05000249/LER-2004-002
| Dresden Nuclear Power Station Unit 3 | |
| Event date: | 01-30-2004 |
|---|---|
| Report date: | 03-30-2004 |
| Reporting criterion: | 10 CFR 50.73(a)(2)(iv)(B), System Actuation 10 CFR 50.73(a)(2)(v)(D), Loss of Safety Function - Mitigate the Consequences of an Accident 10 CFR 50.73(a)(2)(iv)(A), System Actuation |
| 2492004002R00 - NRC Website | |
Dresden Nuclear Power Station Units 2 and 3 are General Electric Company Boiling Water Reactors with a licensed maximum power level of 2957 megawatts thermal. The Energy Industry Identification System codes used in the text are identified as [XX].
A. Plant Conditions Prior to Event:
Unit: 03 Event Date: 1-30-2004 Event Time: 1155 CST Reactor Mode: 1 Mode Name: Power Operation Power Level: 97 percent Reactor Coolant System Pressure: 1000 psig
B. Description of Event:
On January 30, 2004, the Shift Manager decided to swap the Unit 3 Main Turbine Lube Oil Coolers [TD] as the Turbine Oil Continuous Filter Differential Pressure had been increasing for several days. On January 30, 2004, at 1155 hours0.0134 days <br />0.321 hours <br />0.00191 weeks <br />4.394775e-4 months <br /> (CST), with Unit 3 at 97 percent power in Mode 1, an automatic scram occurred due to a Main Turbine trip from low lube oil pressure. The event occurred during a swapping of lube oil coolers. Immediately following the scram, the position of the Feedwater Regulating Valves (FRVs) [SJ] increased from 56 percent (%) open to 63 %. The increase in the position of the FRVs, combined with the post-scram decreasing reactor pressure, caused an increase in total feedwater flow that led to the trip of the "B" Reactor Feedwater Pump (RFP) [P] on low suction pressure. Additionally, subsequent FRVs response to increasing reactor vessel level was not fast enough to prevent the level from reaching the RFP High Level trip set point and resulted in the tripping of the "A" and "C" RFPs. Reactor water level was subsequently restored to normal and the RFPs were restarted. All rods inserted and other than the feedwater response, all other system responded as expected to the automatic scram.
An Emergency Notification System (ENS) call was made on January 30, 2004, at 1335 hours0.0155 days <br />0.371 hours <br />0.00221 weeks <br />5.079675e-4 months <br /> (CST) for the above- described scram event. The assigned ENS event number was 40491.
On February 1, 2004, at 0400 hours0.00463 days <br />0.111 hours <br />6.613757e-4 weeks <br />1.522e-4 months <br /> (CST), subsequent investigations into the January 30, 2004 event determined that the High Pressure Coolant Injection (HPCI) Systems [BJ] for Dresden Units 2 and 3 were inoperable. An evaluation by engineering determined that the Feedwater Level Control System (FWLCS) [SJ] would not maintain the post-scram reactor water level below that which would prevent water from entering the HPCI turbine steam line. Dresden Units 2 and 3 have separate HPCI nozzles in the reactor vessels that are located approximately 50 inches below the main steam nozzles. Technical Specification (TS) 3.5.1, "ECCS-Operating," requires HPCI operable in Modes 1, 2 and 3 with reactor steam dome pressure greater than 150 pounds per square inch gage (psig). At the time of discovery, Unit 2 was in Mode 1 and Unit 3 was in Mode 4.
An ENS call for Unit 2 was made on February 1, 2004, at 0854 hours0.00988 days <br />0.237 hours <br />0.00141 weeks <br />3.24947e-4 months <br /> (CST) for the above-described HPCI event.
The assigned ENS event number was 40494.
The Units 2 and 3 FWLCS post-scram level setpoints were modified on February 2, 2004 and HPCI was declared operable. Unit 3 was synchronized to the grid on February 2, 2004, at 1813 hours0.021 days <br />0.504 hours <br />0.003 weeks <br />6.898465e-4 months <br /> (CST).
These events are being reported in accordance with:
- 10 CFR 50.73(a)(2)(iv)(A), "Any event or condition that resulted in manual or automatic actuation of any of the systems listed in paragraph (a)(2)(iv)(B) of this section." The automatic actuation of the reactor protection system is listed in 10 CFR 50.73(a)(2)(iv)(B).
- 10 CFR 50.73(a)(2)(v)(D), "Any event or condition that could have prevented the fulfillment of the safety function of structures or systems that are needed to mitigate the consequences of an accident." The HPCI is a single train system and the water was in the HPCI turbine steam line for approximately 20 minutes.
C. � Cause of Event:
The root cause of the scram event was incorrect procedural guidance in Dresden Operating Procedure DOP 5100- 04 "Turbine Oil Cooler Operation." The procedure directs the operator to stop filling the oncoming Main Turbine lube oil cooler prior to swapping. This caused air to be induced into the oncoming lube oil cooler from the hot lube oil volume being cooled by cold service water, and resulted in the Main Turbine trip from low lube oil pressure.
This procedural guidance had been in place since 1991 and had been used approximately seven times since 1999. However, system realignment had only occurred once in the month of January.
The root cause of the HPCI inoperability was low margin in the FWLCS to accommodate changes to the post- scram vessel level response. The FWLCS is designed to respond to a scram by adjusting the vessel level set point from +30 inches to +5 inches and then after approximately 2 seconds, to lock the FRVs in place for approximately 15 seconds. After 15 seconds, the valve demand signal positions the FRVs at 30% of their previous position. At that time, the FWLCS reverts to controlling in the normal mode where the FRVs are positioned based on the rate of change in vessel level and the difference between the vessel level and the FWLCS set point.
Following the reactor scram on January 30, 2004, the following occurred.
- The position of the FRVs immediately increased from 56% open to 63% open during the approximately 2 seconds it takes for the FWLCS to lock the FRVs in place for 15 seconds. During this period, the increase in the position of the FRVs, combined with decreasing reactor pressure, caused an increase in total feedwater flow that led to the trip of the "B" RFP on low suction pressure. A RFP had not tripped on previous similar scrams, as the similar scrams occurred prior to the need to operate with 3 RFPs at full power.
- The FRVs began to close from 63% open at approximately 16 seconds after the scram signal due to the pulse down signal from the FWLCS to reposition the FRVs to 30% of their previous position. The FRVs never reached 30% of the previous position because at 24 seconds after the scram, FWLCS signaled the valves to reopen. At approximately 30 seconds after the scram signal the FWLCS signaled the FRVs to close. However, the rate at which the FRVs closed was not fast enough to prevent overfilling the vessel, tripping the "A" and "C" RFPs on high water level, and putting water into the HPCI steam supply line.
The FWLCS operated as designed during this event. The condition that the FWLCS had low margin to accommodate changes to the post-scram vessel level response was not known prior to this event because no analytical model capable of predicting the dynamic interaction between the FWLCS and other factors affecting vessel level was available. This resulted in the failure to adequately evaluate or test the post-scram response of the FWLCS prior to implementation of 3 RFP operation.
The immediate corrective actions for Units 2 and 3 were to lower the FWLCS post-scram vessel level set point from +5 inches to —10 inches. These set point changes provide reasonable assurance that a vessel overfill event will not recur.
The corrective action to prevent reoccurrence is to re-design the FWLCS post-scram response. Exelon Engineering will develop a dynamic model capable of accurately predicting the response of the FWLCS. This model will be benchmarked against the two most recent scrams and used to optimize the re-design. The modifications to install the improved FWLCS design will be implemented if necessary, during the next refueling outage of each unit or outage of sufficient duration after the development of the analytical model to predict the interaction of the FWLCS and post scram vessel level response.
D. Safety Analysis:
The safety significance of the scram event was minimal. All control rods fully inserted and other than the feedwater response, all systems responded as expected to the automatic scram.
The safety significance of the HPCI inoperability event was minimal. For Dresden Units 2 and 3, 2 transients and 2 design basis accidents have the potential for water carryover into the HPCI steam line and assume the availability of the HPCI for redundant long term inventory make-up. For these events, a conservative analysis has been performed using Automatic Depressurization System and low pressure Emergency Core Cooling Systems as an alternate core cooling sequence that demonstrates there is a substantial margin to predicted cladding perforation.
Therefore, the consequences of these events had minimal impact on the health and safety of the public and reactor safety.
E. Corrective Actions:
Procedure DOP 5100-04 has been revised.
The immediate corrective actions for Units 2 and 3 were to lower the FWLCS post-scram level set point from +5 inches to —10 inches.
Exelon will develop an analytical model to predict the interaction of the FWLCS and post scram vessel level response and if necessary, the FWLCS post-scram response will be modified.
F. Previous Occurrences:
A review of Dresden Nuclear Power Station Licensee Event Reports (LERs) and operating experience over the previous five years did not find any similar occurrences associated with the Main Turbine Lube Oil Coolers.
A review of Dresden Nuclear Power Station LERs identified that the most recent LER associated with the FWLCS and a reactor vessel high water level was LER 98-003-00, "Reactor Scram Results from MSIV Closure Caused by a Spurious Group 1 Isolation Signal due to Inadequate Preventive Maintenance." Following the scram, a feedwater transient occurred which resulted in water entering the HPCI steam supply line. The LER corrective actions included modifications to the FWLCS. The actions were successful in preventing water from entering the HPCI steam supply line during subsequent similar scram events when the plant was operated with 2 RFPs.
G. � Component Failure Data:
NA