l (except Indian Point) by orders issued in 1981. Deficiencies i in requirements at Zion include: ]

• Requirements to test - Zion does not appear to be explicitly j required to test PIVs when they are moved from their fully l shut position. In addition, Zion does not appear to have l committed, in response to the February 29, 1980 Confirmatory Order, to test SI PIVs when RCS pressure is lowered to below SI system design pressure plus 100 psig, whereas its response to the February 23, 1980 50.54(f) letter indicates that it does test SI PIVs when RCS pressure is lowered to that poin * Test methodology - Zion tests as many as 16 PIVs at a time in PT-2P, many of which are in series. Series testing will ,

only reveal the leak rate for the better of two series PIV I l

l

l

l

-_ __

..

'o ' !,

.

L The condition of.the worse of the two series PIVs is indeterminat While the NRC's licensing correspondence l frequently refers to "each valve" in discussions of PIV  !

testing, the licensee considers that its series tests l meet that inten * Acceptance criteria - The. February 29, 1980 Confirmatory Order made no mention of a leak rate acceptance criterio Zion's procedural acceptance criterion is 5 gpm. Also,  !

the licensee believes that the intent of the February 29, j 1980 Confirmatory Order is that only gross. leakage  ;

identification is require This position is based on a l January 25, 1980 NRC internal memorandum, which the licensee i has, which stated that in the short term leakage rates up to I the design flow of ECCS relief valves are acceptable. 1981 Orders to other LWR licensees require acceptance criteria  ;

of 1 gpm, unless the increase in leek rate does not exceed '

prior leakage by an amount that reduces the margin between the most recent measured leakage and 5 gpm by 50%. In that case, leak rates up to 5 gpm are permissible. Leak rates above 5 gpm or increases which reduce the safety margin by more than 50% above 1 gpm are not acceptable, and specific actions are require ]

l b. Acceptability of Operations With 2SI-8961 Open

]l PIV testing (PT-2P, 2M and 2N) requires manipulation of a normally i locked closed manual containment isolation valve, 2SI-896 i 2SI- 8961 is the single CIV on the accumulator test line containment l penetration. This valve was also opened on July 28 and September 1,  !

1987 to perform PT-2A, " Safety Injection Pump Test." The resident ]

inspectors were concerned that opening 2SI-8961 appeared to violate containment integrity as defined in Technical Specification (TS)

Definition 1.1 The licensee's FSAR, Section 6, indicates that ECCS testing during power operations will be performed. Although not explicitly stated, this testing would reasonably include opening 2SI-8961. In a Safety Evaluation Report (SER) dated October 6, 1972, the NRC approved the licensee's containment isolation design despite the fact that General Design Criteria 55, 56, and 57 were not met for some containment l penetrations. This acceptance was based on penetrations having either  !

missile protection for affected piping inside containment, low usage factors for the penetrations, or additional valves outside containment which could be closed in the event of a leak. It therefore appears that the intent of both the licensee and the NRC was that opening valves such as 2SI-8961 for short periods of time is permissibl In a telephone conversation between the NRR licensing project manager, NRR reviewers, and the Zion resident inspector and his immediate management, it was agreed that for purposes of interpretation, it would be permissible to allow the licensee to open 2SI-8961 for

1

_ _ . . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

. _ _ _ _ _ _ _ _ _ _ _ - _ -

.- .

short durations during operations, since, after considering the intent of the FSAR and the SER, the closed missile protected pipe constitutes a barrier as good as that specified in section a.2 of .

the definition of containment integrit I The resident inspector informed the licensee that a TS revision is needed to clarify how containment integrity is to be defined as it pertains to " closed systems" inside containment, as in the present case. The licensee agreed to clarify this TS in a future TS change reques . Acceptability of Conducting Tests With 2 MOV-SI8802 Shut IE Information Notice (IEN) 87-01 alerted licensees to a potentially significant event, which may occur when valve misalignment in the ECCS could degrade the system performance. In the examples given, RHR cross-connect valves and discharge valves (all are motor operated valves [M0Vs]) were closed for testing or maintenance, isolating two of four cold leg injection paths from the remaining ,

available RHR pump, while one pump was being tested. For these licensees, ECCS analyses assumed that one RHR pump injected into all four cold leg Closing cross-connect or discharge MOVs therefore placed the plant in an unanalyzed conditio While reviewing PT-2A, " Safety Injection Pump Test," the inspectors learned that to perform PT-2A, the licensee sometimes closes 2 MOV-SI8802 (depending on plant conditions), which isolates both SI pumps from all four SI injection path This appeared to be similar to the event described in IEN 87-01. During the test, the licensee leaves 2 MOV-SI8802 energized. Per Piping and Instrumentation Drawing (P&ID) M-521, this valve receives an "S" signal to open automatically in the event of an ECCS actuatio l l

The inspectors reviewed the licensee's response to IEN 87-0 l The response stated that other valves which are (n the RHR system I correspond to the valves which were noted in the IEN as having l potential for causing degradation of the ECCS if they are close The valves, RHR header cross connect valves (2 MOV-RH8716 A through C)'and RHR discharge valves (2 M0V-RH8809 A and B), are closed in l PT-28, 2P, and 2N. These test procedures (pts) are performed at hot I standby or during power operations. These valves also receive an i

"S" signal to open in the event of an ECCS actuatio The licensee concluded that the IEN did not apply to Zio The IEN response also stated that Zion's ECCS analysis assumes that injection occurs in all four cold legs in the event of an acciden While its response to the IEN fails to consider the closure of ,

'

any of these valves for maintenance or due to equipment failure (non-surveillance conditions), the licensee is correct in that

!

i

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

f q g, j.+ j

. .

J

,

O l

j closing any of these valves does not' render a component, trai '

l~ . system, or subsystem inoperable, because the valves are capable 'q

!,

ofl opening without operator action in-the event lof an acciden j

,

The:1'icensee is revising.' pertinent' pts to limit the closure of these ,

val ve s.- It will consider either clarifications to the Technical

,'

Specifications or guidance to-operators by other means to ensure

.that it is. understood that closure of these valves.will.not be l construed to constitute entry into a;TS Limiting Condition for Operatio Safety Significance According to the EG&G review of.the Zion Probabilistic Risk' Assessment (PRA) (Report No. EGG-EA7304, Contract No. DE-AC0/-76ID01570), the

.

Event-V sequence accounts for 81% of the non-seismic public ris (Seismic sequences' account for 92% of the public risk.) Consequently, PIV integrity is'a significant safety issu The:NRC's-February 23,.

1980 50.54(f)Lletter noted that the average failure probability of check valves would be reduced significantly by periodic testing or by continuous surveillance of each valve, underscoringLthe defense in depth design of having two PIVs in' serie The failure of Zion's PIV testing program as' described above could have led'to a condition in which one of the two series valves was tested and found to be operable,=and the other was not performing its PI function at al Root Cause Analysis

• It appeared that the licensee's PIV testing program is only intended to measure gross leakage. This is inconsistent with NRC Orders to many other licensees, which specify very low acceptance criteri This discrepancy appears to have stemmed from a liberal interpretation of the February 23, 1980 letter and the February 29, 1980 Confirmatory Orde * The licensee's PIV testing program has been unable to test each valve for even gross leakage, as indicated by the procedural deficiencies described abov This problem appears to have resulted from a failure to fully apply the licensee's Quality Assurance Program to testing of safety related PIV * Technical Staff engineers are either unaware of some regulatory requirements regarding testing procedures, calibrated test equipment, and documentation of test results required by Orders, TSs and the licensee's approved Quality Assurance Program, or they do not understand that these requirements apply to their activitie '

• PIV testing is tracked under the licensee's IST program, but the licensee has stated that, while these valves are included in its IST program, PIVs'are not subject to all the requirements of the ASME Section XI Cod The licensee's position is that these valves

. _ _ _ _ _ - - _ _ _ _ _ - _ !'g7 .,y 4-

.s g

' '

e (and others" not mentioned in this inspection) are in the IST program

'

. for. tracking purposes only. Consequently, leak rates are not specified or trended foe individual valves. ; Testing resfor valve leak t conducte ,

byLone' Technical Staff group and-tracked by.another Technical Staff' 4 group,~and results are not evaluated or trended by either grou Inspector Concerns u

4 The licensee's response to GL 87-06, dated June 11, 1987, states, "AllLpotentially: leaking valves are repaired to prevent-

overpressurization of the lower pressure system and consequently lifting'the relief valves." Although PIV leakage was known to be pressurizing.the SI header in mid-1986,- the licensee faile 'to procure replacement valves in time.to make repairs'during the March through August 1987 Unit 2 refueling outage. Its response to GL 87-06 also was sufficiently vague as to permit'a reasonable agency reviewer to conclude that the Zion PIV test response was ,

acceptable, when, in reality, the program was inadequat !

• Some leak rate quantifications were performed visually by the Technical Staff, but the licensee has not been able to show how'

engineers performing visual. leak rate ~quantifications have been  :

trained and qualified. In addition, there exists no: formal program to demonstrate how these measurements are 3erformed, and that the results obtained conservatively estimate t1e sizes of leak Corrective Actions The11icensee has initiated steps to accurately determine the status of untested Unit l'and Unit 2 PIVs. Specific actions involve fabrication and installation of calibrated flow measurement instrumentation for determining PIV leakage, procedure revisions, and reviews for improvements in' test methodology. Actual testing will be performed as soon as preparations are complete. Radiography of some valves will be reviewed in accordance with a) proved arocedures, and if necessary, additional testing or radiograp1y will 3e performed to assure that radiograahed valves are performing their PI function. A review of' visual lea < rate estimation practices will be performe Enforcement Zion Technical Specification 6.2.1 requires that surveillance tests be conducted in accordance with written procedures. PT-2P, " Safety Injection System Backup Check Valve Leak Check," prerequisites require that the reactor coolant system pressure be greater than 1800 psi i Performance of Surveillance Test PT-2P, " Safety Injection System  !

f~

Backup Check Valve. Leak Check," on July 29, 1987, with reactor coolant system pressure below that specified in the test prerequisites is a violation of TS 6. Thisviolationresultedinaninjection of 3500 gallons of water into the reactor coolant syste j i

l

J

-_ _ _

-,

.- ,

l b. 10 CFR 50, Appendix B, Criterion XII, as implemented by Commonwealth j Edison Company's NRC-approved Quality Assurance Program, CE-1-A, I in Quality Procedures 12-1 and 12-2, requires that instruments used to perform activities affecting quality be controlled, calibrated l and adjusted on a periodic basis to maintain accuracy. Flow Instrument 2FI-928A, which was used to perform activities affecting l quality on March 28, 1987, in Procedures PT-2P, " Safety Injection l System Backup Check Valve Leak Check," and PT-2M, "RHR/SI Cold Leg, l SI Hot Leg Injection Check Valves Leak Check," on July 29, 1987, i and July 30, 1987, in PT-2P, and on October 2, 1987, in I Procedure TSSP-74-87, "SI Cold Leg Check Valve Leak Test," was not controlled, calibrated or adjusted on a periodic basis to  ;

maintain accurac c. 10 CFR 50, Appendix B, Criterion XI, as implemented by Commonwealth i Edison Company's NRC-approved Quality Assurance Program, CE-1-A, in '

Quality Procedure 10-53, requires in part that procedures which demonstrate that components will perform satisfactorily shall include appropriate test requirements and acceptance limits and that the test is performed under suitable condition CFR 50, Appendix B, Criterion V, also implemented by CE-1-A, in Quality Procedure 11-51, requires in part that activities affecting quality I be performed using procedures appropriate to the circumstances and i include appropriate acceptance criteria. On March 28, July 29,  !

and July 30, 1987, the licensee performed Test Procedure PT-2P, i

'

" Safety Injection System Backup Check Valve Leak Check." The licensee also performed Test Procedure PT-2M, "RHR/SI Cold Leg, SI Hot Leg Injection Check Valves Leak Check," on March 28, 198 i These tests do not demonstrate that pressure isolation valves will I perform satisfactorily, do not impose appropriate acceptance criteria '

and do not assure that tests are conducted under suitable condition Specific examples include:

• PT-2P fails to adequately leak test valves 2SI-9012 A through ,

D, 251-9004 C and D and 2SI-8905 A and B because the test l lineup equalizes pressure across the valve I

• PT-2M requires that leakage flow be measured at a throttled test pressure of 1750 psi; however, the applied test pressure i is only 600 psi. In addition, one section of the procedure fails to require valving in the pressure gauge used to determine whether check valves are leakin * PT-2M and PT-2P do not test valves individually. Valves are !

either tested in parallel without conservative assignment of leak rates to the parallel valves, or valves are tested in I series, indicating only that one of the two valves is performing its pressure isolation functio * There are no program provisions to test the RHR PIVs when RCS pressure falls to less than or equal to 700 psi, or to test the SI PIVs when RCS pressure falls to less than or equal to 1800 ps l

- - - -

-_

.' ,

)

i i

• There are no program provisions to test RHR or SI PIVs following events which cause the valves to move from their fully closed 1 position I A' Confirmatory Order dated February 29, 1980, required tne licensee ,

to conduct testing to assure that the "LPI/RHR check valves are I in fact installed correctly and functioning as pressure isolation !

barriers . . . whenever RCS pressure has decreased to within 100 psig of RHR design pressure." In a letter dated March 14, 1980, the licensee responded to the Commission's 10 CFR 50.54f letter dated February 23, 1980, stating that "Once each refueling- 1 outage, a leak test is performed on the first check valve [SI-9001 A l through D] of each cold leg injection line (PT-2M) to verify their l integrity. . . [and] both check valves [SI-9012 A through D, l SI-9001 A through D] of each SI cold leg injection line will be tested every time the primary coolant system (PCS) pressure is brought to within 100 psig of the maximum possible pressure en the low pressure side of these check valves." On July 30; 1987, after the RCS pressure had been reduced to atmospheric conditions l during refueling, the licensee failed to perform Section II of 1 PT-2P, " Safety Injection System Backup Check Valve Leak Check," 1 and PT-2M, "RHR/SI Cold Leg, SI Hot Leg Injection Check Valves Leak Check," such that PIVs 2SI-9012 A through D and 2SI-9001 A ]

through D were not effectively tested prior to unit startu Four violations and no deviations were identified. These violations are collectively considered an Unresolved Item pending NRC review for escalated !

enforcement (295/87032-01; 304/87033-01).

3. Licensee Event Reports (LERs) Followup (92700)

i Through direct observations, discussions with licensee personnel, I and review of records, the following event report was reviewed to determine that deportability requirements were fulfilled, whether immediate corrective action was accomplished, and whether corrective actions to prevent recurrence had been accomplished in accordance with Technical Specifications. The LER listed below is considered closed:

Unit 2 LER N DESCRIPTION 87-006 Inadvertent Injection from Safety Injection to the Reactor Coolant System During Periodic Check Valve '

Leak Testing This LER was reviewed in connection with the PIV testing inspection described in Paragraph 2. Failure to meet test procedure initial conditions was noted to be in violation of TS 6. No other violations or deviations were identifie _ _ _ _ _

f J V I

.

4. Lic'ensing History and Associated Licensee Practices (92703)' bnFebruary-23,1980, the NRC issues a 10 CFR 50.54(f). letter to all LWR licensees, identifying concerns related.to Event V valves and _ requesting the following information:

 * Describe the valve configuration at each plant and indicate whether Event V. configurations exist within the Class I boundary of the high pressure piping connecting Primary Coolant System (PCS) piping to low pressure systems, e.g., two check valves in' series, or two check valves in series.with an M0 * For those Event V configurations at the plant, indicate whethe continuous surveillance or periodic tests are being  .

accomplished to ensure integrity. Also indicate whether " testing had revealed any valves lacking integrit (- Indicate whether plant procedure revisions or plant modifications are needed to increase reliabilit The 50.54(f) letter indicated that acceptable methods to assure  ! component integrity included continuous pressure monitoring on the low pressure' side of each check valve and periodic testing of each' valve every time the-plant is shutdown and each time either' check ~ valve (of a series pair) is moved from its fully closed positio,. < 1 The staff expected typical licensee responses to the 50.54(f)' letter i to specifically. identify Event V valves by number, to describe their j configuration and test frequency, and to address test methods used ! to determine individual valve leakag It was also thought that licensees would include past valve leakage testing history and i associated test acceptance criteria if used. Responses were subse-quently forwarded to Franklin Research for a technical evaluation, which if appropriate was used as the basis-for a Safety Evaluation Report (SER). An SER was then attached to each Confirmatory Order issued to those plants whose Technical Specifications were amended " to explicitly require individual PIV testin On March 14, 1980, the licensee responded to the 50.54(f) letter and identified the RHR hot leg injection path and the SI cold leg injection path as Event V configurations. The licensee also stated the following: Once each refueling outage, a leak test is performed on the first check valve [ closest to the RCS] of each cold leg injection line (PT-2M) to verify their integrity. The results of this test indicate that the check valves do not leak or lack integrity. Currently, no tests are performed to verify the integrity of the second check valve of each SI cold leg injection line or either check valve in the RHR hot leg injection lin , 18 ], h __ _ o

p, ..

,

i i In addition, plant procedures at Zion Station have been revised as follows: j

     .c Both check valves of each SI cold leg injection line will- ;

be tested every time the primary coolant system (PCS) l pressure is brought to within 100 psig of the maximum {i possible pressure on the low pressure side of these check valve . Both check valves of each RHR hot leg injection line will l be tested every time the PCS pressure is brought to within ! 100 psig of the maximum possible pressure on the low pressure side of these check valve .

     '

In regard to the discussion above, the inspectors noted the following:

(1) The licensee's check valve tests do not test each valve i separatel (2) The licensee's response to the 50.54(f) letter did not address testing each valve of a series pair each time a valve is moved

. from its fully closed position, and the licensee currently does I not test valves following actuatio ! l (3) There is no specific operating constraint in place to ensure that the first check valve of each cold leg injection line is tested during a refueling outage prior to reactor startu In addition, the procedure currently used to test these valves does not test them individually and was found to be invalid during a recent inspectio (4) The licensee's implemented check valve tests include no requirements to the RHR system PIVs when RCS pressure falls to < less than or equal to 700 psi (100 psi over RHR system design pressure) or to test SI system PIVs when SI pressure falls to less than or equal to 1800 psi (100 psi over SI system design pressure). The licensee requires PIV testing only in GOP-0,

 " Plant Startup Documentation Requirements," at the discretion of the operating engineer prior to leaving hot shutdow b. By Confirmatory Order dated February 29, 1980, the licensee was directed to " conduct testing to assure that the LPI/RHR check valves are in fact installed correctly and functioning as pressure barriers when the plant is at pressure and producing powe Verification. . .shall be performed. . .whenever RCS pressure has decreased to within 100 psig of RHR system design pressure."

By letter dated March 28, 1980, the licensee stated that no immediate actions were required of the station; however, " changes to Zion operating procedures and various periodic tests (pts) are being i

19 l j

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

c , prepared which will require a LPI/RHR check valve test whenever RCS pressure has decreased to within 100 psig of RHR system design pressure. These changes will be implemented prior to plant startup following such an RCS pressure decrease."

l As stated ~in paragraph a above, the licensee has no requirements to ; test check valves when RCS pressure decreases to within 100 psi of l RHR system design pressure, c. On March 13, 1987, the Commission issued Generic Letter (GL) 87-06, ; which requested all operating license holders to submit a list of ' all pressure isolation valves, including the Event V valves, and to .; describe for each valve the periodic tests or other measures performed to assure the integrity of the valve as an independent barrier at the reactor coolant pressure boundary (RCPB), along with the acceptance criteria for leakage, if any, and associated l operational limits, j i The staff expected information specific enough in nature to determine whether verification of the leak tight integrity of each PIV as an independent barrier against abnormal leakage, rapidly propagating failure and gross rupture of the RCS, was being performed. The response was expected to include a list of each individual PIV, the specific test which tests each valve individually, the test frequency imposed and the associate ! individual valve acceptance criteri Operational constraints associated with PIVs were also to be addressed in the licensee : respons ~8y letter dated June 11, 1987, the licensee responded to the GL by providing a list of PIVs, including the Event V valves and the pts in which they are tested. The licensee also stated that: I The above valves are leak tested by monitoring flow or pressure ! buildup on the lower side of the PIVs. The acceptance criteria j for leakage of PIVs is to be below 5 gpm with exception of l M0V-RH8701 and MOV-RH870 Suction pressure of RHR pumps or l less than 200 psig indicates that valves M0V-RH8701 and j MOV-RH8702 have acceptable seat leakage in series per PT-2 j PT-2M and PT-2N are performed if the leakage of valves tested in PT-2P exceeds 5 gpm, to identify the source of leakage. All potentially leaking valves are repaired to prevent overpressure-zation of the lower pressure system and consequently lifting the relief valve The leak rate test results are documented l in Station Procedure TSS 15.5.20V-8, ISI Valve Surveillanc In regard to the discussion above, the inspectors note the following:

 (1) The procedure specified as testing the majority of identified PIVs does not provide a valid test for eight of the valves listed, and those valves which are validly tested are not i

1 e__________ _ _ _

, - .____ _ - __ tested individuall Also, PT-2M and PT-2N as currently , written do not provide a viable method for identifying the ' individual valve which is leaking or for quantification of the valve leakag (2) Zion station currently has evidence that at least two PIVs are leaking at a rate which is sufficient to lift the relief valves on the Unit 2 safety injection pump discharge piping; the j relief valves are in fact preventing overpressurization of the ' non-RCS pressure rated piping. This condition has existed since mid-1986 and was identified prior to the last refueling r outage on Uni (3) The stated leakage acceptance criterion is 5 gp The test  ! methodology employed does not provide a viable means of  ! determining actual leak rates, and for those portions of the test in which a valid leak rate can be determined, calibrated test equipment is not use Clarification of the licensee's interpretation of PIV testing requirements as specified in NRC correspondence and determination of how the current test program ensures regulatory compliance for PIVs are considered to be l an Unresolved Item (295/87032-02; 304/87033-02) l Unresolved Items l ; l Unresolved items are matters about which more information is required i in order to ascertain whether they are acceptable items, items of l noncompliance or deviations. Unresolved Items disclosed during this l inspection are discussed in Paragraphs 2 and j l 6. Exit Interview (30703) l The inspectors met with licensee representatives (denoted in Paragraph 1) throughout the inspection period and at the conclusion of the inspection conducted on October 8, 1987, to summarize the scope and findings of the i inspection activitie The licensee acknowledged the inspectors' comment l The inspectors also discussed the likely informational content of the

  ~

inspection report with regard to documents or processes reviewed by the inspectors during the inspection. The licensee did not identify any such documents or processes as proprietar i ! 21 I _ - _ _ }}