ML19308C016
| ML19308C016 | |
| Person / Time | |
|---|---|
| Site: | Rancho Seco, Crane |
| Issue date: | 06/30/1978 |
| From: | NRC COMMISSION (OCM) |
| To: | |
| References | |
| TASK-TF, TASK-TMR PR-780630, NUDOCS 8001170758 | |
| Download: ML19308C016 (3) | |
Text
,, s,-
>a UNITED STATES
" NUCLEAR-
[ CURRENT EVENTS.
~
' REGULATORY
^'
~~
(commission
?OWER REACTORS THIS COMPILATION OF SELECTED EVENTS IS PREPARED TO DISSEMINA MAT 10tl ON OPERATING EXPERIENCE AT NUCLEAR PCUER PLANTS IN A T THESE EVENTS ARE SELECTED FROM PUBLIC' MANNER AtiD AS OF A FIXED DATE.
NRC HAS, OR IS TAKING CONTINUOUS ACTION ON THESE INFORMATION SOURCES.
ISSUES AS APPLICABLE, FROM Af4 INSPECTION AND ENFORCEMENT, LICENSI!lG AND GENERIC REVIEW STAtiDF0 INT.
1 IRRCh - 30 APRIL 1978 (PUBLISHED JUNE 1978)
LOSS OF NONNUCLEAR IN5TRUMENTS At the Rancho Sec'6 nuclear plant on March 20, 1978, a light bulb was'
~
~
inadvertently dropped into the open light assembly cavity during replace-ment of a burned-out bulb in a backlighted push button switch on the This created a short in the "Y" portion of the 24-volt control console.
DC Non-Nuclear Instrumentation (NNI) buses. During the resulting v
current surge the current-limiting and undervoltage protection for the DC power supplies actuated, causing approximately two-thirds of the NNI signals (pressure, temperature, level, flow, etc.) to provide faulty information to the control room and Integrated Control System (ICS).
The I:5, attempting to match equipment outpu't to the erroneous signals',
This reduction caused the Reactor reduced feedwater flow to zero.
Coolant System (RCS) pressure to increase, with a reactor trip occurring wnen pressure reached the reactor protective system trip setting.
Following the trip, the operators were hampered by the lack of available, instrumentation and by equipment responding to the inaccurate signals.
Tem eratures slowly increased while pressure decayed, maintaining approxi-1 At tne loss of NNI-Y 3C power, tne "A"
steam generator mately 2000 asig.
drifted full-scale during level signal drif tec to zero and tne "B" sicnt' the first nine minutes after the trip.
At tnis time feedwater began to enter tne "A" Once-Througn Steam Generator (OTSG). The auxiliary feed-aater pump had started on the loss of feedwater flow, however, the auxiliary feedwater valves remained closed during tnis initial nine-minute perica in When the level response to tne erroneous OTSG startup level signals.
indication for the "A" 0TSG drif ted below the low level setpoint, the ICS sj-opened tne auxiliary feedwater valves acmitting water to the snell side
( Actual plant conditions showed that botn OTSGS of the steam generator.
went cry during this period.)
This flow of water created a heat sink 80011pg7gg f
-1
'~
~s/
g
.c-},
,h 4
. gggy;yg.yg
- mw
Yb
, ----m anich resulted in @epressurizationLof the RCS, initiating the Safety Features Actuation System;(SFAS) at the 1600 psig setpoint.
The operators also may have increased the main feed pump flow at this time, providing another source of water to the "A" steam generator.
On the SFAS signal both auxiliary feedwater bypass vahes opened, and water flowed to both steam generators.
The operators sontinued the injection _of water that.was started by _the SFAS signal,. maintaining:
-pressure as well.as possible utilizing the prassurizet level indica-itions and the available RCS pressure indicator.
Control was obtained'.by adjusting high pressure injection flow.
The pressurizer heaters were not available due to the NNI power loss.
The continuous injection of auxiliary feedwater resulted in complete filling of both steam generators, after which water began to enter the steam lines.
Throughout this period, th5 operators were unaware of. the.RCS1 temperature.
~
When the NNI-YiDC power supplies were restored one hour and ten minutes ~
into the transient., the RCS temperature had dropped to about 285*F (beyond the technical specification limits).
The operators took immediate action to return the system to within the op; rating limits by spraying the pres-surizer to reduce pressure, keeping three Reactor Coolant Pumps operating to increase temperature, shutting off auxiliary feedwater flow, and draining the 015Gs.
The unit remained shut down while data was gatherei and sent to B&W for analysis.
Further investigations were made at the plant.
The short caused by the light bulb drew excessive current through the 24-volt DC power supplies servicing components in NNI cabinets 5, 6 and 7.
^4
~
The power supplies for these cabinets, designated NNI-Y, are operated current-limited with a setpoint of 7.5 amps.
The subsequent reduction in voltage caused an undervoltage monitor to operate, opening the two shunt breakers through which AC power from the inverters is supplied to the DC power supplies.
Loss of these power supplies meant that every component in cabinets 5, 6 and 7 operating on DC power was not functioning properly.
An NNI signal could have been affected in two ways between its source and the receiving component.
The signal could have been interrupted completely due to a contact opening on being deenergized.
Because most of the signals are -10 volts to
+10 volts, this could have resulted in a mid-scale reading (or in some cases, a reading anywhere between +10 volts), being transmitted to the indicator or sent to the ICS as' an actuaT plant parameter.
If a signal conditionin's com-ponent (buffer amplifier, square root extractor) was affected, this meant that the desired conditioning would not be performed on the signal, or that the component might not pass the true signal, resulting in the erroneous I
values being sent to the indicator or to the ICS.
Since signal paths in the NNI are not restricted to either the X or Y cabinets, about two-thirds l
SAr>
of the signals passed through at least one component in cabinet 5, 6 or 7 and were rendered invalid.
l M
HspWW WuhWu, e #;p 4pyp w y IQ. h @ smSj-Q h W u q t
n
-m m,
r _ m _ p+n
- y u m y g g p 3 ; 3 7 g g g u y __ __
7 1
i I
i a These spurious readings had several effects. Given the changing plant
. conditions. and tne wide variety _of possible errors introduced, it was cifficult for the operators to. ascertain wnich-of their indications were valid. The plant nad to be controlled using the select few param-eters that were known. Another effect was that the equipment was operated automatically, without egard to actual conditions, since the sourious sicnals were fed into the ICS. The first indication of this was the ruhtack of the main feed pumps to zero, causing the reactor trip. Later, the automatic actions involved with adding feedwater to the dry steam generators hindered operator actions, and precipitated the rapid depressurization leading to the SFAS initiation.
Power was finally restored to the fillI-Y when the operators remedied the open shunt breakers situation. The fl!11 was then returned to operation, permitting' proper operator respcnse to the plant condition.
Babcock and Wilcox analyzed the consequences of the temperature transient; imposed upon the primary coolant system.
It was concluded that the reactor could be returned to power, provided limits were placed on maneuvering t
during the first startup, surveillance of loose parts monitors was increased for the' first week of operation, an operability check of on-line and redundant lifil was completed, and a week-long, daily sur-veillance check for leaking components in the primary and secondary radiochemistry was performed.
In addition, a 2255 psig leak test was y
performed on the RCS, the overvoltage trip setpoints on the l4fil-DC power
~
supplies were increased from 27 volts to 29 volts to prevent spurious trips, and the casualty procedures were rewritten to provide required operator actions for restoration of littI power following a trip similar to that experienced. A procedure was also written providing operator 2
instructions if the lit 11 power could not be restored.1' DIESEL MODIFICATIDriS On four different occasions in January 1978, the D. C. Cook Unit 2 diesel generator tripped on overspeed during preoperational testing.
The health and safety of tne general public was not tnreatened since the unit was in the preoperational phase.
The trips resulted from a spurious activation of the overspeed trip device (Dynalco - Model RT-2339 Relay Tachometer), caused by the switching noise on the 250 volt DC power supply system. A design change was implemented, 3*
adding capacitors from the power input terminals of RT-2339 to ground.
F The chance seemed to attenuate the "clitcn" beinc fed into the tachometer,
'7' but additional testing resulted in anotner oversheed trip. A cifferent j
tilter scheme was tried witnout success.
p 7
_g.
121 SYy%,f a,~ z pik.4%
..wg -m
~--
. _ _. _ =
m J_ c w J u d u.
's )
- s
- s
.