IR 05000344/1985025

Insp Rept 50-344/85-25 on 850809 & 12-16.Major Areas Inspected:Effectiveness of Info Exchange Between Vendors of safety-related Equipment & Util.Potential Violations Noted in Areas of Procurement of safety-related Equipment

ML20133F885
Person / Time
Site:	Trojan File:Portland General Electric icon.png
Issue date:	09/25/1985
From:	Baker E, Mcintyre R, Merschoff E, Naidu K, Zech G NRC OFFICE OF INSPECTION & ENFORCEMENT (IE)
To:
Shared Package
ML20133F884	List: IR 05000344/1985025 ML20133F881
References
50-344-85-25, NUDOCS 8510150015
Download: ML20133F885 (18)

Text

.

U.S. NUCLEAR REGULATORY COMMISSION Report No.: 50-344/85-025 Docket No.: 50-344 License No.: NPF-1 Licensee: Portland General Electric Company 121 S. W. Salmon Street Portland, Oregon 97204 Facility Name: Trojan Inspection At: Rainier and Portland, Oregon Inspection Conducted: August 9 and 12-16, 1985 Inspectors:

,/ude T. W. Merschoff, Chi Reactive Inspection Section Date (RIS)

818 E. T. Baker, RIS

&k Date R. P. Mc

- A Date W

• e,SpecialPrpectsInspectionSection K. R. Waidd, RIS

.- L ,

@rkr Date'

Approved by: - Wdv cf g Gary (L_Z_t)ch, Chidf, Vendor Program Branch Date Consultant: P. Farron, NEC l

,

G M 05000344 PDR

_

__ _ _ _ _ _ _ __ . . _ . _ _

'

..

SUMMARY Inspection during the period August 9 and 12-16,1985 (Report No. 50-344/85-025)

Areas Inspected: Five inspectors from the Vendor Program Branch, Office of Inspection and Enforcement conducted an announced inspection of the effective-ness of the information exchange between vendors of safety related equipment and the licensee. The licensee's program for receipt, evaluation, and implementation of actions determined appropriate for vendor technical information was reviewed. Safety related purchase orders were reviewed to ascertain whether they contained appropriate quality and technical requirement The inspection involved 183 hours0.00212 days <br />0.0508 hours <br />3.025794e-4 weeks <br />6.96315e-5 months <br /> by four (4) NRC inspectors and one (1)

consultan Results: Of the areas inspected, two potential violations of NRC requirements were identified. Potential deficiencies were identified in the areas of procurement of safety related equipment and components, and implementation of an effective Part 21 reporting system. In addition, weaknesses were noted in the areas of vendor interface and control of post modification testin .

f i

I

!

!

m

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

__a

..

DETAILS PERSONS CONTACTED Portland General Electric Company (PGE)

• Yundt, General Manager - Technical Functions

• Snyder, Administration Supervisor

• Roller, Manager, Electrical Engineering

• A. Zimmerman, Manager, Nuclear Regulation Branch

• M. H. Malmros, Senior Engineer, Nuclear Safety Branch

• J. Schweitzer, Manager, Quality Assurance Engineering

• H. E. Rosenbach, Supervisor, Material Control, Trojan

• J. Lentsch, Manager, Nuclear Safety and Regulation

• R. Schmidt, Manager, Operations and Maintenance, Trojan

• R. Steel, Manager, Nuclear Plant Engineering

• A. Olmstead, Manager, Quality Assurance J. Piluso, Supervising Electrical Engineer M. Cooksey, Supervising Electrical Engineer R. Greninger, Senior Electrical Engineer F. Krall, Electrical Engineer

,

K. Larsen, Electrical Engineer P. Meyer, Electrical Engineer D. Walters, Plant Electrical Engineer, Trojan D. Bennett, Supervisor, Electrical and I&C, Trojan S. Bauer, Licensing Engineer, Trojan

.

J Nuclear Regulatory Commission

• B. Grimes, Director DQAVT, 0IE

• R. Pate, Chief, Reactor Safety Branch, Region V Kellund, Resident Inspector

.

!

• >

• Denotes those individuals who attended the exit meeting on August 16, 1985.

,

. . . .. . . . - . . ~ _ - . - -. ._ - - - . . - _ . . _ - - _

?

.. BACKGROUND INFORMATION The purpose of this inspection was to ascertain the effectiveness of the Licensee / Vendor Interface by reviewing the implementation of PGE procedures that identify, track, and resolve vendor and NRC identified deficiencies. This review was accomplished by tracking several specific deficiencies through PGE's review and implementation process, and by reviewing safety related procurement actions taken to support modifications perfomed during the most recent outage. In addition, Part 21 reporting procedures, control of nonconforming material, and post modification testing were reviewe . PROCUREMENT Procram Requirements PGE procurement procedures and practices were reviewed with respect to safety related procurements made to support safety related work that was performed during the last outage. In particular, the following safety related procurement actions were reviewed in detai Table 1 Component System RDC # QA Code Vendor Agastat CCW 83-019 D Control Products Control Relay AFW 83-025 D General Electric Control Switch AFW 83-025 0 General Electric Pressure Recorder RCS77-155 0 Branom Instruments RTB Shunt Trip RTS83-042 0 Westinghouse Isolation Limit Switch AFW 84-108 D Conax Flow Switch AFW 84-105 0 ITT Barton Square Root Extractor AFW 84-044 0 Branom Instruments Circuit Breaker RHR 83-009 0 Brown Boveri MSIV Isolation Solenoid MS84-106 0 Steam Supply &

Rubber C Revision 10 of the PGE Nuclear Quality Assurance Program, and the specific implementing procedure, NPEP No. 300-1, " Procurement Procedures for Nuclear Plant Modifications and Additions," establish the methods used by PGE to procure safety related equipment and components. These methods must be consistent with PGE's comitments to the NRC in this area. Specifically, PGE has committed to meet the requirements of ANSI 45.2.13-1976 and Regulatory Guide 1.123 (Rev. 1) " Quality Assurance Requirements for Control of Procurement of Items and Services for Nuclear Power Plants" with the following two exceptions:

(1) Notification points will not be used for all procurement activitie [ _

u

..

(2) Quality-related items and services procured from contractors for which a quality assurance program is required, but has not been reviewed or approved may be utilized by PGE with the specific written approval of the Nuclear Quality Assurance Departmen Program Implementation With regard to implementing these requirements, PGE has established

three procurement quality level codes, A, C, and D. Quality level code D, as defined in NPEP 300-1, is intended to be used for, in general, items of relatively simple fabrication, such as 0-rings, relief valve springs, or piping. For quality level code D items, PGE has elected to use the method of procurement described in ANSI 45.2.13, Section 10. This method allows the use of certifi-cates of conformance (C of C) as the sole basis for accepting an item provided that the item is of simple design, involves standard materials, processes, and tests, and that a means has been provided to verify the validity of the supplier's certificates and the effectiveness of the certification system, such as the performance of audits of the supplier or independent inspection or test of the ite Although PGE procured each of the safety related items listed in Table 1 as quality level code D items, with acceptance of the item based solely on the C of C, insufficient action had been taken by PGE to verify the validity of the suppliers' C of C In particular, neither audits, surveys, nor testing to verify the validity of the C of Cs had been conducted. Currently, PGE relies on NQAP 114,

" Verification of Vendor Certificates of Conformance" to satisfy the requirements of ANSI 45.2.13 and Regulatory Guide 1.123 for verifying the validity of supplier certificates and the effective-ness of the certification system. However, NQAP 114 only requires the QA Department to verify the certification system of a sample of the vendors from which safety related procurements have been mad Further, the QA Department is only required to interview the C of C signer, rather than to perform the required verification of the effectiveness of the certification system by auditing such areas as material control, nonconformance control, and inspectio Consequently, PGE's safety related procurement activities are inadequate in that they fail to meet their commitments to ANSI 45.2.13 and Regulatory Guide 1.123 with respect to verifying the validity of their suppliers' certificates and the effectiveness of their suppliers' certification progra NRC Audit of PGE Suppliers In order to determine if any hardware problems may exist as a result of PGE's improper procurement practices, two of the vendors from which PGE had procured safety related components were inspected by the NRC

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

.*

Vendor Program Branch. These two vendors and the components they have supplied (by way of a middle-man distributor) to PGE are listed below:

Vendor Component Supplied Rochester Instrument Systems, In Square Root Extractor for AFW System Texas Instruments Reactor Coolant System Pressure Recorder The results of these inspections are documented as Inspection Report Numbers 99900222/85-01 (Rochester Instrument Systems) and 99901028/85-01 (Texas Instruments). A summary of the results is provided belo (1) Rochester Instrument Systems (RIS)

Audit findings at RIS revealed that RIS had an effectively implemented quality assurance program that was consistent with the requirements of 10 CFR 50 Appendix B. Although PGE had neither verified that RIS had an effective quality program nor imposed any specific quality requirements on

'

RIS in the purchase order, RIS was familiar with these requirements based on their dealings with other nuclear utilities and implemented their nuclear quality assurance program on the work performed to execute PGE's orde (2) Texas Instruments (TI)

The audit findings at TI revealed that TI did not, nor did they ever, have a documented 10 CFR 50 Appendix B quality assurance program. This finding was confirmed previously by Florida Power

, and Light and the Tennessee Valley Authority, both of which had

, audited TI and declined to approve them as a source of safety I related equipment due to inadequacies in TI's quality assurance program. Nevertheless, PGE accepted a safety related Reactor Coolant System Pressure Recorder which is relied on for monitoring conditions in the Reactor Coolant System during normal operating

'

conditions and anticipated operational occurrences (as stated in section 7.5 of the Trojan FSAR) from TI. This acceptance was based solely on a TI C of C which contained the following disclaimer:

"However, such products are not sold for, nor is the foregoing certification a representation that the products are suitable for applications essential to emergency reactor shutdown, containment isolation, reactor core cooling, containment and reactor heat removal, or otherwise essential in operation of the

,

6 t

I i _

.

..

reactor or preventing a significant release of radioactive material to the environment."

On inquiry, representatives from TI were unable to explain the meaning of the disclaimer. They indicated that they were new to the nuclear business and simply did not have a good feel for the requirements and responsibilities of a supplier. Consequently, their lawyers had recommended use of the disclaime This area has been identified as a potential violation of the requirements of 10 CFR 50 Appendix B (344-85-025-01).

4. 10 CFR PART 21 Program Requirements Nuclear Division Procedure (NDP) 700-4 Revision 1 dated May 17, 1985 establishes the policy and procedures for assuring compliance with 10 CFR 21 to evaluate and report defects in basic components that could cause a substantial safety hazard. Review of this procedure showed significant weaknesses in both adequacy and implementation to the extent that defects identified during the dedication process are not subject to the required review to evaluate their potential significanc Paragraph 3.2 of NDP 700-4 defines a commercial grade item and states that "A commercial grade item is not part of a basic component until after dedication; i.e., until it is drawn from the warehouse or stock and assigned for use or installation." The item is then dedicated for use as a basic component and becomes subject to the requirements of 10 CFR Part 21. Standard industry practice for dedication of comercial grade items may involve one or more quality control measures such as tests, inspections, or nondestructive examinations to determine that-the commerical grade item is free of defects and will satisfy the procurement specification. These tests could reveal inherent defects

! in the operational characteristics of items which are not detected by

! visual examinations performed during superficial receipt inspections.

l However, by PGE's procedure, defects or noncompliances discovered during

'

the implementation of quality control measures are not subject to reporting under the requirements of 10 CFR Part 21. Specifically, I paragraph 5.1.D of NDP 700-4 states " Defects or noncompliances found through implementing established quality control measures (e.g.,

i reviews, inspections, tests or non-destructive examinations) which are

! a result of the work process performed at Trojan and prior to final acceptance of a structure, system, or component for operation, shall not be subject to reporting under 10 CFR Part 21." Thus, any defects revealed during the dedication process are corrected without the benefit of the nonconformance process or the evaluation and subsequent reporting pursuant to 10 CFR Part 21.

l

_

- .- - _ _ . .- . .

.*

b. Program Implementation The NRC inspector reviewed PGE's implementation of 10 CFR 21 reporting requirements by reviewing Request for Design Change (RDC) package

,84-025. This RDC, which was approved on January 22, 1985, involved i'

the replacement of two 400 ampere trip units with 225 ampere trip units in 480 volt circuit breakers 48 and 26, located on Motor Control Centers B 23 and B 24, respectively.

'

The following is a summary of the actions taken to accomplish RDC 84-02 (1) Purchase Order N-29326 dated March 18, 1985 was placed with j Gould Electric, Gresham, Oregon, to supply the following:

,

(a) 4 Class 1E circuit breaker trip units (ITE Catalog JL3-T225)

rated for 600 volts and 225 ampere (b) Certification of seismic qualificatio (c) Vendor certification that the trip units meet the PGE specification requirements.

(2) The trip units were received from Telerrechanique (formerly

Gould Inc.) with the required C of Cs. They were receipt

inspected, accepted, and stored at the warehouse.

(3) In June 1985, the trip units were installed in circuit breakers 48 and 26, and tested. Contrary to both the specification t and the C of C, both circuit breakers failed to trip when 225 amperes continuous current was applie O i (4) The PGE electrical engineer, in consultation with ITE, deter-I mined that both trip units lacked the required thermal overload (5) ITE agreed to replace the four trip unit (6) The replacement trip units were installed in the circuit breakers and tested satisfactoril Although PGE properly tested these components, determined the cause of the malfunction and corrected it, several weaknesses in the overall program are demonstrated by this example. Specifically:

(1) The PGE purchase order did not impose any specific quality assurance or Part 21 reporting requirements on the vendo (2) PGE neither initiated a nonconformance on this item nor were they required, by their procedure, to perform a

'

Part 21 evaluation.

!

_- __ _ , - _ _ _ _ _ _ - - _ _ . . . ._ _ __r, - _ -_ _ -_

.*

W (3) Since neither PGE or ITE reported to the NRC that nonconfor cing Class 1E trip units were supplied to PGE for use at the Trojan Plant nor did PGE enter this problem in the Nuclear Plant Reliability Data System, this potential problem has not

- been made known to the industr This area has been identified as a potential violation of the

,

requirements of 10 CFR Part 21 (344-85-025-02).

5. CONTROL 0F NONCONFORMING MATERIAL While discussing maintenance practices with one of PGE's on-site engineers, the inspector noticed a number of Pacific Scientific snubbers on a table in the engineer's office. When asked why the snubbers were in his office, the engineer replied that the snubbers had failed during inservice testing and were being sent back to the manufacturer for repair. The inspector noted that there were no QC Hold Tags on the snubbers, no paperwork accompanied the snubbers, nor were the snubbers in a nonconformance hold area.

~

Revision 1 of Periodic Engineering Test, PET-9-6, " Mechanical Pipe Snubber Inspection," paragraphs 9.1.1 and 9.3.1, require that the test engineer issue a Nonconformance Report (NCR) per Nuclear Department Procedure, NDP-600-1, for snubbers that fail either the inspection or the tes Paragraph 5.3.c of NDP-600-1, Revision 0, states an NCR is not required for nonconforming items being returned to the supplier to restore the item to conformance with the technical descriptio Paragraph 5.5.M.1 of NDP-600-1, Revision 0 states that the Trojan Plant QA staff may prepare and install a QC Hold Tag at the discretion of the QA Supervisor. However, since an NCR is not required, there is no documented method for informing i the Plant QA staff that nonconforming material exist Section 15 of PGE's Nuclear Quality Assurance Program (NQAP), Revision 10 requires that when feasible, nonconforming items be clearly identified with Quality Control Hold Tags and segregated to indicate their unacceptable status. Section 15 also provides two methods of documenting nonconformances, Maintenance Requests (MRs) and Nonconformance Reports. The principal difference between the use of an MR and an NCR is that an MR is used when the nonconforming item can be restored to the original design requirements l using a rework process at the site, and an NCR is used in all other cases.

- Note that there may be no differences in the actual nonconformance, only in Trojan's ability to perform the rework.

'

The fact that the method of " reworking" the nonconforming items determines the-method of documentation, regardless of the type of nonconformance, brings up an apparent inconsistency. Paragraph 15.2.3 of the NQAP requires that NCRs be evaluated to_ determine if a significant condition adverse to quality,

an unresolved safety question, or defects resulting in substantial safety hazards exist. However, no such requirements exist under paragraph 15.2.2,

" Maintenance Requests," even though the actual nonconformance may be the sam t i

! 9

. -

- . . . . .

..

Additionally, paragraph 5.3.c of NDP-600-1 does not require NCRs to be written on nonconforming items being returned to the supplier for rewor Without an NCR or MR, there appears to be no method for tracking noncon-formig items returned to suppliers for rework and, therefore, no way for these nonconfonnances to be evaluated for significant conditiens adverse to quality, unreviewed safety questions, or defects resulting in substantial safety hazard The fact that 14 nonconforming PSAi mechanical snubbers and 11 nonconforming PSAA snubbers were not tagged with QC Hold Tags would indicate that the intent of Appendix B and the PGE Nuclear Quality Assurance Program is not being fulfilled by NDP-600-1. This is considered an Unresolved Item (344-85-025-03).

6. EMERGENCY DIESEL GENERATOR (EDG) MAINTENANCE Trojan's EDGs are General Motors (GM) diesels which were originally supplied by the now defunct Bruce. Morrison-Kn0dsen has taken over the role of informing Bruce customers of problems with, and maintenance requirements for, GM diesels. The inspectors reviewed Trojan's maintenance procedures for the EDGs, compared the Maintenance Instructions (MI) and Power Pointers (PP) received by PGE with those issued by GM, reviewed PGE's evaluation and implementation of mis and PPs, and performed a valve line up verification of the cooling system for the number one diesel in the east diesel room, Valve Line Up Verification

, An auxiliary operator and a technician from the maintenance department assisted with the valve line up check. The operator provided a copy of the checklist used for operations, 0I-5-1, to assist in performing the valve line up. All valves were in the correct position. The only discrepancy observed was that all the valves listed on the checklist had an alpha-numerical designator, but none of the valves were marked or tagged with the appropriate designator, making identification of the valves difficul Maintenance Based on a review of the maintenance procedures, a comparison between them and the mis and PPs issued by GM, and discussions with maintenance personnel, it was apparent that PGE was not receiving mis and PPs, and that PGE had not established an agreement with Morrison-

Knudsen to receive mis and PPs issued by GM. There was no record to show which mis and PPs had been ree!)ved, evaluated, and implemented by PGE. A review of the mainte d e manual for the EDGs revealed that PGE was using MI 1750, "SchM% o A intenance for Stationary Power -

Units With Turbocharged En,}r 3 " n ted October 1973, which had been

,

superceded by Revision G ot !E ric. " Scheduled Maintenance Program for Domestic Stationary Power Units With Turbocharged Engines," issued in 1984. The latest MI filed with the maintenance manual was dated 1978, although numerous mis have been issued since that time. No PPs were found during the revie .

n 10 /

j

.r ,

.

Several differences were noted between MI 1750 (the basis for PGE's maintenance schedule) and the actual maintenance procedure, MP-12- MI 1750 recomended that the exhaust manifold flange bolts be torqued every six months white MP-12-7 required it only once every three years; MI 1750 recommended that the governor oil be changed every six months while MP-12-7 required it only annually; and MI 1750 recommended a test on the temperature differential between lube oil and cooling water i'to the engine every two years while MP-12-7 did not address the tes PGE maintenance personnel acknowledged that a written evaluation of the differences between the MI and the MP had not been performe The differences between the most recent MI on maintenance (MI 1728, Revision G) ard MP-12-7 are: MI 1728 recommends a monthly lube oil analysis while PGE performs the analysis only quarterly, and MI 1728 recomends an annual check of protective devices and circuit breakers, while PGE does not require these checks to be performed periodicall As a result of PGE's failure to maintain an effective interface with their EDG supplier, the EDG preventive maintenance program at Trojan is predicated on information that is at least 11 years and eight revisions out of dat . TERRY TURBINES Two issues involving Terry Turbines were investigated during the inspection:

(1) inspection of shaft bearing anti-rotation pins and (2) whether the problems associated with conformance to the seismic qualification report identified at Palo Verde Nuclear Station exist at the Trojan plan On May 19, 1983, PGE's maintenance staff performed maintenance in response to a Part 21 report issued by Terry Corporation that notified licensees of the possibility that anti-rotation pins for locating shaft bearings may have been left out when the turbine was assembled. During disassembly and inspection of the turbine by PGE, it was noted that the inboard bearing anti-rotation pin had been installed during factory assembly, but the anti-rotation pin was missing from the outboard bearing. As advised by the Terry Corporation, PGE installed the missing anti-rotation pi During the course of the inspection, PGE received a letter from the Terry Corporation informing them that Terry had recently completed a detailed seismic analysis of the type GS turbine which was performed as a result of deficiencies identified at Palo Verde Nuclear Station. The results of the analysis were that the pedestal to casing bolting, governor end pedestal to base plate bolting, and coupling end pedestal to base plate bolting may require replacement and/or preloading to specific levels. In addition, the bolting material may not have the minimum strength necessary to carry loads imposed by a seismic event. The inspectors and licensee personnel performed an inspection of the Terry Turbine and noted that the required taper pins were missing from the coupling pedestal to base plate joint and that there were no markings on the bolts or studs, as required, to indicate the grade of fastener. The NRC inspectors discussed thesa l \ l l

l

.

..

findings with the licensee and, as a result, the licensee performed a detailed review and determined that in addition to the problems listed above, it was not possible to determine if the governor pedestal to base joints had the required shear pins installe At the completion of the inspection, PGE was still evaluating the proble PGE's engineering staff indicated that a reanalysis using Trojan's actual ground acceleration, 0.25g, rather than the 0.679 used by Terry in their analysis, would resolve the concern and alleviate the need for hardware change The reanalysis is pending PGE's receipt of specific seismic response data from Terry Corporation. However, this reanalysis does not negate the fact that PGE accepted and installed equipment which did not conform to the design analyzed in the seismic analysis. This is considered an Unresolved item (344-85-025-04).

8. TECHNOLOGY FOR ENERGY CORP. (TEC) AC0USTIC MONITORS TEC supplies acoustic monitors for use in monitoring conditions downstream of relief valves to indicate when the valve may be leaking. The inspectors examined two areas of concern involving these acoustic monitors: (1) whether or not PGE's engineering staff was aware of and evaluated a Part 21 report submitted by TEC concerning the failure of TEC Model 914-1 valve flow monitor modules to reset after valve cycling, and (2) whether or not Trojan's maintenance procedures accounted for the limited qualified life associated with TEC components, TEC Part 21 Report PGE had received the July 19, 1985 TEC Part 21 report, had reviewed it, had determined it was applicable to Trojan equipment, and was in the process of taking the appropriate action Limited Life of TEC Components A review of correspondence and maintenance procedures revealed that PGE engineering staff had performed an independent calculation of the qualified life of TEC components. TEC had based their calculations on an activation energy of 0.65eV, while PGE's calculation was based on an activation energy of 1.17eV. PGE justified the use of the higher activation energy by referencing EPRI report NP-1558, "A Review of Equipment Aging Theory and Technology," and previous Bechtel calcula-tions which had used the same activation energy for the limiting component, cross linked polyethylene insulation on the Brand Rex cable to the sensor. The use of a higher activation energy resulted in increased qualified life for the cable, from 4 years to 26.8 year Based on the fact that the acoustic sensor contained no organic components, its qualified life was also increased to 26.8 year No deficiencies were noted in this area.

l

12 l

.

..

9. VEND 0R/ UTILITY INTERFACE The inspectors conducted a review of PGE's vendor information evaluation and tracking system with respect to the receipt and processing of technical infomation provided by Westinghouse (W), INPO, the NRC, and vendors of safety related equipment and component Starting in January 1985, PGE instituted a new system for tracking vendor correspondence. All vendor correspondence is received by the Nuclear Plant Engineering Department and entered into a computerized filing system by categories listed in the Trojan file index. These documents can be viewed on microfich In this same timeframe, a computerized Commitment Tracking List (CTL) was started to track items of potential interest or impact on plant operations and maintenance. These items include Westinghouse Technical Bulletins, INPO Significant Event Reports (SERs) and Significant Operating Experience Reports (SOERs), NRC Information Notices and Bulletins, Vendor 10 CFR 21 Reports, and general vendor correspondenc PGE Nuclear Safety and Regulation Procedure NSRP No. 330-2, " Operating Experience Review Program Administration," and PGE Trojan Nuclear Plant Administrative Order-A0-9-4, " Operating Experience Review Program," define the requirements for review, evaluation, and implementation of vendor technical information. This review and evaluation is done through an Operational Assessment Review (0AR). An 0AR number is assigned and tracked on the CTL by either subject or 0AR numbe From 1981 to 1984, PGE had another formal tracking system, the Required Action Tracking System (RATS) which was also a ccmputerized data base maintained by the Trojan Operational Assessment group. This system functioned almost the same as the current system, except it did not track by 0AR number for all vendor correspondenc Prior to 1981, PGE had no fomal system for the receipt, evaluation, and implementation of actions initiated in response to vendor technical information. For example, PGE had virtually no documentation to indicate that a review and evaluation were performed either for Westinghouse Technical Bulletins affecting Trojan which were issued prior to 1981, or for Westinghouse Data Letters issued prior to 1982. PGE has been in the process of reviewing these older Technical Bulletins and Data Letters for applicability to Trojan. PGE stated in Memorandum DRIC-037-85 that documentation and completion of the required actions with respect to Westinghouse Bulletins and Data Letters is to be completed by September 30, 198 Evaluation of Information Contained In Westinghouse Technical Bulletins A sample of W Technical Bulletins was selected to review the licensee's receipt, evaTuation, and implementation of the information in the bulletins. Technical Bulletins are an advisory notice of recent technical developments pertaining to the installation or operation of l

l 13

..

W supplied Nuclear Plant Equipment. Recipients are advised to evaluate the information and recommendations, and to initiate action when appropriate. The following is a brief summary of the Technical Bulletins reviewed by the inspector (1) TB-80-11 - High Head Safety Injection (HHSI) Pump Flow Verification Deficiency - Degradation of Centrifugal Charging Pump flow performance may result in a considerable change in the pump performance curves and exceed the margin allowed for the safety injection syste Licensee Action - Trojan is monitorin on a periodic basis (at least yearly)gandtheis minimum flow rate-continuing HHSI pump performance monitoring (monthly) per the current inservice testing program. Total runout flow is tested each refueling outage and branch-line balancing is done if a total flow degradation is observe (2) TB-83-07 - Containment Pressure and Containment Sump Water Level Systems Deficiency - During the severe environmental phase (320'F)

of qualification testing of the water filled system, extraneous pressure spikes (ringing) was observed on the containment pressure transmitter output. W solved this problem and recomended that any of the subject systems filled with water be drained, dried, and refilled with specially processed DC702 silicone oi Licensee Action - Per the W recommendation, all affected level transmitters were filled wTth DC 702 silicone oi (3) TB-04-06 - Reduced Auxiliary Feedwater Flow Rates Deficiency - A situation occurred where a W plant throttled its auxiliary feedwater (AFW) control valves to the minimum required flow rate (105 GPM) against a steam generator pressure (700 psig)

which was considerably lower than the pressure (1082 psig)

assumed in the plants loss of normal feedwater analysis. A subsequent hydraulic analysis showed that the AFW flow rate at the higher pressure would have been 27% less than the flow obtained at the lower pressure, and less than that assumed in accident analysis.

i

Licensee Action - The PGE evaluation documented in 0AR-84-121 l

concluded that Trojan's variable speed AFW pump system will provide sufficient flow under all anticipated conditions. The

-

_

.s AFW pump speed can be increased to maintain 100 psid across the throttle valve. However, the AFW system performance and valve inservice test (P0T-5-3) was revised to assure that it is always performed at steam generator pressures greater than 1000 psig in order to be conservativ (4) TB-84-08 - Inverter Capacitor Connections Deficiency - During the 7.5 KVA static inverter qualification testing, it was discovered that two out of the three GE ferro-resonant transformers had two fast-on terminals connected to a solder lug of the tuning capacitors. Electrically the inverters will operate properly, but the mechanical connection could separate during a seismic even Licensee Action - All transformer units currently installed at Trojan were checked for fast-on terminal connections to solder lugs. All units had the correct connection (5) TB-84-11 - Potential Grounding of Ferro-Resonant Transformers Deficiency - During initial startup testing at a plant in late construction, three W 7.5 KVA static inverters failed during a four month period. After a fault analysis was performed, the nature of the failure was determined to be shorting between the coil and core of one of two inductors connected to the secondary side of the ferro-resonant transformer. The manufacturer determined that all three units failed due to insufficient securing of the center leg of the inductor cor Licensee Action - All installed transformers have been in operation at Trojan for greater than six months and were not affected by the bulletin. The newer transformers located in the warehouse were given a high potential leakage current test. All test results were satisfactory. This information was documented in 0AR-84-15 b. Centrifugal Charging Pump Speed Increaser Written correspondence from Westinghouse to PGE, dated February 12, 1985, described a potential problem with the lube oil pump coupling on the centrifugal charging pump speed increaser. This written correspondence was neither a Technical Bulletin nor a Data Lette In December 1984, two facilities suffered failures of the couplings and the intent of this letter was to notify PGE of a potential proble After receipt of the letter, PGE performed an inspection of both charging pump speed increaser lube oil pump couplings for excessive pin wear. After consulting with-Nuttall Gear Corp. (the manufacturer)

_

..

it was determined that the pin wear at Trojan was satisfactory for continued operation. Maintenance Procedure M0P-5-2 was revised to include an annual inspection of these lube oil pump coupling However, during this inspection it was discovered that the inboard coupling half key had almost completely backed out of its keywa There was no means of capture on the key way. Undetected failure of the speed increaser would render the centrifugal charging pump inoperable. This is a different problem than that identified by Since this speed increaser is the same unit that is used on the Auxiliary Feedwater Pump, PGE inspected the AFW pump coupling as well. This keyway was also machined too long, but the key was tigh A modification was made to stake the key in place as a precautio This modification was also made on both centrifugal charging pumps to stake the keys in plac As of the time of the inspection, PGE had not completed their evaluation for reporting per 10 CFR 21 nor had they filed a Licensee Event Report (LER). PGE personnel interviewed indicated an intent to not report this defect pursuant to 10 CFR Part 21. This problem was first identified on April 1, 198 On April 4, 1985, W sent out a letter update on the original charging pump speed increaser lube oil pump problem. They stated they had completed an engineering study of the original two plant failure ,

The results showed that the failure was due to an abnormal wear mechanism, which was most likely the result of a combination of coupling misalignment, excessive lube oil pump torque, and insufficient coupling lubrication. The W Safety Review Comittee

_

concluded that this issue is not a safety concer Notwithstanding the W determination that the initial problem was not reportable, PGE's failure to perform a timely evaluation of the second potentially serious common mode failure is considered another example of a weakness of PGE's Part 21 progra . POST MODIFICATION TESTING Approximately one month prior to this inspection, an event occurred at the Trojan plant in which both auxiliary feedwater pumps tripped on low pump suction pressure following a reactor trip. The low pump suction trip was installed as part of a plant modification. The post modification testing did not reveal this deficiency because the tests performed did not include a demonstration of two pump operation with the condensate storage tank at the minimum acceptable water leve As part of this inspection effort, specific post modification tests were reviewed with respect to whether the testing was sufficient to cover all modes of system operation. In addition, a review was performed to detennine if the testing would prove or duplicate the testing specified in the C of Cs supplied by the vendor !

i i _

.-

The test procedure requirements and acceptance criteria are designated in Detailed Construction Packages (DCPs). Once the DCP testing requirements are received on site, they are transcribed onto Maintenance Requests to be performed by the applicable site departmen A small sampling of As Built Packages (ABPs) and DCPs from the most recent plant outage were reviewed. The ABPs are completed modifications which contain a listing of the testing that was performed and a statement as to the acceptability of the results. Approximately one third of the modifications performed during the outage were closed out and filed as ABPs. The remaining modification packages were in the process of being closed out and the current documentation was contained in the DCPs. The following is a summary of the testing contained in one ABP and two DCPs, and general observations concerning the testing methods and acceptabilit RHR System Modification ABP 84-058 was performed to replace valve motor operators with environmentally qualified units. The valves are safety related and receive isolation signals. The functional test requirements were reviewed in detail and appeared to be acceptable. Maintenance Request (MR)85-181 properly summarized the package test require-ments. The Testing / Installation Results Section of MR 85-181 stated that applicable portions of the Periodic Operating Tests were performed and valves timed within the required specifications. No specific test data was provided in the ABP nor were the revision numbers or applicable sections of procedures identifie MSIV Isolation Modification DCP 84-106 was performed to replace the solenoid-actuated pilot valves for the main steam isolation valves and bypass valves. The functional testing requirements and installation checks for this modification were scattered throughout the package. Field Change Notice (FCN) No. 8 to DCP 84-106 provided the final test requirements to be performed for the modification. MR 85-2101 was the only maintenance request in the DCP and it instructed the plant personnel to perform only a portion of the testing delineated in FCN No. 8. Since the ABP for this modification was not available at the time of the inspection, it was not possible to determine if all the required testing had t,een performed. This is considered an Unresolved Item (344-85-025-05) AFW Flow Control Modification DCP 84-104 was performed to replace the auxiliary feedwater system motor operated flow control valves. The DCP included the functional testing and installation checks to be performed and listed the Periodic Operating Tests (P0Ts) that should be run. The P0Ts were reviewed during the inspection and it did not appear that they would satisfy all the testing requirements. Specifically, the hi-flow I

l

, 17 l

l

~ *

l I

'

..

!

condition valve closure function was not addressed. The MR for this DCP stated that these P0Ts and Temporary Plant Test (TPT)-137 should be performed to satisfy the post modification testing requirement The TPT was not available during the inspection and it could not be determined if the TPT would satisfy the hi-flow condition valve closure function. The ABP for this modification was not complete at the time of the inspection so it was not possible to determine what tests had actually been performed. This is considered an Unresolved Item (344-85-025-06). Control of Testing During the inspection, five ABPs and DCPs from the most recent outage were reviewe In all cases it was difficult to determine what tests and installation checks were required and what was actually perfonne It was also unclear that the testing and checks identified in the originating DCPs would accurately test all the system functions or were properly transcribed to the plant Maintenance Request Additionally, the detailed test results in most cases could not be reviewed because they were not included in the ABPs. The licensee indicated that the process was being reviewed and it was expected that any deficiencies identified would be correcte Verification of C of C Data The records in the modification packages were reviewed to determine whether the parts or components procured, accepted, and installed for the modifications were subjected to additional tests at the site to verify the validity of the C of Cs which formed the sole basis for acceptance of the items. The review determined that the testing and installation checks were functional in nature and not intended to verify the validity of the vendcr's certification syste This issue is addressed in more detail in Section 3 of this report and has been identified as a potential violation (344-85-025-01).

11. EXIT MEETING The inspectors met with Mr. P. Yundt and other staff members on August 16, 1985, to summarize the scope and findings of the inspectio