ML20054A059
| ML20054A059 | |
| Person / Time | |
|---|---|
| Site: | Hatch  |
| Issue date: | 09/22/1981 |
| From: | ORGANIZATION FOR ECONOMIC COOPERATION & DEVELOPMENT |
| To: | |
| Shared Package | |
| ML19240B432 | List: |
| References | |
| FOIA-81-380 NUDOCS 8204150253 | |
| Download: ML20054A059 (5) | |
Text
-
~.
\\;
s,
,. +.,
', T-
.:.;,; %,~ C v.-,.
.. ~. % ~!
. g,
..;.'c~
- s ~
/,~..
t
.A. J. 'w i
3: :
ii!.
=
a ;
e,. 9.,n,. ; -
e
. e
..c.+.
- ... n. o n
,+ r. 3.r. 1 - +.,
^*
,v...-
,. ~.
r y - ' ;i.:4...~:.q
- 3. q ;,.,,, 5 '. A,. '
~.
~
'g
,n;...cy.; +.
,,,,. ;.,. i x c' :
+.. a:..
..y -
c, : -
.. c..
,e ; e.
-u v.. o. t.j f. e)... :. (.t; e\\-
- +.*',
- ,4 3 i;.
,',e.^*(:
~, -.
s*'**
C.; -
- r m
s La~. '..cr.
- '.'i.$). n j g'!l,.. n. :.yl% ~*.$.Cl.N:.:.L '._:iq:l.~ s
.;_ - ',_, ;l;Q{
am
~
M:5.sl.f i.':
.. - -..-.:.. :..r. :
6tP 2 21981 No. IRS 58 RESTRICTED DIFFUSlON RESTREINTE Date of Receipt 8th September 1981 Date de Rdception Name of nuclear power station Nom de la centrale E.I. Hatch 2"(USA)
Date of incident 23rd March 1981 Date de l' incident
.e l
Type of reactor BWR Type de rdacteur l
l Authorized electrical power l
i Niveau de puissance diectrique 784 MWe autorisd l
1 1
1 First commercial operation 1
Date de mise en service September 1979 Included is the extraction from " Power Reactor Events",
USNRC, Vol. 3, No. 3, 1981, with the approval of the IRS i
Coordinator' in the United States 8204150253 811109 PDR FOIA CONNOR B1-380 PDR 4765
4 ISOLATION OF DRYWELL HIGH PRESSURE SWITCHES DURING OPERATION On March 23, 1981, with the reactor operating at near full power, an instrument technician investigating a drywell low pressure indication at E. I. ' Hatch Unit 2* identified eight instrument panel valves as being improperly closed.
These valves isolated safet'y related instruments which are designed to auto-matica11y actuate reactor protection, the main control room environmental control system, emergency core cooling, and primary and secondary containment isolation systems if a loss-of-coolant accident (LOCA) should occur in the containment.
Following identification of the closed valves, the valves were opened and the correct lineup was verified immediately. The reactor continued full power operation.
It was subs ~equently determined that the valves had been closed for a period of 14 days..With loss of the high drywell pressure instrumentation, automatic initiation of the automatic depressurization system (ADS), which functions during a small break LOCA, would not occur and a scram signal would not be actuated by the high drywell pressure.
Had an actual LOCA occurred during this~ 14-day interval, analysis has shown that the diverse initiation signals associated with low reactor water level, in combination with timely.R and correct operator actions, would have mitigated those breaks for which the affected instruments normally provide automatic protective action.
The cause' of this event has been attributed to failure to follow a valve lineup procedure. 'On March 9,1981 the technician responsible for verifying the valve lineup for the reactor protection system / emergency core cooling system (RPS/ECCS) instruments closed.the isolation valves on the instrument panel bel _ieving they were drain valves. The valve lineup verification was performed. pursuant to Hatch Unit 2 procedure HNP-2-1004, following a' shutdown for main steam isolation valve repair work.
Shortly after reactor startup on March 14, 1981, reactor. operators observed that the primary containment low pressure annunciator did not respond to changes in containment pressure. ' The.
pressure switch for this annunciator taps into the same header as the high drywell pressure switches. The switch setting was checked and found to be within specifications.
The instrument isolation valves that previously had been closed during the RPS/ECCS valve lineup verification were not found at this time.
1 1
A second investigation was performed on March 23.
This time, the technician traced down the instrument sensing line and found the panel isolation valve closed, which isolated RPS channel Al high drywell pressure switch.
The isolation valves on.the other three instrument lines associated with high drywell pressure on the panel were checked immediately and were also found closed, as were the four high drywell pressure instrument line isolation valves on the separate redundant instrument panel.
All eight valves were opened immediately, and verification of proper valve lineup for the other RPS/ECCS instrumentation was performed.
^A 784 We BWR located 11 miles north of Baxley, Georgia, and operated by Georgia Power Company.
9
. -.... ~., - - _
2 The technician who had performed the valve lineup verification on March 9 had not performed this particular inspection before.
He had closed what he had believed to be instrument line drain valves.
Typically, on instrument racks the process sample line is routed to the top of the rack while the drain lines and drain valves exit at the bottom.
This is the case on the two panels involved, with the exception of the eight valves that were closed.
These valves are piped up with the process sample line entering the rack from the bottom, and there is no drain valve for these particular instruments. When the valve lineup was first performed, these valves were open, but the technician closed them thinking they were drain valves.
Also, the identification numbers for these valves were not listed on valve lineup procedure HNP-2-1004, although all valves on the panels are tagged with numbers.
To prevent a similar occurrence in the future, the licensee plans to (1) revise instrument valve lineup procedures for safety-related systems to include valve identification numbers and independent verification; and (2) color code valves to safety related instruments with green signifying normally closed valves, and red signifying normally open valves.1,2 In addition, the licensee described the potential consequences of the event,jn a March 30, 1981 report, The Consecuences of the Loss of High Drywell Pressure Sionals as Related to the March 23, 1981, Incioent at HNP -Unit 2.
The "Instru-mentation Logic" section of this report contains a table listing the specific safety-related equipment which initiates on high drywell pressure alone or in combination with other signals.
The equipment functions affected include primary containment isolation (not including main steam isolation valves),
secondary containment isolation (and standby gas treatment system initiation),
high drywell pressure reactor protection system scram, core spray pump start, low pressure coolant injection pump start, automatic depressurization system (ADS) actuation, reactor core isolation cooling initiation, and the main control room environmental control system initiation.
It was determined that the ADS was the only system that could not be automatically initiated by other protective instrumentation.
Another section of the licensee's report covered " Worst Case Effects of Having No Drywell Pressure Instrumentation." The LOCA pipe break inside the drywell
.l is the only postulated event that results in high drywell pressure, which would rely on the high drywell pressure sensors to initiate mitigating systems.
As stated, automatic initiation of the ADS would be completely disabled by the absence of a high drywell pressure signal.
Although it is not needed in the event of a large LOCA, it serves as a back-up system to high pressure coolant injection (HPCI)-for small-to-intermediate size break LOCA.
For those breaks in which high system pressures are sustained, the HPCI system normally would be expected to provide makeup for the inventory lost from the break and the ADS logic would be expected to automatically initiate a rapid blowdown to allow the low pressure systems to refill the vessel.
Howe'ver, with the subject instrument lines isolated, automatic ADS initiation would not occur due to the isolated drywell pressure sensors.
The operator would be receiving signals of a LOCA, such as alarms of high drywell pressure and low water level, and HPCI failure would also be indicated in the control room.
e'
,--~-.~.,y,-,r e-.--
w m~
wwv-
-w- - ~ --
- - = > " < - - - - - " - ~-
' ' " ^
3 Operators would have at least 105 seconds (the assumed ADS timer delay) to realize that ADS was needed and would not function as planned, and to manually initiate it.
The operator could delay actuation of the ADS beyond 120 seconds and still avoid violating the 2200*F peak clad temperature limit in 10 CFR 50 Appendix K.
The delay time related to 2200'F would require further analysis.
For this type of accident, the licensee concludes that the plant would respond as designed if the HPCI system performs; otherwise, operator action would be required.
Another potential concern is for such a small break inside containment that the feedwater system would be able to maintain reactor vessel water level.
In such an event, a low reactor water level condition would not occur. Assuming high drywell pr tions or scram essure were disabled, no emergency core cooling system initia-signals would occur or be received, and the plant might continue to operate at full power while the containment was heating up and pressurizing.
This would not result in a direct threat of loss of core cooling, although containment integrity eventually would be of concern.
The operators would receive enough'information and alarms to warrant a normal plant shutdown.
Thelicensee'sreportalsoreferredtoaGeneralElectriccommon-modefailu'r$
study, An Analysis of Functional Common-Mode Failures in GE BWR Protection and Control Instrumentation (NED0-10189), published in July 1970. One tailure assessed by GE-was caused by a maintenance error which functionally disabled all instrumentation for a given parameter.
Based on the conclusions of that report, which are applicable to the event at Hatch Unit 2, analyses were per-formed for four regions of the pipe break / size / location spectrum inside the primary containment.
The analyses lncluded sequences of protective events.
Region I of the analyses represented steam breaks below 0.001 ft2 and liquid breaks below 0.0007 ft ; Region II covered steam breaks above 0.001 ft 2
2 and below 0.4 ft2 and liquid breaks above 0.0007 ft2 The boundary between Regions I and II corresponds to a leak rate of 40 gpm, which is about the lowest rate t_ hat would result in a high drywell pressure signal.
Region II includes liquid breaks up to the design basis accident.
Steam breaks in Region III (above approximately 0.4 ft ) would result in a high water level 2
signal due to swell.
Steam breaks in Region IV (above approximately 0.7 ft ).
2 would result in a high steam line flow signal.
j Based on conclusions of the GE study, the licensee determined that in all 4
cases the failure to sense high drywell pressure results in other automatic j
protection.
For..small steamline breaks in the HPCI steam supply lines, failure I
would not result in high cladding temperature even though both HPCI and ADS are inoperable.
With depressurization initiated by proper operator action, the consequences will be no worse, over the spectrum of break sizes, than if the system had operated automatically.2 The licensee reacted very responsibly on discovery of this event.
Plant personnel immediately contacted the NRC Resider.t Inspector and the licensee corporate headquarters. A corporate review committee was set up to quickly evaluate and investigate the incident, and corporate senior personnel came to
4 the plant site in response to the incident. The instrument technician involved voluntarily stepped forward with vital information as to how the event occurred.
22, 1981, the NRC published IE Information Noti'ce No. 81-15, "Degrada-On April tion of Automatic ECCS Actuation Capability by Isolation of Instrument Lines,"
which summarized the Hatch event and a similar event that occurred at Peach Bottom Unit 2* on April 1, 1981.
At Peach Bottom, three valves were found closed and redundant channels were verified operable.
No licensee response to the information notice is required, but it is expected that all licensees will review the information for applicability to their operating maintenance and surveillance procedures, paying particular attention to valve alignment check-lists and requirements for independent verification of valve alignments including instrument valves.
e e
D 9
e S'
.