ML20206F998

SER in Support of Conclusion That Util Adequately Addressed Technical & Programmatic Aspects of Cause & Effects of Restrained Thermal Growth of Rcs.Ser Re Util Response to TMI Item III.D.3.4, Control Room Habitability Also Encl

ML20206F998
Person / Time
Site:	Trojan File:Portland General Electric icon.png
Issue date:	06/30/1986
From:	Office of Nuclear Reactor Regulation
To:
Shared Package
ML20206F917	List: ML20206F998
References
TASK-3.D.3.4, TASK-TM TAC-61405, NUDOCS 8606250046
Download: ML20206F998 (41)

Text

.

Attachment 1 SAFETY EVALUATION OF INOPERABLE STEAM GENERATOR SNUBBERS ON THE TROJAN NUCLEAR PLANT REACTOR COOLANT LOOP by U.S. Nuclear Regulatory Comission Office of Nuclear Reactor Regulation June 1986 i

t h-i 8606250046 860616 PDR ADOCK 05000344 1 S PDR

1 TABLE OF CONTENTS Pace

1.0 INTRODUCTION

I 2

2.0 BACKGROUND

AND DESCRIPTION OF LICENSCE ACTIONS....................

6 3.0 STAFF EVALUATION OF LICENSEE ACTI0NS..............................

3.1 Evaluation of Supplemental Inspection Resul ts. . . . . . . . . . . . . . . . . . 6 3.2 Evaluation of RCL Thermal Expansion and Hot Leg Elbow Analyses...................................................... 20 3.3 Evaluation of Steam Generator Hydraulic Snubbers............... 27 3.4 Evaluation of RCS Thermal Expansion Monitoring Program......... 31 3.5 Evaluation of Independent Peer Review. . . . . . . . . . . . . . . . . . . . . . . . . . 33

4.0 CONCLUSION

S........................................................ 35

5.0 REFERENCES

......................................................... 37

.k 9

i

1 i l

l

1.0 INTRODUCTION

In LER 85-13 dated January 7, 1986, the licensee for the Trojan Nuclear Plant described an event related to high failure rates of mechanical and hydraulic snubbers found during snubber inservice testing conducted in 1985. In parti-cular, the 900-kip Anker-Holth steam generator hydraulic snubbers were found in a significantly degraded condition. The degraded condition of the steam generator snubbers may have contributed to a high stress condition of the reac-tor coolant system (RCS) hot leg to "B" steam generator pipe elbow for the worst case. In addition, structural failures found in RCS hot leg pipe whip restraints and erratic thermal movements of the pressurizer surge Ifne may have been a direct effect of the inoperable steam generator hydraulic snubbers.

Trie actions taken by the licensee and the results of its programs for addressing the potential safety concerns resulting from the inoperable steam generator snubbers were provided to the staff in letters from B. D. Withers (Portland General Electric) to S. A. Varga dated May 9, 1986, May 21, 1986, June 3, _

1986, and June 6, 1986. The staff has completed its review of the licensee's actions and the details of our safety evaluation of the impact of the inoperable steam generator hydraulic snubbers on the Trojan reactor coolant system is provided herein.

l \

5 I

1 l

1 l

2.0 BACKGROUND

AND DESCRIPTION OF LICENSEE ACTIONS 1 In June 1982, the pressurizer surge line at the Trojan Nuclear Plant was modi-

! fied at the surge line connection to the reactor coolant loop (RCL). Subse-quent inspections and the need for continual reshirmning of the clearances between the surge line and its pipe whip restraints identified erratic thennal movements of the surge line. A walkdown of the surge line at the beginning of the 1985 refueling outage revealed that erratic thermal movement continued. In May 1985 the licensee hired a consultant (Impell Corporation) to evaluate and analyze the surge line in order to determine the cause of the inconsistent thermal movements.

Impell reviewed possible causes of the displacement pattern and' concluded that small bending rotations of the RCL could produce the observed surge line dis-placements. However, the surge line evaluation did not conclusively identify the cause of the RCL bending rotation.

During the 1985 refueling outage, testing done on two steam generator snubbers on Loop D of the RCL found the snubbers to be in a degraded condition. The snubber seals were worn and the hydraulic fluid was heavily contaminated with seal material and rust. At that time (June 1985) the apparent failure of the two steam generator snubbers to pass initial testing was attributed to a sensi-f tive lock-up rate and an overly restrictive test acceptance criterion. Based on marks found on the snubber cylinder walls during overhaul the licensee con-cluded in June 1985 that the snubbers had been moving and would not have re-stricted thermal growth of the reactor coolant system. Consequently, the snub-bers were rebuilt, retested, and reinstalled, and the plant resumed operation.

Additionally, during the 1985 refueling outage, the licensee found that the Loop B hot leg pipe rupture restraint had been damaged. ,

i During discussions with the licensee in November 1985, Impell learned o' the de-graded steam generator snubbers. Based on analyses completed in N'ovember 1985, Impell detennined that degraded snubbers restricting RCS thermal growth could 2

i have produced the erratic surge line movements and caused the hot leg rupture restraint damage. A worst-case thermal expansion analysis of the RCL was sub-sequently perfomed by Impell assuming locked-up steam generator snubbers. A linear-elastic analysis was perfomed in accordance with the ASME Boiler and Pressure Vessel Code (1980 Edition)Section III Subsection NB-3600. The analy-sis results revealed an overstressed condition of the hot leg elbow at the steam generator. Accordingly, a simplified elastic-plastic analysis was performed for this elbow in accordance with paragraph NB-3228 of the ASME Code Section III. A l maximum strain of 0.8 percent was calculated for the elbow. In addition, the i themal expansion analysis identified a potential overstress condition of the bolts in a steam generator column support. The results of the analyses were documented in Licensee Event Report (LER) 85-13 dated January 7,1986.

3 In LER 85-13 (Revision 1) dated April 1, 1986, the licensee supplemented its previous LER with the results of the steam generator support evaluation. The support evaluation further identified an overstressed condition of the column support beseplate bolts and cap screws. As a result, the if censee comitted to

]

inspect the baseplate anchorage and cap screws in the steam generator column supports during the 1986 refueling outage. The licensee also ccm itted to inspect the rupture restraints on the Loop B and Loop C RCS hot leg.

In a letter from B. D. Withers to S. Varga dated May 9, 1986 (Reference 4), the licensee provided to the staff additional clarification of the events dis-i cussed in LER 85-13 Revision 1. In particular the licensee concluded that the root cause of 1) the damage to the Loop B RCS hot leg rupture restraint, 2) the abnormal pressurizer surge line movements, and 3) the overstress condition of

! the RCS hot leg to steam generator elbow was the inoperability of the steam j generator hydraulic snubbers. Additionally, in the May 9,1986 letter (Ref. 4), the licensee provided to the staff the results of its preliminary thermal expansion analysis (Reference 1) and the basis for using a 1 percent i strain acceptance criterion for the plastic analysis of the RCS hot leg pipe 4

elbow. The limit of 1 percent strain had been previously used as nn acceptance criterion for the seismic evaluation of stainless steel piping at the San Onofre Nuclear Generating Station (SONGS) Unit I under the Systematic Evaluation Program and was found to be acceptable by the staff as reported in 3

i

!.--_______._ . _ . - _ _ - _ - . _ _ _ _ . _ _ - . _ _ _ _ _ . . . . _ . _ . ~ _ . _ _ . _ _ . - - . _ _ _ _ _ _ _ - _ _ . . . _ . . _ _ , - . - . . . , _ . .

Reference 2. The staff evaluation of the acceptability of a 1 percent strain acceptance criterion for the Trojan Nuclear Plant is provided in Section 3.2.2 of this SER.

In the May 9,1986 letter, (Ref. 4) the licensee also committed to perform sup-plemental inspections and corrective actions to verify the adequacy of the pip-ing systems and components which could have been impacted by the restricted thermal growth of the RCS.

In the May 9,1986 letter (Reference 4) the licensee informed the staff that Bechtel Power Corporation was retained to perform an independent review of the RCS thermal expansion analysis perfortned by Impell Corporation and to provide an assessment of the overall program of the licensee. Westinghouse Electric Corporation was also retained by the ifcensee to provide verification of the "as-designed" RCS stress analysis criteria, to provide recomendations regard-ing the "as-found" conditions, and to provide final hot clearance requirements consistent with the system design basis.

In a letter from B. D. Withers to S. Varga dated May 21, 1986_(Reference 5), the licensee provided to the staff the results of the supplemental inspection pro-gram described in the May 9,1986 letter (Reference 4). The results of the sup-plemental inspection program, the corrective actions taken, and the staff evalua-tion are provided in Section 3.1 of this safety evaluation report.

In the May 21, 1986 letter (Reference 5), the licensee also reported that during the current 1986 refueling outage, all 16 steam generator hydraulic snubbers were retested for freedom of movement of the snubber piston and operability of snubber control valve. All the snubber pistons were found to move freely.

However, 11 of the 16 snubbers failed the control valve test. The control valves for all 16 snubbers were subsequently replaced with new style control valves. The licensee comitted to conduct further testing and inspection of the old snubber control valves to ensure that the cause of the inoperable snubbers was positively identified and corrected. The results of the additional snubber testing and inspection and the staff evaluation.of the inoperable steam generator snubbers are provided in Section 3.3 of'this SER.

4 i

~

l l

In order to verify that the RCS will thermally expand and contract as predicted, f the licensee established a supplemental monitoring program to be implemented du.'ing plant heatups and cooldowns. The RCS themai expansion monitoring pro-gram was described in the May 21, 1986 letter (Reference 5). The licensee has comitted to continue its monitoring program for each successive heatup and cooldown of the RCS until it has been clearly demonstrated that the thermal movements are as predicted. A description of the program and the staff evalua-tion of the program is provided in Section 3.4 of this SER.

In a letter from B. D. Withers to S. A. Varga dated June 3, 1986 (Reference 6),

the licensee provided to the staff 1) a description of the postulated scenario of the events leading to RCS restrained themal expansion, 2) additional infor-mation regarding the root cause of the steam generator hydraulic snubber failures and corrective actions taken, 3) results of additional inspections of the Loop B pipe whip restraint anchorage, 4) a copy of the Temporary Plant Test for the RCS Thermal Expansion Monitoring Program, and 5) additional information regarding the ultrasonic tests conducted on the RCS cast stainless steel hot leg elbows.

In a letter from G. Zimerman (PGE) to S. Varga dated June 4,1986 (Reference 8),

the licensee provided the staff the results of its RCL thermal expansion stress analyses and fatigue evaluations. A description of the analyses results and the staff evaluation are provided in Section 3.2 of this SER.

Lastly, in a letter from B. D. Withers to S. A. Varga dated June 6, 1986 (Reference 10), the licensee transmitted to the staff the conclusions of the independent peer review conducted by Bechtel Power Corporation of the RCS thermal expansion issue. A sumary of the Bechtel conclusions and cur evaluation are provided in Section 3.5 of this SER.

I h

5

3.0 STAFF EVALUATION OF LICENSEE ACTIONS The staff reviewed the actions taken by the licensee to assess the impact of inoperable steam generator snubbers on the Trojan Nuclear Plant reactor coolant I system and other potentially affected piping systems. The staff evaluatinn was 4 based on a review of the licensee's submittals, discussions with the licensee, a visual walkdown of the reactor coolant system and audits of the analysis methodologies used for the reactor coolant loop and components. Our safety evaluation addresses the acceptability of the licensee's corrective actions and the reactor coolant loop structural integrity.

3.1 Evaluation of Supplemental Inspection Results The licensee provided in Reference 5 the results of its supplemental inspections performed on the Trojan reactor coolant system during the 1986 refueling outage.

The results of the licensee's inspections and the staff evaluation are provided below.

3.1.1 RCS Hot Leg Pipe Whip Restraints

a. A Loop Inspection Results - The RCS hot leg pipe whip restraint was found to have a gap between the pipe and the restraint. Most of the graphite shims in the restraint saddle were damaged, however, indicating at some time the pipe had contacted the whip restraint. No damage to the pipe whip re-straint structure was observed.

Corrective Action - The broken graphite shims in the restraint will be re-placedwithcarbonsteelshimsandtherequiredhotwhiprestfaintgapwill be established. .

l 6

. l l

l l

l Staff Evaluation i Based on a visual walkdown by the staff of the RCS Loop A pipe whip restraint, the staff concludes that the damage to the pipe whip restraint shims was caused by contact with the RCS hot leg piping. Based on discussions with the licensee, the pipe whip restraints were shimmed during initial plant heat-up to allow a one-tenth inch gap. It is unlikely that the contact of the shims with the RCS piping could have been attributed to installation errors. The staff believes that the observed damage is likely to have occurred as a result of reactite downward forces on the shims due to the uplift of the Loop B hot leg on the opposite side of the reactor vessel. The uplift of the Loop B hot leg is further discussed in the next inspection item (RCS Hot Leg Pipe Whip Restraint - B Loop).

The staff has evaluated the corrective action by the licensee of replacing the broken graphite shims in the rupture restraint with carbon steel shims.

Although the licensee attributes the broken graphite shims to restrained move-ment at tha steam generator upper seismic support ring, the temperature of ac-tual expansion of the ring girder is uncertain at this time. Thus, there is some degree of uncertainty of the events leading to the contact between the hot leg and the shims. Consequently, the use of carbon steel shims in all four RCS hot leg rupture restraints in lieu of graphite shims needs to be reevaluated to assess the long-tenn consequences of the steel shims on the RCS should con-tact between the shims and the hot leg recur in the future. However, for the 1986 startup, a monitoring program to measure the gaps between the hot leg and

~

the pipe restraint provides assurance that neither thermal binding nor contact will occur. Thus, the actions proposed by the licensee are considered to be acceptable.

b. B Loop InspectionResults-Thepipewasfoundtobeincontactwith'gthewhip restraint and more than half of the graphite shims were damaged. No abnormalities were observed with respect to the pipe whip restraint structure. There was no evidence of damage to the concrete in the

! 7 l

.I

vicinity of the anchors for the horizontal member of the whip restraint which had previously been pulled approximately 5/8-inch away from the shield wall (discovered in 1985).

Corrective Action - The broken graphite shims in the restraint will be re-placed with carbon steel shims and the required whip restraint gap in the hot condition will be established. Further inspection of the whip restraint anchors will be performed.

Staff Evaluation Based on the staff walkdown of the RCS Loop B hot leg pipe whip restraint, the staff concludes that both the broken graphite shims and the pull out of the hori-zontal member anchors were a result of contact between the hot leg and rupture restraint.

The staff walkdown identified a slight rise in the concrete at the baseplate of the vertical rupture restraint members. The raised concrete could have been indicative of concrete shear cone failure. Because the horizontal member of the Loop B rupture restraint pulled out of the shield wall approximately 5/8 inch, a large load is believed to have been exerted on the anchorage of the rupture restraint vertical member. As a result, the licensee initiated an additional inspection of the Loop B RCS hot leg rupture restraint anchorage.

The results are discussed in Item 3.1.11 of this SER.

The staff concludes that the pull out of the baseplate for the horizontal rup-ture restraint member was a result of frictional binding of the RCS Loop B hot leg piping with the pipe rupture restraint shims during RCS thermal expansion.

The frictional binding is likely to have occurred as a result of locked-up steam generator snubbers in Loop B. Locked-up steam generator snubbers will cause a tipping of the steam generator towards the reactor vessel, thus causing a downward displacement of the Loop B and C hot leg piping. Thef[ictional binding in Loops B and C likely occurred in the earlier stages of the RCS sys-tem heatup. The frictional binding is believed to have resulted in the pull out of the anchors of the horizontal members in Loops B and C and a leaning of 8

the pipe rupture restraint. The leaning of the rupture restraint is believed to have then resulted in a uplift of the Loop B hot leg piping. The uplift of the B loop and the resulting see-saw effect on the entire reactor coolant sys-tem would have resulted in a downward displacement in the A and D loops, thus breaking the shims there, and an upward displacement in the C Loop, thus relieving the frictional binding there.

The staff has evaluated the corrective actions proposed by the licensee. The staff evaluation of the replacement of the broken graphite shims with carbon steel shims has been previously discussed in the staff evaluation of Loop A inspection results. Based on the results of the concrete inspection for the Loop B RCS hot leg rupture restraint anchorage as reported in Reference 6 and as supplemented by the RCS thermal expansion monitoring program, the staff con-cludes that the structural failures found in the Loop B hot leg rupture re-straint have been adequately addressed and the corrective actions taken by the licensee are acceptable.

c. C Loop Inspection Results - The C pipe whip restraint had one damaged shim. A gap between the pipe and the whip restraint was confimed except at one location on the side of the pipe where horizontal contact with the whip restraint occurred. It was at this location that the damaged shim was found. There were no abnomal' ties observed on the pipe whip restraint

, structure.

i l

Corrective Action - The graphite shims will be replaced with carbon steel j l

shims and the required whip restraint gap in the hot condition will be established.

Staff Evaluation i 1 During staff walkdown of the RCS Loop C hot leg pipe whip restrain't, the staff l identified a slight pull-out of the anchor of the Loop C horizontal pipe rup- l ture restraint member (similar to Loop B but not as severe). No damage to l

9 l

the concrete surrounding the Loop C vertical member anchor was apparent. The Loop C rupture restraint shims were not damaged as found in Loops A, 8, and D.

As previously described, the staff concludes that downward displacement of Loops B and C hot legs due to Loop B steam generator snubber lockup caused frictional binding in both Loop 8 and Loop C between the hot leg and the pipe whip restraint, thus causing both the Loop B and Loop C horizontal member anchors to pull out. Subsequently, the leaning of the Loop B rupture restraint reversed the downward hot leg displacement and caused an uplift of the Loop B hot leg piping and, consequently an uplift of the Loop C hot leg. Thus, the Loop C shims did not experience a downward force of sufficient tragnitude to cause the shims to crush. However, the staff believes an uplift of the Loop B hot leg caused a large reactive downward load on the Loop A and C rupture re-straint shims, resulting in their damage. Frictional bindin; would not have occurred in Loops A and D to cause pull out of the horizontal member anchors because contact between the Loops A and D hot legs and their respective rupture restraint shims is believed to have occurred in the later stages of RCS system heat-up.

Our evaluation of the corrective action concerning the use of carbon steel shims is the same as discussed in our evaluation of the corrective action for Loop A.

i Based on the consistency of the behavior pattern of the failures found in the Loop C hot leg pipe rupture restraint with the postulated scenario described above, the staff finds the corrective actions taken by the licensee for Loop C to be adequate and acceptable,

d. D Loop Inspection Results - A gap was found between the Loop D RCS hot leg and the whip restraint. Most of the graphite shims were damaged, however, indicat-ing the pipe had contacted the pipe whip restraint. No abnonnalities were observed with respect to the pipe whip restraint structure.

I CorrectiveAction-Thedamagedgraphiteshimswillbereplacbdwithcarbon steel shims and the required gap between the pipe and the whip restraint will be established.

10

.a __ ..- . _ . - , - - - , , , , _ _ _ -

Staff Evaluation The staff evaluation of the licensee's inspection results and corrective actions

' is the same as our evaluation discussed above for the Loop A inspection results.

3.1.2 Steam Generator Supports and Restraints

a. Steam Generator Supports Inspection Results - The base plate anchor bolts and support pad cap screws on all steam generator vertical column supports were inspected and no abnormalities were discovered which could be at-tributed to restrained RCS thermal growth. There were several cap screws which had a slight gap between the head of the cap screw and the steam generator support pad on the A steam generator. There was no evidence of cap screw elongation, however. The slight gap between the cap screws and the support pad appears to be a construction anomaly.

Corrective Action - The cap screws which had gaps were torqued to the required value.

Staff Evaluation Based on the staff review of the steam generator support inspection results,

- the staff finds the observed gaps in the cap screws to be consistent with the staff's evaluation of the events associated with restrained RCS thermal growth.

1 The staff further explored the possibility that the observed gaps in the cap screws could have been caused by uplift of the hot leg pipe. Based on bounding analyses perforned, the stresses in the cap screws were found to be less than yield stress. The staff concludes that the structural adequacy of the cap screws have not been adversely affected. The corrective action taken by the licensee is, thus, deemed acceptable and adequate. [

11

b. Steam Generator Seismic Support Rings Inspection Results - The upper and lower steam generator seismic sup-port rings were inspected for proper gaps between the bumper pads.

The upper seismic support ring bumper pads serve as guides as the steam generator moves during thermal transients. The gaps measured on all four generators appear to be too small to accommodate the ex-pected thermal growth of the support ring. In this case, the steam generator movement would be restrained when the support ring expands and contacts the guides pads. The amount of thermal growth of the upper seismic support ring is not precisely known since the tempera- 1 ture of the support ring under hot conditions is not known.

The lower seismic support ring has stop pads which limit movement of the steam generator away from the reactor vessel. In the cold condition the gap between the bumper pad on the steam generators and the stop pads on the support ring appear to be too small to accomodate the expected thermal movement. Paint residue found on the bumper pads and the support ring stop pads for some of the steam generators indicate contact had occurred.

Corrective Action - The shims will be adjusted to provide adequate clear-ance on both the upper and lower seismic support rings. The temperature of the upper support ring will be monitored during power operations to detemine precisely the thermal growth of the support ring.

Stress analyses have been perfomed to detamine the effect on the RCS if the steam generator thermal movement were restrained by the stop pads on the lower seismic support ring. Two analyses were performed; one assuming themal restraint occurred at both the upper and lower seismic support ring locations and a second assuming steam generator movement is only restrained at the lower support ring. In both cases, the analyses demonstrated the strain in the hot leg elbow would be less th'an 1 percent.

f 12

en Staff Evaluation .

Because the temperature of the support ring under hot conditions is not known, it is uncertain whether the binding in the support ring caused restraint of thermal growth of the steam generators. The monitoring of the temperature of the upper support  ! by the licensee during hot conditions will verify the degree of restraint of the steam generator thermal growth. Thus, the staff finds the actions taken by the licensee to conduct temperature monitoring of the upper support ring to be appropriate.

J The bounding stress analyses performed to assess the impact of the restraint of steam generator thermal growth due to binding of the support ring provides as- ,

surance that the structural integrity of the RCS has not been adversely affected.

Based on our review of the fatigue evaluation of the reactor vessel nozzle, the staff concludes that the analyses for the thermal restraint of the steam  !

generator due to binding of the support ring is bounded by the analyses for the locked-up steam generator snubbers. Thus, the staff finds the corrective actions taken by the licensee to be acceptable.

3.1.3 Pressurizer Surge Line Pipe Hangers and Whip Restraints Inspection Results - The inspections completed in the cold condition reveal there has been no repeat of abnormal surge line movement observed during previous years. This movement was not expected to occur this year because one surge line pipe whip restraint was modified during the 1985 refueling outage to serve as a rigid seismic restraint. This restraint is being converted back to a pipe whip restraint during the current refueling outage.

Corrective Action - All gaps on pipe whip restraints and loads on hangers have been reset to original design requirements. '

f 1

13

Staff Evaluation Based on our review of the licensee's inspection results, the staff concludes that the results are consistent with our evaluation of the events associated with restrained RCS thermal growth. Thus, the corrective action taken by the licensee to restore the restraints and supports on the surge line to original design requirements is deemed acceptable.

3.1.4 Reactor Coolant Pump Supports Inspection Results - The supports for all four reactor coolant pumps were in-spected and no abnormalities were found which could be attributed to restrained RCS thermal growth.

Corrective Action - None Staff Evaluation Based on our review of the licensee's inspection results, the staff concludes that the results are consistent with our evaluation of the events associated

! with restrained RCS thermal growth. Thus, no corrective action for the reac-tor coolant pump supports is deemed necessary.

3.1.5 RCS Crossover Pipe Whip Restraints Inspection Results - The two pipe whip restraints on each RCS crossover leg were inspected and the bearing plates for several pipe whip restraints were found cocked slightly. Additionally, in the cold condition, the gap in the crossover whip restraints appears to be too small to accommodate the expected thermal movement.

Corrective Action - The plates will be adjusted to eliminate the cbcking, i

and the gap clearances will be adjusted to accommodate expected thermal growth.

The effect of the inadequate gap clearances on the RCS piping and components

has been analyzed.

14

Staff Evaluation l Based on the staff walkdown of the RCS Loop B crossover pipe whip restraints, I the staff concurs with the licensee's inspection results. The staff concludes that the direction of cocking of the Loop B pipe whip restraint base is con-sistent with the direction of tilt of the steam generators under snubber lock-i up conditions. The staff finds the corrective action to adjust the bearing plates to be acceptable and adequate. The effect of the inadequate gap clear-ance has been addressed as part of the final piping analyses results as dis-cussed in Section 3.2 of this SER.

, 3.1.6 RTD Bypass Manifold Piping Inspection Results - No abnormalities were found on the RTD bypass manifold piping.

Corrective Action - None Staff Evaluation Based on our review of the licensee's inspection results, the staff concludes that the results are consistent with our evaluation of the events associated with restrained RCS thermal growth. Thus, no corrective action for the RTD Bypass Manifold Piping is deemed necessary.

~

3.1.7 Main Steam Line Pipe Whip Restraints Inspection Results - The gaps were measured between the four main steam lines and the pipe whip restraints at the exit of the steam generators. The gaps were within allowable tolerances.

1 Corrective Action - None i h

I 15

.- . ~ . _ ._ - . . . - - ._ -

l i

(

Staff Evaluation Based on our review of the licensee's inspection results, the staff concludes i that the results are consistent with our evaluation of the events associated with restrained RCS thermal growth. Thus, no corrective action for the main steam line pipe whip restraints is deemed necessary.

i 3.1.8 RCS Hot leg to Steam Generator Elbow Weld i

Inspection Results - Liquid penetrant examinations of the welds (including the l

heat affected zones) connecting each of the RCS hot leas to the respective l steam generator elbows were performed and no indications were found.

Corrective Action - None Staff Evaluation Based on the results of the liquid penetrant examination, the staff finds the I integrity of the RCS hot leg to steam generator elbow weld to be acceptable. ,

No further corrective action is required.

i 3.1.9 Other System Connections to the RCS inspection Results - Each of the piping systems attached to the RCS were in-spected for any abnormal indications. The first 10 feet of all pipes 2.5 inches in diameter and less were inspected. The first 20 feet of all pipes greater than 2.5 inches in diameter were inspected. Pipe hangers and whip restraints, as applicable, were included in the inspections. No abnormalities were found during these inspections which could be attributed to restrained RCS thermal i

growth.

l l

! Corrective Actions - None 1

16 i

l

. [

i

- The UT data sheets indicate that there was limited contact between the '

search unit and the surfaces due to the "as-cast" condition. However, the licensee states that this refers to the surface contour of the elbow and not the smoothness of the surface.

- The licensee states that the examination performed can detect a 20-30 per-cent minimum sized through-wall indication using the 41 degree refracted L-wave technique on the cast stainless steel elbow material.

- The licensee states that attenuation differences between the calibration blocks and the cast elbows were not measured because the materials are the same.

l The staff finds the UT examination performed on the hot leg elbows lacked the quality which should have been instituted to provide an acceptable examination.

Measurement of the acoustical properties of the elbow material and calibration block material is essential for cast material even when the materials are the same specification and type. The staff also finds the signal-te-noise ratio was not measured and documented, thereby eliminating the reference to determine whether the response from a flaw could be distinguished from background noise.

Based on the above findings, the staff has determined that the UT examination performed on the RCS hot leg elbows was not adequate to detect inside diameter flaws less than thirty percent (30%) through-wall in depth. However, given the areas on which the examination was performed, the examination covered those areas most likely to contain flaws initiated by the constraint of RCS thermal expansion. In addition, a dye-penetrant examination was performed on the elbow-to-pipe weld and on a significant portion of the base metal. The staff l concludes that although the UT examination lacked the quality expected, the l combination of examinations (surface and volumetric) and the results of the  :

fatigue evaluation provided in Table 3-1 of this SER provides assurance that gross structural defects are not likely to exist in the hot leg elbows.f How- l 1

ever, the staff requires that additional examinations be performedion the RCS l Loop B hot leg elbow during the next refueling outage to ensure that small cracks which could contribute to long-term fatigue effects were not initiated 18

.-e - - -- - - - - . a .. -- - - -,- . - - , , - - - - - - - , - - - - - , - - . - -

I i

on the inside diameter of the pipe. The staff requires as a confirmatory mea-sure that the license conduct:

(1) Ultrasonic examination (with full documentation as discussed above) on the l subject elbows after properly preparing the examination surface, or 1

(2) Remote visual and liquid penetrant examinations on the inner diameter (ID) surface of the subject elbows.

3.1.11 Inspection of Steam Generator Insulation Support Frames and Supplementary Inspection of RCS Hot Leg Whip Restraint B Embedments Inspection Results - Following NRC staff inspection on May 11, 1986, the staff suggested to PGE that deflection of insulation support frames located below the steam generator channel heads may be a further indication of steam generator rotation / tilting. The staff also suggested that a sloped floor around the "B" hot leg whip restraint may have been caused by a conical pullout of the re-straint anchorage. PGE investigated these observations and has concluded:

(1) that the insulation frame (which is not rigidly supported from the steam generator or any other structure) was deflected due to the normal horizontal shift of the steam generator when going from cold to hot conditions; and (2) the sloped floor around the "B" hot leg whip restraint reflects an as-built condition since (a) there is no sign of concrete cracking or degradation and (b) an extremely high uplift load and/or abnormal load far beyond any antici-pated would be required to cause such a pullout.

Corrective Action - Additional inspection of the Loop B hot leg rupture j restraint anchorages was performed.

Staff Evaluation i

Based on further review by the staff of the insulation framing dethils, the staff concludes that the deflections of the insulation support frabs were not likely caused by steam generator tilting. However, the staff evalGation of the licensee's inspection results concerning the adequacy of the concrete 19

, . - - ,,- . - , - - , - ,m-,- -- - -- --4y--_m. , ,_ - - ,

. - + - - - --- e y ,,-,q

surrounding the Loop B hot leg rupture restraint vertical member found that possible cracking cannot be readily detected because the concrete floor is covered with a layer of epoxy paint. Furthermore, because of the potential for an extremely large displacement-controlled tension load to occur in the anchor bolts due to the binding of hot leg and pipe rupture restraint during thermal axial growth of the pipe, the staff requested that the licensee conclusively verify the condition of the concrete surrounding the Loop B pipe whip restraint vertical member anchorage. In Reference 6, 'the licensee provided the staff with the results of its additional inspection. For the B Loop hot leg rupture restraint, portions of the epoxy coating around the baseplates of the vertical and horizontal members were renoved and the concrete was examined. No concrete cracking or other visible signs of damage was observed. The apparent rise in concrete, as identified by the staff, was found to be grouting that when covered with epoxy, gave the appearance of uplift of the concrete.

Thus, based on the results of the additional inspection of the Loop B rupture anchorages as reported in Reference 6, the staff finds the corrective actions taken by the licensee to be acceptable.

3.2 Evaluation of RCL Thermal Expansion Analyses and Hot Leg Elbow Analysis As reported in LER 85-013 (Revision 1), the results of a preliminary themal expansion analysis perfomed in 1985 by the licensee revealed an overstressed condition of the elbow where the RCS hot leg enters the steam generator. Sub-l sequently, a simplified elastic-plastic analysis was performed in accordance

~

with paragraph NB-3228 of the ASME Boiler & Pressure Vessel Code Section III (Reference 7). The results of the preliminary RCL thermal expansion analyses were provided to the staff in Reference 4.

In May 1986, the licensee initiated further analyses of the RCL. The addi-tional analyses which superseded the analyses performed in 1985, included non-linear analyses using a more detailed modelling of an RCS sing)e loop, linear elastic analyses of the four loop RCS configuration, and a hlastic l (shell)analysisoftheRCShotlegelbow. The NRC staff consultahts from l Brookhaven National Laboratories conducted an audit of the methodology and i

20 l l

i acceptance criteria on May 22-23, 1986. The final results were provided to the i staff in a letter from G. Zimeman (PGE) to S. Varga (NRC) dated June 4,1986 (Reference 8).

3.2.1 Evaluation of the Analytical Methodology a The method of analysis used by the licensee for the evaluation of the reactor coolant system under restrained thermal expansion included ifnear-elastic and non-linear analysis techniques. The analyses which were perfomed by Impe11 Corporation included 1) a preliminary two-dimensional elastic analysis of the RCS hot leg and steam generator, 2) a four loop linear-elastic analysis using the SUPERPIPE computer code, 3) a one loop non-linear analysis using the ANSYS l

I computer code, and 4) a three-dimensional finite element plastic analysis of j

the hot leg elbow using the ANSYS computer code. The evaluations also included i

component fatigue evaluation and component support qualification.

3.2.1.1 Preliminary Two-Dimensional Elastic Analysis A two-dimensional model of the RCS hot leg from the reactor vessel nozzle up to ,

and including the steam generator and its supports was initially used to

! perform the preliminary thermal expansion analysis of the RCL. The analysis calculated the loads and stresses due to steam generator snubber lock-up. The model also included modelling of the hot leg rupture restraint to evaluate the effects of contact between the hot leg piping and the rupture restraint. The

results of the preliminary thermal expansion analyses were reported in j

~

LER 85-13. Based on our review of the preliminary thermal analysis report (Reference 1), the staff finds that the calculated secondary stresses for the

! hot leg elbow exceeded the ASME Code-specified 3 Sm ifmit for secondary stress (Equation 12 of NB-3653.6). The ASME Code does not pemit the use of a

! simplified elastic-plastic analysis when Equation 12 cannot be satisfied. The f simplified elastic-plastic analysis initially performed by the licensee to obtainthestraininthehotlegelbowwasthus,technicallyinapppopriate.

The Code intent of the 3 Sm limit is to assure that elastic shaked6wn 'will

(

occur under design themal loading conditions, thus, preventing excessive l

plastic deformation which could lead to incremental collapse, and validating 21

-- _ -~ _ _ _

the application of elastic analysis when perfoming fatigue evaluation.

Although the Code intent of the 3 Sm limit is a desirable requirement from a design standpoint, the stresses resulting from constrained RCS themal expansion due to snubber lockup were not produced by design loading conditions but rather occurred as a result of an unanticipated event. Thus, the staff 1 finds that the design intent of the Code as required by Equation 12 of tiB-3653.6 to prevent excessive piastic defomation is not a prereauf site for unanticipated events nccurring after the fact. The staff concludes that the exceedance of Equation 12 stress limits was not a Code violation. However, the significance of not assuring elastic shakedown and the assessment of plastic defomations which may have occurred due to unanticipated events must include a proper evaluation for fatigue considerations to assure the adequacy of the component structural integrity for the remaining life of the plant. A fatigue evaluation vas subsequently performed by the licensee using appropriate analyses n.ethods and is discussed later in this section of the SER.

3.2.1.2 Four-Loop Elastic' Analysis A four-loop elastic analysis was perfomed to determine crossover effects of steam generator snubber lock-up from one loop to the other loops. Each loop model included the hot leg, steam generator, cross-over piping, reactor coolant pump, cold leg and their respective supports. Lock-up of the Loop B steam gen-erator snubbers was assumed and maximum vessel movements were detemined. The staff finds the four-loop analysis to be an effective method to predict the cross-over effects caused by constrained RCS themal expansion. Furthermore, based on the small movements calculated at the reactor vessel, the staff con-cludes that the results of the four loop analysis provides adequate justifica-tion for the licensee's application of a single loop model with a fixed reactor vessel centerline to calculate the bounding loads and-stresses in its non-linear analyses. l 3.2.1.3 One-loop Nonlinear Analysis l A nonlinear analysis of a single loop was performed by the licensee to more accurately calculate the stresses, displacements, support loads, and gap 1

22 l l

l l

effects due to a worst-case constraint of RCL thermal expansion. A finite-element model was developed using plastic pipe elements, plastic elbow elements, elastic straight pipe elements, and gap elements. The support were modelled as elastic elements. Material properties for the RCL analyses were based on Reference 9.

The loading conditions were based on several postulated scenarios of RCS ther-mal constraint in order to develop bounding results. RCS thermal constraint assumptions were based on minimum gaps and clearances measured in the plant during the supplemental inspections described in Section 3.0 of this SER.

Other assumptions used to establish bounding conditions included 1) assuming fully locked snubbers in the hot and cold RCS position and 2) using various gaps between the RCS hot leg and the rupture restraint including zero gap, 0.25 inch cold gap, and uplif t of hot leg of 0.125 inch due to restraint contact and subsequent rotation.

Based on our review of the methodology, the staff finds the use of the nonlinear analysis provides an acceptable approach to predict the effects of gaps and clearances on the behavior of the RCS during thermal expansion.

Furthermore, the nonlinear analysis provides an acceptable methodology for utilizing the material stress-strain curves presented in Reference 9. The staff finds the material stress-strain curves (Reference 9) provide an acceptable means of defining stainless steel properties at normal operating RCS temperatures. Because the material curves used in the analysis give a yield

- strength at 0.2 percent offset strain which is less than the actual yield strength of the hot leg elbow material, the staff finds the material stress-strain curves to be acceptable and results in a conservative approach.

The analyses of various loading conditions provide assurance that the worst-case scenario related to RCS thermal constraint has been evaluated and that the analysis assumptions are consistent with the physical conditions found in the plant. Based on our review of the loading conditions considered, the staff finds that the combination of loading conditions which have been ahalyzed pro-vide a reasonable basis for concluding that the actual loading condition 23

l

. 4 1

experienced by the plant has been bounded by the analyses perfomed. Thus, the l staff concludes the loading conditions analyzed by the licensee are acceptable.

3.2.1.4 Plastic Analysis of the Hot leg Elbow The nonlinear analysis discussed above indicated that the location of highest stress in the RCL was at hot leg to steam generator elbow. Accordingly, a de-tailed plastic analysis of the hot leg elbow was performed to confirm the strain results predicted by the plastic piping elements used in the single loop nonlinear model. The ANSYS finite-element model utilized 20 node three-dimensional isoparametric solid elements. Geometric data were obtained from spool piece design drawings. The steam generator nozzle connected to the elbow was assumed to remain round because of large stiffness values of the nozzle. At the hot leg side, the model was extended to preclude end effects on the elbow results. The loading on the elbows consisted of internal pressure and applied displacements obtained from the nonlinear single loop piping analysis.

The staff concludes that the plastic analysis performed by the licensee for the hot leg elbow is an acceptable method for confirming the strain results pre-dicted in the non-linear single loop analysis. Our finding is based on a de-tailed staff review of the appropriateness of the input data, assumptions, and modelling techniques used in the plastic analysis of the hot leg elbow.

3.2.1.5 Component and Component Support Evaluation Fatigue evaluations were performed for critical RCS components and included the reactor vessel hot leg nozzle, reactor vessel hot leg bimetallic weld, hot leg elbow, steam generator hot leg bimetallic weld, steam generator hot leg nozzle, steam generator crossover leg nozzle, and pump crossover nozzle. A fatigue evaluation of the hot leg elbow was perfomed by using the strains predicted by the single-loop nonlinear analysis (and verified by the plastic analysis of the ,

hot leg elbow) and calculating the increased usage factor due to restrained RCS thermal expansion. The evaluation was performed in accordance with para-graph NB-3228 of the ASME Code Section III for plastic analysis.

24

The bimetallic welds were evaluated using loadings from the nonlinear analysis and combining these with temperature mismatch effects to detemine usage due to the restrained thermal growth event. The ASME simplified elastic-plastic analysis method was used since the average strains at the bimetallic weld were only slightly above yield strains.

The nozzles were evaluated using Bijlaard analysis techniques. Maximum loadings were obtained from the nonlinear analysis and the usage factor was calculated based on the ASME Code simplified elastic-plastic analysis method.

The component supports were evaluated by using standard linear elastic techniques for static loadings which are considered to be acceptable.

Based on our review of approach used by the licensee for the fatigue evaluations of the critical RCS components, the staff concludes that the methodology is consistent with the applications of plastic analyses described in paragraph NB-3228 of the ASME Code Section III and is, thus, acceptable.

3.2.2 Evaluation of Acceptance Criteria 3.2.2.1 Pressure-Retaining Components The acceptance criterion used for the fatigue evaluation of pressure-retaining components was a total cumulative usage factor of less than 1.0. The total usage factor was calculated by adding the fatigue usage due to all loadings associated with 28 cycles of restrained RCS thermal growth to the usage factor calculated in the design basis. In addition, a strain limit was used to ensure

~

that the structural response of the system was capable of withstanding the loadings imposed by the restraint of RCS thermal expansion. A strain limit of 1 percent (0.01 in/in) was used for the restraint of RCS thermal growth.

Based on our review of the acceptance criteria for fatigue and strain, the staff concludes that the limits used are acceptable. The basis fo'r the acceptability of a 1 percent strain range is provided in Reference!'3. A1 percent strain range for austenitic steels corresponds to a stress ~ amplitude of

^

141,500 psi using the ASME Code design fatigue curve (Figure I-9.2.1). The 25

corresponding number of allowable design cycles is 366. Accordingly, a usage factor for 1 percent strain appifed 28 times is 0.77. However, because the ASME Code fatigue evaluation requires the stresses due to all applicable loadings be added for the calculation of the usage factor and because of the non-linear nature of fatigue usage factors as a function of stress or strain, the calculation of usage due tu only the restraint of RCS thermal expansion would not provide meaningful results. Thus, the staff utilized a dual acceptance criteria stipulating that: (a)theexcess(overdesign) usage factor not exceed 0.100, and (b) the total usage factor (up to present and anticipated future) not exceed 1.0. Because of the difficulty in isolating the

" excess" usage factor from the " design" usage factor, the 0.100 limit was established only as a comparative guideline to assess the relative usage factors between design and unanticipated loadings. The total usage factor limit of 1.0 is an ASME Code limit for its fatigue evaluation of design loadings. The combination of the fatigue usage due to the RCS thermal restraint and the fatigue usage due to the design loading conditions satisfying the 1.0 cumulative usage factor limit assures that the component structural integrity is acceptable for the remaining life of the plant. The acceptability of the 28 cycles is based on the actual number of heat-up/ cool-down cycles _

experienced at Trojan since it began operation in 1976.

3.2.2.2 Component Supports The acceptance criteria for component supports are based on Subsection NF of the ASME Code. Exceptions to the Code were taken for several support subcomponent parts (e.g. bolts and pins), but were limited to less than yield stress. Because the loading condition related to restraint of RCS thermal growth is not a design condition, the staff finds the use of higher-than-Code design allowables to be appropriate for an evaluation of the significance of the event. The limitation of the stresses to less than yield stress under this  ;

unanticipated loading condition is acceptable because the basic characteristic j of secondary stresses (such as is developed by the restraint of th) RCS thermal expansion) is that it is self-limiting. The combination of assuring that support stresses remained less than yield for an unanticipated event such as constraint of RCS thermal growth and the supplemental inspections performed en 26 l

the RCS component supports to identify potential damage to the support provide a reasonable basis to conclude that the functionality of the RCS component supports has not been adversely affected. Thus, the acceptance criteria and the exceptions taken to the Code design allowables when supplemented by inspections are considered to be acceptable for assuring the functionality of component supports.

3.2.3 Evaluation of Results The results of the RCL thermal expansion analyses and fatigue evaluations were provided to the staff in the June 8,1966 letter (Reference 8). A sumary of the results for component fatigue has been extracted from Reference 8 and is presented in Table 3-1 of this SER.

Based on our review of the results, the staff concludes that the reactor cool-ant loop satisfies the acceptance criteria established for the evaluation of restrained RCS thermal growth. Because the analysis results bound the postu-lated events which actually may have occurred, the adequacy of the RCL has been confirmed. Thus, the staff finds the results of the Trojan RCL thermal expansion analyses and hot leg elbow analysis provide an acceptable basis for confirming the structural adequacy of the reactor coolant system for the restraint of RCS thermal growth. The combination of the fatigue evaluations, corrective action taken to pret.lude future restraint of RCS thermal growth, and the RCS Thermal Monitoring Program provide adequate assurance that the stresses and strains in the components of the reactor coolant system will not exceed the allowable design basis limits for the remaining life of the plant.

~

3.3 Evaluation of Steam Generator Hydraulic Snubbers I

In a letter dated May 21, 1986 (Reference 5) the licensee described the root causes of the failures related to inoperable steam generator snubbers and the corrective actions taken to preclude further failures. Additionally,theli-censee presented to the staff the results of further testing perfoI1ned by Wyle Laboratories on the snubber control valves in a meeting held at thi NRC offices (Bethesda, MD) on May 29, 1986. The Wyle test results were documented in 27

Table 3-1 RCS COMPONENT FATIGUE RES'JLTS COMPONENT U, Wu U TOT RPV Outlet Nozzle 0.2 0.02 0.22 RPV Outlet Nozzle Safeend Weld 0.63 0.14 0.77 Not Leg / Elbow Weld 0.2 0.1 0.30 Hot Leg Elbow 0.1 0.03 0.13 Safe Weld SG Inlet Nozzle 0.88 0.004 0.884 SG Outlet Nozzle 0.88 0.001 0.881 RCP Outlet Nozzle 0.20 0 0.20 RPV Inlet Nozzle 0.1 0.004 0.104 NOTES: U,,= original design usage factor Wu = Usage factor due to restrained thermal expansion event UTOT = original design usage factor plus usage factor due to restrained thermal expansion event i

f 28

Reference 6. The results of the licensee's evaluation of the steam generator hydraulic snubber are provided below. The staff evaluation of the results follows.

Background

During the 1985 refueling outage, two 900 kip Anker-Holth steam generator hy-draulic snubbers were found in a degraded condition and subsequently all 16 snubbers were overhauled. The degraded condition was caused by deteriorated seals and hydraulic fluid contaminated with rust, water, and seal material.

The contaminants in the fluid apparently caused plugging of the bleed orifice in the snubber control valve. The springs on the ball check valves in the con-trol valve assembly were also worn which could cause the check valves to seat prematurely resulting in snubber lock up.

Following overhaul of the snubbers and control valves and replacement of the hydraulic fluid, the snubbers were bench tested prior to reinstallation. This testing was difficult to perform due to the sensitive nature of the control valves. The ball check valves were very sensitive to snubber velocity causing the snubbers to lock up at low velocities. The snubber drag test criteria were revised to a minimum velocity of 7.6 mil / min under a 5000-lb load in order for the snubbers to pass this test.

In light of the difficulties encountered in testing the snubbers, the decision was made to replace the control valves in 1986 with control valves used in more recent hydraulic snubber designs. Whereas the old control valves were designed

~

to lock up above 25 mil / min, the newer control valves would cause snubber lock up when velocity exceeds 6-8 in./ min. The new control valves were procured prior to the 1986 refueling outage for installation during the outage.  ;

Results of Snubber Tests Conducted During 1986 Refueling Outage j i

During the 1986 refueling outage, when the snubbers were tested in the "as-found" condition (i.e., with the old control valves), the snubber piston was found to be free to move but the control valves failed the test for 11 of the l

29 i

e 16 snubbers. The failures were due to high bleed rate and/or excessive drag force. The effect of the failures due to high drag force would have been to restrict thermal growth of the RCS similar to what has been analyzed. The effect of the high bleed rate would have been a failure to fully restrict dy-namic motion when the snubber would be acting as a rigid restraint. An addi-tional abnormality observed during this year's testing was 550 ppm of water was found in the hydraulic fluid. This water contamination could have been intro-duced through the vent port on the hydraulic fluid reservoirs or could have occurred during fluid replacement in 1985. As corrective action, desiccant filters are being considered for installation on the reservoir vent ports during the next refueling outage. The hydraulic fluid used to fill the snubbers this year was also verified to have less than 150 ppm water.

Following the "as-found" testing, the old control valves were preserved for root cause failure analysis. Three of the control valves were disassembled onsite to determine the cause of failure. No apparent cause for the failures could be found. The remaining control valves were sent to Wyle Laboratories for further testing and inspection in an effort to determine the root cause of failure.

The conclusion reached by Wyle Laboratories was that various factors caused the performance of the snubber control valves to be inconsistent. The tight ball check tolerances, weak ball check springs, change in fluid viscosity, and de-sign intent to activate under very low differential pressures were all found to be contributing factors. However, testing in series demonstrated that even if

~

two of the control valves were overly restrictive, by-pass flow of fluid from I one control valve to another was possible through the interconnected tubing.

i Staff Evaluation Based on the results of the licensee's testing, the staff concludes that the likely cause of snubber lock up could be attributed to the sensitijity of the old-stylesnubbercontrolvalve(ballcheckvalve)designandtheIendencyfor the bleed orifice to plug up. However, based on the results of the series testing perfonned at Wyle Laboratories, the likelihood of full thermal lock-up 30

occurring would require that the various contributing factors would have had to affect three or four of the hydraulic snubbers on a single steam generator.

Thus, the staff concludes it is unlikely that full thermal lock-up of the steam generator snubbers occurred for more than one steam generator. Although the old style control valves exhibited a tendency to lock-up when tested individually, the interlinking (series) connection between the four hydraulic control valves prevented a single or dual snubber failure to cause full thermal lockup of the system.

The old style snubber control valves have been replaced by new style snubber control valves which are more widely used with hydraulic snubbers in other nuclear facilities. Although the licensee's action to replace the old style snubber control valve will assure thet snubber lock-up attributed to the old control valves will not recur, the staff will continue to evaluate the accept-ability of the new style snubber control valves. In the meantime, the monitor-ing program established by the licensee to ensure proper thermal movement of the steam generator near the steam generator hydraulic snubbers (i.e., to de-tect snubber lockup) will provide assurance that the snubbers with the new con-trol valves are thermally functioning as designed. Thus, the staff finds the _

corrective actions taken by the licensee to preclude further lockup of the steam generator hydraulic snubbers to be acceptable.

3.4 Evaluation of RCS Thermal Expansion Monitoring Program l In the May 21, 1986 letter (Ref. 5), the licensee provided a brief description

~

of its monitoring program. The program was established to verify that the reactor coolant system is expanding and contracting during heatups and cool-downs as predicted. The monitoring program will include measuring the move-  ;

ments of the RCS piping and steam generators and comparing those movements to predetermined acceptance criteria. The monitoring program is described in Temporary Plant Test TPT-166 dated June 1986 and the details were provided to the staff in the June 3,1986 submittal (Ref. 6). Thefollowingi}abrief description of the monitoring program: I 31 l

1. RCS Piping

a. The movement of the RCS hot legs at the pipe whip restraints will be monitored to ensure the proper gap is maintained between the pipes and the restraints,

b. The gaps at the eight crossover leg pipe whip restraints will be measured. This measurement will establish movement of the RCS crossover leg.

2. Steam Generators

a. The movement of the steam generator upper seismic support ring will be monitored. This movement will be a direct indication that the steam generator snubbers are expanding and contracting and that no binding is occurring at the ring bumpers. Additionally, the tempera-ture of the upper support ring will be monitored to detennine the thennal growth of this ring.

b. The gaps between the steam generator lower seismic support ring and the 4 guide pads and 2 stop pads will be measured. This measurement will be used to determine if the steam generator is moving as designed.

l The monitoring program will commence with the plant heatup at the end of the

~

1986 refueling outage and will be continued for each successive heatup and cooldown of the RCS until it has been clearly demonstrated that the movements are as expected.

l Staff Evaluation Basedonourreviewofthemonitoringprogramdetailsprovidedinfeference6, the staff finds the program provides a reasonable basis for assuring that the RCS is thermally expanding and contracting as predicted. The staff will con- l tinue to review the development of the procedures for the monitoring program, 32

the implementation of the procedures, and the results of the program. The staff intends to closely monitor and conduct confirmatory followup audits of the results of the first heatup as part of our review of the licensee's imple-mentation of the RCS Thermal Expansion Monitoring Program. Furthermore, the staff requires that the licensee provide to the staff adequate notification of the temination of its monitoring program.

3.5 Evaluation of Independent Peer Review In a letter dated June 6, 1986 (Reference 10), the licensee transmitted to the staff the results and conclusions of an independent peer review e.onducted by Bechtel Power Corporation of the RCS thermal expansion issue.

An interim report was issued by Bechtel Power Corporation in a letter from R.

W. Fosse (BPC) to R. L. Steele (PGE) dated May 23, 1986 (Reference 11) entitled, " Interim Report of the Independent Review of Trojan Nuclear Plant Reactor Coolant Loop Thermal Movements Evaluation Program," (May 23,1986).

The interim report identified various analytical anomolies associated with the preliminary Impell thermal expansion analysis (Reference 1). The report concluded that the reactor coolant loop themal expansion irregularities were likely caused by malfunctioning steam generator snubbers and by insufficient pipe whip restraint gap clearances. As a result, Bechtel recommended 1) the l preliminary themal expansion analysis be supplemented by a worst case loading scenario for each critical RCS component to determine the acceptability of l

, stresses, strains, and fatigue levels, 2) the establishment of a program to monitor key system thennal expansion points, and 3) the establishment of a long-term program to assure that the steam generator snubbers continue to perform their intended function.

In a letter frem R. W. Fosse (BPC) to R. L. Steele (PGE) dated June 2, 1986 (Reference 12), Bechtel supplemented its May 23, 1986 letter (Reference 11) withconclusionsregardingtheRCSThermalExpansionTestProgramhPT-166). j Bechtel concluded that the program presented in TPT-166 provided a' satisfactory l method to monitor and evaluate critical reactor coolant system moveinents and 33

parameters during RCS heat-up including corrective actions to be taken if unpredicted thermal movements were to occur.

Additionally, in a letter from R. W. Fosse (BPC) to R. L. Steele (PGE) dated June 6, 1986 (Reference 13), Bechtel presented the findings of its review of the Impell analyses performed subsequent to its preliminary thermal expansion analysis. Based on its review, Bechtel concluded that 1) the credible worst case loading condition had been identified for each critical RCS component, and

2) the stress / strain levels and fatigue usage to date for the critical components for the worst case loading condition are acceptable.

Based on the staff review of Bechtel's conclusions as discussed above, the staff finds that the conclusions reached by Bechtel are consistent with the evaluation findings reported by the staff in this SER. Thus, the staff concludes that the independent peer review provides additional confidence that the actions taken by the licensee in its evaluation of the RCS thermal movements resulted in adequate corrective actions and acceptable results.

i i

34

4.0 CONCLUSION

S The staff concludes that the bounding measures taken by the licensee to assess the potential safety impact of a thermally constrained RCS and the corrective actions implemented to preclude future unanticipated thermal constraints, pro-vide a reasonable basis to allow continued operation of the Trojan Nuclear Plant. Our conclusion is based on the following:

- The supplemental inspections and walkdowns of the RCS conducted during the 1986 refueling outage verified the extent of damage to the RCS due to thermal constraint and provided physical evidence of the actual behavior of the RCS in order to assure that the assumptions used in the RCS stress analyses would yield consistent or bounding results.

- The stress analyses and fatigue evaluation performed by the licensee assures the adequacy of the RCS structural integrity for the worst-case loading experienced by the RCS up to the present and for the remaining anticipated life of the plant.

- The root cause of the steam generator hydraulic snubber lockup which likely resulted in a themally constrained RCS has been identified and corrected.

- Other potential causes of a thennally constrained RCS have been identified and evaluated. Appropriate preventive measures have been taken by the licensee to preclude the likelihood of their causing restraint of RCS

~

thermal growth.

- A monitoring program has been established to ensure that the RCS is expanding and contracting in a predictable and consistent manner.

Anindependentpeerreviewofthetechnicalactionstakenbyphelicensee to address the RCS thermal expansion anomolies resulted in findings consistent with those of the staff.  :

35

I In order to ensure that the long-tenn considerations of RCS thermal constraint have been adeauately addressed, the staff requires as a confinnatory measure the following:

(1) The licensee shall provide an assessment of the impact of using carbon steel shims in lieu of graphite shims prior to terminating its RCS thermal monitoring program (TPT-166). The assessment shall consider the potential impact of the shims on the RCS pressure boundary in the event of contact between the shims and the hot leg.

(2) The licensee shall provide prior written notification to the staff of when it intends to tenninate the RCS thermal monitoring program (Temporary Plant Test-166) in order for the staff to assess the need for any long-term operability requirements for the steam generator snubbers.

(3) During the next refueling outage, the licensee shall perfonn additional examinations of the RCS Loop B hot leg elbow to ensure that small cracks have not been initiated in the interior surface. The examinations shall consist of: _

(a) ultrasonic examination with full documentation on the RCS Loop B hot leg elbow after properly preparing the examination surfaces, or (b) remote visual and liquid penetrant examinations on the inside diameter surface of the RCS Loop B hot leg elbow.

PRINCIPAL CONTRIBUTOR:

D. Terao I

b l

l l

36

5.0 REFERENCES

1. " Preliminary Thernal Expansion Evaluation of the Reactor Coolant Loop for Trojan Nuclear Plant," by Impell Corporation, dated January 1986, Impell Report No. 01-0300-1471 Revision 0.

2. NRC Safety Evaluation Report, "Long Term Service Plan - SEP Seismic Re-evaluation Criteria and Methodology," transmitted in a letter from H. Thompson (NRC) to K. P. Baskin (Southern California Edison Company)

I dated September 19, 1985.

3. Letter from E. C. Rodabaugh (consultant) to M. Reich (Brookhaven National Laboratory), dated May 26, 1986.

Subject:

Trojan LER 85-013, Revision 1, Meeting at Impell (Walnut Creek, CA), May 22 and 23, 1986.

4. Letter from B. D. Withers (Portland General Electric Company) to S. A.

Varga (NRC) dated May 9, 1986 with attachments.

5. Letter from B. D. Withers (Portland General Electric Company) to S. A.

, Varga (NRC) dated May 21, 1986 with attachments.

6. Letter from B. D. Withers (Portland General Electric Company) to S. A.

Varga (NRC) Dated June 3,1986 with attachments.

7. ASME Boiler and Pressure Vessel Code,Section III, Division 1, " Nuclear Power Plant Components," American Society of Mechanical Engineers (1983 Edition).

8. Letter from G. Zimmerman (PGE) to S. Varga (NRC) dated June 4,1986.

9. ASME Paper, "Isochronous Stress-Strain Curves for Austenitic Stainless Steel," L. D. Blackburn (1972). I h

10. Letter from B. D. Withers (PGE) to S. A. Varga (NRC) dated June 6, 1986 with attachments.

37

,- - - - .

• _ , _ , - . ~ . _ - , , . . , . - - , , - - - , . , , - - , ,

r. , _

d

11. Letter from R. W. Fosse (Bechtel Power Corporation) to R. L. Steele (Portland General Electric) dated May 23, 1986.

12. Letter from R. W. Fosse (BPC) to R. L. Steele (PGE) dated June 2, 1986.

13. Letter from R. W. Fosse (BPC) to R. L. Steele (PGE) dated June 6, 1986.

I e T

38

UNITED STdTES ff y

(Agj 9

NUCLEAR REGULATORY COMMISSION WASHINGTON, D. C. 20555

\...../

Attachment 2 SAFETY EVALUATION FOR THE TROJAN NUCLEAR PLANT f

PORTLAND GENERAL ELECTRIC COMPANY DOCKET NO. 50-344 CONTROL ROOM HABITABILITY ISSUES INTRODUCTION On January 28-30, 1986 and February 25-27, 1986 NRC representatives visited the Trojan Nuclear Plant to discuss with Portland General Electric (PGE) the NRC's control room habitability survey at Trojan, and a Technical Specification change request involving the performance of a surveillance test of the emergency air conditioning unit.

Based on information gathered during the two visits, the team members concluded that the manner in which the control room ventilation systems functioned did not appear to be consistent with the PGE TMI Task Action Plan Item III.D.3.4 submittal or the updated FSAR. The team also identified deficiencies in the Technical Specifications and the plant procedures, instructions, etc. utilized to demonstrate control room habitability in

. accordance with III.D.3.4.

~

Because of deficiencies noted in the plant visits between the original TMI Action Plan III.D.3.4 analyses dated January 2 and March 2,1981 and the as-found condition, the staff concluded that the conclusions reached in the staff's Safety Evaluation dated February 17, 1982 may have been invalidated.

By letter dated May 7,1986, the staff issued a request for additional information identifying items which needed to be addressed in order for the staff to assess control room habitability (III.D.3.4) at Trojan. By letter dated May 19, 1986, PGE responded to the staff's requests. On May 28, 1986 a meeting between the staff and PGE was held to discuss the May 19,11986 submittal and the improvement program for the Trojan control room! ventilation system in general. By letter dated June 4, 1986, PGE submitted their revised TNI action item III.D.3.4 analysis as requested by the staff for review prior to plant startup. The staff had further discussions with PGE on the above submittal's in a telephone conference call on June 9,1986.

i

l DISCUSSION AND EVALUATION We have reviewed the May 19, 1986 letter from Portland General Electric (PGE)

Company which outlined their responses to the NRC concerns on the operation of the Trojan control room ventilation system in accordance with their previously submitted III.D.3.4 analysis and the adequacy of the Trojan Technical Speciff-cations and procedures. In addition, we have reviewed the revised III.D.3.4 analysis submitted by PGE in their June 4, 1986 letter. Based upon our review of the revised analysis, we have determined that the doses calculated utilizing the Trojan plant specific assumptions of 525 cfm of makeup flow, 2475 cfm of recirculated flow, 10 cfm of inleakage into the control room due to ingress /

egress and no inleakage from dampers, ductwork or ventilation system components such as air handling units meet the acceptance criteria of General Design Criterion (GDC) 19.

It should be noted that PGE has committed to an evaluation of toxic gas threats (ammonia, sulfur dioxide and chlorine) on the control room. The evaluation will include consideration of the leakage of various components. PGE states that this evaluation will be completed by September 30, 1986. In addition, PGE has committed to submit by September 30, 1986 a number of changes in the plant Technical Specifications that reflect resciution of several of the concerns identified in the Trojan inspections. We will review the toxic gas evaluation and the Technical Specification changes upon their receipt.

CONCLUSION Based on the above considerations, we consider that the Trojan III.D.3.4 control room habitability issue has been satisfactorily resolved for restart.

Followup items which are not a prerequisite to restart, as discussed in PGE's May 19, and June 4,1986 submittals, the May 28, 1986 meeting and the June 9, 1986 telephone conference call will be evaluated when they are submitted.

PRINCIPAL CONTRIBUTOR:

J. Hayes l