IR 05000293/1993020
| ML20059G101 | |
| Person / Time | |
|---|---|
| Site: | Pilgrim |
| Issue date: | 01/10/1994 |
| From: | Cerne A, David Kern, Macdonald J NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION I) |
| To: | |
| Shared Package | |
| ML20059G098 | List: |
| References | |
| 50-293-93-20, NUDOCS 9401210116 | |
| Download: ML20059G101 (21) | |
Text
.
.
U. S. NUCLEAR REGULATORY COM' MISSION
REGION I
Docket No.: 50-293 Report No.: 93-20 Licensee: Boston Edison Company 800 Boylston Street Boston, Massachusetts 02199 Facility: Pilgrim Nuclear Power Station location: Plymouth, Massachusetts Dates: November 2 - December 6,1993 Inspectors: J. Macdonald, Senior Resident inspector A. Cerne, Resident inspector D. Kern, Resident inspector Approved by: O 9 E. Kelly', Chief 'Dat'e Reactor Projects.Section[3A Scope: Resident safety inspections in the areas of plant operations, maintenance and surveillance, engineering, and plant support. Initiatives selected for inspection included balance of plant feedwater heater leak repairs, HPCI/RCIC reliability and maintenance programs, and a review of an independent assessment of the BECo quality assurance progra Inspections were performed on backshifts during November 3-5, 8-10,16, 24-25 and 29 and December 1-2, and 6. Deep backshift inspections were performed on November 8 (10:00 a.m. -
11:30 p.m.), November 9 (10:00 a.m. - 10:20 a.m.) and December 5-6 (8:30 p.m. - 2:45 a.m.)
Findings: Performance during this five week period is summarized in the Executive Summar PDR ADOCK 05000293 G PDR L .-
'
.
s EXECUTIVE SUMMARY l
l Pilgrim Inspection Report 93-20-Plant Operations: Control room operators promptly identified the degradiag performance of a low pressure feedwater heater, and management developed a standing order that addressed plant and personnel safety in the event of further heater degradation prior to plant shutdown for l
necessary repairs. Operators also promptly identified minor steam leaks in balance of plant systems located in the condenser bay. Potential adverse effects from steam impingement were evaluated while repair plans were developed, and excellent controls were evident during reactor power manipulations to facilitate the repairs However, licensed operators failed to recognize and respond to diverging reactor vesd water level indication during preparations for the ]
November 8,1993 reactor restart which, in part, caused a suberitical reactor protection system j actuation without control rod motion, i
!
Maintenance and Surveillance: Repairs to the feedwater heater and minor balance of plant i steam leaks were well controlled and implemented in accordance with established work plan j A multi-disciplined analysis team was established to evaluate the heater degradation, and effectively utilized past plant experience and heater manufacturer expertise to accurately predict probable failure mechanisms and approximate tube impact. Testing during the plant shutdown !
confirmed anticipated heater performance conditions. A potential weakness was identified in the conduct of high priority, emergent maintenance in that the implementing procedure may not )
l ensure all necessary aspects of post maintenance testing are specified. Routine surveillance testing was conducted in accordance with established procedures and was consistent with master l surveillance test program periodicitie )
Engineering: Corrective actions to install a level instrument on the diesel fire pump fuel oil tank that conforms with applicable National Fire Protection Association codes were adequate to resolve a previously cited NRC violation. However, during subsequent modifications to enhance the level indication, an uncontrolled operator aid was installed as an apparent breakdown in the modification turnover process. A detailed review of event reports and problem reports related !
'
to the high pressure coolant injection system and the reactor core isolation cooling system indicated that the systems are well maintained and Betterment Program initiatives have effectively resolved past deficiencies, resulting in improved availabilities for the system Incorporation of the increased system availabilities as well as decreased system demand failure data into probabilistic risk assessment calculations has resulted in an overall decrease in calculated core damage frequenc Safety Assessment / Quality Verification: An independent semi-annual assessment of the BECo quality assurance program was conducted by a team comprised of industry peers. The assessment reflected a sound performance-based perspective, action item responsibilities were positively established, and the exit meeting was well attended, ii
i
.
I
-
Executive Summary Plant Support: The security program continues to be implemented in accordance with the station Security Plan, appropriate compensatory postings were stationed during periods of inclement weather. Proper radiological postings were also noted, and radiological protection technician presence was evident at the radiological control area access control desk as well as in the process i buildings. Portal and hand held monitors were in typically good working order and within required calibration periodicitie )
!
l
,
i lii
-_ _ .__ _ _ _ _ . ._. . _ _ _ _ . - _ _ - - _ _ _ _ _ _ _ _ - _ _ _ _ _ - _ _ _ _ _ _ _ _ _
.
.
SUMMARY OF FACILITY ACTIVITIES ........................ I PLANT OPERATIONS (71707, 40500, 90701) ..................... 1 Plant Operation s Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I Reactor Vessel Low Water L.cVel Scrata Signal While Shut Down ..... 2 MAINTENANCE AND SURVEILLANCE (61726,62703,71710,90712) .... 4 Repair of Feedwater Heater Leak ......................... 4 Balance of Plant Steam Leak Repairs ....................... 5 Routine Surveillance ................................. 5 Maintenance Program implementation ........... .......... 6 ENGINEERING (37828, 71707, 92700, 92701) . . . . . . . . . . . . . . . . . . . . 10 HPCI and RCIC Turbine Control Valve Position Indication . . ...... 10 i HPCI and RCIC System Problem Report Analysis . . . . . . . . . . . . . . I1- HPCI and RCIC System Licensee Event Report Analysis . . . . . . . . . . I1 HPCI and RCIC Performance Data and IPE Input .............. 13 Review of Previously Identified Violation ................... 14 PLANT SUPPORT (71707) ................................ 15 SAFETY ASSESSMENT / QUALITY VERIFICATION . . . . . . . . . . . . . . . . 15 Combined Utility Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Licensee Event Report Review .......................... 15 NRC MANAGEMENT MEETINGS AND OTHER ACTIVITIES (30702) ... 17
' Routine Meetings .................................. 17 Management Meetings ............. ............. .. 17 Other NRC Activities . . . . . . . . . . . . . .............. ... 17-Note: Procedures from NRC Inspection Manual Chapter 2515, " Operating Reactor Inspection Program" which were used as inspection guidance are parenthetically listed for each applicable report sectio l iv
.. .- . . .
. _ _ _ - - _ _ - - _ _ _ _ _ . - _
._ _ _ _ _ _ _ _
l
.
!
DETAIL I SUMMARY OF FACILITY ACTIVITIES l
At the start of the report period Pilgrim Nuclear Power Station was operating at approximately 100% of rated power. On November 5, the reactor was shut down to facilitate repairs to a leak l in the fourth point feedwater heater (see section 3.1). Other repairs completed following the i shutdown included replacement of the pilot valve assembly to safety-relief valve 203-3A, i
!
replacement of the failed intermediate range "C" nuclear instrumentation detector, replacement of control rod drive accumulator 30-07, and repair of various minor steam leaks. On November 8, during preparations for reactor startup, a reactor vessel low water level full scram signal was experienced (see section 2.2). All control rods were previously fully inserted and the event was of minimal safety significance. Reactor startup was conducted on November 9 and full power l was achieved on November 1 The reactor water cleanup system was isolated for a planned maintenance outage from November 16 to 20. On November 18, reactor power was briefly reduced to 60% for preliminary repair of leakage from two feedwater system flanges located within the main condenser bay. Reactor
,
power was again reduced to 60% on November 21 to perform a thermal backwash of the main ,
'
condenser and complete repairs of the feedwater system Ganges (see section 3.2). The reactor l was returned to full power following completion of planned maintenance on November 21. The reactor was at full power at the close of the report perio .0 PLANT OPERATIONS (71707,40500,90701)
i Plant Operations Review The inspector observed the safe conduct of plant operations (during regular and backshift hours)
in the following areas:
Control Room Fence Line Reactor Building (Protected Area)
Diesel Generator Building Turbine Building l
Switchgear Rooms Screen House l Security Facilities
~
'
Control room instruments were independently observed by NRC inspectors and found to be in correlation amongst channels, properly functioning and in conformance with Technical Speci5 cations. Alarms received in the control room were reviewed and discussed with the operators; operators were found cognizant of control board and plant conditions. Control room and shift manning were in accordance with Technical Speci6 cation requirements. Posting and control of radiation, contamination, and high radiation areas were appropriate. Workers complied with radiation work permits and appropriately used required personnel monitoring i
device ,
-- . .
- ,
2 Plant housekeeping, including the control of Hammable and other hazardous materials, was observed to be acceptable. During plant tours, logs and ecords were reviewed to cosore compliance with station procedures, to determine if entries were correctly made, and to verify correct communication of equipment status. These records were found to be acceptable, and included various operating logs, turnover sheets, tagout, and lifted lead and jumper log ; Reactor Vessel Low Water Level Scram Signal While Shut Down On November 8 at 12:06 p.m. . while nearing completion of a five day maintenance outage, the plant received a reactor vessel (RV) low water level full scram signal. The reactor coolant system was at approximately 230 degrees F and 20 psig and slowly heating up in preparation for reactor startup at the time of the scram. All control rods were previously fully inserted, therefore no rod motion occurred as a result of this event. The reactor building isolation system i (RBIS) and group 2 and 6 primary containment isolation systems (PCIS) actuated as designed with the exception of one RBIS damper (A-N-70) which failed to fully close. A priority one maintenance request was promptly initiated to repair the RBIS damper. The cause of the scram ,
signal was actual RV water level lowering to the protective setpoint of +12 inches (corresponds !
to 140 inches above the top of active fuel zone). Operators verified actual RV water level (+ 12 !
inches), raised RV level to approximately +50 inches, and reestablished shutdown cooling j pending event revie I Reactor operators have five separate redundant and diverse RV level indicators available at the C-905 panel in the control room. Four of the indicators (two connected to the feedwater ,
condensing pots and two connected to the emergency core cooling system (ECCS) pots) provide j instantaneous level indication, while one indicator (connected to the feedwater condensing pot) l records level permanently on a chart recorder which provides both' instantaneous and trend ,
information. The five separate level indicators should indicate similar RV level values. During l
'
plant heatup operators monitor these level indicators while adjusting reactor water cleanup (RWCU) system reject flow to control RV level. The ECCS RV level instruments provide protective action initiation signals to several safety systems while the feedwater RV level instruments provide input to the feedwater control syste The inspector reviewed computer recorded (EPIC) data following the event and observed the following. Shortly after the RV head vents were shut in preparation for reactor startup, the three RV level indicators connected to the feedwater condensing pots began to slowly increase as RV pressure rose. Reactor vessel water level appeared to be increasing due to control rod drive water input and slowly rising RV temperature. The two RV levelindicators connected to the ECCS condensing pots remained constant. Operators increased RWCU reject flow six times to maintain RV level within the desired band (+30 to +35 inches) during the forty minutes immediately preceding the scram signal. The feedwater RV level instruments indicated l approximately 20 inches higher than the ECCS RV level instruments when the scram signal occurred. The RV level instrument divergence developed slowly, but was clearly evident when l reviewing the EPIC printout !
J
, _ . _ _ . -
-,. _ _ ,
_ . _ _ - _____ - -_ _ _ _
.
The licensee conducted an event critique and initiated problem report (PR) 93.9462 to deter mine the cause of the low RV level scram signal. Personnel error was identified as the root cause in that the reactor operator did not effectively monitor all available RV water level indication during pre-startup evolutions. Response to other control room annunciators during pre-startup activities distracted the operator from periodic coiNparison of RV level indications. The operator did not realize the time lapse between level monitoring and focused more closely on the feedwater RV level chart recorder than on the other RV level indicators. At the time of the scram signal, the operator did not recognize the level divergence. Following review of event data, operator logs, and personnel interviews, the critique team concluded that control room ,
staffmg was appropriate and the number of pre-startup activities was not excessive. However, operator performance was less than adequat The critique also identified the presence of non-condensible gases, generated at power and remaining within the reference legs of the feedwater RV level instruments at the time of the previous reactor shutdown, as a contributing cause to the event. The reactor had been near full power for 58 days prior to the November 5,1993 shutdown. The non-condensible gases expanded during RV depressurizat;on causing water to be displaced from the instrument reference leg. Any non-conde.iole gases remaining in the instrument reference leg piping during subsequent RV pressurization are compressed which in turn reduces the height (and weight) of the vertical column of water contained within the reference leg. During RV pressurization from cold shutdown conditions, the limited amount of boiling in the RV is not sufficient to refill the condensate pot and instrument reference leg. These combined effects directly resulted in false high feedwater RV level indication during RV pressurization prior to !
reactor startu Operators had backfilled RV level instrumentation lines prior to several previous reactor startups based upon operating experience. However, this action was not specified by procedure 2.1.1, ;
"Startup from Shutdown," and was not performed prior to the reactor startup on November 8, l
!
1993. Operators did not observe RV level spiking during the November 5 reactor shutdown and did not, therefore, expect non-condensible gases to be present in the instrument reference leg piping. Subsequent licensee review of EPIC printouts indicated only minor (1 inch) feedwater icve1 spikes during the shutdown. An ECCS RV level instrument reference leg backfill modification (PDC 93-24) had been installed in July 1993 as documented in NRC Inspection Report 50 3 3/93-14 to prevent the buildup of non-condensible gases. That modification affected only he ECCS RV level instruments. The problem report corrective actions included con 4 Won of feedwater reference leg backfill prior to each reactor startu Immediate corrective actions included a calibratior check of the ECCS RV Icvelinstruments to verify their accuracy, backfill of the feedwate; RV level instrument reference leg piping to remove non-condensible gases and restore water volume, and counseling of the operating cre l Longer term actions include revision to procedure 2.1.1 to verify feedwater RV level instrument ]
reference legs full prior to closing the RV head vents, engineering evaluation to determine ;
feasibility and benefit of installing a hardware modification similar to PDC 93-24 for the feedwater RV level instrument reference legs, and training on this event to all operating crew !
l
.
The mspector attended the licensee critique and determined all pertinent event information was presented and appropriately evaluated. However, independent inspector causal analysis of the event concluded primary root cause was the indicated vessel water level error on the feedwater level instrumentation, with the failure mechanism being the piesence of noncondensible gases in the associated reference legs. 'lhe inspector also concluded that operator error contributed to the reactor protection system actuation in that the diverging level indications were not identilied and evaluated prior to receipt of the scram signal. Licensee corrective actions addressed both reference leg backfill and operator performance concerns adequately. Overall, the licensee demonstrated good self assessment, with active senior management and onsite review committee presence evident. The inspector had no further questio .0 MAINTENANCE AND SURVEILLANCE (61726,62703,71710,90712) Repair of Feedwater IIcater Irak On November 5,1993, a reactor shutdown occurred to inspect and repair a suspected feedwater heater leak in the "A" string fourth-point heater located internal to the main condense Symptoms of this leak were first observed earlier in the week when plant operators noted a significant mismatch in condensate /feedwater flow, along with a drop in reactor feedwater pump header suction pressure. Licensee analysis of this problem suggested that the fourth-point heater in the "A" train low pressure feedwater heater string had developed tube leaks, estimated to be less than ten of the 1,012 tubes in the heat exchange The licensee convened a multi-disciplined analysis team (MDAT) to evaluate this problem and assess the available corrective action and repair options. The inspector attended MDAT meetings and discussed assumptions, historical evidence, vendor recommendations and equipment limitations with the cognizant BECo engineering and operations personnel. The MDAT adequately assessed both the immediate concerns and short term operating constraints, along with longer term plans for repair and recover The licensee conducted additional heater drain analysis / testing during cooldown which confirmed the leak was in the fourth-point heater. After the heater was drained and opened, eddy current testing and tube plugging activities were commenced. A total of 44 tubes were plugged, some based upon wall loss and location, along with the identified leakers. The heater manufacturer had indicated that up to 100 tubes could be plugged without adversely affecting heater performance. The cluster of tube failures and resultant tube plugs was consistent with the theorized failure mechanism, i.e., a cascading impingement effect of a single tube failure upon the surrounding tube The inspector reviewed the problem report (PR 93.9450) documenting this feedwater heater leak and noted that a thorough review of past maintenance records was included in the root cause analysis. The inspector also reviewed the Standing Order (93-10) issued to the operators to
.
.
!
address the degraded feedwater conditions, before the plant shutdown commenced, and attended on-shift training / discussion of contingencies needed to be considered by the operations staff j should additional turbine and/or thermal limit problems develo I
'
Plant startup commenced on November 9 after completion of the feedwat:r heater repairs, and
- in conjunction with other maintenance conducted during this four day outage. The inspector !
evaluated the planning, preparation and corrective action relative to the identined heater tube leaks and determined that the licensee's response to this problem was thorough and effective and demonstrated a good safety perspective relative to personnel protection and plant component and operational concern .2 Balance of Plant Steam Leak Repairs The inspector observed the repair of two blank flanges from the "A" and "B" trains of the number two feedwater heaters. The repair technique was an injection of a sealant material into the flange seating area, bounded by a mesh wire outer barrier, to stop the steam leakage. Pre-work radiological protection, heat stress, and maintenance briefings were excellent. The inspector reviewed work instruction packages including a determination of the maximum volume of scalant material permitted for use in this repair application. Work packages were complete and the inspector had no technical concerns. Initial repair efforts were not fully successful. On November 20 reactor power was again reduced, fabricated clamps were installed, and sealant was injected within the clamp ring to stop the leak. Repairs were successful in stopping the <
steam leak and the reactor was returned to full power. First line maintenance supervisors and radiological protection technicians closely monitored all wor .3 Routine Surveillance The inspector observed portions of surveillance to verify proper calibration of test instrumentation, use of approved procedures, performance of work by qualified personnel, conformance to limiting conditions for operation (LCO), and correct system restoration following testing. The following activities (with comments as noted) were observed:
- Procedure 8.B.1, " Fire Pump Test," was performed with satisfactory results on
-
November 12, 1993. This procedure is performed weekly to demonstrate that the fire water delivery system provides a minimum of 2000 gallons per minute flow at 125 psig header pressure.- During verification of system prerequisites to run the diesel driven fire pump, operators questioned the fuel oil day tank level indicatio The units of measurement on the level gauge differed from those called for in procedure 8.B.1 (see Section 4.5 for further discussion). This discrepancy was properly resolve * Procedures 8.5.5.1, "RCIC pump Operability Flow Rate and Valve Test at Approximately 1000 psig" and 8.5.5.4, "RCIC Motor Operated Valve (MOV)
Operability Test Monthly / Quarterly" were performed and satisfactorily completed on November 29, 1993. System engineers closely monitored the performance of these
,
_ _
-
,
.
surveillances, in part, to observe turbine governor control valve (HYD-1301-159)
position indication response which was the subject of a previous problem report (PR 93.9437) which remains active for resolution. Operatorr. do not directly operate valve HYD-1301-159, but the valve repositions to control steam flow to the turbine. After completing the turbine run, operators returned the RCIC system to the normal standby lineup. The Nuclear Watch Engineer (NWE) observed that control room indication showed valve HYD-1301-159 to be closed (green light indication only) instead of open (dual light or red light only indication). The final position of this valve is not specified in procedures 8.5.5.1 or 8.5.5.4. However, procedure 2.2.22, "RCIC" requires valve '
HYD-1301-159 to be open when the system is in standby. The RCIC turbine may be unabic to start if the valve is actually in the fully closed position when the system receives an initiation signal. The NWE and system engineers locally verified that the actual position of the valve was approximately two-thirds open. . System engineers reviewed applicable system technical documents and confirmed that the correct standby valve configuration is partially open in a mid stroke position. The NWE verified that PR 93.9437 addressed the discrepant remote (control room) valve position indicator (VPI), and annotated the surveillance documents accordingly prior to returning the RCIC system to service. Actions taken by the NWE to resolve the VPI discrepancy and verify the correct system standby lineup prior to completing the surveillance were appropriat .4 Maintenance Program Implementation The inspector reviewed several aspects of the maintenance program as they pertain to the availability, reliability, and performance validation of the high pressure coolant injection (HPCI)
and reactor core isolation cooling (RCIC) systems. The master surveillance tracking program (MSTP) was reviewed to determine whether technical specification required periodic surveillances were identified, scheduled, and implemente The Master Surveillance Trackine Program The inspector reviewed station Techrdcal Specifications (TS) and the Final Safety Analysis Report and concluded HPCI/RCIC system surveillances were properly scheduled and tracked in MSTP. All required .surveillances were curren The Inservies Testing (IST) Program The inspector reviewed IST program instructions, HPC1/RCIC surveillance data for the past two years, and discussed component performance analysis with IST engineers. Procedures 8.I.3,
" Control of IST Pump Reference Values" and 8.I.l.1, " Inservice Pump and Valve Testing Program" provided well defined instruction for the establishment of component reference, alert,-
and required action performance levels. Required action levels established for HPCI/RCIC valve operation were well within the limits specified in sections 5.2-4 and 7.3-1 of the Pilgrim Final Safety Analysis Report (FSAR). Engineers closely monitored and evaluated pump flow and i
> = . . - - . _
.
vibration, and valve operation timing. The inspector noted that each time a component reached an alert or required action parformance level, the licensee effectively communicated test results throughout the organhation and initiated approprian corrective action ,
During review of the MSTP the inspector questioned the periodicity of IST surveillance procedure 8.I.27, "IST Check Valve (CV) Sample Disassembly and Exercise Program" for CV 2301-39, located in the HPCI torus suction line. Proecdure 8.I.27 established test groups of similar CVs within several safety related systems. Check valve 2301-39 was the only CV in its group and had initially been assigned a test interval of once per refueling outage (RFO). The IST program engineer stated that based upon the positive results of two consecutive inspections, the disassembly and inspection interval for CV 2301-39 was revised to once every other RF The inspector reviewed the inspection results with the IST engineer and concluded that the surveillance interval had been appropriately assessed and revise The Preventive Maintenance Procram All but three of the HPCl/RCIC preventive maintenance actions (PMs) are currently up to dat The three overdue PMs were deferred in accordance with Administrative Procedure 3.M.1- The inspector noted each of the three PMs, originally scheduled for RFO 9 in April 1993, were deferred following the completion of the outage. Each deferral received the appropriate approval signatures, including that of the system engineer for technical assessment. But the reason for request (or technical justification) was not consistently documented. Previously, an unresolved item was issued in NRC Inspection Report 50-293/93-13-01 as a result of similar inconsistency in the deferral of outage PM The inspector questioned the scheduled due date of several PMs which appeared incorrect based upon the periodicity and documentation of the last completion date contained within the PM data base. The licensee acknowledged the apparent conflict and explained that it was the result of ;
the limitations of the computer software which maintains the PM data base. However, in each case, PM schedulers had compensated for this limitation by manually recalculating and entering the corrected due date. Subsequent inspector review indicated appropriate due dates had been established, however this process was labor intensive and required close scrutiny by PM program managers to avoid potential errors. The inspector noted that the responsibility for PM scheduling will be transferred to the MSTP coordinator effective December 10, 199 Discussions with the MSTP coordinator indicated the individual was aware of the existing PM :
software constraint and that this would be addressed during an upcoming conversion of the PM -l scheduling data base, j The Reliability Centered Maintenance (RCM) Pwgam l
_
The HPCI/RCIC system reviews for development of an RCM Program have been completed and revised component preventive maintenance (PM) recommendations have been submitted. The majority of the PM recommendations are pending fm' al review for implementation. Itis, therefore, too early to assess the effectiveness of RCM program recommendation !
l
!
- . , . . _ _ _
'
.
,
Maintenance Reauest (MR) Backlog The nspector reviewed the MR backleg for the HPC1/RCIC systerre. None of the existing l material conditions for which the MRs were written posed an operability or reliability concer Most of the MRs were written to implement scheduled overhauls and material upgrades to system metor operated valves. Inspector review of MRs performed during the past two years ;
indicated that with the exception of scat leakage on selected steam supply valves, corrective maintenance was effective without problem recurrence. Final correction of repeated seat leakage "
on steam supply valves is being addressed by a planned redesign and replacement of those valves during RFO 10 in 1995. The inspector concluded that the MR backlog is properly manage Work Planning ;
'I The inspector reviewed the work planning process with speci0c emphasis on post maintenance test (PMT) determination. Procedure 1.5.3.5 " Maintenance Request Planning," provides ;
guidance for work package planning, including reference to satellite Procedures 1.13.1, " Post Work Test Matrices and Guidelines," 3.M.1-30, " Post Work Testing Guidance," and 8.I.1.1 for detailed guidance for PMT identification. Maintenance work planners and work coordinators ;
receive training on procedure 1.5.3.5 during the M-1 module of Maintenance Section general l l
indoctrination training. Work planners and work coordinators demonstrated detailed knowledge of procedure 1.5.3.5 during interviews with the inspecto There is one alternate work planning process. Procedure 3.M.1-34 " Generic Troubleshooting and Maintenance Procedure," may be used in place of 1.5.3.5 for emergent priority one or emergency corrective maintenance with the permission of the Chief Operating Engineer or the Operations Section Manager. The inspector observed that procedure 3.M.1-34 does not discuss identincation or inclusion of PMT requirements. Work coordinators stated that although PMT identification is not speci0cally discussed in 3.M.1-34, the expectation and general work practice is to specify PMT requirements and include them at the end of the 3.M.1-34 work plan. The inspector noted that procedure 3.M.1-34 is not covered during the M-1 training module and, based on interviews with maintenance personnel, concluded that training on 3.M.1-34 is not formalized. The inspector expressed concern that additional actions may be necessary to ensure the controls which establish post maintenance testing criteria in accordance with procedure 1.5.3.5 are maintained for maintenance performed in accordance with procedure 3.M.1-34, which by its very nature would be emergent and of high priorit A specific example of the inspector's concern was evidenced in the October 25,1993, RCIC turbine governor control valve remote servo hydraulic relay replacement, that was completed using the 3.M.1-34 work planning process (Priority 1, MR 19303408). Use of this process minimized the amount of time that the RCIC system was inoperable. However, the work plan did not include an American Society of Mechanical Engineers (ASME)Section XI code required position indicator verification (PlV) test. Following completion of the component replacement specified in the work plan, the work package was returned to the Nuclear Operations Supervisor (NOS) for identification of Technical Specification PMT requirements prior to returning the
l c
.
system to service. The NOS appropriately speciGed portions of procedure 8.5.5.1, but did not recognize the need to perform the IST program PIV test. In this case, neither the maintemtnce work plan process nor the NOS review for PMT veri 5 cation identified the PIV test. Inthis specific instance, not discretely identifying the PlV test as a PMT requirement did not affect system operability. This was because the actual direction of travel and the as-left standby position of the valve were locally verified to be correct during the system operability run. The PIV was performed as a portion of the November 29,1993 routine RCIC surveillance testin The inspector discussed this concern with maintenance, operations, and engineering section personnel. The licensee initiated problem report (PR) 93.9499 to clarify IST program test requirements and to improve the determination of PMT requirements for emergent work. The inspector discussed the intended corrective actions, including procedure revisions and training, with the responsible licensee managers and concluded that the PR was appropriat Vendor Ma.lahn The inspector reviewed eleven vendor manuals associated with the HPCI/RCIC systems. Each manual was properly maintained and updated. The inspector concluded that the vendor manual'
upgrade program was effective in maintaining current vendor information available on-site for l component evaluation and maintenance work package planning activitie !
Motor Operated Valve (MOV) Program Undate NRC Inspection Report 50-293/93-19 previously addressed an effective MOV design basis review by licensee engineers that identified an industry practice which could result in misapplication of run versus pullout efficiency factors when calculating MOV available thrus Consideration of this factor created sufGcient uncertainty in the calculated performance of the HPCI outboard steam supply isolation valve (MO-2301-5) to cause the licensee to declare the valve inoperable to rebuild its actuator. Further, NRC Generic Letter (GL) 89-10 MOV program design review has resulted in proposed improvements to four RCIC and seven HPCI MOVs. The MOV improvements are intended to upgrade valve thrust performance to fully account for current industry ' uncertainties such as valve factor, thermal degradation, motor performance uncertainties, rate of loading, spring pack relaxation, torque switch repeatability, and test measurement accuracies. BECo MOV program engineers have worked closely with the HPCI/RCIC Betterment Team to effectively integrate proposed enhancements with recommended l corrective actions to existing problems, resulting in the planning and scheduling of eleven HPCI/RCIC MOV improvements by RFO 10 in 199 I The most risk signincant components of the Pilgrim HPCI/RCIC systems are the speed control circuit and steam supply MOVs. The inspector reviewed MOV design calculations and planned GL 89-10 program upgrades. Several minor discrepancies in design torque calculations were j previously discussed in NRC Inspection Report 50-293/93-19. Although none of these specific 1 errors resulted in MOV inoperability, they were characteristic of incomplete control of design )
documents. The licensee initiated a problem report (PR 93.9479) to address this concern. The i problem report identified additional design calculation and drawing errors. Each discrepancy l I
j i
l
.
-
L l
was promptly evaluated and had no impact on MOV operability. Recommended corrective action includes a broad review of design documents by licensee engineers, which is expected to last several months. The inspector discussed the matter with the quality assurance (QA)
department manager who stated that a series of QA surveillances on MOV design document controls was also planned. The inspector expressed concern that a stronger emphasis be placed upon maintaining the integrity of MOV design document controls, considering the significant amount of activity associated with the MOV design basis review and associated GL 89-10-program upgrade .0 ENG INEERING (37828, 71707, 92700, 92701) IIPCI and RCIC Turbine Control Valve Position Indication The inspector reviewed Plant Design Change (PDC) 91-55 and all field revision notices. This modification addressed a problem identified during the Detailed Control Room Design Review at PNPS to verify compliance with the applicable provisions of NUREG-0737. Specifically, the limit switches for the HPCI turbine stop valve and both the HPCI and RCIC turbine governor control valves were reconfigured to provide intermediate position indication of valve travel. The inspector verified that the modiGcation activities were classified as safety-related work, even though both system governor control valve limit switches perform indication only, (i.e., a nonsafety function).
As discussed in NRC Inspection Report 50-293/93-19, the inspector had reviewed three problem )
reports related to RCIC system inoperability. One of these problem reports (PR 93.9437)
concerns position indication for the RCIC turbine control valve which, as indicated above, had been modified by the implementation of PDC 91-55. Further followup of this problem report, as it relates to the quarterly surveillance and inservice testing requirements for remote position indication of the RCIC turbine control valve (HYD-1301-159), is discussed in section 3.4 of this inspection report. However, since the valve position indication test is a post maintenance -
activity required by PNPS procedure 8.5.5.1, the inspector also checked how this requirement was fulGiled during the post modification testing of PDC 91-5 The inspector reviewed PNPS Temporary Procedure (TP 92-013), which established the post ,
work test conditions and acceptance criteria for the valves modified in accordance with PDC 91-55. The inspector confirmed that functional testing was conducted to verify satisfactory-operation of the position indicating lights for the affected HPCI and RCIC turbine valves. The inspector also examined the governor linkage on the RCIC turbine and checked the position indication lights for the HPCI and RCIC turbine valves on panels C903 and C904 in the control room. The BECo records reviewed by the inspector provided documented evidence of adequate preoperational testing of the HPCI/RCIC components modified by PDC 91-55. The inspector had no further questions regarding implementation, control and testing of this design chang .
. . . .
.
.
l 11 IIPCI and RCIC System Problem Report Analysis j
The inspector requested a list of all problem reports generated in 1992 and 1993 that discuss !
HPCI and RCIC system issues. The inspector then reviewed approximately tu percent of these ,
problem reports to determine if the issues were properly handled, if reportability was correctly - ;
considered, and if some trend of system performance was recognizable from the problem report j statistic For one problem report (PR 92.9049) involving set point drift on the HPCI turbine steam supply j low pressure isolation switches, the inspector noted an error in the 10 CFR 50.73 Evaluation :l Form relative to consideration of the static head of pressure inherent in the elevation of the i isolation switch. A new problem report (PR 93.9498) was initiated to document this error, but ,
'
discussion with cognizant licensee personnel indicated that the basis for the reportability-determination was not affected by the corrected data. Furthermore, the affected pressure isolation switches were subsequently removed from service by a plant design change (PDC 91-75) which added reactor low pressure signals to the HPCI Group 4 isolation relay logic. The inspector reviewed PDC 91-75, the elementary (relay logic) diagram for the HPCI system, and Revision 167 to the PNPS Technical Specifications to verify consistency in certain controlled documents affected by this design change, as implemented during RFO No. 9 in April 199 ,
inspector review of the other sampled problem reports identified no safety questions or concerns ;
regarding HPCI or RCIC system performance. However, two pair of apparently related problem reports were noted relative to motor operated valve (MOV) engineering evaluations. Although the MOVs in question are part of the RCIC system, the programmatic issue reflected in these problem reports is not a RCIC-specific concern, but rather one involving design margin and NRC Generic Ixtter 89-10 guidance. Programmatic inspection of this area is documented in NRC Inspection Report 50-293/93-19, along with NRC specialist rt /iew of BECO's MOV testing and surveillance program, to be documented in Inspection Report 50-293/93-2 j l
The licensee has issued a problem report (PR 93.9479) to document various discrepancies identified relative to MOV calculations, drawings and vendor data sheets. Cont NRC i resident inspection of MOV issues at PNPS is documented in section 3.4 (Page S; m this l inspection report. However, other than the MOV issues, no generic safety concerns were l identified in the HPCI/RCIC performance data and trends which were gleaned from the problem i report review. The inspector concluded that the licensee corrective action to the identified !
problems was technically adequate, timely and commensurate with the significance of the I reported problem .3 IIPCI and RCIC System Licensee Event Report Analysis The inspector reviewed all Licensee Event Reports (LERs) issued from January 1990 to the present (December 6,1993), which discuss HPCI and RCIC system problems or conditions of inoperability. The start date for this event analysis was selected to coincide with the closure of Confirmatory Action Letter 86-10 and its supplements and to complement the analysis of the
!
I
. - .
- . - . . .
.
'
HPCI system LERs submitted between 1980 and 1989 which was documented in the NRC High Pressure Coolant Injection System Risk-Based Inspection Guide (NUREG/CR-5924) published in October 1992. Thus, the current review not only examined the LERs for regulatory compliance to the criteria of 10 CFR 50.73, but also evaluated the effectiveness of licensee corrective measures to preclude repetition of similarly caused event A total of 24 LERs were evaluated and HPCI/RCIC system inoperability events were categorized relative to both the cause of the degraded condition and the effect upon system availability to perform its safety function. This analysis revealed an improving trend in HPCI and RCIC system availability and no correlation of recent reportable events to past problems with turbine speed and control fault It appears that licensee corrective actions in the 1989-1990 time frame have been effective in addressing the most risk-significant system failure modes. As documented in NRC Inspection-Reports 50-293/90-20 and 90-22, along with the closure of unresolved item 90-20-01 t documented in inspection Report 50-293/92-08, NRC followup of sequential RCIC turbine trips and HPCI flow oscillations verified that the licensee's Multi-Disciplined Analysis Team (MDAT), formed to investigate the September 1990 feedwater control system transient, appropriately assessed the root causes and directed the proper remedial measures. The inspector confirmed that with respect to the HPCI system, where NUREG/CR-5924 identified a high risk importance associated with turbine speed control problems, the licensee's MDAT review and resultant corrective actions have proven effective in preventing recurrence of the 1990 event anomalie Overall, inspector analysis of the 24 LERs issued from 1990 to the present time has revealed little correlation with pre-1990 HPCI/RCIC system problems, except as noted above regarding turbine controls. Several recent LERs relate to surveillance issues, setpoints, or engineering questions which would not have rendered the affected HPCI/RCIC system incapable of performing its safety function. The LERs were written with sufficient detail to cover event cause, consequences and corrective action, while considering the appropriate regulatory and safety criteria. While system unavailability was adequately documented, the conditional nature of inoperability and its relation to system unavailability was not fully explained in each LER write-up. The inspector notes that a full understanding of each reported event is required to assess how declared system inoperability reflects the overall system performance, in summary, no adverse trends in the performance of either the HPCI or RCIC systems were identified. The inspector indicated to the licensee that a clearer description of the basis for inoperability would assist the NRC's understanding of the relevant importance of any reported event. However, the inspector noted that current LER's have been found to be in compliance with the requirements of 10 CFR 50.73. The inspector concluded that no unresolved safety issues or concerns were identified as a result of this revie !
. .
.
i
'
13 IIPCI and RCIC Performance Data and IPE Input i
In accordance with USNRC Generic Letter 88-20, BECo submitted an Individual Plant Examination (IPE), addressing the severe accident vulnerabilities at PNPS, to the NRC on September 30,1992. The IPE and its results were founded in a series of risk-based studies for the PNPS, culminating in an evaluation report which BECo described in its submittal to the NRC as a full scope Level 1 and Level 2 probabilistic risk assessment (PRA). The Pilgrim PRA uses a plant-specific data base (e.g., component failure rates) that selected an end date of September -
30, 1989 for analysis. However, since PNPS was shutdown for most of the period from 1986 l to 1988 to implement a long-term improvement program, any improvements reflected in the l system performance and component data since plant restart in 1989 would not be considered in l the static PRA data base.
l However, with respect to the HPCI and RCIC systems, the licensee elected to evaluate, in a pilot application, the importance of maintaining a "living" (dynamic) data base. This was accomplished through the use of a five year moving average of data input into the PRA methodology, with this moving average kept current to the latest completed calendar quarte This pilot program has demonstrated from 1989 to the present time a signincant improvement l in HPCI performance and slight improvement in RCIC performance.
l The licensee developed an integrated action plan for HPCI/RCIC system improvement and has been tracking original action item completion and adding new items as they are developed. This I l approach has resulted in over a 25% decrease in the calculated core damage frequency contribution of the combined HPCI/RCIC high-pressure injection systems, considering the latest two year moving performance average versus HPCI/RCIC data included in the IPE submitta Using a five year moving performance average results similarly in approximately a 17% 1 reduction in the combined core damage frequency contribution of these systems using PRA l techniques. This represents not only a significant improvement in the pre-1989 performance ,
trend, but also provides evidence of a continuing positive trend in risk reduction relative to the i HPC1/RCIC systems since plant restart in 198 The inspector noted that the above PRA data were consistent with the HPC1/RCIC system LER analysis discussed in section 4.3 of this inspection report. The inspector interviewed BECo Nuclear Engineering Department (NED) personnel involved with the BECo IPE submittal and independently cross-checked the HPCI/RCIC LERs against the system unavailability data used by the engineers in maintaining the HPCl/RCIC pilot PRA program updates. The LER and IPE i unavailability data were generally well correlated and in agreement, however, the inspector I
identined two instances in which HPCI system unavailability had not been properly accounted for. The addition of this data had only minor impact on the core damage frequency calculations ;
~
for the two year moving performance average ending November 30, 1993, which decreased l somewhat when compared to the results at the end of October 1993.
l l
l
, - l
._ _ .- .
..
The inspector concluded that the pilot program fcr tracking HPCI/RCIC performance, using PRA methodology, has proven effective in illustrating the results of the licensee integrated action plan for system improvement. The inspector did note, however, that the value of such risk analysis is naturally lin:ited to the sophistication and fidelity of the data utilized. Therefore, as evidenced by the HPCI unavailability data that had not been considered by recent calculations, continual and effective dialogue between the system engineers and the NED personnel working with the IPE program is necessary to ensure the maximum benefit of such data analyses in providing proper future program and action plan directio .5 Review of Previously Identified Violation (Closed) Notice of Violation (NOV) S0-293/92 27-01.2, Inadequate Corrective Action - Diesel Fire Pump Fuel Oil Day Tank Level Instrument This violation pertained to the diesel fire pump fuel oil day tank level instrument which did not conform the National Fire Protection (NFPA) code standards. This was a long standing discrepancy, that was identified by several licensee fire protection program audits, but had not been resolved in a timely manner. Specifically, the existing tank mounted gauge glass was not j an acceptable level indication as defined by NFPA code. As an interim measure, the licensee had valved out the gauge glass and maintained the tank level above the control room high tank level annunciator alarm to ensure suf6cient fuel oil volume existed. Subsequently, plant design change (PDC) 92-23 was implemented to install a NFPA code conforming level instrument. The tank mounted site glass was removed, connections to the day tank were capped, and a new level ,
gauge which satisfied NFPA code was installed. The inspector reviewed PDC 92-23 and the l associated completed work package (MR 19203736) and determined that this modification had been properly implemented. This NOV is therefore close Subsequent to PDC 92-23, several additional design changes were implemented to further improve the human factors readability of the day tank level instrument. The inspector observed the level gauge on November 19, 1993, during performance of a routine surveillance. The inspector noted that an apparently uncontrolled operator aid was in the place of the permanent level gauge face plate. Although the aid, which was a piece of annotated adhesive tape, was !
ledgeable, it was not controlled or authorized as required by procedure 1.3.34. " Conduct of )
Operations." During the surveillance witnessed by the inspector, this discrepant condition was properly reported and resolved by operations personnel. Inspector review determined that the uncontrolled operator aid had been installed during implementation of PDC field revision number j (FRN) 92-23-04 as a temporary measure while a permanent face plate was being manufacture However, the discrepant level gauge had not been noted during the two weekly diesel fire pump '
surveillances which had been performed since installation of PDC FRN 9~2-23-04. The inspector expressed concern that the presence and apparent use of the uncontrolled operator aid indicated a breakdown in the implementation and turnover of PDCs from engineering to operation _ _ _
.
15 PLANT SUPPORT (71707)
l The inspector reviewed security program performance during routine plant tours. Security force i members were observed to be alert and aware of posting requirement Appropriate compensatory postings were observed to be in place during heavy weather conditions on the l
cvening of December 5,1993 through the following morning when the weather subside Positive radiological controls were also observed during routine plant tours. Radiological protection technician presence provided positive controls at the radiological control area access point. Technician presence was also noted within the process buildings. Proper postings were noted and survey maps reviewed were observed to be current. Portal and hand held monitors were observed to be in good condition and within proper calibration periodicitie .0 SAFETY ASSESSMENT / QUALITY VERIFICATION Combined Utility Assessment I
The inspector attended the exit meeting on November 5,1993 of the team which performed a combined utility assessment of selected portions of the BECo Quality Assurance (QA) Progra '
The audit team, consisting of personnel from other nuclear utilities, reviewed procurement, QA audits of training and qualification, and internal audit processes, as well as corrective action to ;
i previous combined utility assessments. The inspector also reviewed audit report 93-13, l
'
documenting the results of this Combined Utility Assessment. The QA Department has assigned responsibility for each audit recommendation and appears to be tracking these findings to appropriate resolution. The inspector considers such external audits to represent evidence of compliance with the technical specifications relative to the independent review and audit function at PNPS, and to demonstrate a working self-assessment proces .2 Licensee Event Report Review The inspectors reviewed Licensee Event Reports (LERs) submitted to the NRC to verify accuracy, description of cause, previous similar occurrences, and effectiveness of corrective action The inspectors considered the need for further information, possible generic implications, and whether the events warranted further onsite followup. The LERs were also reviewed with respect to the requirements of 10 CFR 50.73 and the guidance provided in NUREG 1022," Licensee Event Report System" and its associated supplement * LER 93-20 LER 93-20, " Anticipated Transient Without Scram (ATWS) Division 1 Pressure Transmitters Found Valved Out of Service," dated August 23, 1993, describes an inadvertent valve lineup crror which resulted in one train of ATWS being inoperable following reactor startup on July 26. The root cause of the event was procedural weakness. Procedures did not uniformly identify instrument isolation valves and did not sufficiently verify ATWS instrument operability-l
!
I
. - _
_
.
.
prior to placing the reactor mode select switch in the RUN position. This event was of minimal safety significance, and corrective action was promptly completed as further documented in NRC Inspection Report 50-293/93-14. The LER correctly addressed all reporting criteria, o LER 93-21 LER 93-21, " Automatic Closing of isolation Valves during Reactor Core Isolation Cooling (RCIC) System Pressure Control Operations," dated September 22, 1993, describes the unanticipated RCIC high steam flow isolation on August 24,1993. The RCIC system was being operated in the pressure control mode with the reactor being maintained at less than one percent power and all four main steam lines isolated for corrective maintenance to an outboard main steam isolation valve. While controlling reactor pressure, the RCIC pump was discharging about 375 gallons per minute at a discharge pressure of 1350 psi to the condensate storage tank via the full flow test line. Reactor vessel pressure was 487 psig when a RCIC turbine steam supply line high Cow signal caused the Group 5 primary containment isolation. The high pressure coolant injection system remained operable as required by Technical Speci0 cations to permit continued ,
reactor operation at pressure above 150 psig. Technicians calibrated the high flow sensors with i satisfactory as-found results. The RCIC system was successfully retested in accordance with procedure 8.5.5.1, "RCIC Pump Operability Flow Rate and Valve Test at approximately 1000 ,
I psig" and returned to servic Technical Specifications require the RCIC high steam Dow isolation setpoint to be .5;. 300 percent of rated turbine steam Cow to detect and provide isolation in the event of a RCIC steam line break. Engineers verified that the steam Dow instrument sensed 105 inches of water differential pressure which is the isolation signal setpoint. This setpoint was calculated to correspond to high steam flow in the RCIC steam supply line at a peak reactor vessel pressure of 1135 psig. The licensee noted that steam is less dense at lower pressures, resulting in a higher differential pressure at a given mass flow rate. This corresponded to a higher sensed differential pressure (steam flow) when the RCIC turbine was operating to control reactor pressure at approximately 470 psig. An additional factor affecting steam flow was the high pump flow demand via the throttled full flow test line which resulted in a high pump discharge pressure. Although actual steam flow was within normal parameters, the sensed steam flow was higher and reached the protective isolation setpoint. The inspector attended a HPCI/RCIC betterment review meeting where this topic was discussed. An action item was assigned for engineers to review the high steam flow setpoint calculation to determine whether a greater operating margin could be established. The licensee also initiated an impact review of system operation at reduced steam pressures and at increased turbine speeds. The licensee plans to review both the HPCI and RCIC systems for these concerns due to system similarities. The inspector concluded that appropriate corrective action had been initiated to address this even The LER was detailed and addressed all reporting criteri .
l l
_;
I l
17 NRC MANAGEMENT MEETINGS AND OTIIER ACTIVITIES (30702) Routine Meetings At periodic intervals during this inspection, meetings were held with senior BECo plant management to discuss licensee activities and areas of concern identified by the inspectors. At ;
the conclusion of the reporting period, the resident inspector staff conducted exit meetings on l December 8 & 10, 1993, summarizing the preliminary findings of this inspection. No l proprietary information was identified as being included in the repor .2 Management Meetings j l
On November 30, 1993, licensee representatives met with NRC, Region I emergency preparedness (EP) specialists at King of Prussia, Pennsylvania to discuss preparations for the upcoming full participation EP exercis On December 6,1993, Mr. Tnomas T. Martin, NRC Region 1 Administrator; M James C. Linville, Chief Reactor Projects Branch 3; and Mr. Eugene M. Kelly, Chief Reactor Projects Section 3A, toured the station and met with licensee management to discuss current licensee performanc .3 Other NRC Activities On November 29 - December 2,1993, one NRC Region I Senior Operations Engineer and an NRC contractor administered initial operator license examinations to three senior reactor operator candidates. Inspection results will be documented in NRC Inspection Report No. 50-293/93-2 !
!
l
-- - -