05000499/LER-2002-003
| South Texas | |
| Event date: | 07-07-2002 |
|---|---|
| Report date: | 08-01-2004 |
| 4992002003R01 - NRC Website | |
DESCRIPTION OF EVENT
On July 7, 2002, Unit 2 was operating at 100% reactor power. At approximately 2300, a pre job brief was conducted in the control room in preparation for removing the Channel II battery charger number 1 from service and placing the number 2 battery charger in service. Following the pre-job brief, battery charger number 1 was removed from service and approximately one minute later battery charger number 2 was placed in service. When charger number 2 was placed in service, it exhibited an unexpectedly high float voltage of approximately 135 volts direct current (DC). The plant operator contacted the control room and received direction to secure the number 2 charger and return the number 1 charger to service. Channel II charger number 2 was secured and the charger number 2 DC output breaker was opened. Approximately four seconds later, inverter 1202 failed when the DC input fuse blew de-energizing distribution panel 1202. A voltage transient is believed to have been initiated when the battery chargers DC output breaker was opened. When the distribution panel lost power, it resulted in the loss of all Channel II instrumentation. The Unit Supervisor recognized a loss of distribution panel event, and directed the primary and secondary Reactor Operators to monitor the control boards to determine the impact to the plant. All steam generator level controls were selected to Channel II, which caused a loss of level signal to all steam generator (SG) main feedwater regulating valves (MFRV) control circuits and the SG level indicator/recorders which are the instruments operators normally use for SG level control. In response to the loss of the level signals, all MFRVs opened to the full open position. In response to the reduction of feedwater header pressure caused by the opening of the MFRVs, the main feedwater pumps went to maximum speed further increasing feedwater flow to the SGs. The Unit Supervisor observed SG water levels increasing and directed the secondary Reactor Operator to take manual control of MFRVs and control level. Failure of all the SG level indicator/recorders required the secondary Reactor Operator to use the analog indication. The secondary Reactor Operator noted the levels were highest in SG D and C.
The operator took manual control of SG D and C MFRVs and adjusted demand to successfully reduce flow and halted the level increase in SG D and C. After adjusting the feedwater flow to SG D and C, the operator took manual control of SG A and B. Steam generator 2B reached its high-high level set point resulting in the main turbine trip and the feedwater isolation signal. At approximately 2313, the Unit 2 reactor tripped automatically as a result of the main turbine trip.
Failure of the DC input fuse resulted in the loss of SG level signal to all four SG MFRV control circuits because they were all selected to the same channel. Operating the plant with all four steam generator level controls selected to the same channel allowed the SG level control function to be affected by a single point failure. STP design permits operation of SG level control from two separate channels to mitigate the effects of single point failures.
EVENT SIGNIFICANCE
This event resulted in no personnel injuries, radiation exposure, offsite radiological releases or damage to important safety related equipment. The event is reportable because it resulted in actuation of reactor protection system and a reactor trip. This was a normal reactor trip and the plant performed as expected. The SG B high-high level trip occurred at the expected setpoint.
The reactor trip resulted from main turbine trip following de-energization of vital Channel II instrumentation due to loss of power from its distribution panel. The probability of core damage is 5.7E-07 for an excessive feedwater initiating event with a non-functional instrument channel. The Conditional Core Damage Probability (CCDP) for an excessive feedwater event with all risk significant equipment available is 1.9E-07. The change in CCDP for this event is 3.8E-07. Based on these values for CCDP, the risk significance of this event is considered low.
CAUSE OF EVENT
1. The technical cause was a change in breaker E2D11/3A characteristics. As the breaker has aged, the time differential between opening of the breaker contacts has increased. As a result, the contact opening has become more sequential and less simultaneous. The change in breaker opening characteristics is age-related, which explains why the failure did not occur until 2002 even though the breaker has been installed since the plant was built. Conventional breaker testing is performed on one pole at a time. Such testing would not find any change in how the contacts open. This particular aging mechanism in a breaker is extremely rare. Normal aging effects both contacts equally such that there is little or no change in how the contacts open.
The increased time between opening of the breaker contacts allowed the inductive spike, created when relay K2 de-energized in charger E2D11-2, to impact the gating board in inverter 1202.
A technical contributing cause was the lack of suppression diodes across the DC relays in charger E2D11-2. Suppression diodes would have quenched the spike before it formed preventing the event. Suppression diodes are a common design practice for DC relay coils.
Another technical contributing cause is that the inverter is susceptible to DC noise causing the gating circuit to malfunction. Tripping of the inverter during charger swapping is an intermittent failure mode. The noise pulse created in the battery charger must form at a time during the inverter cycle when the gating control logic is susceptible to causing the oscillator circuit to go into saturation. When this occurs, the gate timing is extended and a simultaneous gate occurs.
2. The second cause of the reactor trip was having all four steam generator level control switches aligned to a single control channel coupled with the loss of power to instruments on that channel, thus, allowing the SG level control function to be affected by a single point failure.
CORRECTIVE ACTIONS
1. The distribution panel power supply was transferred to the regulating transformer and Channel II instrumentation was restored to service. Completed July 7, 2002.
2. The blown fuse in the inverter was replaced. Completed July 8, 2002.
3. The inverter was returned to operable status. Completed July 8, 2002.
4. The SG level channels used for control of the MFRVs were split between two separate control channels to minimize plant impact when a loss of a distribution panel occurs. Completed July 11, 2002.
5. Operations procedure, ESF (Class 1E) DC Distribution System, will be revised to include a statement to make operators aware that the routine changing of battery chargers could result in the loss of the associated distribution panel and to deselect the channel affected by the charger swap at the steam generator level controls. Completed.
6. A trouble shooting plan is being developed to perform testing, during the next Unit 2 outage, to measure the effects of transferring battery chargers on inverter operation to determine the root cause of the fuse failure. Due date: Complete.
7. Issue Unit 2 Design Change Packages to: 1) install suppression diodes in all class lE chargers on relays K1, K2 and K4 and 2) lift a coil lead on K3. Complete.
8. Install suppression diodes in all Unit 2 class 1E battery chargers. Complete.
9. Issue Unit 1 Design Change Packages to: 1) install suppression diodes in class 1E chargers on relays K1, K2 and K4 and 2) lift a coil lead on K3. Complete.
10.Install suppression diodes in all Unit 1 class 1E battery chargers. Complete.
11.Add diode testing to battery charger maintenance program.
12.Replace breaker E2D11/3A during 2RE11.
13.Ensure replacement inverter specification includes requirements that the new inverters are immune to both low voltage and high voltage short duration noise spikes. Complete.
ADDITIONAL INFORMATION
A similar event occurred on August 7, 2002 during operations to change the battery charger. This blown fuse did not result in a reactor trip. The corrective actions associated with this event were effective in preventing a reactor trip.
I The potential for changing battery chargers to cause a loss of the inverter was also not recognized based on a prior STP event. The previous event in 1992 was not addressed as a station problem report because it occurred with the plant shutdown. I