ML20115A138
| ML20115A138 | |
| Person / Time | |
|---|---|
| Site: | 05000000, Dresden |
| Issue date: | 03/23/1984 |
| From: | Caroline Hsu NRC OFFICE FOR ANALYSIS & EVALUATION OF OPERATIONAL DATA (AEOD) |
| To: | Seyfrit K NRC OFFICE FOR ANALYSIS & EVALUATION OF OPERATIONAL DATA (AEOD) |
| Shared Package | |
| ML20114G054 | List:
|
| References | |
| FOIA-84-616 AEOD-E406, NUDOCS 8404200375 | |
| Download: ML20115A138 (3) | |
Text
',
e.
a
~
j#amm
'o,,
NUCLEAR REGULATORY COMMISSION UNITED STATES
{
g WASHINGTON, D. C. 20555 5
.j k...+/
MAR 2 3 1984 AE0W E W MEMORANDUM FOR: Karl V. Seyfrit, Chief
~
Reactor Operations Analysis Branch Office for Analysis and Evaluation of Operational Data THRU:
Earl J. Brown, Lead Engineer Engineering Systems Reactor Operations Analysis Branch FROM:
Chuck Hsu, Engineer Engineering Systems Reactor Operations Analysis Branch
SUBJECT:
MECHANICAL SNUBBER FAILURE The enclosed Engineering Evaluation report is forwarded for your information and further consideration. This evaluation illustrates that in addition to dynamic transient overload, three other mechanisms can also damage inter-nal parts of a mechanical snubber and cause the snubber to fail in a locked-up condition similar to that which occurred at Dresden 2.
These three failure mechanisms are plant induced piping vibrations, twisting of the snubber, and excessive' bending force on the snubber.
Twisting and excessive bending fcree are related to installation and/or maintenance. The cautions against twisting and mishandling in both the manufacturer's and the licensee's installation procedures may not be adequately addressed or implemented to provide the required protection.
With regard to the piping vibration, it appears that the fatigue strength for the vibration has been overlooked in the selection of mechanical snubbers used for seismic or shock restraint. This may be related to insufficient knowledge about the plant induced piping vibration when the piping system was designed.
None of the damaged mechanical snubbers were detected by the visual in-service inspection program in plant technical specifications.
This suggests a potential inadequacy with current acceptance criteria of visual inservice inspection for safety-related mechanical snubbers with the likelihood there are undetected, locked-up mechanical snubbers in operating plants.
\\
Chuck Hsu, Engineer Engineering Systems Reactor Operations Analysis Branch phg \\
Enclosure:
As stated hpL44'6%'Nj g'
E 1
1
&w. ?F 6? l%i-bKh"dgi
'$%7
- r-
=
bec: D. Guzy, MSEB J. O'Brien, MSEB J. Richardson, MSEB R. Li, MEB-H. Shaw,' MEB R. Bosnak, MEB W. Anderson, IE R. Kiessel, IE R. Baer, IE I. Ying,-RIII W. Little, RIII G. Bagchi, NRR V. Noonan, NRR R. Lagrange, NRR D. Reiff, NRR S. Hou, NRR DISTRIBUTION:
DCS AE0D RF ROAB RF i
ROAB CF
-i CHsu EBrown..
TIppolito CJHeltemes KYSeyfrit w
i
.9
. -N
/.
/
s
'\\
/
/
onerp
- u,es) un>
.....=..... *.
w c r:x u p a no,seinac u e n o
/
OFFICIAL RECORD COPY
- " 3 " 2 " '-*~
2
A_EOD ENGINEERING EVALUATION REPORT
- UNIT:
Dresden 2 EE REPORT N0. AEOD/E406 DOCKET NO.:
50-237 DATE:
March 22, 1984 LICENSEE:
Canmonweaith Edison EVALUATOR / CONTACT:
C. Hsu NSSS/AE:
General Electric /Sargent & Lundy
SUBJECT:
MECHANICAL SNUBBER FAILURE EVENT DATE(S):
Februa ry 10, 1983 (LER 83-012/01T-0)
~_
SUMMARY
During the Dresden Unit 2 refueling outage, on February 10,1983, while performing snubber stroke tests, five safety-related mechanical snubbers instal-led on the main steam line headers adjacent to safety relief valves (SRV) in-side the drywell were found to be damaged and in a locked-up condition. The failed snubbers are Model PSA-10 manufactured by Pacific Scientific Company and had been installed during the previous refueling outage. The failures were not found on other Commonwealth Edison units utilizing the same type of mechanical snubbers.
The licensee conducted an extensive program to monitor and verify the operational transients in the four month period between March and June of 1983.
The actual cause of the failures was not determined by this program; however, it does not appear that a dynamic transient overload was the cause.
A search of the Sequence Coding Data files from 1980 to the present for failure mechanisms of mechanical snubbers made by Pacific Scientific revealed that, in addition to dynamic overloads, the effects of a snubber being twisted, piping vibration, or excessive bending forc'e have caused snubber failures.
similar to that at Dresden 2 in which the snubber became locked-up as a result of internal canponent damage.
There were a total of 11 mechanical snubbers found to be defective in this search, seven due to twisting at Virgil C. Summer 1, two due to piping vibration at Sequoyah 2 and Duane Arnold (one for each plant), and two due to excessive bending force at LaSalle County 1.
Addi tion-ally, more than 100 mechanical snubber failures were reported in response to IE Bullet.n 81-01, but were not in the LER data system.
None of the damaged snubbers were detected by virual inservice inspection during plant operation. This may indicate that the current NRC requirements for visual in. service inspection for safety-related mechanical snubbers are not adequate to verify operability of these components with the likelihood there are undetected, locked-up mechanical snubbers in operating plants.
This report suggests the following actions:
- 1. NRR should review the adequacy of the requirements for design, installa-tion, and the current visual inspection of mechanical snubbers and modi fy them as necess a ry.
' 2. IE should expedite the evaluation of the responses to IE Bulletin 81-01 and provide the results to NRR and AEOD.
- 3. IE should issue an Information Notice to alert licensees of these
-issues related to potential inoperab'ility of mechanical snubbers.
~
This document supports ongoing AEOD and NRC activities and does not represent
,1he-poshiottoriteqqents of the responsible NRC prc'; ram of fice.
\\ ct ; 4 (y \\qggyp 3 y \\
\\\\ o f '
_y INTRODUCTION LER 83-012/01T-0 reported that during the plant refueling outage on February 10, 1983, while conducting routine functional testing of a 10% sample, mechan-ical snubber f50 o'n main steam line "C" failed a stroke test.
In accordance with technical specifications, an additional 10tsof the snubbers _were stroke tested.
Four additional snubbers were found failed or inoperabl~e after a total of thirty out of thirty-nine safety related mechanical snubbers installed on the four main steam lines were stroke tested.
Among the four additional snubbers found inoperable, two were on main steam line "C" and one each on main steam lines "B" and "D".
These snubbers had been installed during the previous refueling outage. All five failed snubbers were installed in the drywell on the main steam line adjacent to the safety relief valves (SRVs) and are Model PSA-10 manufactured by Pacific Scientific.
Subsequent disassembly of these damaged mechanical snubbers revealed that except for one without damage indication (Ref.1), the damage to the other four snubbers was very extensive and included a dented dust cover, broken inertia mass snap ring, inertia mass stuck to ball screw shaft, broken torque transfer drum snap ring, inner capstan spring retainer out of place,-bent ball screw shaf t, thrust bearing inner race stripped out, and end piece' not secured to the ball nut.
The mechanical snubbers which are installed in the drywell locations are designed to accanmodate the thermal growth of the main steam line during normal operation and to provide a restraint to pipe movement during
. M-
?
transient shock and seismic events.
These snubbers do not provide a pipe support function during normal operation.
Snubbers failed in a locked-up condition will prevent the piping system, to which they are attached, from moving freely during the normal thermal heatup and cooldown associated with plant operations.
Restraining this thermal motion results in higher
-than normal stresses which, if high enough and/or repeated frequently enough, can lead to premature failure of the piping system.
5 As a result of this snubber damage on the main steam lines of Unit 2, Unit 3 was shutdown for inspection and functional testing of all PSA-10 mechanical
- snubbers on the main steam system during the first weekend of March 1983.
No snubbers were found failed in this inspection and functional test.
The cited damage appeared Ito be a type which could have resulted from the snubbers being overloaded due to a dynamic transient event from the operation of the main steam system.
These failures were not found on other Commonwealth Edison units utilizing the same type of snubbers.
In compliance with the confirmatory action letter (Ref. 2) issued by Region III, the licensee instituted a program to determine the cause(s) for the i
f ail ures.
The program included the following:
=
1 A search of the operating history to determine what operational i
transients had occurred.
2.
An analytical evaluation of the forces exerted on the snubbers by operational transients.
.i 4
h 6
O
---c.
X
~
-W n_
.y 3.
An evaluation of how potential installation deficiencies or improper materials of construction could affect snubber performance.
4.
A test program to verify the load on the snubbers by the worst operational transient (lifting of pressure relief valves).
L The program was conducted in the four month period of operation fran March
-th" June 1983, in which the following tests and evaluation were performed i
'on the main steam piping inside the drywell.
l In-Plant = Test (Refs. 3, 4, and 5)
Piping. thermal displacement was monitored fran ambient conditions a.
i prior to startup to operating tenperature along with hanger reading, i
snubber piston reading,'and snubber behavior during expansion.
b.
Piping and snubber response was monitored during the operability check of the main steam safety relief valves with the reactor at low pressure and.at normal operating pressure.
The response of the main steam system was monitored continuously from the beginning of startup until the stable power operation was attained in order to record the response of the system in the event an unexpected transient occurred in the system during power ascension.
The in-plant test results showed that:
(1) There was no installation interference or binding.
All snubber thermal movements agreed closely with expected design values and were within the working range.of the snubbers, and all movements were in the
-expected direction.
~(2) No measured response that could be considered as contributing to the failure of the five main steam snubbers was identified.
The largest:
measured snubber load was equal to 29% of the snubber operating capacity and equal.to_19% of the snubber faulted load capacity.
Based on the inalysis,of the architect engineer, the resulting piping support loads caused by the M51V closure event are bounded by the SRY discharge loads and are insignificant relative to the SRY transient loads.
(3) During thermal expansion, there was no measurable drag load on the snubbers and there was no evidence of the snubber locking up during thermal expansion.
06t-of-Plant Test (Ref. 6)
Out of plant testing was conducted at Nutech Testing Corporation.
A series of dynamic tests were performed on three mechanical snubbers, Model PSA-10, manufactured by Pacific Scientific to detennine whether or not 4
e
-4_
l r
the snubber damage was caused by an excessively high dynamic loading or l
the consequence of a possible incorrect snubber installation which results in a moment across the snubber in addition to a dynamic loading. The test results indicated that the PSA-10 mechanical snubbers sustained several dynamic impulse load events, without degradation of performance, up to two times its rated operating capacity with 10 cycles per test.
I.n addition, the unit sustained these dynamic events in combination with a 500 ft-lb L
moment.
On the other hand, the PSA-10 mechanical snubber was permanently damaged when a moment of 3000 ft-lbs was applied across the snubber, causing a lateral deflection of 0.2 inches, with a 9,900 lbs peak dynamic load. The damage observed in the test was similar to that which occurred at Dresden 2.
However, it was the licensee's evaluation that this condition could not have existed at Dresden 2.
Based on the test program results (Ref.1), the licensee concluded that:
SRV loads alone are not of sufficient magnitude to cause snubber failure, no kl.own transient that occurred during unit operation could produce the loads needed to fail the snubbers, and bending in a snubber can result in snubber lock-up and permanent degradation.
The actual cause of the snubber failures could not be determined in.the licensee's program since none of the tests or analytical measures implemented in the program could be shown to simulate or duplicate the type (s) of loading that caused the five mechanical snubbers to fail.
DISCUSSION E
Because of the bending response of the piping system to the impulse of a
(
water hammer wave in the system, water hmmner events are considered the
=y
,. most likely cause for damage to piping support restraints and joints. There-fore, a dynamic overloading was assumed as the possible cause of the snubber failure at Dresden 2 prior to the performance of the licensee's dynamic te;t and piping monitoring program. Generally, there are three types of water hammer events that have occurred in a BWR main steam line between the reactor pressure vessel (RPV) and the main steam isolation valve (MSIV) at the contain-i ment penetration (Ref. 7). These are the dynamic transients induced by SRV x.
lifting, MSIV closure, and MSIV sudden opening during the startup valve timing test.
The; impacts of dynamic transients induced by SRV lifting and MSIY closure were considered to have occurred in the licensee's program.
Since the analytical transient ldads of MSIY closure are insignificant in comparison to y
that of SRV lif ting, the snubber response loads to the dynamic transient induced by the SRV lifting was monitored. The dynamic testing results showed
- j the force magnitude to be far below the analytical predictions. A subsequent operational event of MSIY closure also indicated that there was no evidence of r
snubber f ailure following the water hammer shock as 'a result of MSIV closure.
l' The third water hammer is a steam-water entrainment type water hammer which 9
occasionally will occur when the MSIY is suddenly opened during startup
.E' valve timing test.
If this type of water hammer had ever occurred in the J]i J
main steam line, the MSIY would also have been damaged. Since there was no evidence of MSIY damage in this event, it does not appear that the snubber l
damage was caused by this type of water hammer.
Therefore, unless an unknown j !
transient had occurred, dynamic overload seems unlikely as the cause of these L
snubber failures.
M i
w s
io j.
l A search of the Sequence Coding Data from 1980 to the present was conducted for failure of mechanical snubbers manufactured by Pacific Scientific.
The search revealed that in addition to dynamic transient overloading, the i
failure mechanisms listed below have caused mechanical snubber damage similar to that found at Dresden 2.
1 Snubber being twisted due to mishandling and/or improper iristallation.
2 Plant-induced vibration.
3.
Excessive bending force.
t An event in which mechanical snubber failure was caused by improper handling and installation was reported to have occurred at Virgil C. Summer 1 (Ref. 8).
While perfonning snubber inspection per IE Bulletin 81-01, seven mechanical snubbers in safety-related systems were found to be inoperable. The affected snubbers were manufactured by Pacific Scientific Company.
The snubbers were in a locked-up condition as a result of internal parts damage. There was no evidence to indicate that overloads could have been caused by a system opera-tional transient. The licensee detennined that the probable cause of snubber lock-up was mishandling, and/or abuse during construction.
The manufacturer specifically noted that the defective snubbers appeared to have been twisted.
The telescoping cylinder is not designed to be twisted against the snubber housing.
When telescoping cylinders were twisted during installation, mechan-l ical damage occurred that caused the snubbes to lock and prevented free movement of piping.
External dents on the snubber housings also indicated that excessive force may have been applied during installation.
These failures at l
Virgil C. Summer 1 involved snubbers of relatively small sizes (PSA-1/4, l
PSA-1/2, and PSA-3) which may be more susceptible to the damage caused by mishandling or abuse. Although the damaged snubbers at Dresden 2 were a larger
(
size (PSA-10), they still.could be twisted with the use of a regular wrench according to the manufacturer. There was a possibility that the snubbers may have been twisted to facilitate attachment to the mounting brackets. The external dents on those damaged snubbers seem to support this rationale.
Discussions.with the manufacturer and the licensee revealed that a caution against twisting is included in the manufacturer's manual of installation 3
and maintenance; a special caution dealing with mechanical snubbers to avoid l
abusing or mishandling also is specified in the licensee's installation pro-i cedures.
However, usually the work of installation or handling of mechanical snubbers is perfonned by contractors.
In order to prevent or reduce the t-occurrence of snubber twisting, the caution in the procedures should be strengthened by indoctrinating personnel in the proper handling of snubbers when they work with or around them.
Two events involving Pacific Scientific mechanical snubbers with excessive piping vibration as the possible cause of snubber failure occurred at two plants, Sequoyah 2 and Duane Arnold. The snubbers were found to be binding while perfonning functional tests during a refueling outage. The event at Sequoyah 2 was reported in LER 82-62.
bu,.og an inspection and functional testing of safety-related mechanical snubbers (Pacific Scientific) in May 1982, I
47A437-4-10 (Containment Spray), 2-RHR-503 (RHR), and 2-RCH-143 (Reactor Coolant) snubbers inside containment were discovered to be inoperable. Two.
of the snubbers (47A437-4-10 and 2-RCH-143) were inoperable due to bent rod and bearing assemblies attributed to improper handling during installation and i
adj ustment. The other snubber (2-RHR-503) had a broken rod and bearing assembly attributed to excessive pipe vibration. This support was redesigned.
LER 83-10 reported the event at Duane Arnold which occurred on April 11, 1983.
One mechanical snubber DBA-5-SS-31 (Pacific Scientific) located on the head spray line of the RHR system was found binding and inoperable during. functional testing of four snubbers. This testing was being conducted because the failures had been identified during a 1981 inspection pursuant to IEB-81-01. This snubber had been inspected early in this outage and detemined to be operable in accordance with the visual inspection acceptance criteria of the plant technical specification. The functional test, however, revealed that the snubber was binding as a result of damage to some internal parts. The snubber was replaced with a new one.
In the conclusion of the licensee's engineering
- evaluation, the cause of binding of this snubber was that possibly vibration in conjunction with thermal effect results in slow deterioration.
Mechanical snubbers are susceptible to fatigue failure re'sulting from steady state low-amplitude, high-frequency vibration.
If mechanical snubbers are installed for seismic restraint alone, it is relatively easy to establish the maximum potential loading and rate the snubber accordingly. On the other hand, when water hammer or other disturbances must be anticipated in the system operation, it i's difficult to determine the loadings and establish a rating for -
_The reason for this is that the identification of cyclical or
'the snubber.
periodic disturbance and the prediction of an accurate number of operational cycles are difficult at the time of system piping design. Additionally, the NRC Standard Review Plan may not adequately define conditions in which a snubber (shock arrestor) shall also be used to accommodate adverse effects of load cycles- (vibration).
It appears in most cases that even though a potential water hammer and. its load magnitude are determined, the fatigue strength for the vibration may not be considered in the selection of a snubber.due to the difficulty in determin.ing the causes of vibration.
- The damaged mechanical snubbers on main steam headers at Dresden 2 were installed adjacent to the junction of SRV piping connections in the drywell 4
where the area is subject to plant-induced vibration.
In addition to the vibration, the high ambient temperature - up to 260*F - in the drywell could contribute significantly to the deficiency of mechanical snubbers on main steam headers. The Pacific Scientific mechanical snubber-is designed for normal
~' operation within a temperature range of -20 to 300*F. The local temperature of mechanical snubbers attached to main steam pipe in the upper drywell area might even' reach or exceed the upper design limit as a result of heat conduction through the steel pipe clamp from the main steam pipe with a temperature more than 550*F. The failures may be related to piping vibration in combina-tion with high local temperature which results in slow deterioration of fatigue strength of the snubbers.
Damage of mechanical snubbers as a result of, or related to excessive lateral l
force being applied, was indicated in two incidents which occurred at LaSalle 1
County 1.
The two events were reported in LER 82-90 and LER 83-09.
~
~
n-*s
--e-n- - - -,,
..m
,-v.---r,4,-r.
,,,e,y
,,,,--v-
-,,.,,---,.._,w v
y
-,,,n
---,-e,----
7-The' mechanical snubbers involved in these two events were manufactured by Pacific Scientific.
LER 82-90 reported that, with.the plant in startup mode at 60 MWe, mechanical snubber MS 14-1056S was found damaged during the perfomance of a hot line walkdown required by the startup test pro-cedure on August 20, 1982.
In the test procedure, a preservice inspection program was included to verify the expected snubber movement due to themal expansion of the piping at specified system temperature intervals during initial plant heatup.. The snubber was found not moving in this preservice test. The damaged snubber required an excessive amount of force to initiate movement-in a subsequent functional test.
An evaluation of the damaged snubber revealed that the failure was caused by application of excessive 4
forces in the lateral direction, thus causing the snubber to bend and/or twist. This bending and twisting resulted in excessive internal friction which caused the snubber to bind. The other similar event described in LER 83-09 occurred on February 8,1983. During the inspection of' main steam drain line piping with the unit in cold shutdown, mechanical snubber MS20CB-1 1/2-H01S was found excessively hard to stroke.
AnL evaluation of the damaged snubber revealed that the ball screw shaft was slightly bent. This was most likely the result of forces being applied in the lateral direction such as by stepping on or using the snubber as a rigging point.
The bent shaft caused excessive friction between internal parts which in turn caused the snubber to fail.
The affected piping was re-analyzed with the damaged snubber modeled as a rigid restraint.
The analysis of the first event revealed that stress and loads were dthin allowable code limits whereas the results of analysis for the second event revealed that one weld had exceeded the aTiowable usage factor and one n
additional weld may be approaching this factor.
Bending moments at snubbers are nomally generated as a result of improper installation related to alignment or inadequate clearance from surrounding structures, binding occurring during the piping thermal growth, or snubber mi sal ignments.
None of these phenomena was observed on the main steam lines inside the drywell of Dresden 2 during the licensee's perfomance of themal.
growth monitoring program in their in-plant testing (Ref. 3, 4, 5, and 6).
Therefore, excessive lateral force appears unlikely as the possible cause of
.the five snubber failures at Dresden 2.
The mechanical snubbers which were installed in the drywell location (at Dresden 2) are designed to accommodate the thermal growth of the pipe line during normal operation and to provide a restraint to pipe movement during dynamic flow transients and seismic events.
These snubbers do not provide a pipe support function during normal operation.
A survey of past snubber failures (Ref. 9) indicates that mechanical snubbers generally fail in a i
locked condition. The nomal method to provide initie.1 indication of snubber deterioration is the visual inservice inspection requirement in-
. plant technical specifications (Ref.10).
The visual inspection depends on observations to give warning of degradation; however, none of the damaged
)
snubbers in this review were identified by a visual inservice inspection.
Since the ~ damage which occurred was to. internal snubber compone'nts and could i
not. be observed, it indicated that visual inservice inspection may not be i
adequate to detect damage that could render mechanical snubbers inoperable.
c A recent Daily Event Report (Ref.11) together with infomation from the
- resident. inspector, ' identifies that a large number of Pacific Scientific mechanical snubbers failed in the functional stroke test.at Quad Cities 2.
n
,-a
,.e.-
,,w~m-n
~,
-,-~ -
^
The failed mechanical snubbers ranged in size from 1 Kip to 100 Kips. This may be indicative that damage to mechanical snubbers may be widespread and not yet discovered. <
Near completion of this evaluation, we became aware of the infomation reported by licensees in response to IE Bulletin 81-01 on surveillance of mechanical snubbers.- We also understand that IE is planning to initiate an evaluation :of the responses to the Bulletin.
Although the reports were not in great detail pertaining to the particular damage, AE0D has performed a quick review of these responses and we are able to identify approximately 115 Pacific Scientific mechanical snubber failures.- In general, failure was in a locked-up condition, and most of the damaged snubbers were discovered during the functional stroke tests.
We also observed that fewer than ten of these failed snubbers were reported by LERs, presumably because at the time the Bulletin was issued, most' plants did not have technical specification requirements to cover surveillance of mechanical snubbers. Therefore, the results of our study may not represent the true extent of the failures experier.ced by mechanical snubbers and there are likely to be undetected, locked-up mechanical snubbers in operating plants.
iThe preceding discussion has cited failures of mechanical snubbers during plant operation that may be related to functional qualification of such equipment to operate under these conditions.
In this context, there have been ongoing NRC efforts pertaining to design qualification and acceptance tests for snubbers in systems'important to safety. The proposed draft regulatory guide of February 1981 (Ref.13) developed in conjunction wi'th this effort would be applicable to all types of snubbers.
It appears that the proposed regulatory guide would address some of the failure mechanisms observed with mechanical snubbers during this evaluation.
The specific areas-of-interest in the proposed draft guide are: (1) functional speci-fication relative to the vibration environment under nomal, upset, emergency, and faulted plant condition ( Appendix A), (2) installation requirements-that could address bending and twisting ( Appendix A), and (3) qualification testing to include low-amplitude vibration for not less than 5X10 6 cycles ( Appendix B). Altfiough we do not have specific evidence that the proposed guide would have solved or prevented the operational' failures, it seems plausible that some of the problems which have developed in service would have been detected if the guidance in the
' proposed regulatory guide had been available and in place.
FINDINGS AND CONCLUSIONS The initial evaluation was based on a review of five failed mechanical snubbers in a locked-up position at Dresden 2.
An additional search identified 11 more mechanical snubber failures at four other plants with a similar locked-up failure mode.
Based on our evaluation, the following finding, and conclusions are provided:
i 1.
The licensee's program was unable to determine the failure mechanism T of the five defective mechanical snubbers at Dresden 2.
e a
-e
-,,w,,
s
F
-9_
_2 Although dynamic overloads from plant operational transients would nonnally be the expected source of excessive snubber loadings with
' subsequent damage, this study found that three other mechanisms can also cause damage to mechanical snubbers. These are plant-induced piping vibrations, twisting of the snubber, and excessive -bending force on the snubber (the latter two appear related to installation and/or maintenance).
3.
At Dresden 2 five mechanical snubbers were found failed (in a locked-up condition) due to internal part damage.
'4.
Eleven more mechanical snubbers at four other plants were found with damage similar to that found at Dresden 2.
Among these 1.1 mechanical
. snubbers, the damage to seven was due to twisting, two due to plant-induced piping vibration, and two due to excessive bending force.
5.
The damage to all mechanical snubbers described in this evaluation was discovered during the functional stroke test.
None of the damage was identified by the visual inservice inspection program in plant technical specifications (Ref.10) for these components.
6.
There may be a relationship between the damage of mechanical snubbers and local ambient temperatures which possibly exceeds their design temperature limits.
However, we were unable to identify specific data which supports this concern.
( After.this report was completed, LER 373/
83-145 identified mechanical snubber damage caused by excessive heat'.in LaSalle Unit 1 drywell.)
7.
If the proposed regulatory-guide for snubbers.had been in place and implemented, it would appear that the requirement for vibration testing might have identified the adverse effects on mechanical snubbers due to vibration during service.
It is evident from this study that mechanical snubbers can be damaged by improper installation, maintenance, or mishandling in addition to dynamic transient overloads.
Although both manufacturer's installation manuals and licensee's installation procedures provide a caution against twisting, the guidance may not be adequately covered to provide the needed protection.
Also, it appears that vibration sources have not been identified such that' fatigue strength has been overlooked in the selection of mechanical snubbers used for a seismic or shock restraint.
Some of the failures seem to suggest: that extensive' vibration was present even though the primary function of snubbers was a seismic or shock restraint.
In this respect,
-SRP 3.9.3 (Ref.12) does not appear to adequately define conditions in j
' which the snubber must also accommodate vibration effects.
As a result of this evaluation, we conclude that there is a need for NRR
~to review specific issues related to failures of mechanical snubbers to determine the adequacy of design, the procedures for installation, and the ' visual inspection requirements, and modify them as appropriate. The
~ issues related to locked-up mechanical snubbers also appears to be within the
. scope of the NRC Piping Review Coamittee (Ref.14).
These issues are as follows:
e
~ ~.
1.
Since none of the damaged mechanical snubbers were detected by the visual inservice inspection program, this may indicate that current requirements for visual inservice inspection for safety-related mechanical snubbers may not be adequate to verify the operability of these components.
We suggest that NRR review the adequacy of current requirements for visual inservice inspection'for safety-related mechanical snubbers and suggest modifications as appropriate.
- 2. _ Yibration effects appear to have been overlooked in the selection of mechanical snubbers used for a seismic or shock restraint. This may be related to insufficient knowledge about loads and the number of load cycles from plant induced piping vibrations at the time the piping system is designed. However, the failure data suggests that adverse effects are related to vibration and the issue should be reviewed.
- 3. - Although both the manufacturer's installation manuals and the licensee's installation procedures provide a caution against twisting, it appears that such cautions are not as effective as they should be. The twisting of mechanical snubbers during initial installation, maintenance, or re-installation after inservice testing 'could be prevented with proper attention to administrative controls for these procedures and manual s.
This should be investigated and implemented as appropriate.
s We recognize that this evaluation is based on a limited sample size.
However, a quick review of licensee responses to IE Bulletin 81-01 identi-fied over 100 mechanical snubber failures.
We feel confident that an evaluation of these responses would further substantiate the findings and conclusions of this evaluation.
In this regard we suggest that IE consider the following actions:
1.
Issue an Information Notice covering the salient points in this report.
2.
Expedite-the evaluation of the bulletin responses and provide the results "to NRR and AE00.
e u__
- s
-. + '
REFERENCES 1.
Letter from W. S. Little, NRC Region III, to C. Reed, Commonwealth Edison Company,. dated August 12, 1983,
Subject:
Special Safety Inspection Relative to the Main Steam Mechanical Snubber Problems Observed at Dresden Unit 2.
2
-Confinnatory Action Letter fran J. G. Keppler, NRC Region II, to C. Reed, Canmonwealth Edison Company, dated March 17, 1983.
3.
Sargent and Lundy, "Dresden Unit 2 Main Steam Monitoring Procedure, Seven-Day Data Evaluation," Rev. O, Calc. No. EMD-043449, CE Co.
5486-23, May 9,1983.
4 Sargent and Lundy, "Dresden Unit 2 Main Steam Monitoring Review of Sixty-Two-Day Outage Action," Rev. O, Calc. No. EMD-044320, Proj. No.
5486-24, July 13,1983.
5.
Sargent and Lundy, "Dresden Unit 2 Main Steam Procedure Test - Analysis Correlation," Rev. O, Report No. IMD-043846, CE Co. 5486-23, June 18, 1983.
6.
Letter from B. Rybak, Commonwealth Edison Canpany, to J. G. Keppler',
NRC Region III, dated June 6,1983,
Subject:
Dresden Unit 2 Main Steam Line Snubber Failure.
7 R. A. Uffer, Quadrex Corp., Evaluatio'n of Water Hammer Events in Lig.ht Water Reactor Plants," USNRC Report NUREG/CR-2781, July 1982.
8
. Letter from 0. W. Dixon, Jr., South Carolina Electric and Gas Co., to J. P. O'Reilly, NRC Region II, dated April 23, 1983.
Subject:
Vi rgil C. Summer Nuclear Station, Inspection per IE Bulletin 81-01.
9.
A. T. Anesto, Energy Technology. Engineering Center, " Snubber Sensitivity Study," USNRC Report NUREG/CR-2175, July 1981.
- 10. Letter from D. G. Eisenhut, NRC, to all power reactor licensees (except SEP lic'ensees), dated November 20, 1980,
Subject:
Technical Specification Revisions for Snubber Surveillance.
- 11. I&E Daily Event Report, Daily Report-RIII, December 7,1983.
- 12. USNRC, " Standard Review Plan - LWR Edition," USNRC Report NUREG-0800, Section 3.9.3, "ASME Code Class 1, 2, and 3 Component Supports, and Core Support Structures," Rev.1, July 1981.
.13.* USNRC, " Qualification and Acceptance Tests for Snubbers Used in Systems Important to Safety," USNRC Draft Regulatory Guide SC708-4, Rev.1, Februa ry 1981.
- 14. Memorandum from H. R. Denton and R. B. Minogue, NRC, to W. J. Dircks,
" Proposal for Reviewing NRC Requirements for Nuclear Power Plant Piping," July 13, 1983.
+