ML20198Q928

Forwards Test Rept 17822, PG&E - Trojan - 1986 Snubber Valve Failure Analysis, Rev 0 to Temporary Plant Test TPT-166, RCS Thermal ... & Westinghouse Amend 3,Rev 2 to Procedure ISI-205,per NRC 860529 Meeting Request

ML20198Q928
Person / Time
Site:	Trojan File:Portland General Electric icon.png
Issue date:	06/03/1986
From:	Withers B PORTLAND GENERAL ELECTRIC CO.
To:	Varga S Office of Nuclear Reactor Regulation
Shared Package
ML20198Q932	List: ML20198Q928 ML20198Q962 ML20198R016 ML20198R066
References
IEB-79-14, TAC-61405, NUDOCS 8606090420
Download: ML20198Q928 (11)

Text

o g

June 3, 1986 Eat D W hers Vca Presmrt Trojan Nuclear Plant Docket 50-344 License NPF-1 1

l Director of Nuclear Reactor Regulation Attn: Mr. Steven A. Varga Director, PWR-A Project Directorate No. 3 U.S. Nuclear Regulatory Commission Washington DC 20555

Dear Sir:

l RCS Thermal Expansion Evaluation During our meeting with you on May 29, 1986 to discuss the results of our RCS thermal expansion evaluation, additional information was requested.

The requested information is attached.

Attachment A is a description of the postulated scenario of the events leading to the RCS restrained thermal expansion. The actions being taken in response to this scenario are summarized and the bounding conditions for the thermal expansion analyses are described.

Attachment B consists of four enclosures and further discusses the cause of the steam generator hydraulic snubber failures. Additionally, further justification for reliance on the new control valves to resolve the snubber failures is provided.

In order to ensure the RCS is operable in accordance with Trojan Technical Specifications, specific inspections of the hot leg pipe whip restraints were performed. Attachment C contains a description of these supplemental inspections of the RCS hot leg pipe whip restraint anchorages and the results of these inspections.

Attachment D is a preliminary copy of the Temporary Plant Test as it has been submitted to Plant management for approval. This test will be used to monitor thermal expansion of the RCS during the forthcoming heatup and initial power operation. Monitoring will be continued for each successive heatup and cooldown of the RCS until it has been clearly demonstrated that movements are as expected.

Attachment E will be the summary report of the results of the RCS thermal expansion analyses. The summary report is not yet complete and will be forwarded as soon as it is available. Details of the analyses, including assumptions and quantitative results, have been provided separately to your consultants from Brookhaven National Laboratory.

8606090420 860603 V PDR ADOCK 05000344 g P PDR p .

12: S W Samon S:reet Pomand. Oregon 97204

a N Ge 1 erd M Coir,;eniy Mr. Steven A. Varga June 3, 1986 ,

Page 2 Attachment F is a response to your request for additional information con-cerning the ultrasonic tests of the cast stainless-steel RCS elbows.

l The independent review being conducted by Bechtel Corporation is still in {

progress. With the delay in the issuance of the summary report of the analyses, Bechtel has been unable to complete its review. The results of their independent review will be submitted prior to power operation.

During the meeting on May 29, a question was raised as to why the condition 3 of the graphite shims was not determined as part of the piping inspections and reanalysis performed in response to IE Bulletin 79-14. The Bulletin specified that inspections should be performed on ". . . elements to be used in verifying that the seismic analysis input information conforms to the actual configuration of safety-related systems". Our review of the documentation associated with that Bulletin indicates that the RCS hot leg pipe whip restraints were not inspected because they are not part of the seismic design.

We trust this information will suffice to resolve your questions and concerns on this issue.

Sincerely, ,

Bart D. Withers

• Vice President Nuclear Attachments c: Mr. John B. Martin Regional Administrator, Region V U.S. Nuclear Regulatory Commission Mr. Lynn Frank, Director State of Oregon Department of Energy

l Trojan Nuclear Plant Mr. Steven A. Varga Docket 50-344 June 3, 1986 License NPF-1 Attachment A Page 1 of 6 POSTULATED REACTOR COOLANT SYSTEM THERMAL EXPANSION SCENARIO Background Information The Trojan Nuclear Plant is a Westinghouse-designed PWR with four Reactor l Coolant System (RCS) loops. The design concept for the RCS (relative to I thermal expansion) is a freely expanding system outward from the center- ,

line of the reactor pressure vessel (RPV). A conceptual drawing of a I typical loop is shown in Figure 1. The RCS is supported at the RPV, the steam generator (SG) and the reactor e.colant pump (RCP). Additional seismic and pipe whip restraints are located on the piping and some equipment. These seismic and pipe whip supports are designed, however, to have adequate clearance so that no restraint is applied to the RCS during normal evolutions or events.

The normal thermal growth of the RCS is explained by reviewing Figure 1. i The expansion of the system is radially outward. The hot les grows l axially, moving the SG away from the RPV. The steam generator vertical l column supports are pinned to allow movement radially from the RPV with minimal lateral movement. Upper SG support is provided by snubbers which also allow radial movement along the axis of the RCS hot leg. The RCP

" floats" on its supports and accommodates movement from the cold leg and l crossover leg. The RCS is very flexible under normal design conditions l and provides relatively unrestrained thermal growth of the system.

Stresses in the system due to thermal expansion are very small during normal operating conditions. '

The hot leg between the RPV and SG represents the highest stiffness com-  ;

ponent in the system. The hot leg will expand axially with insignificant vertical and rotations 1 motion. This action is permitted because'the SG can move outward. Conditions which would cause vertical motion or rota-tion of the hot leg during normal conditions would be abnormal movement at the RPV nozzle or at the SG nozzle. The hot leg connects to the RPV  !

and.SG at locations (elevations) where these components are supported vertically. Therefore, no significant vertical displacements will occur at these locations.

During Plant inspections, three conditions have been observed at the Trojan Plant which indicate that unusual movement of the hot leg has occurred. These indications are:

• Movement of the pressurizer surge line.

j Trojan Nucliar Plant Mr. Steven A. Varga-Docket 50-344 June 3, 1986 License NPF-1 Attachment A Page 2 of 6

+ Contact between the hot les piping and the hot les pipe whip restraints.

• Cold gap measurements at various reactor coolant loop (RCL) sup-ports (upper and lower SG seismic support rings, crossover leg pipe whip restraints) which are less than the gaps predicted by thermal expansion analyses.

A review of the hot functional and startup reactor thermal expansion test results obtained in 1975 indicated that the gaps at the various reactor coolant loop supports were set consistent with the measurements taken during performance of the tests. Measurements were taken during hot functional testing in November 1975 followed by shimming after the plant had been cooled down and a reverification during the next heatup in December 1975. As noted in Attachment G, the Bechtel Corporation inde-pendent review report, there is speculation that the gaps obtained at hot conditions (550*F) may not have been extrapolated to operating conditions (617'F) prior to shimming. This speculation cannot be confirmed, how-ever. Nevertheless, during the forthcoming heatup, as a part of the monitoring program provided by Attachment D, additional data will be taken which should provide confirmation as to whether the original gaps were correct.

The surge line is attached to the hot leg on Loop B and is a very flexible line. The surge line flexibility provides an indicator of unusual hot leg movement because movements at the surge line/ hot les connection will be magnified in the surge line displacements. During the past three out-ages, the surge line has moved upward to contact surge line pipe whip restraints. This movement of the surge line, although unusual, was not a problem because the stress placed on the line was well below allowable.

Evidence of contact between the hot legs and the hot leg vertical pipe whip restraints has been obierved. Gaps between the restraints and piping were reestablished bt cold conditions for Loops A, C and D. For three loops, crushing of some of the graphite shims on the hot leg pipe whip restraints has been seen. The crushing of the shims in Loops A and D helped reestablish the gaps. The B Loop remained in contact with its restraint. Additionally, a horizontal stabilizing member on the restraint on Loop B has been pulled out from the wall. While not a major structural component, the stabilizing member pullout indicates high frictional forces between the pipe and restraint could have occurred producing the required horizontal load.

Description of Event Based on the RCS design, the event postulated to explain the observed Plant conditions is restrained thermal growth of the RCLs. The only

Trojan Nuclear Plant Mr. Steven A. Varga Docket 50-344 June 3, 1986 License NPF-1 Attachment A Page 3 of 6 component which can rer:trict growth of the RCL is the SG through inopera-bility of the snubbers or inadequate gaps on the lower seismic support ring. The postulated events are:

+ Steam generator lower support gaps are inadequate. As thermal growth of the RC3 occurs, the SGs contact the lower support bumpers such that small rotations of the SG are produced which result in movement of the hot leg. Flexibility in the lower seismic support ring reduces the effect of these inadequate gaps.

I

• The SG snubbers are degraded and become locked in the cold posi-tion. This condition causes a tilting of the SG and a significant

]

rotation at the SG-hot leg nozzle. This rotation causes the hot leg to move downward onto the hot les pipe whip restraint. For complete snubber lockup, the force at the hot leg pipe whip restraint is large enough to crush the graphite shim plates and to produce frictional forces between the pipe and restraint resulting in pull out of the horizontal stabilizing mem-ber anchors. The movement of the hot leg will also produce move-ment and rotatior. at the surge line/ hot les connection.

Due to the inadequate lateral gaps between the SG and its seismic support rings, the scenario can be further complicated to account for frictional and slippage effects between the SG and its lateral direction supports.

These factors, however, produce similar effects to those postulated above. Also, since the SG moves outward radially from the RPV at the same time it expands radially from its centerline, the effect of SG bind- i ing at the lateral suppcrts is less severe than the effects of locked l snubbers or back bump +ar contact at the SG lower seismic support ring. '

Corrective Action PGE has comprehensively addressed all causes and effects of the restrained thermal expansion of the RCS. This action has been concentrated in three major areas:

+ Returning the RCS and the snubbers to the original design condi-tions to preclude recurrence of the event.

Monitoring of RCS thermal movements during forthcoming heatups and cooldowns.

• Evaluation of the event to determine if the RCS structural inte-grity has been adversely affected.

New, corrected gaps on all key supports (hot leg pipe whip restraints, upper and lower SG seismic support rings, crossover leg restraints) have

_ _ _ _ . - _ _ _ - _ _ _ _ _ . - . _ _ . . _ _ _ = . _ _ __ ____--_ _ - - - _ - _ _ -

d

. Trojan Nuclear Plant Mr. Steven A. Varga Docket 50-344 June 3, 1986 License NPF-1 Attachment A Page 4 of 6 i

been developed, and the snubbers have been restored to an operable condi-tion. The monitoring will ensure that the snubbers operate properly and that normal thermal movements are occurring in the RCS. The monitoring l will be conducted during subsequent heatup and cooldown cycles and will ensure that restrained thermal growth of the RCS does not occur.

The RCS structural integrity is being verified by inspection and analysis.

] The analysis consists of a nonlinear evaluation of the RCL to demonstrate the strains produced by the worst-case event results in fatigue usage

]~

that satisfies current ASME Code criteria. The fatigue evaluation will combine the usage-to-date with the design usage factor from the original-Trojan stress report to demonstrate fatigue acceptability. Inspection of )

the RCS, including nondestructive examinations of the hot leg to SG noz- l

zie elbows and hot les pipe to SG elbow welds, have been performed to j i demonstrate no loss of structural integrity.

Discussion of Bounding Conditions i

The postulated scenario and inspection results were utilized to identify

the bounding conditions for further analysis of the RCL. The expansion

! of the system is a finite value dependent on temperature and geometry.

I For. the hot leg, the maximum rotation at the hot leg-SG nozzle occurs j when the snubbers are assumed locked in the cold position during normal J operation and when the snubbers are assumed locked in the hot position l during ambient conditions. Additionally, the snubber lockup assumption has a greater effect than all other possible scenarios involving poten-

! tial binding of the upper SG seismic support ring.

f Assuming snubber lockup in the cold position, the contact of the hot leg pipe whip restraint is important because the assumed gap affects the

analytical results at the hot leg-RPV nozzle and the hot leg-SG elbow.

The various combinations of hot les restraint gaps listed below were i explored to bound the results

1 1 + Gap is zero.

J l

l

• Pipe whip restraint pushes pipe upward due to rotation of-the I restraint.

!

• Gap varies from 0-0.25-in.

{' These condition: bound the effect of the hot leg pipe whip restraint gap for the analytical results.

Finally, the bumper gaps on the SG lower seismic support ring were varied

! from the minimum gaps measured to the design gaps to consider these

! effects. The load cases evaluated are shown in Table 1, and bound the worst-case occurrence.

SAB/mr/0585P f

Trojan Nuclear Plant Mr. Steven A. Varga Docket 50-344 June 3, 1986 License NPF-1 Attachment A Page 5 of 6 TABLE 1 LOADING CASES ANALYZED LOAD CASE NO.

CONDITION 1 2 3 4 5 6 7 i

CRAVITY X X X X X X X PRESSURE I X X X X X X_

HOT TEMPERATURE I X X X SNUBBERS LOCKED - COLD X X X X SNUBBERS LOCKED - HOT I HOT LEG WHIP RESTRAINT CONTACTED X X X X RADIAL S.G.

BUMPERS CONTACTED I X X j CROSSOVER LEG WHIP l

RESTRAINTS CONTACTED I I I

SAB/me 0585P l

Trojdn Nuclear Plant Mr. Steven A. Vcrga Dockst 50-344 . June 3, 1986 Lictnse NPF- I

• Attachment A

,Page 5 of 6

% ws .

8

. N. ,. .

y a:,- .

CL ~

me1 cor, y,3., -1

+----.z'. 3._

[i>f .= ,

< ~

U O

,g 9. ...

-)

, . _ L c', } .- '

f

~

T '. ~

h , ,9,,-

df  :

/S$,s

. . ...-  %.c-

. .~.

h l ,-

%.J ,

^ y l

s

,,s og h y ,

S.G. Hydrau Lic qj

, Snubbers

,,,sd,p ~'

. - D j ,,,4 b ./

W ~~ n

$ 6Y Q?A Cff ,

ff k

ik

• e Pipe Whip 1

d '(

l I" 'm ,

I

{ lh Yusia ss's

(' Restraint i , , s *%.,

c.] ..cs M'h,,'

. l t x

l ./

/

uj,' x"%

e

$$0Aff TAIC Ytt9Y O' W92 #4tMAMf L 00$ $737tM

(*rar**r scos ** r,s I e su 3 awser) ee ACd4J

g - - - , .

- - = - _ - - - - - - - - - .

Trojan Nuclear Plant Mr. Steven A. Varga Docket 50-344 June 3, 1986 License NPF-1 Attachment B Page 1 of 2 STEAM GENERATOR SNUBBER FAILURES

.The snubbers installed on the steam generators at Trojan are 900 kip units (four on each steam generator) that were manufactured by McDowell-Wellman Engineering Company (Anker-Holth). The model number of the snubber is 25.12620.008. As described previously in our May 21, 1986 letter, the ball check valves in the control valves on the snubbers were very sensitive to lockup and were determined to be a contributing cause of the snubber functional test failures this year. The control valves were included as an integral portion of the original snubbers and are identified only as part number 25.00000.016.

Our May 21, 1986 letter indicated the original control valves had been sent to a snubber test facility for further evaluation of the root cause of

! their failure. The failure analysis is now complete, and the results are

provided in Enclosure 1 to this Attachment. The report concludes the con-i trol valves perform <d inconsistently due to inadequate design. The control l valve is highly ser sitive to very small changes in differential pressure

(

due to small clears tees and weak springs in the ball check assemblies. A contributing factor to the sensitivity of the control valve was determined to be a change in hydraulic fluid viscosity when.Versilube F-50 was replaced with SF-1154 in 1985.

The snubber hydraulic fluid was changed in 1985 from Vers 11ube F-50 to SF-1154 in order to standardize the hydraulic fluid used throughout the Trojan Nuclear Plant and due to better qualities of the SF-1154 for use in a radiation environment. At the time, the viscosity differences between the two fluids was analyzed and determined to have an insignificant effect on snubber operation. In light of What is known now concerning the extreme sensitivity of the old control valves, it appears the viscosity change may'*

not have been as insignificant as originally. determined.

A significant finding of the failure analysis is that the control valves, though individually inoperable, would have allowed snubber movement when connected in series as they were installed in the Plant. This finding reveals that there was most likely no restraint to RCS thermal growth by the snubbers during the last operating cycle.

As stated in earlier correspondence, the original control valves have been replaced with a new poppet-valve style control valve. The differences in the two control valves is described in Enclosure 2, a letter from Wyle Laboratories Which describes the self-cleaning nature of the bleed grooves in the poppet valve seat to prevent the new snubbers from permanently locking up. Enclosures 3 and 4 demonstrate that the new style control valves are used extensively in the industry, and that these valves have a proven history.

1 l

Trojan Nuclear Plant Mr. Steven A. Varga 1 j . Docket 50-344 June 3, 1986 l

License NPF-1 Attachment B Page 2 of 2 i

, Our letter of May 21, 1986 also indicated water contamination (550 ppm) was

! found this year in the hydraulic fluid for the control valves. Although the source of the water contamination could not conclusively be determined, it is suspected to be the high humidity in the Containment. For that i reason, as a part of the monitoring program in Attachment D, temperature measurements will be recorded in the area of the snubbers in order to determine if dessicant filters could be installed on the hydraulic reservoir vents as a design change during the next refueling outage. As an interim

) corrective action, the hydraulic fluid was replaced with fluid verified to J

have less than 150 ppm water. In order to monitor the water content of the i fluid, we intend to sample the fluid prior to, and following, performance

] '

of the Containment integrated leak rate test, as well as during any forced outages which occur during the 1986-87 operating cycle. The fluid will also be monitored for any particulate contamination as evidence of internal corrosion of the snubbers. The new control valve clearances are signifi-cantly larger than those in the old control valves and, therefore, there is

, little reason to suspect degraded snubber performance if particulate )

contamination occurs.

l l

Enclosures

1. Snubber Valve Failure Analysis - Test Report.

2. Wyle Laboratories Letter to Mr. Arlen Wogen of Moy 15, 1986.

3. Wyle Laboratories Letter to Mr. Rodger Wehage of May 30, 1986.

,} 4 . Paul-Munroe, Inc. Letter to Mr. Arlen Wogen of May 30, 1986.

1 i

\

i i

1 l

t l

l i

i SAB/mr 0586P l

l

. I 1

Trojan Nuclear Plant Mr. Steven A. Varga Docket 50-344 June 3, 1986 License NPF-1 Attachment C Page 1 of 1 OPERABILITY DETERMINATION OF RCS HOT LEG PIPE WHIP RESTRAINT ANCHORAGES In order to ensure the RCS is operable in accordance with Trojan Technical Specifications, various inspections of the RCS were performed under cold shutdown conditions. The results of these inspections were documented in Attachment A to our May 21, 1986 letter.

Subsequent discussions with representatives of the NRC indicated a con-cern with the hot leg pipe whip restraint for the B Loop. Supplemental inspections of the RCS hot leg pipe whip resLeaint anchorages for all four loops were performed to verify the integrity of these anchorages.

On the B reactor coolant loop pipe whip restraint, portions of the epoxy coating around the base plate of the vertical and horizontal members was removed and the concrete examined. No concrete cracking or other visible damage was observed. No uplift of the concrete in the area beneath the l

baseplate of the vertical member of the B Loop was observed.

Since the hot leg pipe whip restraints were installed near the end of j construction at the Trojan Nuclear Plant, allowances for anchorages had not been considered. Instead, anchorages were prepared by cone drilling and the use of expansion bolts. Grouting was applied beneath the base plates which, when covered with epoxy, gave the appearance of uplift of the concreto.

! Finally, the anchor bolts on the horizontal members for all four reactor coolant loop pipe whip restraints were torque-checked to verify anchor integrity. It was verified that installation torque values were met in all cases.

It is PGE's position that the RCS and associated supports and restraints are operable in accordance with Trojan Technical Specifications.

CAZ/SAB/me 0588P f

- --