IR 05000237/1997016
| ML20217E475 | |
| Person / Time | |
|---|---|
| Site: | Dresden  |
| Issue date: | 10/02/1997 |
| From: | NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION III) |
| To: | |
| Shared Package | |
| ML20217E468 | List: |
| References | |
| 50-237-97-16, NUDOCS 9710070041 | |
| Download: ML20217E475 (19) | |
Text
.
.
U.S. NUCLEAR REGULATORY COMMISSION REGIONlli
,
l l-Docket No:
50 237 l
License No:
DPR 19 l
Report No:
50 237/97016(DRS)
Licensee:
Commonwealth Edison Company
>
FatSty:
Dresden Nuclear Station Unit 2 Loation:
6500 No,th Dresden Road Morris,IL 60450 Dates:
July 28 though August 1,1997 Inspectors:
D. McNeil, Reactor Engineer, Rlll J. Munro, Senior Reactor Engineer, NRR -
D. Pelton, Resident inspedor (in training), Braidwood, Rill D. Roth, Resident Inspector, Dresden, Rlli Approved by:
M. Leach, Chief, Operator Licensing Branch Division of Reactor Safe'.y
9710070041 971002 PDR ADOCK 05000237 G
.
,
e EXECUTIVE SUMMARY Dresden Nuclear Station Unit 2 NRC Inspection Report 50 237/97016 Ooerations This inspection report covers the events associated with a manual reactor scram that occurred on July 28,1997.
The root cause of this event was the failure of the NSO to correctly assess the condition
of the FWCS (Section 01.1.c).
The inability of the NSO to control feedwater flow manually (Section 01.1.c), the
'
.
procedures used to start the RFP and respond to the RPV level changes (Section 03.1.c), the material condition of the 2B RFP RV (Section M2.1.b) and three element i
control of the FWCS (Section 02.1.c) were contributing causes.
An additional contributing cause to the ev<.nt was the lack of substantialInformation
.
regarding FWCS operating characteristics provided to the operators during continuing operator training. (Section OS.1.c)
The licensee response to the event was good. The prompt Investigation team formed
.
by the licensee performed a thorough review of the event and of the contributing factors to determine the preliminary root causes. Management review of the prompt investigation team's finding: was also thorough and challenging. (Section 01.1.c)
Maintenance The cause for the failure of the RFP RV was a loose connection on a terminal board.
.
Maintenance performed during recent outages did r.ot appear to have contributed to the failure. The licensee was unable to locate any record of maintenance that involved the terminal board connection. (Section M2.1.c)
The prompt investigation team did not initially plan on checking for other loose
.
connections in the cabinet. However, management directed that all terminals in the same location be checked prior to restart of Unit 2. (Section M2.1.c)
Engineering The degr6ded condition of the three element control of the FWCS contributed to the
.
event. (Section 02.1.c) The Unit 2 FWCS only could be used in sir.gle-element control, it could not be determ!ned if this would have prevented the RPV water level from '
exceeding a five inch change in water level. (Section E2.1.c)
,
-
,
.
s
.
Reami.Qnialla Summary of Plant Status Unit 2 entered a forced outage (D2F29) after the manual scram due to a feedwater translent on July 28. A plant startup was conducted and the reactor taken critical on July 31. By August 1, the licensee had closed the turbine generator output breaker and was continuing power ascension.
Unit 3 was at 95% of full power during this inspection. Load was limited due to a problem maintaining correct levelin a high pressure feedwater heater. At the end of the inspection
'
period, the licensee had not completed corrective actions on the feedwater heating system and
,
power remained at 95% Unit 3 was not impacted by this event.
N
!
A detailed sequence of events is included at the end of this report. (Enclosure 2)
l 1. Operations
Conduct of Operations 01.1 { Unit 2) Transient Descriotion a.
Insoection Scone (93702)
Using the guidance found in inspection Procedure (IP) 93702, " Prompt Onsite Response to Events at Operating Power Reactors," and IP 71707," Plant Operations,"
inspectors reviewed the significance of the 'svent, performance of safety systems, and actions taken by the licensee. While the inspectors were in the control room they reviewed the plant's statut with operators and observed plant parameters inspectors interviewed the Unit Supervisor and both NSOs involved in the feedwater pump (FWP)
shift and rev;ewed the Unit 2 Emergency Plan requirements, inspectors reviewed plant status and licensee manaCement plans to initiate a subsequent plant startup.
Inspectors reviewed station logs, attended a Plant Operations Review Committee meeting, and reviewed the licensee's prompt investigation team results, b.
Observations and Findinas The plant (Unit 2) was found to be in a stable, cold shut down condition. All control room indications were normal for the wrrent plant conditloh. Operators were atten'ive to control room Indications at,d were performing surveillances required prior to restart of the Unit 2 reactor.
Interviews with the crew involved with the reactor scram were candid. The oparators were willing to accept responsibility for the events that had occurred, but indicated their lack of faith in the ability of the FWCS to control RPV water level during transient conditions. One operator interviewed declared the Unit 2 FWCS to be the." ultimate operator work around."
o
_-__
.
'
A complete time line of the sequence of events had been developed by the prompt Investigation team. Their findings were presented to the inspectors at the inspection entrance meeting. A sequence of events recorder printout was made available to the inspectors to confirm the findings of the prompt investigation team. The following is a description of the events leading to the reactor scram, including a precursor event that occurred on the previous shift.
On the afternoon of July 27,1997, the Unit 2 operators responded to a failure of the 2B RFP recirculation valve (RV). The RFP RV went from full closed to midway-open, back to full closed, then to full open in approximately two minutes.
The operators diagnosed the falling RFP RV and took positive control of the RFP
'
RV by moving its control switch from automatic to open. During the cycling of the RFP RV, the operators assessed the status of the feedwater control system (FWCS) and stabilized reactor vessellevel by lowering power to within the capability of the feedwater system, and by taking manual control of the feedwater regulating valves (FRVo).
-
At one point during the transient the crew observed a high level alarm (+35")
while the FWCS demand signal to the FRVs was still 100% This led the crew to believe there was something wrong with the FWCS. The feedwater system engineer was called in to investigate the RFP RV behavior and the FWCS response to the transient. The system engineer determined that the FWCS was operating correctly in the single element mode. A loose lead was found in the control circuit for the RFP RV, causing the RFP RV cycling noted on the afternoon shift.
The licensee concluded tha'. the 2B RFP needed to be secured so that the loose lead could be tightened without causing an additional feedwater transient. The midnight shift crew was informed during shift turnover of the problems with the RFP RV, and the need to swap operating RFPs. The crew held a pre evolution br!af to discuss the evolution. During the brief the nuclear station operator (NSO) and an extra operator (Aux NSO) assigned to assist the NSO were told by the Unit Supervisor (US) that they should expect a 2" to 3" rise in Reactor Pressure Vessel (RPV) level after 2C RFP was started.
The NSO started the 2C RFP early on July 28, and reactor vessellevel increased by more than +5"instead of the expected 2" to 3". After RPV level had increased above +35", the NSO took manual control of the FRV and rapidly closed the FRV. The Aux NSO alerted the NSO that feedwater flow had been significantly reduced (from 8 million pounds mass / hour to less than 2 million pounds mass / hour). The NSO then rapidly opened the FRV to raise feedwater flow. Feodwater flow increased to such a high value (greater than 11 million pounds mass / hour) that the RFPs experienced low suction pressure resulting in the automatic trip of 2C RFP The RPV low level alarm annunciated as a result of the momentarily lowered feedwater flow entering the RPV. The US observed the rapid sequence of the high RPV level alarm, low feedwater flow, trip of an RFP, and o :ow RPV level alarm in approximately 15 seconds, and concluded
.
.
'
that the NSos did not have control of the FWCS (and RPV water 13 vel) and ordered a reactor trip. The crew complied, and the reactor was shutdown. The operating crew followed Dresden General Procedure (DGP) 02-03, * Reactor Scram," to cool the reactor after the trip.
The lleansee's prompt investigation team presented their findings at a Plant Operations Review Committee meeting held to determine the status of the subsequent plant startup. The team's presentation at the committee meeting was essentially the same as the presentation given at the NRC inspection entrance meeting. The committee questioned the team concerning operator training, status of the FWCS, and repairs associated with the failed RV.
l c.
Concluslang l
The inspectors concluded that the root cause of the event was the failure of the NSO to
'
correctly assess the condition of the FWCS. The FWCS was responding to the high level when the NSO took manual control of the system, if the NSO had not placed the FWCS in the manual mode, the FWCS would have automatically restored the RPV level to its normallevel. The Inability of the NSO to control feedwater flow after taking manual control was a contributing cause of the event.
The licensee's investigation was thorough, completed in a timely manner, and the root causes of the event were determined. Management review of the prompt investigation team's, findings was also thorough and challenging.
Operational Status of Facilities and Equipment 02.1 Feedwater System a.
insoection Scone.f,93102)
Inspectors reviewed licensee event reports (LERs), past NRC inspection reports, and interviewed plant personnel to determine the performance record of the feedwater system.
b.
Observations andfindings The transient that occurred on July 27 was caused by the cycling of the 2B RFP RV.
The cycling was caused by a loose connection in part of the feedwater recirculation valve logic. The licensee identified no work related periormance items on the feedwater system that might have caused the loose connection.
,
Prior to the July 28 transient, the Unit 2 FWCS was being operated in single-element control. As such, the FWCS used only RPV water level to position the FRVs. The original design of the FWCS used three-element control, which used the difference between feedwater flow and steam flow in addition to RPV level to position the FRVs.
'
,
.
'
However, startup testing in the early 1970s demonstrated that three-element control performed poorly, and the licensee decided to use single element control, in 1994, the licensee reviewed and essessed open engineering requests and decided to upgrade the FWCS. In early 1996, modifications were made to the FWCS and extensive testing of the FWCS using three-element control was completed, in May 1996, a software error in the FWCS caused RPV water level to decrease outside of a pre ret band. Operators manually tripped the reactor in response to a low RPV level.
i
'
The level transient and manual scram were documented in LER 50 237/96-009-00, and diccussed in inspection report 96009.
The licensee assessed the information from the 1996 transient, as well as inforr11ation from other nuclear plants, and decided not to operate Unit 2 in three element control until additional system improvements were made. Unit 2 has remained in single element control. The continued inability of operations / engineering to implement three element control in the Unit 2 FWCS fosterM a "no confidence" attitude toward the FWCS by control room operatorc. Operators believed the FWCS would not respond properly to transients and would quickly take manual control of the system during transients that could be easily handled by the FWCS if properly operating in three-element control, c.
Conclusiong
'
The inability of Unit 2 FWCS to operate in three-element control impacted the FWCS'
response to the transient. In single element contrcl, the predictive response of the FWCS was removed, which significantly delayed FWCS response to the event.
Operatione Procedures and Documentation 03.1 Review of Procedure Adequacy a.
insoection Scone (93702)
The inspectors reviewed the procedures used by the operators to respond to the level-translents and to attempt to perform the swap of feedwater pumps. These procedures were assessed for adequacy based on interviews with operators and actual operator performance. Procedures reviewed included:
Dresden General Pacedure (DGP) 02-03, " Reactor Scram," Rev. 35 Dresden Operating Abnormal (DOA) Procedure 0600-01," Transient Level
Control," Rev. 20
Dresden Emergency Operating Procedure (DEOP) 100, " Reactor Control (Unit 2)," Rev, 06
Dresden Annunciator Procedure (DAN) 902(3) 5, "RPV LVL HI, E-8" Rev. 05
DAN 902(3) 5,"RPV LVL LO, F 8" Rev. 05 Lesson Plan (LP) 259L S1, "Dresden Emergency Operating Procedures,"
Rev.1
.
,
.
.
.
Dresden Operations ProceNre (DOP) 3200-03,"Startup of Second Reactor Feed Pump or Shifting to Attemate Reactor Feed Pump," Rev.18 b.
Observations and Findings Operators we, e using DOP 3200-03 to start the 2C RFP. The DOP did not refer to any anticipated change in RPV water level when starting an additional RFP. ")e DOP had three locations whcre the statement, " Verify reactor water level is stable" appeared.
The procedure did not direct control room operators to monitor or compare feedwater and steam flows during the RFP start process. A proposed revision to DOP 3200-03 did not contain Instructions to monitor feodwater or steam flows when starting a feodwater
,
pump.
The first annunciator energized during the transient was RPV LVL Hl. DAN 902(3) 5 E 8, directed operatore to enter DOA 0600 01, " Transient Level Contro!," and indicated there were no automatic actlans associated with the annunciator. Procedure DOA 0600 01 directed operators to monitor leycl (emphasis added), but did not direct operators to monitor the difference between feedwater flow and steam flow while controlling level manually. A revision to procedure DOA 0600-01 was issued prior to startup of Unit 2 that directed operators to match steam flow with feedwater flow to l
stabilize RPV water level when manually controlling feedwater flow.
I c.
Conclusions The inspectors concluded that operating procedures (DOA 0600 01, DOP 320C-03 and DAN 902Z(3)-5 E 8) lacked specific guidance for operators to monitor parameters other than RPV level during various feedwater system operations, but an NSO should have been able to successfully complete the assigned pump shift by following the provided proHures. The lack of specific procedural guidance was considered a contributing cause to the event.
Operator Knowledge and Performance 04.1 0) nit 2) Ooerator Pedormance
a.
Insoection Scoce (97302)
The inspectors used the procedural guidance found in IP 93702 to determine the adequacy of licensee training programs addressing the precursors to this event. The inspectors assessed the operators' performance and procedural compliance through interviews and review of data recorded during the transients. The inspectors reviewed the following:
DGP 02-03, " Reactor Scram," Rev. 35 DOA 0600-01, " Transient Level Control," Rev. 20
DEOP 100, " Reactor Control (Unit 2)," Rev. 06
DAN 902(3)-5 E-8, "RPV LVL HI," Rev. 5
'
l
,
_
_
,
_-
.
.
.
LP 259L S1,"Dresden Emergency Operating Procedures," Rev.1 DOP 3200-03, Start'ap of Second Reactor Feed Pump or Shifting to Attemate
Reactor Feed Pump, Rev.18 Portions of the Unit 2 Sequence of Events printout for July 27 and July 28.
Prompt investigation for Unit 2 B Reactor Feed Purrp minimum flow valve failure.
b.
Observations and Findings Response to Fallure of RFP Recirculation Valve The tranunt on July 27 was first noticed by the operators when the RPV LVL LO annunciator alarmed. The Unit 2 Unit Supervisor directed the operating crew to assume appropriate control stations to monitor and control the reactor. The NSOs monitored
'
feedwater reg ilating v,1ve position and took manual control of the FWCS when the expected system respi nse was not obseived. The Aux NSO took manual control and i
operated the FWCS w'ille the NSO reduced reactor power, averting actuation of any i
reactor protection sys an: circuits. During the event the FWCS was observed to l
demand 100% FRV r.osition while a nigh RPV level condition existed. This caused
)
operators to believe tiiere was something wrong with the FWCS.
The system engineer was called to the station to investigate the RFP RV and the apparent failure of the FWCS. The system engineer ario maintenanco technicians were (
able to locate the cause of the RFP RV failure. The system engineer confirmed that the l
FWCS responded correctly and indicated thst the high FRV position demand accompanied by the high RPV level happened because the FWCS was still ramping-down from a high demand caused by the previous low PPV level signal (RPV level was below the RPV low level alarm setpoint for several minutes). The operating crew was unaware of the time constants built into '.he FWCS that caused the indicated conditions and did not raalize the system was operating correctly.
Heightened Level of Awareness Brief The operating crew held a heightened level of awareness (HLA) brief and discussed the p'an to swap operating feedsvater pumps. During the brief the US directed the NSO to trip the reauor if RPV level.ncreased to +40" cr decreased to +20". These values were chosen because they werr previously used during the feedwater system testing completed during 1996. The values were well withiri th6 uand of the automatic feedwater pump trips and reactor trips. The US indicated thet the NSOs should expect a 2" to 3" change in RPV level after the RFP was started. No additional discussions were held concerning NSO contingency actions.
Following the HLA brief, the NSO decided to take manual control of the feedwater system if the level varied by more than five inches. A five inch deviation correspondori to the first lesel-induced control room alarms (RPV LVL HI), and required entry into DOA 0600-01, " Transient Level Control." The immediate operator actions were to: " Evaluate feedwater system performance. IE necessary, THEN...: Shift REG VLV STATION (S),
as applicable, to MAN ANQ control level manually." The NSO's decision was based on
)
- - _ - -
- - - - - - - - - - - _ _ _
,
.
.
.
the belief that a five inch level deviation demonstrated that the FWCS was not performing correctly. The NSO discussed the decision to take manual control of the
.
FWCS with the Aux NSO, but did not discuss the decision with the US.
Manual Coritrol of F'edwater System Following the 2C RFP start RPV level exceeded +35", and the NSO took manual control of the feedwater regulating valves. The NSO did not evaluate feedwater system performance in a systematic manner prior to placing the FWCS in the manual mode.
The NSO slowly, then rapid!y, closed the FRVs. The Aux NSO alerted the NSO that feedwater flow had decreased below 2 million pounds per hour (down from an original 8
,
million pounds / hour) and ths NSO rapidly increased feedwater flow above 11 million pounds per hour. The rcpid increase in flow caused 2C RFP to trip on low suction pressure. The NSO failed to properly monitor feedwater flow, steam flow, or FRV position.
Manual Reactor Trip The US had directed the NSO to scram the reactor if RPV levelincreaad to +40" or l
decreased to +20". RPV level did not reach either of the specified values. The US was unable to verifiy the manual actions being executed by the NSO. The indications available to the US were the control room annunciators, in approximately 15 seconds the US noted an RPV high level alarm, an RPV low level alarm, and a FWP trip alarm due to low suction pressure. Based on these Indications, the US determined that the NSO did not have adequate control of the FWCS and ordered a manual reactor trip.
c.
Conclusions The July 27 afternoon operating crew on Unit 2 demonstrated good diagnostic and operating skills by determining the cause of the feedwater transient and by rubsequently reducing reactor power to stabilize the plant. The crew demonstrated good procedural compliance with DOA 0600-01 by diagnosing the problem, placing the approp< ate controllers in manual and controlling the plant. Manual manipulation of the FWCS was correct, timely, and accurate.
The afternoon crew's determination that the FWCS was not operating correctly indicated a lack of understanding of FWCS response when the FWCS controllers have been saturateo. The weakness did not result in any complication of the event because the NSO was able to control level manually.
Operators were very conservative when they chose setpoints for initiating action. The US selected +40" as the point where the crew would manually scram even though the afternoon shift had reached +46" during their response to the RPV RV failure without a reactor protection system actuation. The NSO's decision to take manual control of the FWCS with a 5" deviation in water level was also conservative.
l
'
.
.
.
Communications between crew members were weak and contributed directly to the reactor trip on July 28. During the HLA the US failed to provide the NSO with contingency actions in the event a response other than the expected response occurred.
The NSO decided to take manual control of the FWCS after a 5" level deviation without consulting with the US. This removed the US from the decision making process and prevented the US from providing additional guidance to the NSO.
The NSO failed to diagnose the condition of the FWCS during the 2C RFP start. The
failure to diagnose the condition of the FWCS was the first precursor leading to the l
reactor trip. An accurate diagnosis of the FWCS would have revealed that feed flow was decreasing and would have been less than steam flow within a few seconds. This would have caused RPV water level to begin decreasing and return to its normal setpoint. The inspectors determined the failure of the NSO to accurately assess the condition of the FWCS was the root cause of the event.
The NSO's control of the FWCS was inaccurate and resulted in a rapidly changing RPV water level and a decision by the US to scram the reactor. The decision by the US to scram the reactor was justified.
Operator Training and Qualification 05.1 hst in Time Eyent and Continuing Feedwater Level Control System Training a.
Insoection Secoe (97302)
The inspectors observed one session of "Just in Time" (JIT) training regarding the feedwater transient events. The inspectors also reviewed the licensed operator training material regarding operation of the condensate and feedwater, and feedwater level control systems to determine if adequate operator training had been provided. (JIT is training required by the Dresden Station Training and Operations Departments for operators prior to operating the plant after an event of this type.)
Material Reviewed:
Initial License Training (ILT) Lesson Plans;
- 216L-S1 Vessel Instrumentation-259L S1 Condensate and Feedwater
- 259LD01 Condensate and Feedwater Simulator Demonstration / Performance
- 259L-S2 Feedwater Level Control
- 259LD02 Feedwater Level Control Simulator Demonstration / Performance
- lLT S033 Low Water Level #1 Continuing Training:
S966-01 RFP Vibrations, FW Line Break in DW, Loss of MCC 281, ATWS, ED w/ failed SRV
,
-
.
.
.
.
e
.
-
x
'
. S955-04 ATWS, Failure of RWCU lsolation, Loss of all Feedwater
- OPEXW4 t Isolation Condenser Level Instrumont Failure, Feedwater System Rupture, Failure to Scram b.
Observations and Findinos On July 30 and 31,1997, JIT training was provided to licensed shift personnel prior to the re-start of Unit 2. The classroom training sessions included a genere' review of the l
circumstances and causes surro nding both feedwater transients, a dircussion of severallessons leamed, and a review uf the proposed procedural changes as a result of the events. The simulator training sessions involved instructor discussion and demonstration of FRV controller operation including a simulated walk through of the July 28 event.
Although the inspectors considered the tra!ning provided to be generally informative, the overall quality of the JIT training was considered weak as it did not addres,s the following issues and concemo:
The training offered the opportunity for the operators to practice FWCS manual
control. Since the training was not required none of the operators from the first session and only approximately 50% from the remaining sessions (as indicated by a training representative) availed themselves of this opportunity. This weakness was considered particularly noteworthy in light of feedback provided by two operators during the first JIT session in which the two operators indicated frustration because of the absence of simulator and control board continuing training for approximately the last four months as a result of the outage.
The training provided discussed the difference between the expected reactor
water level response of 2 to 3 inches versus the actual response of approximately 5 Ir'ches but provided little explanation for this difference. (The difference was attributed to the higher power level selected for the pump transfer.) Also, no detail was provided or available regarding the delayed response of reactor water level to changes in feedwater flow. Similarly, no feedback was provided regarding the potential for a simulator model response different from Unit 2 regarding the rates of change of feedwater flow and reactor water level during transient operation.
The training provided no explanation regarding the operating crew's concern with
the response of the FWCS to the RFP RV transient on July 27.
l The potential for negauve simulator training was evident as the reactor water
~
level increase modeled on the simulator is less than that experienced during the event.
The ILT leston plans reviewed provided good overall detail and explanation regarding component and instrument functions, indications, controls, and operation. The continuing training simulator lesson plan (S966-01) required the crew to demonstrate
_
.
.
.
.
proficiency on the simulator in shifting to the alternate RFP - manual FRV control not required per DOP 3200-03. However, the training lessons discussed above provided little detalled information on the integrated FWCS operation during transient conditions in single element control and the associated reactor water level response, including any delayed response, to changes in feedwater flow and/or steam flow. The lessons l
provided no operator direction to match steam flow and feedwater flow to stabilize reactor water level when in manual FRV control.
!
The inspectors reviewed the licensee's continuing training schedule for the last two years. The lessons for the FWCS and Condensate /Feodwater syvems !.ed not been provided in the two years prior to cycle 3/97. FWCS system trcining (259L S2 and 259LD02) was conducted in cycle 3/97 (on-going training cycle at the time of the event)
but had not been provided to the crew on shift during the July 28 reactor trip. FWCS modification training was conducted for all licensed operate:s in cycles 5/95 and 5/96.
The inspectors confirmed attendance for the modification training and for continuing simulator training to shift RFPs (S966-01) by the crew on shift during the reactor scram,
c.-
Conclusions The inspectors concluded that the nuality of the JIT was weak and could havo been significantly enhanced if operalors would have been requirtid to demonstrate successful l
manual control of the FWCS in the simulator. The continuirg training Information for the
!
FWCS failed to provide substar,t:al and detailed informatior regarding FWCS operating
{
characteristics during transient conditions particularly with FRV controllers in manual.
The inspectors determined that the failure to provide this level of training detail was a
!
contributing cause for this event, ll. Maintenance j
M2 Maintenance and Material Condition of Facilities and Equipment M2.1 RFP RV Failure a.
Insoection Scone (93702)
Using the guidance provided in IP 93702, inspectors reviewed plant status and interviewed plant personnelinvolved in determinbg the cause of the failure of the RFP RV.
b.
Observations and Findinas A thorough investigation of the RFP RV was conducted by the maintenance department in conjunction with the feedwater system engineer. The cause, for the failure of the RFP RV was determined to be a loose connection on a terminal board inside the controller for the RV.
i
i i
.
.
.
..
.
i
'
.
.
Maintenance Department personnel reviewed all work packages associated with the feedwater system but were unable to locate records of any work being performed inside i
the controller cabinet that may have involved the terminal board connection. The cause of the loor,e connection was attributed to vibration.
When the prompt investigation team determined the loose connection was the cause of the RV cycling, they directed maintenance personnel to tighten the connection and l
checked other similar connections within the controller cabinet. When this Information was presented to the Plant Operations Review Committee, the committee chairman directed the team to have all comparable controner cabinets checked for loose
'
connections, c.
Conclusions The inspectors agreed that the cause of the RV cycling was a loose connection in the terminal board controlling the operation of the RV. Maintenance performed during recent outages did not appear to have contncuted to its failure. Management exhibited conservative decision making by directing maintenance to check all comparable cabinets for loose connections prior to plant startup.
Ill. Engineedng E2 Engineering Support of Facilities and Egilpment E2.1 (Unit 2) Outstanding Engineering Reauests it r the Feedy* 'gr_ Control Sistem a.
insoection Scong, The inspectors reviewed planned modifications to the Unit 2 feedwater system (ER 9605207, " Modify Unit 2 FW Control System to be the Same as Unit 3") and assessed the,nodification's effect on feedwater transients, b.
Oks.ervations and Findings The purpose of the modlilcation was to upgrade the Unit 2 FWCS to make it identical to the Unit 3 FWCS. The cited benefits were to provide the same "look and feel" of the FWCS on both units, and to simplify instrument maintenance. The change included installation of three element level control. Three element level control was installed on Unit 3 during the most recent refueling outage and has been performing acceptably, c.
Conclusions
.
The inspectors were unable to determine if the modification chan0e would have made the RPV level stay within a S inch band given the same conditions,
<
1 s
.
i i
i
V. Management Meetings
!
X1 Exit Meeting Summary
{
The inspectors presented the inspection results to members of license management on t
~ August 1,1997. following the conclusion of the inspectiori;1riod. The licensee acknowledged the findings presented. No proprietary information wa. identified during the inspection or at the exit meeting, f
I r
l i
-
i
. _ - _ _ _ _ -. _ _. _.
.___. __._ _ __.
_ _ _ _. _ _ __ _ _ _ _ _.-____ _
,
.
,
- .
PARTIAL LIST OF PERSONS CONTACTED
Licensee
'
- D. Ambrer, Executive Assistant
- S. Barrett, Operations Manager
- T. Bezouska, Staff Assistant
,
'E. Carroll, NRC Coordinator i
]
'T. Eason, Operations Training Supervisor
'M. Heffley, Plant Manager
- R. Holbrook, Training Manager
- S. Kuczynski, Shift Operations SLpervisor t
i
- S. Stiles, instrument Maintenance Supervisor i
'R. Whalen, Performance Monitoring
'T. Yarbrough, Acting Maintenance Manager
'
NBC
'D. McNeil, Reactor Engineer, Rill
- J. Munro, Senior Reactor Engitorar. NRR
'D. Pelton, Resident inspector (in training), Braidwood
'D. Roth, Resident inspector, Dresden
' Denotes present at August 1,1997, exit meeting
,
,
'
INSPECTION PROCEDURES USED I
IP 93702:
Prompt Onsite Response to Emnts at Operating Power Reactors IP 71707:
Plant Operations
'
ITEMS OPENED, CLOSED, AND DISCUSSED
Opened
.
None Closed
,
None Dissunted -
None
'
3
,,
- - -.
-
--
..., - -
-
,-,-e, e-
,, -
- -- -
,,
'
.
.
.
..
LIST OF ACRONYMS USED n-fj CFR Code of Federal Regulations v;
DAP Dresden Administrative Procedure E
DEOP Dresden Emergency Operating Procedure E
DGP Dresden Ganeral Procedure DOA Dresden Operating Abnormal DOP Dresden Operations Procedure i
FRV Feedwater Reguiating Valve FWCS Feedwater Control System FWP reedwater Pump HLA Heightened Level of Awareness ILT Initial License Training IP inspection Procedure (
JIT Just in Time L
LER Licensee Event Report MWe Megawatts Electric NSO Nuclear Station Operator RFP Reactor Feed Pump RG Regulatory Guide h
}
RV Recirculation Valve SER Sequence of Events Recorder
<
_
E.
K
-
-
_
-
-
-
'
- -...
, __
.
.
.
(Unit 2)Seguence of Events Enclosure 2 There were two events that resulted in operator intervention to cor trol level. Tha first started at 15:18 on July 27, the second started at 01:30 on July 28. This serluence of events was constructed based on the control room logs, interviews with operntions staff, review of the sequence of events recorder (SER) printout, and review of the licensee's prompt investigation report. Times are based on the SER printouts.
The timeline for the translent of July 27 covers from initial observation of the reactor level perturbation to the time when operations censidered the reactor system to be stable. The timeline for the transient of July 28 begins with the HLA brief to the exit of the Dresden Emergency Operating Procedures (DEOP).
Initial Conditions:
Approximately 100 percent power (804 MWe).
Normal offsite power and all Emergency Core Cooling S stems available.
/
Reactor Feedwater Pumps A and B running.
Feedwater Control System in Master Automatic, single-element, maintaining +30" level with A and B feedwater regulating valves (FRV)in automatic and the low-flow regulating valve full closed in automatic.
15:18 2B RFP recirculation valve partially opened and diverted about 6% of feedwater flow to the condenser. No control room annunciato;9 or alarms to indicate occurrence.
t 318 Reactor Low Level alarm (+25").
i Unit 2 auxiliary nuclear station operator (Aux NSO) and Unit 3 Aux NSO report to Unit 2
" horseshoe" area in front of controls.
Unit 2 Unit Supervisor (US) directed Unit 3 Aux NSO to monitor and control the feedwater system.
Unit 3 Aux NSO and Unit 2 NSO diagnose that both FRVs are full open and that steam flow (reactor.wer)is beyond the capacity of the feedwater system. The NSOs request and receive permission from the US to decrease reactor power, 15:19 Unit 2 NSO initiated power decrease.
The 2B RFP recirculation valve closed and restored flow to the reactor.
Reactor vessel level increased to +40".
Unit 3 Aux NSO diagnosed that the FRVs were demanding 100% open. Unit 3 Aux NSO requested and received permitsion from the US to control the FRVs manually.
Unit 2 NSO ceased reouction of power to reduce changes to the reactor.
Unit 3 Aux NSO manually closed 2B FRV to 50%
..
.,
.
.
Unit 3 Aux NSO observed that level was stillincreasing, and took the 2A FRV to manual and closed it to 50%
Reactor vessel level peaked at +46", and started to decrease.
The 2B RFP recirculation valve opened and diverted flow away from the reactor.
Reactor vessellevel decreased.
Unit 3 Aux NSO opened the FRVs to restore level, Lowest level reached was +25".
The 2B RFP recirculation valve fully opened and sounded the control room annunciator.
The operator moved the 2B RFP recirculation valve control switch from " AUTO" to
"OPEN" to attempt to prevent valve cycling, The recirculation valve ceased cycling, 15:22 The Unit 3 Aux NSO stabilized level by opening the A and B FRVs to balance steam flow and feedwater flow, and allowed the low flow regulating valve to control in l
automatic.
15:27 Reactor level was stable at +30".
The next shift was tasked with swapping running feedwater pumps.
Initial Conditions:
Unit 2 was at 690 MWe.
2B RFP recirculation valve open.
All other conditions the same as the first event.
01:30 An HLA brief was performed in preparation for the transfer of RFPs (starting of 2C and securing 28), The US ordered that a manual trip (scram) be performod if level exceeded the +20" to +40" band.
After the HLA brief, the Unit 2 NSO Informed the Unit 2 Aux NSO that the Unit 2 NSO would take manual control of the FWCS if level exceeded the +25" to +35" band.
01:38:18 Upon starting the 2C RFP, feedwater flow increased from 8.3E6 to 9.0E6 lbs/hr and reactor water level began to increase.
The FWCS automatically responded to the level increase by FRV demand, but the response was not rapid enough to prevent level from increasing above +35".
01:38:49
"RPV Level Hi" recorded on SER as levelincreased above r35".
01:38 Unit 2 NSO announced and placed the feedwater control in manual (master manual). The announcement was acknowledged by the Unit 2 US.
The Unit 2 NSO placed the feedwater control in master manual and began closing both FRVs. The Unit 2 NSO did not see an instant decrease in level and therefore reduced FRV position demand more rapidly.
'
,
.
01:38:58
"RPV Level Hi-OK" recorded on SER as level decreased' below +35".
- The Unit 2 Aux NSO announced that feedwater flow had been decreased to 2 million pounds per hour.
The Unit 2 NSO responded to the announcement by rapidly opening both FRVs.
As feedwater flow increased to 11 million pounds per hour, RFP suction pressure decreased.
01:39:04
"Cond Booster PP Suct Press Lo" recorded on SER.
01:39:06
"RFP Suct Press Lo" recorded on SER.
"RFP Trip: PP 2C" recorded on SER as 2C RFP automatically trippad on low suction pressure.
01:39:07
"RPV LVL LO" recorded on SER as level decreased below +25" due to previous operator actions of closing feedwater regulating valves.
The Unit 2 US ordered a manual trip (scram).
01:39:10 Manual reactor trip was inserted, 01:39:13
"CH A/B RPV LVL LO" recorded on SER as post-scram level dropped below +8".
Group 2 and Group 3 Isolation automatically initiated.
01:39:44 Level restored above the Low level group isolations and standby gas treatment initiation setpoint.
01:40:00
"RPV LVL LO - OK" recorded on SER as level increased.
01:40:09
"RPV LVL HI" recorded on SER as levelincreased past +35".
01:4i20 Feedwater Pumps A and B both trip on high level as level increased to +51.5".
02:10-Operator manually started 2A RFP for level control.
02:25-Operator manually started reactor water cleanup system.
02:40 Exited DEOP 100
-
-
_
.
-