ML20137B123
| ML20137B123 | |
| Person / Time | |
|---|---|
| Site: | LaSalle, 05000000 |
| Issue date: | 09/16/1985 |
| From: | NRC OFFICE FOR ANALYSIS & EVALUATION OF OPERATIONAL DATA (AEOD) |
| To: | |
| Shared Package | |
| ML20136E841 | List: |
| References | |
| FOIA-85-668, TASK-AE, TASK-E513 AEOD-E513, NUDOCS 8511260195 | |
| Download: ML20137B123 (8) | |
Text
~
AE00 ENGINEERING EVALUATION REPORT
- UNIT:
LaSalle County Station EE REPORT N0.: AE0D/E513 Units 1 and 2 DATE: September 16, 1985 DOCKET NOS: 50-373, 50-374 EVALUATOR / CONTACT:
S. Salah NSSS/AE:
General Electric / Stone &
Webster
SUBJECT:
HIGH PRESSURE CORE SPRAY SYSTEM RELIEF VALVE FAILURES
SUMMARY
In 1984, LaSalle Units 1 and 2 had three failures of the High Pressure Core Spray System (HPCS) pump discharge line relief valves within a four month perind. As a result of these events, an AE0D investigation was initiated to evaluate the cause, consequences and safety significance of these failures. The results of this review indicated that excessive hydraulic forces (back pressure) from the discharge side of the relief valve caused the relief valve's internal bellows seal assembly to fail.
At'the LaSalle plant, the discharge side of the HPCS relief valve is piped into the HPCS minimum flow and full flow test return lines. The return line penetrates primary containment and returns flow in these lines back to the suppression pool. The relief valve is' located in the piping between the drywell and suppression pool, outside primary containment.
Therefore, the T
failure of the HPCS relief valve bellows seal asspmbly and subsequent
'j leakage from a 1/2 inch weep hole in the valve body caused a degradation of the primary containment boundary. The fluid leakage from the valve bellows seal also had the potential to increase the primary containment leak rate beyond the maximum value allowed rf Technical Specifications.
fo' reduce the potential for bellows failure,' the licensee has modified the
~
HPCS relief valve discharge path.
Instead of piping the discharge from the valve to the HPCS minimum flow line, the discharge will be piped to the HPCS room sump. This change will essentially eliminate the hydraulic
~~
(back) pressure applied to the bellowt seal assembly. However, the planned change will result in the relief valve discharging outside of the primary containment whenever the valve lifts. From a review of the other BWRs with a HPCS system, it is concluded that River Bend Unit 1 is the only S
other BWR potentially susceptible to the kind of relief valve failures which occurred at LaSalle.
Both plants utilize similar relief valve discharge piping arrangements and Crosby relief valves with an internal bellows seal.
- This report supports ongoing AE00 and NRC activities and does not represent the position or requirements of the responsible NRC program office.
~
8511260195 851031 PDR FOIA EMW3LLY85-668 PDR
)
1
,s j
INTRODUCTION Between June 15, 1984 and October 2,1984, there were three HPCS discharge relief valve failures at LaSalle Units 1 and 2.
These failures occurred in the internal bellows seal assembly section of the valve which provided a leakage path from the primary containment boundary through the weep hole in the relief valve bonnet.
The leakage resulted in the degradation of the primary containment boundary and could have violated the technical specification requirements for the maximum allowable leak rate out of the primary containment. The LaSalle units are the first operating Boiling Water Reactors (BWRs) to utilize a HPCS system for high pressure emergency core cooling and the first BWR to report such HPCS system problems.
Accordingly, AE00 initiated an investigation of the causes, consequences and corrective actions taken for these events. The potential safety significance and the need for follow-up actions also were evaluated.
The following sections of this, report present the results of these investiga-tions and evaluations.
f DISCUSSION i
t A brief description of the events are discussed below:
f LaSalle Unit 2 (June 15,1984)
~
(
)
At 0215 on June 15, 1984, the LaSalle Unit 2 HPCS. system was declared l
/
inoperable after fluid was observe ~d discharging from the weep hole of the HPCS discharge line relief valve (2E22-F035). Since the valve provides a boundary for primary containment integrity and the amount of leakage could not be' determined, an " Unusual Event" was declared at 1715 on June 15, and 1
a.' plant shutdown was initiated. The unit reached hot shutdown at 0200 on June 16 in accordance with Technical Specification 3.6.1.1. A subsequent investigation by the licensee determined that this leakage was due to a failed internal bellows seal in the relief valve.
)
A work request- (L37865) was written by the licensee to repair the relief valve.
The bellows seal was replaced and leak tested (no leakage) and the valve was reinstalled in the HPCS system at 1500 on June 16, 1984. The 4
HPCS system was subsequently returned to service.
~
LaSalle Unit 2 (September 6, 1984)
At 0845 on September 6,1984, the LaSalle 2 HPCS system was declared inoper-able. At the time of the occurrence, the Unit 2 HPCS syst,em was discharging l
to the suppression pool as part of a full fhw test. After less than a minute, leakage in the form of a water spray was observed emanating from the Unit 2 HPCS discharge relief valve (2E22-F035). Since the valve provides a boundary for primary containment integrity, and the amount of leakage could not be readily determined, an " Unusual Event" was declared at 00CJ. The plant
)
operating staff initiated an orderly shutdown per Technical Specification
/
3.6.1.1. - At the time of the occurrence, the reactor was in hot shutdown.
k
,I s
6
. The failure of the HPCS discharge relief valve was caused by hydraulic back pressure forces on the discharge side of the relief valve. The hydraulic forces were responsible for rupturing the internal bellows seal.
Following the failure of the HPCS discharge relief valve, a work request (L40709) was written by the licensee to repair the damaged valve.
The bellows seal was replaced and leak tested (no leakage) and the valve was reinstalled in the HPCS system. Work request L40709 was completed on September 8,1984 and the HPCS system was subsequently returned to service.
The HPCS pump was run in the minimum flow mode to verify no leakage from the relief valve weep hole.
LaSalle Unit 1 (October 2,1984)
On October 2,1984, at 0520 hours0.00602 days <br />0.144 hours <br />8.597884e-4 weeks <br />1.9786e-4 months <br />, during the performance of HPCS System Inservice Test Surveillance Procedure LOS-HP-Ql, the LaSalle Unit 1 HPCS discharge relief valve, 1E22-F035, failed, allowing leakage from the primary containment into the secondary containment. At the time of the event, the unit was in cold shutdown and primary containment was not required to be operable.
The cause of this failure was attributed to excessive back pressure on the valve bellows due to the piping arrangement on_the discharge side of the relief valve.
)
ANALYSIS AND EVALUATION In each of the three cases, the cause for the HPCS discharge relief valve failure was traced to the internal bellows seal assembly. The seal assembly
, failed as a result of excessive hydraulic}(back pressure) forces on the d'ischarge side of the relief valve.
The bellows seal is designed for a maximum back pressure of 90 psi while the actual back pressure at the seal was detennined to be about 100 psi. The excessive hydraulic forces are inherent to the original design of the LaSalle HPCS piping system and the
~
placement of the relief valve discharge. The relief valve discharge is hard piped to the suppression pool along with the HPCS minimum flow bypass line and HPCS full flow test return line. All three lines share a common line to the suppression pool as shown in Figure 1.
As a result, when the 2
system is operating with the relief valve closed, dynamic hydraulic (back pressure) forces are applied to the outside of the relief valve internal bellows seal. This was shown by a LaSalle special test (LST-84-175). The,
result of the test showed that approximately 100 psi pressure'is applied at the discharge side of the relief valve when the system is operating in the suppression pool-to-suppression pool full flow test mode. This magnitude of pressure combined with the effects of dynamic hydraulic forces resulted in the failure of the internal bellows seal which has a design pressure 4
rating of only 90 psi.
The internal structure of the LaSalle HPCS relief valve, including the 4
- 'T internal bellows seal, is shown in Figure 2.
Wnen the bellows seal ruptures, 4
I fluid is discharged from the 1/2 inch weep hole. Fluid entering the inlet
]
i.
l v,
1 IfTAltiLIENT e
4 e
f. DIlVWELL SPnAY SPARGER
\\
A0 I
a IllJECTIOff LINE g
gg CONDENSATE
'NM b
'l' SUPI'nESSION
~
it!JIICTIOff STonAGE POOL TANK 9
';)
g TEST VALVh U
t.
LillE
+-
a y
uo -Y 5
.{
h
~
TO'n CIC 7.
?.
II I f
~
,h SUCTION
~'l' e
k 11 0
'MO g
(j.
Q
_ __ g J
s::
- c y
}
7 m:
5
_\\-
j-
{
r i
s HPCS REl.fEF Y#LVE no IAlt!! MUM FLO'.'! TINE f
gli Q A.
'l' f,
1 : T g.
L'gJ
~
V i
.r-IIPCS LC
- L' PUMP MO -N.
N g
g g
lil'C5 CST TEST,
{
WATER LEc - gjo LINE i
ruur i
X UO k
i FIGUnE 1 181G11 PRESSUltG COltE SPn AY SYSTEM j
i l
e t
,I l
i
=
j t
e i
g ug
t
\\
j 1
1 e
r U,1 ti.
(g,N t.
N W. p%
s
.a w
Yz.THCH 3
g,..s/
WEEP Holt G.f~
3
=
n,.+. -
INTERIVAL SELLOWS Qlg s e,W; i
c SEAL pP([
l!.
.T o
- C 1.
wa u w....A is 1 OUTLET l
sl,!c,af 8
r
.[
ri
! i r
- 3.
1d
,y,
,i.,1
~
a e
'~:
INLe r 1
3 m.
F GURE 2.
HPCs PILiEF VALVE.
a*w
.---,m_,
- - - ~. -..
1 y...
j '
i port of the valve normally does not exert a hydraulic force on the l
inside of the bellows seal. However, fluid at the outlet port (connected L
to the minimum flow line) does exert a hydraulic back pressure force on the outside of the internal bellows seal. The excessive hydraulic back pressure from the fluid flowing in the interconnected downstream piping j
caused the internal bellows seal to fail.
The failure of the internal bellows seal did not affect the ability of the HPCS system to perform its design flow delivery function.
The HPCS i
system was still capable of meeting its flow requirement if challenged.
j A bellows seal failure does affect plant operation it, that it allows a j
leakage-path from the primary containment boundary via the weep hole in the relief valve bonnet.
The leak rate through the failed bellows was
]
evaluateil by the licensee and found to be less than the maximum allowable i
primary containment leak rate. However, if the plant were to operate with other leaks, it is possible that the combined total leak rate could exceed i
the technical specification limit.
To correct this problem, the licensee has modified the piping arrangement on the discharge side of the HPCS relief valve.
Instead of piping the relief l
valve discharge to the minimum flow line, the discharge will be piped directly into the clean radwaste sump in the HPCS room. This change will j
)
eliminate the hydraulic back pressure which was previously applied by the fluid flowing in the minimum flow line.
It is expected that this change l
will eliminate the recurrence of HPCS relief valve bellows seal assembly failures.
A' review of the High Pressure Coolant Injection system relief valve dis-charge piping arrangements at several BWR/4 plants (i.e., Browns Ferry, Pilgrim and Dresden) found that the relief valve Jischarge is also piped into the clean radwaste sump. These plants have been licensed with these arrangements which have been accepted by NRR. On this basis, it is concluded that it is acceptable to repipe the HPCS relief valve discharge to the sump.
l To evaluate the generic applicability of the LaSalle experiences to other BWRs with HPCS systems, the HPCS flow diat. rams for the Grand Gulf, Washington i
Nuclear Plant 2 (WNP-2), Nine Mile Point 2, Perry 1, Clinton 1 and River Bend 1 plants were reviewed. The respective plant diagrams indicate that each of these plants have a HPCS system with a relief valve which discharges into' the suppression pool similar to the original arrangements at LaSalle.
i Therefore, the discharge side of the relief valves at these plants may be subject to an elevated back pressure. However, an LER search of the HPCS system at the operating BWR plants with HPCS system (i.e., Grand Gulf'and WNP-2).found that only LaSalle Units 1 and 2 have reported HPCS relief I
valve internal bellow seal failures. An LER search for plants such as Nine Mile Point 2, Perry 1, Clinton 1 and River Bend 1 cannot be made at the
)
present time because of their pre-operational status.
1 4
$ p.;
~_. - e m.p,p.m9,..v.g3v_m,,_,_g,.
-,-c_%_,,.
-.---w 7.m.%,,.,,,p
,m.
,m_
mwg,_..-y_,
..g_,_9mw.y,
.opy.,,w-,,..,,,,,,.y,.
.,._,e,.w_,,
\\p]
v A detailed investigation of the valves installed at the respective operating plants was conducted to evaluate design differences which might explain the lack of failures at WNP-2 and Grand Gulf.
The investigation found that while the LaSalle plant utilizes a Crosby relief valve, both WNP-2 and Grand Gulf have relief valves manufactured by J. E. Lonergan Co.
The Lonergan valves do not utilize an internal bellows seal, which was the component which failed in the Crosby valves installed at LaSalle.
Fi nally,
the Lonergan valves have a significantly higher discharge back pressure design rating (approximately 260 psi) than the Crosby relief valves (90
-psi). The higher pressure rating and lack of a bellows seal would appear to explain the absence of reported HPCS relief valve failures at WNP-2 and Grand Gulf.
A detailed investigation of the HPCS relief valves installed at pre-operational BWR plants was also conducted.
The investigation found that two of the pre-operational plants, Perry 1 and Clinton 1, utilize, respectively, Target Rock and Dresser relief valves. Neither the Target Rock nor the Dresser relief valves have an internal bellows seal.
The Target Rock and Dresser valves also have a significantly higher maximum discharge back pressure design rating.
Therefore, for the Perry and Clinton plants, the higher back pressure rating and lack of a bellows seal in the relief valves would rule out the type of relief valve failures which occurred at the LaSalle plants.
N,,)
U Both Nine Mile Point 1 and River Bend 1 utilize Crosby valves similar to those used at LaSalle.
The Nine Mile Point 1 HPCS system relief valve discharges to a common header which also serves the RHR system relief valve discharge lines. However, the RHR system relief valve discharge 4,
l.ine would not be expected to generate a back pressure large enough to h
rupture the internal bellows of the HPCS relief valves.
Therefore, the d
potential for HPCS relief valve failures at Nine Mile Point I due to f.
rupture of the internal bellows does not appear to be significant.
There-fore, only River Bend Unit 1, with Crosby valves which discharge into the suppression pool via the minimum flow line, is considered susceptible to
~
internal bellows failure similar to those which occurred at LaSalle.
FINDINGS 1.
Excessive back pressure from the HPCS minimum flow line caused the LaSalle HPCS relief valve bellows seal assembly to fail.
2.
The failed HPCS relief valve provided a leakage path outside of primary containment.
3.
The licensee has eliminated the excessive back pressure on the dis-charge side of the HPCS relief valve by piping the discharge to the clean radwaste sump.
This appears to be an effective and acceptable corrective action.
Q)
I 4.
A review of the BWR plants with HPCS systems indicate that River Bend Station Unit 1 is the only other BWR plant which may be susceptible to HPCS relief valve failures similar to those which occurred at LaSalle.
(
CONCLUSIONS AND SUGGESTION l
=The cause of the re ief valve internal bellows seal assembly failures at LaSalle Units 1 and 2 were determined to be excessive hydraulic forces applied to the bellows seal from the discharge side of the relief valve.
Any leakage fecm the bellows seal represents a leakage path from the primary containment.
As a result of these relief valve failures, the licensee has modified the discharge piping of the relief valves for both Units 1 and 2.
The modifi-cation resulted in the relief valve discharge being directly piped into the HPCS room equipment sump. This modification will eliminate the excessive hydraulic back pressure on the bellows seal assembly and is expected to prevent bellows seal assembly failures in the future.
As a result of this study, it is suggested that Region IV be notified about the potential for HPCS relief valve failures at River Bend Unit 1.
REFERENCES 1.
Licensee Event Report 84-030, LaSalle Unit 2, Docket 50-374, July' 9, 1984.
2.
Licensee Event Report 84-068, LaSalle Unit 2, Docket 50-374, October 3 k
1984.
3.
Licensee Event Report 84-059,~LaSalle Unit 1, Docket 50-373, October 24, 1984.
6 89 6
e d
e
- ge -*
- -,*e.gw=m..
,