IR 05000354/1995015
| ML20149L227 | |
| Person / Time | |
|---|---|
| Site: | Hope Creek  |
| Issue date: | 02/14/1996 |
| From: | Calvert J, Ruland W NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION I) |
| To: | |
| Shared Package | |
| ML20149L220 | List: |
| References | |
| 50-354-95-15, NUDOCS 9602230411 | |
| Download: ML20149L227 (16) | |
Text
. - . . . .. .
. . - _ - . - - -- -.- - ,
.
.
l U. S. NUCLEAR REGULATORY COMMISSION l
REGION I
. DOCKET / REPORT NOS: 50-354/95-15 LICENSEE: Public Service Electric and Gas Company
Hancocks Bridge, New Jersey 08038 w
FACILITY: Hope Creek Nuclear Generating Station I
LOCATED AT: Artificial Island INSPECTION DATES: September 5-November 9, 1995
. INSPECTOR: ( / &# 'N '
ohn A. Calvert, Reactor Engineer Date Electrical Engineering Branch Division of Reactor Safety APPROVED BY: Y William H. Ruland, Chief Date Electrical Engineering Branch Division of Reactor Safety Areas Insoected: The areas inspected were the safety-related reactor vessel water level modification and the corrective actions for the digital feedwater control system (DFCS) and the reactor core isolation cooling
'
(RCIC) jockey pump corrective action process.
)
Results:
e The reactor vessel water level instrumentation backfill modification was adequately designed, installed, and tested.
e The problem identification, root cause, and corrective action were effective for the DFCS potential software common mode error. The system engineer interaction with, and support of, operations was excellent, as was the interaction of the design engineer in support of the system engineer. The technical understanding of the installed system was excellent, as shown by the engineer's timely analysis of the historical logs to arrive at the failure details and by the depth of detail in the engineering evaluation.
9602230411 960215 PDR ADOCK 05000354 G PDR
,
.
- . . - , - .- --
.
- The problem identification, root cause, and corrective action for the RCIC jockey pump low NPSH was effective. The root cause was thorough, accounted for actual pump installation configuration, identified the design calculation error, and identified the mechanism for the internal pump recirculation flow. The possible interaction with the ECCS strainers was evaluated, and recent industry operational feedback was considered.
- Engineering management acted proactively regarding the quality of software for an analog-to-digital upgrade project in the design stages concerned with a turbine EHC system.
.
,
ii
._ -
._ .
.
DETAILS 1.0 PURPOSE AND SCOPE The purpose of this inspection was to evaluate the design change process, root cause, corrective action, and management oversight at Hope Creek Nuclear Generating Station (HCNGS). The scope involved the safety-related reactor vessel water level modification and the corrective actions for the
'
digital feedwater control system (DFCS) and the reactor core isolation cooling (RCIC) jockey pump corrective action process.
2.0 REACTOR VESSEL WATER LEVEL BACKFILL MODIFICATION (IP TI 2515/128)
NRC Bulletin 93-03, " Resolution of Issues Related to Reactor Vessel Water Level Instrumentation in BWRs," was issued to address a concern that noncondensible gases may become dissolved in the reference legs of boiling water reactor (BWR) vessel water level instrumentation. These gases could result in false high level indications after a plant depressurization event.
The bulletin requested that licensees implement hardware modifications during the next cold shutdown beginning after July 30, 1993. HCNGS
installed the safety-related reactor vessel water level instrumentation i backfill modification during the last quarter of 1993.
The inspector reviewed the design / evaluation / test documents and 10 CFR 50.59 safety evaluation.
2.1 Modification Description '
n l Four reference legs provide a constant reference pressure to the reactor vessel level instrumentation. The piping for each reference leg is maintained full by condensed steam provided from a condensing pot located at the top of the reference leg. This design has been found to be potentially susceptible to the accumulation of noncondensible gases that may result in the elimination of steam flow to the condensing pot, thus resulting in the loss of makeup water to the reference leg. Another
- potential effect is that the water in the reference legs may absorb
'
significant amounts of dissolved gases. The dissolved gases may then come out of solution during a depressurization of the reactor vessel and
' introduce voids in the piping and, during rapid depressurizations, may expel significant amounts of the reference leg water inventory. The loss of inventory or the introduction of voids can result in false high level
,
indications.
HCNGS Modification 4EC-3407 provides four independent flow-regulating stations, each providing continuous makeup water to the respective reference leg, ensuring they remain full, and also prevent the accumulation of significant amounts of dissolved gases in the reference leg water. The backfill water enters a reference leg at the connection point outboard of where the instrumentation tubing connects to the containment isolation excess flow check valve.
I
.
.
The source of the backfill water is the control rod drive (CRD) system that is maintained at approximately 260 psid above reactor pressure during normal plant operations. The water from the CRD is filtered to 20-30 microns in each backfill system to prevent the introduction of particulates in the piping that could interfere with the operation of system components.
Two fine control needle valves are used in each line to maintain the flow rate to the reference leg of 0.5 gallons per hour (gph). A flow indicator was in each of the four individual supplies to the reference legs.
Prior to the connection point outboard of the containment isolation valve, two spring-loaded check valves provide a boundary between the nonsafety-related CR0 cvstem piping and the safety-related reactor vessel instrumentation system. At the backfill flow rates, the check valve plug will float off the seat and is designed to close on no flow and/or reverse differential pressure to insure that the reference leg inventory is not drained.
2.2 Manual Isolation Valve Evaluation The licensee stated that the manual isolation valves, that could result in the pressurization of a reference leg, are normally locked open in accordance with plant procedures, are in a high radiation area, and are not operated by any plant procedures during power operation. Furthermore, the valves must be accessed by a ladder. Based on this information, the inspector concluded that the potential to induce a severe plant transient ,
'
by inadvertent closure of a manual isolation valve by an operator during normal plant operations should be very low. l l
I 2.3 Flow Rate Evaluation I I
The licensee evaluated failures that could result in excess or reduced backfill flow. An example of excess flow evaluated was if the CRD line should fail high, a pressure-reducing regulator limits the pressure so that the flow rate is limited. Examples of reduced flow rate were instrument tubing failure between the spring-loaded check valves. The inspector reviewed the evaluations and found them adequate.
'
2.4 Maintenance and Surveillance of Check Valves The inspector reviewed the documents that showed the check valves in the backfill line were included in the licensee's inservice test (IST) program.
The inspector concluded that the backfill line check valves were included in the IST program.
2.5 Procedures The licensee had procedures for the following:
- place system in service;
- adjust the flow rate with metering valves;
- take a regulating station out of service; i
i
.
A 3 q l
- backflush the regulating station, e clean and/or replace the in-line filter. ;
The procedures were well written, provided detailed instructions, and included precautions for operation of the system.
The licensee calculated an allowable outage time (A0T) for the backfill ;
'
system.. The A0T for a single system was 30 days, and for two systems, the A0T was seven days. The A0T calculation was based on a GE report, which was applied to HCNGS. The inspector reviewed the calculations and found them adequate. 1 2.6 Low Cycle Fatigue
' A General Electric (GE) analysis determined that high cycle fatigue of inlet steam piping and condensing chambers were not a concern for flow rates of less than approximately 45 lb/hr and that high cycle fatigue of the reactor vessel nozzles is not a concern for such low flow rates, as long as the differential temperature between the backfill water and the nozzle is less than 50 F. j GE also determined that low cycle fatigue of the inlet steam piping and the condensing chambers is possible; and, therefore, plant-specific calculations are necessary to confirm that the thermal stresses associated with low flow rates will not be of concern for the life of the plant. The >
GE analysis stated that piping failures due to low cycle fatigue would take several operating cycles; and, therefore, interim operation was determined to be acceptable during the period when plant-specific calculations are being performed.
i The licensee had an interim stress calculation and has planned a plant-specific calculation for December 1995. The inspector audited the interim calculation and found it adequate. !
'
2.7 System Testing:
!
. System testing involved calibration, flushing, leak test, check valve test, l and transient tests. The transient tests involved monitoring indicated
'
level for acceptability for the following:
j * CRD pump swap;
- * movement of a single rod in and out; !
'
- CRD water pressure increase to move a stuck rod;
< * simulation of a sudden decrease in CRD pressure;
- simulation of an instrument tubing rupture in the backfill line.
The inspector evaluation was that the testing adequately covered the design j objectives, operating transients, and adequately showed system operability. l
, \
,
l__________.________ J
.
4 2.8 Walkdown The inspector walked down the installation of the four independent regulating stations and observed that the metering valve dial settings were different, although the indicated backfill flow rates were within tolerance. The system engineer stated that, if the rated flow rate could not be achieved with the metering valves fully open, then corrective action would be required. This meant that the metering valves settings could be different due to drive water pressure and particle entrapment in the filter.
In response to the inspector's observations, the system engineer found that the flow rate values were recorded twice daily, but the corresponding metering dial settings were not. This meant that trending for possible degradation was incomplete. The system engineer then took action to insure that metering dial values will be recorded for trending purposes. The system engineer also took action to backflush each regulating station and replace the filter during plant startup following the refueling outage.
2.9 Conclusions The inspector concluded that the reactor vessel water level instrumentation backfill modification was adequately designed, installed, and tested. The system operating procedures were in place and contain adequate directions to support system operation.
3.0 DIGITAL FEEDWATER CONTROL SYSTEM (DFCS) CORRECTIVE ACTION (IP 37550)
'
3.1 DFCS Problem Identification Sequence The inspector examined Nuclear Incident Report No. 950605215, dated June 6,1995, " Digital Feedwater Control System (DFCS) Alarms," and related
,
documents for engineering participation in plant activities, root cause determination and corrective action effectiveness. The incident involved operator discovered anomalous DFCS system conditions and failed signal indications on computer graphic displays, but with no corresponding annunciator alarm. The feedwater system remained stable.
The licensee's off-duty system engineer was contacted and decided to come in to the station to support operations. He scanned the various displays, verified proper feedwater controls, verified no system error messages, and verified that there were no obvious hardware failures. He then reset the fault-tolerant applications processor (FTAP) and one of the fault-tolerant control processors (FTCP). The DFCS system alarm on the graphics display then cleared.
The design engineer reviewed error messages in the DFCS historical log and found numerous error messages that identified failures along with the steps that the DFCS went through to maintain control.
i
.- . _ _ _ _ _ . . _ . - _ _ _ _ . . _ _ _ ._
'
.
'
,
The system engineer'and the design engineer realized that, if the FTAP i fail / restart could cause error conditions to exist in one FTCP, then it
could also cause failure in both FTCPs, so they were concerned about a -
potential software common mode failure. The potential common mode failure could result in an unnecessary turbine trip and feedpumps trip, with
! subsequent challenge to the plant protective systems.
3.2 DFCS Computer System and Data Communications
The DFCS consists of a fault-tolerant application processor (FTAP), two fault-tolerant control processors (FTCP), and other equipment. The FTAP provides the initial program loads for the FTCPs, monitors the health of
,
j system hardware, monitors process conditions, and the reporting of alarm messages to the annunciator. Except for initial program load, the FTCPs
- are designed to not require the FTAP for proper control of the feedwater
- system.
primary and one is the shadow AP. The FTAP is intended to function in the
- event of failure of either the primary or shadow AP, or the associated hard S
disk drive.
,
There are two FTCPs used to handle the processing load of the DFCS. Each FTCP is a primary and shadow CP configured as a fault-tolerant pair. One
! FTCP (3001) contains the majority of feedwater flow, feedpump speed, steam flow, and reactor water level algorithms. The other FTCP (3002) contains e the feedpump minimum flow control algorithm and miscellaneous functions.
< The inputs and outputs needed for the control of the DFCS are connected to the FTCPs through a separate redundant data link connected to field
! interface modules. If a FTCP fails, the outputs fail to preprogrammed
! " fail-safe" state, so that system upset is minimized until a rollback and automatic restart is complete.
t The FTAP communicates with the control processors (CP) through a redundant l data link called the node bus and the cps share certain information with
'
each other over the same data link. Data collisions can occur on the node
- bus, but each affected processor is designed to sense the collision, wait a random time interval, and then rebroadcast its data.
,
3.3 Review of Evaluation, Root Cause, and Corrective Action
The inspector reviewed document No. H-1-AE-EEE-0992, Revision 0, 4 November 8,1995, entitled, " Engineering Evaluation of the Dual Control
- Processor and Application Processor Failures Associated With The Hope Creek
) Digital Feedwater Control System." The event analysis indicated that the FTAP and one of the FTCPs (3002) failed and then were automatically restarted according to the design. The automatic restoration process took
. 105 seconds. The analysis also indicated that, because the FTAP failed,
, the signal to the annunciator was not generated.
-
,
.
.
'
l
I Engineers from the licensee's digital systems group and their consultant traveled to the vendor's facility, where it was determined that the failure l of FTCP 3002 was caused by a vendor software design problem. The failure occurred when data communication software incorrectly handled out-of-sequence data message acknowledgements from the FTAP.
.
The data message acknowledgements got out of sequence when the FTAP became .
I unsynchronized. Normally, the two constituent APs in a FTAP are synchronized together, both receive messages, but only one will send acknowledgements. What happened was that, for an undetermined reason, the two APs became unsynchronized. This should have caused self-diagnostics to j be performed, resynchronization to take place, and one AP sending message acknowledgements to be reestablished. But in this case, both APs were healthy (but unsynchronized), and both transmitted message acknowledgements to FTCP 3002 at slightly different times. This time skew is what caused the acknowledgements to arrive out of sequence.
The FTCP software design error interpreted the out-of-sequence message acknowledgements in such a way as to cause a larger value to be stored in
,
-
memory. The larger memory value was detected by the FTCP as a memory violation. Subsequently, the FTCP stopped the synchronization of its two constituent cps, detected that both cps had memory violations and therefore considered both cps failed. The CP failure detection started the automatic
- restart sequence. The output values were designed to maintain the values prior to the failure detection for the time required for the restart. The constant output values allow for the " graceful degradation" of the control function, and not a sudden loss of control.
e Failure Evaluation and Corrective Action for FTCPs The data communication protocol with the FTAP caused conditions for a FTCP to detect a failure and take itself off line until it could reestablish correct operation. The inspector determined that the time that the FTCP would be off line represents a potential challenge to the effectiveness of the DFCS control system and, consequently, to the plant and plant protective systems.
The random unsynchronization of the FTAP caused out-of-sequence acknowledge data packets to be sent to the FTCPs. This caused one of the FTCPs to generate a memory violation, which, in turn, caused the affected FTCP to unsynchronize and reboot. The cause of the FTCP memory violation was a
,
software design error in the data communications protocol. The software design error was common to both FTCPs and, therefore, a potential common mode failure.
The licensee evaluated the effects of FTCP 3001 and 3002 common mode failures on the plant operations. The extensive results indicated that the DFCS would maintain its analog and digital outputs at the values prior to the failure for 10 seconds. After 10 seconds, all three feedpumps and the main turbine would trip. The evaluation stated that, if a transient were
'
to occur during the 10 seconds, the plant would be within the bounds of the safety analyses because of turbine trip and reactor protection system trip.
- . . - . .. - - - . . . - - . . .- .. . - - - - . - - - . --
,
'
,
- 7
$ The vendor designed a software modification (Revision 3.3/3.3.1) that was designed to remove the susceptibility of the FTCPs to respond incorrectly to improperly-sequenced acknowledge data packets from the FTAP, which could 1 cause a FTCP to unsynchronize and restart.
'
Modification No. 4EC-3529 replaced the software with vendor Revision
,
3.3/3.3.1, which was designed to remove the potential for simultaneous FTCP
. failures. The method was to reload the FTAP with the new software revision and then reboot all the FTCPs, which will then receive the downloaded revision from the FTAP. The licensee witnessed the vendor factory
,
l acceptance test that demonstrated that the new software revision corrects l
the software error that causes out-of-sequence data packets to cause FTCP i
unsynchronization, with subsequent rebooting. The licensee stated that f
selective tests were run after site installation to verify proper system d
and FTCP functioning. The inspector concluded that the FTCP corrective j action was effective.
.
,
- Failure Evaluation and Corrective and Action For FTAP
! The FTAP problem that caused the FTAP to become unsynchronized was not
! addressed in software Revision 3.3/3.3.1. The licensee stated that, with
the revised FTCP software, the effects of FTAP unsynchronization would
'
continue to be assessed to uncover the source of the apparent random
, unsynchronization.
.
The licensee evaluated the effects of FTAP failures on plant operations and l l concluded that there would be no direct impact on feedwater control. The '
operators would be unable to request new displays. There would be loss of
- historical / diagnostic data and inability to load initial programs of other l
processors. The licensee stated that, while this loss of function would be i inconvenient, there would be no effect on the control of feedwater. There !
l would be no overall DFCS system health-monitoring and associated i
annunciator alarm, but this would require the loss of both constituent APs, l which the licensee considers a remote occurrence.
i The root cause of the random FTAP unsynchronization was not discovered.
L The licensee plans to continue to track the random FTAP problem so that the
root cause can be identified and corrective action formulated.
! 3.4 Vendor Failure Notification System
.
'
The licensee found that the vendor system of notification for problems was
flawed because notifications were sent to the purchaser of the equipment
,
and not the user of the equipment. The FTCP memory violation problem due i to out-of-sequence data acknowledgement packets was reported to the vendor
- by two other users in 1993. The vendor sent out a safety advisory to 1 purchasers on April 1,1994, which described the problem and offered
'
corrected software. The vendor advisory was sent to PSE&G's nuclear i training center, which purchased a DFCS training set, and also to another j PSE&G vendor, which was the actual purchaser of the _ equipment for PSE&G.
PSE&G verified that the vendor changed their problem notification list to ;
ensure that PSE&G user' personnel would be notified of any future problems.
'
I
I
!
! ,
_,
i
. . . - . .. . - - . - - . - _ .. . . .
'
.
'.
I 8 3.5 Safety Evaluation i The inspector reviewed the completed Form NC.NA-AP.ZZ-0059-2, "10 CFR 50.59 i Applicability Review," for the modification package that updated the DFCS
'
software. The form consisted of screening questions, the answers to which j indicated that a formal evaluation according to 10 CFR 50.59 was not j required.
l The licensee stated that the change corrected identified software problems i
and allowed the software to operate in accordance with the original design.
l The original DFCS design modification had a 10 CFR 50.59 evaluation. The inspector reviewed documents that evaluated the effects of the software problems on the equipment and also the results of the corrected software.
l The inspector concluded that the licensee evaluation as reviewed in Section
3.3 did not invalidate the original 10 CFR 50.59 evaluation and that the
- system should be within the bounds of the UFSAR safety analysis.
-
3.6 Conclusions
The inspector concluded that the system engineer interaction with, and support of, operations was excellent, as was the interaction of the design
< engineer in support of the . system engineer. The technical understanding of
' the installed system was excellent, as shown by the engineer's timely analysis of the historical logs to arrive at the failure details and by the depth of detail in the engineering evaluation.
.
The inspector concluded that the root cause determination was technically
'
sound and showed excellent, aggressive pursuit and solution convergence of possible common mode software problems. The licensee verified that the corrective action software changes for the control processors were
- effective by direct observation at the vendor facility, before plant
' installation.
,
The inspector concluded that the licensee failure evaluation did not
,
invalidate the original 10 CFR 50.59 evaluation.
-
.
l The inspector concluded that the licensee's aggressive problem recognition and corrective action for the potential software common mode failure in the DFCS software was a proactive step in the identification and minimization
. of unnecessary plant trip challenges. The problem identification and
- corrective action were effective.
4.0 REACTOR CORE ISOLATION COOLING (RCIC) JOCKEY PUMP SUCTION PIPING l i
MODIFICATION The inspector reviewed LER 95-019-00, "RCIC System Jockey Pump Suction Piping Inoperable Since Plant Startup," dated October 2,1995. The LER described a low differential pressure failure of the RCIC jockey pump when
'
aligned to the torus during an IST run. The IST test was previously run successfully aligned to the CST. The RCIC jockey pump was declared l
>
_ _ _ _ _ _ . . ._ _ _ __ _ _ _ _ _. . _ _ _ _ _ . _ _ _. _.
l
<
inoperable. Effective corrective action was taken that involved removing a one-inch suction line and replacing it with a two-inch line per Modification 4HE-0262. The pump was disassembled and no cavitation damage was identified. The other ECCS-jockey pumps have two-inch suction lines.
.
'
>
4.1 Jockey Pump Function L
The normal function of the RCIC jockey pump, which, under alignment to the 1 CST, is to maintain the RCIC discharge piping full to prevent water hammer
'j on pump start, would not have been challenged. The combined ability of the RCIC and/or HPCI jockey pumps to provide the long-term water seal on the
feedwater check valves after a LOCA would have been challenged. The long-
!
term water seal is required by the technical specification in Table g
3.6.3-1, Note 2, which allows either the RCIC or the HPCI jockey pumps to j establish the seal. The licensee stated that the loss of the water seal J would not challenge the current dose analysis, since no credit is taken for
the water seal during leak rate testing. Also, the HPCI jockey pump was operable, so the alternate path for long-term feedwater seal would have been available. The RCIC system is not part of the emergency core cooling j. system as described in the UFSAR, Section 6.3.
- 4.2 Root Cause Evaluation The inspector reviewed " Root Cause Report for IB-P-228 RCIC Jockey Pump i 4.0-5-P Failure," dated September 21, 1995. The licensee stated that the l RCIC jockey pump would have cavitated when aligned to the torus or at the lower levels of the CST in post-LOCA operation with the one-inch line. The "
pipe was sized incorrectly because pump internal recirculation flow was not added to the design flow in the design calculations. This increased the
head losses and reduced the NPSH. The pump internal recirculation flow was
.
from the combined effects of a piping reducer on the suction side and an
'
orifice plate on the discharge side. The report stated that recirculation
induced cavitation, but at low flow rates the cavitation is considered low energy and pump damage is not expected, although pump performance can be f affected. The original design calculations for NPSH did not include
,
I consideration of operating the pump in a region that could result in
,
.
internal recirculation.
I The licensee stated that an analysis was performed for the two-inch suction i piping that indicated a sufficient margin above the required NPSH to support the maximum flow rates for the feedwater seal.
4.3 Particulate Identification and Evaluation
- The one-inch pipe torus suction was inspected after removal. No obstructions were found, but a black film was noted. The 40-gauge mesh
.
(0.015 inch grid) suction strainer also had a black film. The Chemistry
'
Department identified the material as fine grain magnetite, a common iron corrosion product. The strainers are planned to be inspected after the next test of each jockey pump. The Chemistry Department has been assigned
'
the task to determine the source and potential impact of the fine grain black material (Problem Report No. 9590808066).
,
l
_
.
.
When the RCIC jockey pump was disassembled, a black film coated the internals that could be easily removed or swiped from the surface by hand.
The internals were expected to be clean due to cavitation, and it was assumed that the black material settled onto the internals during the idle
,
period after the test. The root cause report evaluated any effects of the film on system resistance or effective pipe diameter and found that any effects were negligible.
The licensee stated that the torus was inspected and vacuumed during the last refueling outage (April 1994), that no fibrous material was found, and that sections with appreciable accumulation were cleaned by vacuum or by a diver. NRC Information Notice 95-47, " Unexpected Opening of a Safety Relief Valve and Complications Involving Suppression Pool Cooling Strainer Blockage," was evaluated by the licensee. They contacted Limerick engineering and the SRV vendor and planned to conduct an underwater camera inspection for the next refueling outage.
The ECCS suction strainers in the torus are stacked washers with one-eighth inch openings. Since the black substance passed through the 40-gauge mesh of the RCIC jockey pump strainers, the licensee evaluation concluded that the substance, if present in the torus, would not pose a blockage concern for the ECCS strainers.
- 4.4 Conclusions The inspector concluded that the corrective action for the RCIC jockey pump low NPSH was effective. The root cause was thorough, accounted for actual pump installation configuration, identified the design calculation error, and identified the mechanism for the internal pump recirculation flow. The extension of the possible interaction with the ECCS strainers was
,
evaluated, and recent industry operational feedback was considered.
5.0 MANAGEMENT OVERSIGHT i The inspector determined through interviews that the engineering management acted proactively regarding the quality of software for an analog-to-digital upgrade project in the design stages concerned with a turbine EHC system. The licensee's Digital Systems Group (DSG) performed a critical
, digital review of a potential vendor and identified a critical software error along with a total lack of software documentation. The DSG recommended that proper documentation be purchased so that the software could be verified. The engineering directors approved a course of action that involved purchase of the necessary software documentation, so that the DSG could determine the quality of the software. After the quality determination, the decision to purchase the digital upgrade would be made.
- . .
.
. 1 11 l l
6.0 CLOSURE OF PREVIOUS INSPECTION OPEN ITEMS 6.1 Closed - Item 50-354/95-02-01 This open item involved the suitability of the RPS loads to possible change ,
in total harmonic distortion (THD) caused by the change to a ferroresonant I transformer in the alternate feed. The licensee measured the THD on the l bus, while utilizing the alternate supply, in order to determine the (
'
suitability for the RPS loads. The measured THD was 6% in the loaded case and 2% in the unloaded case. The inspector concluded that the THD was acceptable in this case, since the alternate feed is designed for maintenance, not operational use. This item is, therefore, closed.
6.2 Closed - Item 50-354/92-80-08 This item concerned a relay that was connected to a Class lE dc bus, but not hooked up to any alarm or actuation system. The licensee spared the relay in place under Design Change Package 4EC-3368. This item is, therefore, closed.
7.0 EXIT MEETING An exit meeting was held on November 9,1995, with members of the licensee's staff noted in Attachment 1. The inspectors discussed the scope and findings of the inspection. The licensee had no disagreements with the findings. The inspectors received and reviewed proprietary material during the inspection and used the material only for technical reference. No part of the material was knowingly disclosed in this inspection report.
Attachments:
1. Persons Contacted 2. State of New Jersey Letter
. . __ _ _ _ _ . . . . ___ __ .
.
.
ATTACHMENT 1 l Persons Contacted l
"
Public Service Electric and Gas M. Abdullah Senior Staff Engineer, Digital Systems C. Atkinson Supervisor, Hope Creek Electrical Engineering
'
R. Binz System Engineer, IST M. Bursztein Manager, Nuclear Electrical Engineering
,
M. Cirelly System Engineer 4 J. Clancy Manager, System Engineering, Hope Creek J. Defebo Supervisor, Hope Creek Plant Support Assessment
- P. Duke Engineer, Licensing and Regulation C. Fuhrmeister Supervisor, Mechanical Systems D. Garchow Director, Nuclear Systems Engineering
- M. Idell Design Engineer
'
A. Kao Principal Engineer, Mechanical design Engineering S. Karimian Technical Consultant, Nuclear Engineering Department
- D. La Mastra Manager, Hope Creek Engineering Design (Acting)
i C. Lambert Manager, Projects
A. Narayan Engineer, Mechanical Engineering T. Nicholson Engineer, Stress Engineering
- *G. Overbeck Director, Nuclear Design Engineering
>
J. Pollack Manager, Quality Assurance
- J. Priest Engineer Licensing and Regulation
- M. Reddemann General Manager, Hope Creek Operations M. Reeser Senior Staff Engineer, Engineering Mechanics J. Rosas System Engineer B. Simpson Senior Vice President, Nuclear Engineering J. Thompson Engineer, System Engineering C. Tully Principal Engineer, Nuclear Reactor Regulation
. *C. Waite Supervisor, Digital' Systems New Jersey Deoartment of Environmental Protection Suren Singh Engineer
- Indicates those present at the exit meeting on November 9, 1995.
t I
s
,
-
-- _ _ _ _ _ _ _ . - _r_
_.q ( ArrACmNr ~2
.
" '
g
$ tate of piefs 3erseg l Christine Todd Whitman Department of Environmental Protection ' Robert C. Shinn. Jr. I:
Csvernor Cornininionir l
Division of Environmental Safety, Health, l Analytical Programs !
Radiation Protection Programs- I Bureau of Nucle'ar Engineering )
CN 415 ,
'
Trenton, NJ 08625-0415 January 9, 1995
.
Mr. James T. Wiggins, Director Division of Reactor. Safety
'
U.S. Nuclear Regulatory Commission 475 Allendale Road King of Prussia, PA 19406
Dear Mr. Wiggins:
Subject: NRC Engineering Inspection Hope Creek Generating Station (HCGS)
In accordance with the provisions of the July 1987 Memorandum of Understanding between the Nuclear Regulatory Commission (NRC)
and the New Jersey Department of Environmental Protection (DEP),
the DEP is providing feedback regarding the NRC's Engineering
,
Inspection of HCGS. This inspection was conducted by Mr. John Calvert of NRC Region I from September 5 to 8, 1995, which then continued from October 2 to 5, 1995. A New Jersey DEP's Bureau of Nuclear Engineering (BNE) representative obser/ed only the second week of this inspection. In keeping with the agreement between the DEP and the NRC, the DEP will not disclose its inspection observations to the public until the NRC releases its final report.
The NRC inspection of the week of October 2, 1995, focused on the RCIC jockey pump design modification. Its implementation, as PSE&G stated, was necessary because the pump was cavitating with excessive noise when aligned to the torus. During the inspection of this modification, PSE&G also stated that the pump was disassembled, inspected and found to be in satisfactory condition.
A close inspection of the pump and strainer revealed that the pump internal casing was clogged with black powder / slurry material. The source of this material was a concern to the NRC, which PSE&G staff could not address satisfactorily at the time of this inspection.
~.
e Newire=1m Evaloppomaer twkrw ww n=
. . - - - . . .
in l
. * ~
.
,
-l
'
I However, based on BNE discussions over the telephone with the NRC,
'
PSE&G had provided technical justification prior to the November 9, 1995 exit meeting, which BNE could not attend. NRC considers this a resolved issue, however, BNE plans to review PSE&G's technical report to perform an independent analysis of the source of this black material clogging.the pump.
The BNE staff appreciates the opportunity to review the engineering calculations for the Backfill system modification. The
- calculation process specifically used for this modification was
- found to be effectively performed. .
This information has been communicated to the DEP management.
If you have any questions, please contact me at (609) 984-7700.
l
Sincerely,
.
'
Kent Tosch, Manager Bureau of Nuclear Engineering I
c: Dave Chawaga, SLO, NRC Jill Lipoti, Ph.D., DEP Dennis Zannoni, DEP
,
[
a 6 k
(
. ._ _._