ML19256E610

Safety Evaluation Supporting Amend 20 to License DPR-70

ML19256E610
Person / Time
Site:	Salem
Issue date:	10/30/1979
From:	Office of Nuclear Reactor Regulation
To:
Shared Package
ML19256E607	List: ML19256E606 ML19256E610 ML19256E612
References
NUDOCS 7911080141
Download: ML19256E610 (30)

Text

.

[

UNITED STATES o,7.

NUCLEAR REGULATORY COMMISSION

{

,E WASHINGTON. D. C. 20555

\\

/

SAFETY EVALUATION BY THE OFFICE OF NUCLEAR REACTOR REGULATION SUPPORTING AMENDMEni NO. 20 TO FACILITY OPERATING LICENSE NO. DPR-70 PUBLIC SERVICE ELECTRIC AND GAS COMPAN,Y2 PHILADELPHIA ELECTRIC COMPANY,

DELMARVA POWER AND LIGHT COMPANY, AND ATLANTIC CITY ELECTRIC COMPANY SALEM NUCLEAR GENERATING STATION, UNIT NO. 1 DOCKET NO. 50-272 Introduction By letter dated March 2,1979 (Reference 1) and supplemented by letters dated April 30,1979 (Reference 2), August 8,1979 (Reference 3), and August 9,1979 (Reference 4), Public Service Electric and Gas Company (the licensee) proposed changes to the Technical Specifications for Salem Generating Station Unit 1 for Cycle 2 operation. This document represents the NRC staff's evaluation of these proposals.

During the refueling outage the licensee notified toe NRC that fuel assembly grid strap damage (Reference 5) and broken RCCA rodlets (Reference

6) had been observed in the Cycle 1 core. The effect of these structural problems on the Cycle 2 core has been reviewed as part of our reload evaluation.

The refueling outage for Salem Unit No.1 also coincided with a period when the staff was reviewing the safety of all operating nuclear power plants in light of several potential safety problems identified by the NRC Office of Inspection and Enforcement (IE). These concerns have been described in IE Bulletins 79-02 (Reference 7),79-06A (Reference 8), 79-07 (Reference 9), 79-13 (Reference 10) and 79-14 (Reference 11) and are being reviewed by IE and/or NRR. The status of these potential safety problems is also included in this evaluation.

Discussion Technical Specifications Salem Unit 1 completed its first cycle of operation on March 31,1979 and immediately began preparing for initiation of Cycle 2 in June 1979. The only itens identified by the licensee for review for Cycle 2 related to 1294 344 7511080l+/ 4

. the reload core which was proposed to consist of 40 new Wstinghouse 17 x 17 fuel assentlies. Two of these assemblies are of the optimized fuel assembly design as part of the Westinghouse " Optimized Fuel Assembly Demonstration Program." The licensee has reviewed the "as loaded" Cycle 2 core in relation tc the Cycle 1 core that was reported in the FSAR and, as a consequence, propose t.ie following changes to the Technical Specifications for Cycle 2 operation:

1.

Increase in radial peaking factor (Fxy)*

2.

Revision of the normalized heat flux hot channel factor (K(z)) curve third l'ine segment.

3.

Restriction of axial flux difference for the first 2700 MWD /MTU.

4.

Revision of the nuclear enthalpy hot channel factor (FAH) to take credit for currently approved rod bow penalty Our review of these proposed Technical Specification addresses the licensee's

- earlier request to change the FjH limit to account for the reduction in departure from nucleate boiling ratio due to fuel rod boiling (Reference 12).

Grid Strap Damage The reloading outage for Cycle 2 has been extended far beyond the original schedule because of several reasons. The first unexpected problem arose when the licensee observed that some of the Cycle 1 fuel assemblies were damaged when removed from the core. This problem and its satisfactory resolution is addressed in our overall evaluation of the Cycle 2 core.

Broken Rodlets The outage and review of Cycle 2 operation have been extended further because of structural failures that were observed in six RCCA's that contained eight broken rodlets. The licensee discussed this problem in detail with the staff (Reference 13) and, subsequently, Westingnouse provided its findings and guidelines for early detection of dropped rodlets (Reference 14). Because of the uniqueness of this problem, we have reviewed the analyses perfonned by Westinghouse and the licensee's resolution for Cycle 2 to assure safe operation of the Cycle 2 core.

IE Bulletins As part of its continuing review of nuclear plants that have been licensed to operate, the O'.fice of Inspection and Enfor::ement identifies safety 1294 345

. problems that may be of generic nature to all or specific types of nuclear reactors. Five of these Bulletins have been considered to be pertinent to the Salem Unit No.1 plant and of sufficient potential to warrant discussion or evaluation before pemitting initiation of Cycle 2.

IE Bulletin 79 Pipe Support Base Plate Designs Using Concrete Expansion Anchor Bolts As the result of its concern over structural failure of piping supports of safety equipment, the NRC has described actions required of the licensee to determine if this potential exists at Salem Unit 1.

Before achieving Mode 4, Hot Shutdown, the licensee has agreed to meet the requirements of IE Region I.to:

1.

Complete a test program to ve 1fy correct installation of wall and ceiling mounted concrete anchors, in safety-related systems in inaccessible areas, 2.

Submit a test program outline for testing of anchor bolts and base plates in accessible areas.

(This testing s expected to be completed by November 15,1979.),and 3.

Perfom an evaluation detailing the basis for resumption of operation (Reference 15).

The licensee responded to these requirements by letter of September 24, 1979 (Reference 16) and is presently performing the verification tests.

IE Bulletin 79-06A - Review of Operational Errors and System Misalignment identified During the Three Mile Island Incident This Bulletin identifies certain actions to be taker, by the licensee to review design and operational aspects of Salem Unit No.1 that may be similar to those that were in affect at TMI. The licensee responded to this Bulletin by letters of April 25, 1979.(Reference 17), May 11, 1979 (Reference 18), July 13,1979 (Reference 19) and August 14, 1979 (Reference 20). These responses have been reviewed by a special NRC task force and have been found to be. acceptable to pemit return of Salem Unit No.1 to power. This task force is continuing to review specific long-tem provisions of this Bulletin.

IE Bulletin 79 Seismic Stress Analysis of Safety Related Piping In the course of evaluation of certain piping designs, significant discrepancies were observed between the original piping analysis computer code used to analyze earthquake loads and a currently acceptable computer code developed for this purpose. The licensee wcs notified on August 28, 1979 (Reference 21) that, prio. to achieving Mode 4, hot shutdown during 1294 346

, Cycle 2, the licensee must provide the basis for return to service prior to completion of all the requirements of this Bulletin. The licensee responded by letters of September 21,1979 (Reference 22) and of October 1979 (Reference 23). Our review of these resporses is included in

. this evaluation.

IE Bulletin 79 Cracking in Feedwater System Piping After discovery of cracking in the feedwater lines of other Westinghouse steam generators, the NRC required the licensee to determine if similar problems existed at Salen Unit No.1.

In response the licensee made the necessary' inspections and informed the NRC that cracks had been found (Reference 24). Additional information was provided by letters of June 14,1979 (Reference 25) and August 24,1979 (Reference 26) and during a meeting with the staff on July 12, 1979. The licensee also provided the staff with samples of the faulty sections of the steam generator for analysis. Our safety evaluation of the licensee's actions is enclosed.

IE Bulletin 79 Seismic Analysis for As-Built Safety Related Piping Systems By means of this Bulletin, the NRC recuested the licensee to take certain actions to verify that seismic analyses are applicable to plants as-built.

The licensee is implementing the requirements of this Bulletir as part of its program to respond to Bulletin 79-07. By letters of August 16, 1979 (Reference 28) and September 14,1979 (Reference 29) the licensee described the field walks of safety-related piping systems that are being made. Sample walks have cafirmed that actual configurations conform to the stress isometric drawings.

Inspection of all inaccessible

• areas will be performed before the p. ant returns to power.

Mss of Eddy Current Template Plug Assembly Eddy current testing of the steam generator tubes was perfonned during the reload outage for Salem Unit No.1. At the conclusion of this testing, one of 24 plug assemblies brought onsite for this purpose could not be accounted for and is assumed to be lost inside the primary coolant system. The licensee notified the staff to this effect on May 9,1979 (Reference 27).

An analysis of this potential problem by Westinghouse indicates that the plug is of insignificant mass and physical size to cause any effect upon the plant safety analysis or operation (Reference

).

Tests have shown that the 21 gram plug (appre ximately 2 inches long) and 1 inch in diameter) will undergo mechanical disintegration from the turbulence of the Primary Coolant System and will undergo thermal decomposition at the operating temperature of the Primary Coolant System. A chemical analysis of the residual components of a plug heated to a Reactor Coolant System operating temperature would n'Jt be detrimental to the integrity of the system operations or equipment.

g 747

. Our review of this problem indicates that if a plug is located in the hot leg area of the steam generator it will remain there until it has been mechanically fragmented or decomposed to such an extent that it will pass through the steam generator tubes. The size of the plug will probably be reduced by both mechanisms while the Primary Coolant System is being heated through operation of the coolant pumps during startup for Cycle 2.

The composition of the residue from the plug has been

.shown by Westinghouse to be a low density pliable mass with a maximan weight of approximately 5 grams. The fate of such a semi-solid mar, would depend on whether it would adhere to a portion of the Primar)

Coolant Boundary before it was ronverted into colloidal state as a'r, emul sion. The total mass and size of the eventual residue is constared to be too small to result in a blockage in the Primary Coolant Systen, such as has been analyzet. in the FSAR. Therefore, we agree that the possible cresence of a template plug in the steam generator does not pose a safety nroble. during Cycle 2.

Evaluation 1.

Proposed New Technical Specification Changes Nuclear Design The Cycle 2 loading consists of 36 region 1 fuel assemblies (16.6 MWD /MTU average bxnup), 60 region 2 fuel assemblies (17.0 MWD /MTU average burnup), 57 region 3 fuel assemblies (12.4 MWD /MTU average burnup), and 40 fresh region 4 fuel assentlies. The grid damage and rodlet drops discovered during the refueling outage prompted reanalysis of the core. The results of the as-loaded core analysis are discussed here.

Cycle 2 operation is designed with a peaking factor envelope limit of 2.32.

'he large break LOCA analysis provided in Reference 29 was perfomed with a peaking factor of 2.32 using the February 1978 model.

This has been reviewed and approved by the staff (Reference 30).

For Cycle 2 operation, the licensee provided a reanalysic (Reference

2) of the small break LOCA to justify a revised third line segment of the nomalized operating envelope. This was perfomed with approved methods and is acceptable.

~

An "18 cases" analysis was provided by the licensee to show that the worst peaking factors encountered for postulated load follow maneuvers during Cycle 2 are within the bounds of the proposed nomalized operating envelope. The analysis was perfomed with a radial peaking factor (Fxy) of 1.65 and was done according to the riethods described in Reference

17. The results are acceptable for axial offsets ( I) of +5f. from the target band..

For Cycle 2, the licensee proposed a Technical Specification which limits the axial flux difference to less than positive 7.5 percent for the first 2700 MWD /MTU. The licensee asserts that analysis perfomed 1294 348

. for the first 2700 MWD /MTU. The licensee asserts that analysis perfomed by approved methods shows that although the target flux difference at 80.: is greater than 7.5 percent, it will be necessary to limit the axial flux difference to less than 7.5 percent to assure that the Fq limit of 2.32 is maintained. We find this acceptable for Cycle 2.

The rod bow penalty Technical Specification has been updated for Cycle 2 and conforms to the provisions of Reference 34. We find this acceptable.

For Cycle 2, the licensee is loading two Westinghouse 17 x 17 demonstration

" optimized fuel assemblies." These assenblies are similar to current 17 x 17 design except that zircaloy grid straps are used and the fuel pins are slightly smaller in diameter. Loading criteria develoned by Westinghouse, based on nuclear and thermal-hydraulic analyses and presented in Referenca 16, require that the demonstration assemblies be placed in locations such that FaH is at least six percent lower and Fq is at least G.10 peaking factor units lower than the maximum allowed for standard assemblies. For Cycle 2 at Salem, the locations of the demonstration assemblies are at the core periphery where both these criteria are met. One assembly is instrumented with a thermocouple and the other with a movable incore flux detector. The licensee, in cooperation with Westinghouse, will follow the fuel surveillance program proposed in Reference 35. Because the use of the demonstration assemblies is limited to two assenblies loaded in low power regions of the core, we find their use acceptable for Cycle 2.

Fuel Design With the exception of the two optimized fuel assemblies, the Cycle 2 reload fuel assenblies are of the same mechanical, nuclear and thermal hydraulic design as the Cycle 1 fuel assemblies. The Cycle 2 fuel rod internal pressure design criteria limit the internal pressure of the lead rrd in the reactor to a value below that which could cause (1) the diam.tric gap to increase due to outward cladding creep during steady. state operation and (2) extensive DNB propagation to occur.

The NRC has accepted this design basis (in Reference 31). Calculations of the clad flattening time predict no clad flattening during Cycle 2.

2.

Fuel Assembly Grid Anamolies During the refueling operation at Salem, it was noted by the licensee that some of the assemblies that were removed had suffered grid mechanical damage. This was reported to the NRC in Reference 5.

Subsequent to this discovery, all fuel assemblies were removed from the core for examination. The degree of the damage to the grid straps was classified in three categories: small pieces missing (15 assemblies), grid 1294 349

. material ripped and laid over (5 assemblies), larger sections missing and fuel pins exposed (11 assemblies). No damage to the fuel pins was observed. A total of 31 assemblies suffered some grid damage.

The damage ippears to be the result of corner to corner interaction of the grid straps of diagonally adjacent fuel asseslies during the vertical loadtag and unloading movements. No corralation of the damage to core location, grid elevation, or manufacturing and shipping batches has been identified.

The licensee and the fuel manufacturer established the following guidelines' for Cycle 2 for reloadirg damaged assedlies:

(1) those assemblies with full width pieces missing will not be reloaded for Cycle 2; (2) those assemblies with defomed edges and those with chips missing will be reloaded with special procedures to prevent further damage.

Westinghouse has noted similar damage of this type in several other plants, However, in none of these cases were the number of asses 11es danaged as great as at Salem. Examination of some of the damaged assemblies indicated that they had operated through the previous cycle with no detrimental operational effects.

The nuclear and,hermal-hydraulic effects of operation during Cycle 2 with 19 assemblies which have bent or chipped grid spacers is expected to be minimal. The loss of Inconel metal from the grids is insufficient to have a discernible effect on the core neutronics.

It should be noted that there was no observable grid compression or defomation. The damage was restricted to the grid straps and teb s.

Nevertheless, if any minor defomations had been observed, they would have resulted in effects that would have been bounded by the rod bow considerations which were included in the Salem reload analysis and Technical :. -ifications and have been found to be acceptable.

There was some concern that pieces of grid strap which were not recovered would result in either flow blockage of an assembly or in jamming to prevent scram of an individual control assembly.

With respect to flow blockage, the Salem FSAR (Reference 32) describes the results of analyses of complete blockage of an assembly nozzle and of partial flow blockage in the subchannels..For complete blockage of an assembly inlet nozzle, the analysis with the THINK-IV code shows that the flow is restored to nomal within 30 inches of the nozzl e.

For those locations where the flow is disrupted, the DNBR does not approach 1.30 at full power conditions because these are not the peak power regions of the core. Examination of all damaged assemblies shows that a total of approximately 25 square inches of grid material was broken off. The licensee estimates that after recovery of some of the larger pieces, no more than seven pieces 1294 350

. larger than two square inches each remain somewhere in the RCS.

All these pieces together would be insufficient to totally block an assembly nozzle. Therefore, this event is considered impossible.

The FSAR also describes tests with partial flow blockage in the coolant channels, wnich show that with as much as 41 percent of the subchannels blocked, flow recovers to nomal within five inches of the blockage.

It is estimated that at Salem for full power steady state conditions, a reduction in local mass velocity of approximately 70 percent would be required to reduce the DNBR to 1.30. The mass velocity effect on the DNS correlation was based on the assumption of non-turbulent flow along the channel length.

In reality, a local flow blockage is expected to promote turbulence which would lesson the effect on DNB.

For pieces of grid strap which are free in the RCS and large enough to cause even minor blockage, the most likely place for the blockage would be the bottom nozzle or the first grid assembly elevation. These are relatively low power locations.

Because of the limited effects of flow blockage from small pieces of grid strap in the fuel channels and because this blockage is expected at low power elevations, we believe the consequences of this type blockage do not endanger the public health and safety.

With respect to jamming of a control assembly to prevent scram, the likelihood of a piece lodging where it could cause a problem is extremely remote. Most of the chips would be expected to settle out in the stagnant regions of the lower plenum. However, if a piece were entrained in the RCS flow and lifted through the assembly inlet nozzle, it would be very unlikely to complete the torturous path up through the fuel assembly to the upper internals where it could lodge to interfere with a scram. Technical Specifications call for periodic exercising of control asse211es which would alert the operators of any control rod binding should it occur.

In addition, all accident and transient analyses which result in reactor scram are done assuming that the most reacti fe control assembly does not scram. Because of the very small likelihood that an unrecovered piece could prevent scram, because rods must be periodically exercised, and because the consequences of one stuck control assedly are acceptable in all accident and transient analyses, we find that it is acceptable to operate for Cycle 2 without recovery of all small grid fragments.

3.

Safety Evaluation of Broken Rodlets During the current refueling at Salem, some reactor control cluster assemblies (RCCAs) were observed with individual rodlets which had broken from the main assembly. This was reported to the NRC in Reference 6.

Subsequent to this initial discovery, all RCCAs.in 1294 351

. the Salem core were inspected. Six RCCAs with a total of eight detached rodlets (four RCCAs with one broken and two RCCAs with two broken) were found. The detached rodlets remained inserted in their respective fuel assembly guide tubes.

Examinations of the failed RCCAs have shown that the failures occurred in the threaded area of the (female fitting) fingers into which the (male fitting) rodlets are threaded, torqued and pinned. The

. failures were complete circumferential cracks in the fingers at a location adjacent to the topmost threads of the rodlet endpiece.

All the dropped rodlets were traced to two receiving lots of fingers from a manufacturing subcontractor. All the fuel asseablies containing dropped rodlets and all the RCCAs (25) with fingers from the two suspect receiving lots have been removed from the reactor.

3.1 Materials Considerations Each of the 53 RCCAs in a Westinghouse reactor with 17 x 17 fuel contains 24 individual rodlets for a total of 1272 rodlets. Earlier Westinghouse 15 x 15 fuel designs use RCCAs with 90 rodlets each.

In both cases, the finger designs are similar witn only slight variations in dimensions. The past performance of Westinghouse RCCAs has been satisfactory. A total of 1382 RCCAs are in 33 operating plants with individual service times that range from a few months to 140 months. Through December 1978, only ten RCCAs have experienced some operational problem.

In six of these, the rodlets became detached from the spider hub due to vane separation. This separation has been attributed to faulty braze joints. Two RCCAs required repair due to galling of single rodlets in each assembly. One RCCA was discharged due to bent rodlets. One RCCA experienced a single rodlet separation in a manner somewhat similar to Salem Unit No.1. About half of these reported events occurred during the initial cycle of reactor operation while the remainder occurred randomly over a period of 2 to 9 years.

In general, each event involved only 1 or 2 RCCAs.

A review of the design of the finger at the location of the break showed that worst case loads were no more than 20 percent of the design values. In fact, the point of failure is designed as one of the strongest points in the RCCA. Thus, the failures are not likely due to stress or fatigue.

All failures occurred in fingers in the outer row. Within the outer row of fingers, the failures occurred in random positions. No failures occurred in fingers of the inner rows. The physical positions of the RCCAs with failed fingers were random with respect to core location.

~

1294 352

, In an effort to determine the cause of the. failures, Westinghouse reviewed manufacturing records of the affected fingers. This review included materials records, procurement records, deviation records

~

and manufacturing processes. The data is presented in the attachment to Reference 8.

The results of this review of materials records and procurement records determined the following:

1) All failed fingers were machined from two material heat lots and were contained in ta sequential receiving lots of outer fingers from one supplier.

2) Many other receiving lots of both outer and inner fingers e;ch did not exhibit failures were also machined from the same heat lots by the same supplier.

3) The total population of outer fingers in Salem Unit No.1 for Cycle 1 was machined from three material heat lots and was comprised of 11 receiving lots from three ifferent suppliers.

The conclusic,ns drawn from the materials records are that the problem is limited to the two receiving lots of fingers at Salem and does not extend generically to other Westinghouse plants.

Westinghouse asserted that review of the deviation records associated with finger machining and spider d3sembly showed nothing of significance.

The review of the manufacturing processes for finger machining and spider assembly was not conclusive but provided these observations:

1) Any contaminants left by the finger machining process would probably be cleaned out by the high temperature and vacuum applied to the finger during brazing to the spider.

2) During final assembly of rodlets to fingers, retapping of threads and shoulders is sometimes required for final fitting and a contaminant could have been introduced at this time ;The cleaning process is not repeated after retapping. Since the time of manufacture of the Salem fingers in 1976, manufacturing change notices were issued by Westinghouse to change tapping specifications so that fewer fingers needed retapping and also to remove a certain threading lubricant from the assembly area.

1294 353

. The conclusion from the review of the manufacturing processes are that an undetermined contaminant may have been introduced in the threaded area of the finger after initial cleaning along with high residual stresses but that changes in the fabrication process have since eliminated the problem.

It should be noted that one difference between the fabrication process for 15 x 15 RCCAs and 17 x 17 RCCAs was found.

It is sometimes necessary to retap the fingers to allow proper fit of the threaded rodlet into the finger. Because of a slightly shallower threaded area on 17 x.17 fingers, a bottoming tap is used to retap the threaded area rather than a tapered tap. This could result in increased stress concentration in the 17 x 17 fingers. However, due to the large esion margin to stress at this point, and the absence of failut et in fingers from other receiving lots, it was concluded that the effects of this difference in tapping would be insuffucient to cause the failure.

As part of the materials investigation, hot cell work was perfomed on two RCCAs removed from Salem Unit No.1. After Cycle 1, RCCA R-31, which contained two failed fingers, and RCCA R-37, which contained no failed fingers and no fingers from the two susper.t receiving lots, were examined. Three damage types were discovered and the damage was all characterized as stress-corrosion cracking. The three finger damage types were described as:

(1) larger circumferendal threaded region; (2) minor axial cracks in the cracks in the top (3) local cracking in the shoulder area.

thread area; and Eleven of sixteen outer fingers from R-31 were sectioned. Only three were clear of damage. Of the remaining eight that were sectioned, three (in addition to the two that had failed during cycle 1, one additional rodlet finger failed out of the reactor prior to hot cell testing) had failed completely, six exhibited 20 to 80 percent circumferential cracking in the top threaded area, and three exhibited local cracking in the shoulder area. All four middle fingers were clear. Of four inner fingers, two were clear and two exhibited local cracking in the shoulder area.

RCCA R-32, which contained no fingers from the suspect lots, showed nine of nine sectioned outer fingers clear of damage, four of four middle fingers clear, and three of four inner fingers clear. One inner finger showed local cracking in the shoulder area.

1294 354

. According to Westir.ghouse, tests for elevated chloride levels in the outer finger of R-32 appeared to show greater concentration than the inner fingers; however, the results of this testing were not conclusive.

Further metallurgical examination of the crack surfaces of the circumferential cracks indicated that the cracks. were "old" and had probably occurred early in the cycle. Further evidence of this is provided by the appearance of the flux tilt early during Cycle 1 operation, and with hindsight, by examinations of Cycle 1 flux maps. As an estimate of the time frame for failure due to stress corrosion, Westinghouse estimated that with a saturation concentration of chloride ions at 550*F, with stress concentrations in the threaded area, failure could occur in less than one hour.

The conclusions reached by Westinghouse and the licensee with respect to the failed RCCA fingers are:

1) Failures do not represent a structural inadequacy or generic design weakness.

2) Failures are the result of stress corrosion cracking and were contained within the two receiving lots of outer fingers.

3) Indications of stress corrosion cracking on other than the two receiving lots are located in the shoulder area, are of a different composition and severity, and would not lead to dropped rodlets.

4) A review of the flux maps of operating reactors and successful refueling of two 17 x 17 cores shows that no positive evidence of broken rodlets exists for other plants.

5)

A. review of Salem Cycle i flux maps shows that dropped redlets occurred prior to low power operation and were present throughout Cycle 1.

6) Elimination of all RCCAs containing fingers from the suspect lots should prevent recurrence.

The staff agrees that the evidence presented by the licensee supports the conclusions stated above.

1294 355

. 3.2 Nuclear and Themal-Hydraulic Considerations Calculations were perfomed by the licensee and by Westinghouse to provide estimates of the nuclear and thermal-hydraulic effects of broken rodlets. The information was provided in Reference 8 and in various telephone conversations with the licensee and with Westinghouse.

The reactivity of a dropped rodlet in the core is estimated to be worth about 10 pcm per; dropped rodlet. Using this estimate, and the value for the excess shutdown margin (SDM) available during cycle two of 0.5% k/k, at least 50 rodlets dropped randomly into the core would be required to cancel the excess SDM. The value of excess SDM is calculated assuming the rapid cooldown of the moderator due to a steamline break and failure of the most reactive RCCA to scram.

Relative to shutdown margin requirements to accommodate the postulated steam line break at end of cycle, the licensee has demonstrated ample shutdown margin to accommodate all other postulated transients.

The combination of low probability events required to potentially endanger the public health and safety are:

(1) large steamline break; (2) most reactive RCCA stuck; and (3) more than 50 rodlets dropped.

As discussed later in this evaluation, core surveillance would make it unlikely for such a large number of dropped rodlets to go undetected.. Because of the unlikelihood of the combinatfoe of low probability events and the likelihood of detecting 50 dropped rodlets, we believe that loss of shutdown margin due to dropped rodlets is not a significant safety concern.

Cycle 1 was operated with excess SMD of 1.60% k/k. This was equivalent to the worth of at least 160 rodlets. Thus with respect to SDM, safe operation during Cycle 1 wa: not. jeopardized with eight dropped rodlets.

The presence of a detached rodlet in the core could be of concern with respect to mechanical movement of a loose part.

In the case of Salem, all c,f the detached rodlets remained in the guide tubes of the respective RCCAs. It is expected that because of the upper guide structure templates, a rodlet which fell from a withdrawn position would be guided without binding into its respective RCCA 1294 356

. RCCA guide tube. The hydraulic uplift forte on an individual rodlet due to normal reactor coolant flow through the guide tubes is estimated by Westinghouse to be less than one half the force required to lift a rodlet from its fallen position. Thus, the staff concludes that there is no danger of loose rodlets moving about in the reactor coolant system.

The effectiveness of the Technical Specifications for detection of power anomalies such as dropped rodlets was demonstrated during Salem Cycle 1 operatian when a flux tilt in excess of the Technical Specification allowances was encountered. Although the cause of the tilt was not discovered at that time, in retrospect the cause has been identified as dropped rodlets.

Flux maps taken at that time also showed the presence of the dropped rodlets. In any case, the licensee was required to analyze the tilt to show that safety limits were not jeopardized. Flux maps taken at that time also show the presence of the dropped rodlets.

Detennination of values of Fxy, Fq and Fg are required to be made at least once every 31 days of operation. The values of these peaking fators are determined from incore instrumentation measurements which would inc;ude the effects of the flux depressions due to the dropped rodlets.

If the allowed values of these peaking factors are exceeded duri1g power operation, power must be reduced. The accident and trantient analyses are valid only if the peaking factor limits are mainta'ned. Since the peaking factors can be measured regardless of the presence of dropped rodlets, it possible to maintain the core in a safe condition by observing the current applicable Technical Specifications.

3.3 Augmented Surveillance and Startup Program All evidence from Cycle 1 at Salem indicates that the failure of the. fingers resulting in the dropped rodlets occurred prior to going to power. With hindsight, hot zero power flux maps at beginning of Cycle 1 show dropped rodlets. Also, Westinghouse estimated that with saturation chloride levels at 550'F, the failures could occur within less than one hour.

During startup testing, the licensee is required to measure shutdown margin and to measure the critical soluble baron concentration for comparison with calculated values.

Ideally, it should be possible 1294 357

. to detect dropped rodlets at tnis time. The accuracy of the titration methods used for measurements of soluble boron concentration is estimated to be approximately +10 ppm boron. This converts to approx-imately +0.1 k/k reactivity. This measurement error is about 20%

of the magnitude of the reactivity worth of 50 dropped rodlets.

The licensee is required by Technical Specifications to report deviations of greater than +100 ppm of the design value of critical boron concen-tration. This w uld detect the presence of approximately 100 randomly

~

dropped rodlets. And although not required by Technical Specifications, a discrepancy of +50 ppm from design is used by the licensee as the cri.teria for Tnitiating a design review. This criteria would detect the presence of approximately 50 dropped rodlets.

The licensee has also submitted (Reference 33) an augmented surveillance program for Cycle 2 startup to detect dropped rodlets. Flux map analysis will pay particular attention to flux depressions. Acceptance criteria for flux maps require discrepancies of less than +10% of design for assertly powers greater than 0.9 nominal and +1T% of design for assembly powers less than 0.9 nominal.

If the design acceptance criteria for flux maps is exceeded, the licensee will measure rod worths to the N-1 condition.

In this way, if flux maps indicate a problem, measurements on individual RCCAs would be used to localize possible dropped rodlets.

In addition to the flux maps, the licensee will continuously monitor certain plant parameters to detect any changes that might indicate a dropped rodlet: 1) the core reactivity computer has sufficient sensitivity to detect a change of approximately 10 pcm (the worth of one rodlet); 2) the primary coolant temperature instrumentation is expected to detect the occurrence from one to three dropped rodlets due to moderator temperature feedback; 3) control rod motion, turbine load, xenon, baration, and dilution will be monitored to separate intentional reactivity changes from unexpected deviations. Any indications of dropped rodlets will be reported to the NRC. The staff agrees that the additional startup tests and augmented surveillance program proposed by the licensee provide sufficient assurance that failure of a large number of rodlets will be detected.

These additional tests and the augmented startup program are in addition to the standard physics startup test program which has also been reviewed. Low power physics include: baron endpoint, isothermal temperature coefficient, rod worth and flux map measurements.

9 1294 358

The high power physics test include power coefficient and flux map measurements. Acceptance and design criteria as well as remedial actions for these tests have been approved. The staff considers this total program to be appropriate and adequate.

4. Seismic Stress Analysis of Safety-Related Piping (IE Ccetin 79-071 PSE&G has used PIPDYN II cory.Jter code for pipestress analyses. The computer analysis involved calculation of piping responses due to X-component earthquake, Y-component earthquake, Z-component earthquake, X and Y earthquake, and-Y and Z earthquake. During the X and Y and

'Y and Z earthquake evaluation, however, the intramodal piping responses were inadvertently calculated by use of the algebraic summation method.

This is considered unacceptable as it may predict unconservative results in the seismic piping analysis. This code with the intramodal summation method was used in the seismic analyses of most of the safety-related systems at the facility. The licensee has identified the seismically analyzed (Seismic Category I) systems at the facility analyzed with PIPDYN II and the algebraic sumation technique. It has also identified portions of the Control Air System as the only system not seismically analyzed (i.e., static method). Furthemore, the licensee has reported the results of reanalyses using an acceptable earthquake response summation technique. This latter technique consists of utilization of the individual X, Y, and Z earthquake responses which were previously computer calculated using PIPDYN II and the hand calculation of the root-sum-square (SRSS) of intramodal responses due to the three com-ponents of earthquake loading.

We have evaluated the results of all the methods of pipe stress analysis previously utilized and used in the reanalyses for the facility. Technical information required for this evaluation is provided in the licensee's submittals of August 28, 1979, September 21, 1979 and October 11, 1979.

1) Systems The following 15 systems were identified by the licensee as having been analyzed with PIPDYN II with the algebraic summation technique:

Residual Heat Removal Chilled Water Reactor Coolant Chemical and Volume Control Safety Injection Control Air System Steam Generator Feedwater Steam Generator Blowdown Systems Component Cooling Spent Fuel Cooling Service Water Main Steam Auxiliary Feedwater Containment Spray System Diesel Generator Starting Air and Fuel Systems 1294 359

. The licensee has reanalyzed all 823 pipe stress problems originally involved in the algebraic summation calculation.

In addition, the licensee has stated that the pipe stress problems, which were analyzed by hand calculation, did not sum earthquake responses algebraically and are acceptable.

The licensee's letters of October 11, and October 18,1979 entails three phases of program as described below:

Phase I Prior to entering Modes 3 and 4, the following work will be accompli shed:

(a) Completion of pipe stress analysis (for both OBE and DBE) on safety-related systems required for safe shutdown.

(b) Re-evaluation of the associated supports, nozzles, and perietrations, within the inaccessible area.

(c) Re-evaluation of the supports, nozzles, and penetrations for entire Auxiliary Feedwater System.

(d) Re-evaluation of the supports for the Reactor Coolant System Pressure Boundry.

(e) Field modification to supports and penetrations evaluated in (2), (3) and (4) that fail to meet the criteria stated in September 21, 1979 submittal. Field modification to nozzles which fail to meet manufacturer's acceptance criteria.

(f) Re-evaluation of the supports, nozzles, and penetrations of the following systems:

(1) High pressure safety injection using the Chemical and Volume Control System.

(2) Low pressure safety injection using the Safety Injection System.

(3) Main Steam System up to the isolation valvas to include the steam supply to the steam driven auxiliary feed p.np.

(4) Containment Spray and Recirculation.

(g) Field modification to supports ard penetrations evaluated in (f) that fail to have a factor of safety of at least 2.

Field modification to nozzles which fail to meet manufacturer's acceptance criteria.

1294 360

N x

2

\\

n N

\\

~:

~

s

^

mN

~

N a

(a

.\\

'\\

\\

m g

y x

3

.\\

x

,4 s

c s

N

.E (f

' x N

\\

a. -

u

.,\\

N jVy%'M[Asic5,4Sy,O-bg4?}(?a

[

df m p.

ES s

I %r g lp a

m Mme~a$jhfh*,i.s%@

\\

y

?

b t

b W

i {[h uF h j. qu W4

';p [,

y ie

. ag g MSMr%IAk %.,Q #cyt

\\

a# 3 a kk W@y$,xkQ 4ygd %y.

h' t 8

Qwf'p, vg S

)

f

~ MhWi3 f &ghp

/'

jy@3L %.ed hsb ff Sg; bg+ggft uQny' N s gQ' fh Mf>D:%pf%

g

?p G

g E u%

'4*

K

%r.h N

1

&f "7

MMrr n'

KM 31 b[(vu$j$ r m p-xpgx~'

w

'M (Ew N'

.p h

M.m,hlh, 2

~

\\

$w[w q

q Of h,.

N

. e; N

• Q,R 9

,.hhhh

, L 7

p A fy5g,

v4..

sh lh*g 7 $'s 2

'.s' a

y yr2 y,

s

A+++

g&o>

s'4

,.e....<e 1,.

TEST TARGET (MT-3) l.0

's 8E RM EEE I.I

[,'8 LE M

I.25 1.4 g

6" 4%

+4

• S;f$/

$,4h4 5,,

.gy 7

%5%$>

gh('AS 4

e

%'#+*%

4#

4

e. ev <e 1,.

TEST TARGET (MT-3) 1.0 5BEBE e j,4 DE

[5 llle I.I i!

l.25 1.4 1.6 l

=

s..

=

1. • o4 4$+4

• %f h/

h,+p

?

WO my,,

i

. Phase II Prior to entering Modes 1 and 2 the following work will be accomplished:

Field modification and corresponding modifications associated with IE Bulletin 79-02 to supports and penetrations evaluated in item f of" Phase I that fail to meet the design criteria as stated in September 21, 1979 submittal. Modifications will be made within the time constraints of the action statements of the Technical Specifications if re-evaluation shows that system operability is affected.

Phase III The licensee shall complete reanalysis of the remaining pipa supports, noz71es and penetrations outside containment and shall propose a schedule for implementation of all identified modifications, both within 60 days of the date of plant scartup.

For each modification identified as a result cf reanalysis of the supports outside containment after resumption of facility operation, when the overall margin of safety of the support to ultimate capacity is determined to be less than 2, the NRC shall be notified within 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> after making each such determination. The affected system shall be considered inoperable as that term is used in the facility Technical Specifications until the necessary mcdifications are implemented within the time frame alicwed by the facility Technical Specifications.

Of the 823 pipe stress problems re-evaluated,. requirement for hardware modifications have been identified on 248 individual suppcrts to bring the pipe stresses in the inaccessible area and auxiliary feedwater system within allowables.

It has been revealed, during reanalysis that most of these modifications can be attributed simply to the original support design errors, rather than due to use of the algebraic summation method. Classification of these unacceptable supports to-gether with the proposed modifications are typically as foilows:

(a) 97 U-bolts used in most systems as anchors or guides were inadequate to withstand lateral loads and moments. Add structural steel or use heavier material.

(b) 34 structural teel members used in most systems contain certain members that could not withstand torsional or bending moments.

Add steel members.

(c) 57 straps in most systems used as anchors or guides were inadequate to withstand lateral loads and moments. Add structural steel, plates and heavier material.

1295 001

. (d) 31 welds in supports in most systems were overstressed.

Strengthen the weld itself and/or add bracing to relieve the stresses on the weld, or eliminate that particular support through stress calculat-ion.

(e) 22 trunions used mostly to anchor 6" diameter piping in component cooling system were overstressed. Add structural steel beams, straps, gusset plates and/or additional welds.

(f) 4 undersized rods. Replace with correct size rods.

(g) 1 undersized snubber in residual heat removal system. Replace with one of proper size.

(h) 2 improperly embedded anchor bolts. Replace

• Tse plate and use a new bolt pattern with additional bolts.s carry the design load. Also add new bracing.

2) Verification of Analysis Methods We have reviewed the acceptability of the analytical methods which are currently a basic for the facility piping desing. The licensee has identified the following computer code / analysis methods as applicable:

PIPDYN II (used only for response calculations for individual earthquake components)

Static Analysis Methods PIPDYN II In response to IE Bulletin 79-07, the licensee has submitted documentation of the positions of the computer code which was used in the piping reanalysis of Salem 1.

The code is called PIPDYN II, wh;ch originated at the Franklyn Institute Research Laboratory,(FIRL).

FIRL has stated that this code perfoms response spectrum and time-history analysis and that it calculates intramodal and intermodal responses according to the provisions of Regulatory Guide 1.92. A review of the code listing, as well as direct comunications with FIRL personnel, has confirmed this.

The FIRL is presently solving a set of NRC designed benchmark piping problems, using the response spectrum analysis method.

A preliminary comparison of FIRL and NRC solutions indicate good agreement.

1295 002

. The licensee has also submitted recently a piping problem and solutions for confirmatory calculations by the Brookhaven National Laboratory. These calculations are scheduled to be completed in the immediate future.

Based on these considerations we find the use of the present version of the code PIPDYN II provisionally acceptable for seismic analysis by the response spectrum method.

Static Analysis Some of the 1-1/2" or smaller safety-related low temperature field run piping in the control air system at Salem Unit 1 was analyzed using' simplified static methods. For such piping the seismic support spacing was detemined by assigning a rigid frequency (say 30 cps) to an equivalent simply supported straight beam. The approach for support sizing was to support the system rigidly at specific intervals already determined. The support loads and piping stresses were then calculated by the statir load method assuming the highest peak acceleration over the entire frequency ranges of the floor response spectra.

3) Reanalysis Methods and Results The safety-related piping systems at Salem Unit 1 have been reviewed to detemine the method of analyses. Eight Hundred and Twenty-three (823) computer stress problems of safety-related piping have been identified where the analysis used the computer code PIPDYN II which used an algebraic intramodal sumation of responses to earthquake loadings. The problems where an algebraic intramodal response combination technique was used in the design have been reevaluated using an acceptable method. The method uses the individual X, Y and I earthquake responses previously calculated by PIPDYN II and then uses hand calculations to combine the above intramodal responses by the SRSS method.

Th'e floor response spectra used in the reanalysis was the original amplified response spectra.specified in the FSAR. The peaks in the amplified floor response spectra were broadened by +10% to account for variation in material properties and approximatTons in modeling.

1295 003

. The piping systems were modeled as three dimensional lumped mass systems which included consideration of eccentric masses at valves and appropriate flexibility and stress intensification factors.

The' dynamic analysis procedur2s meet the criteria specified in the plant FSAR and are acceptable. The resultant stresses and loads from the reanalysis were used to evaluate piping, supports, nozzles, and penetrations.

All of the 823 PIPDYN II pipe stress problems have been reanalyzed and verified by the licensee's Quality Assurance Program. This reanalysis completed the entire scope of piping stress reanalysis.

Based on the infomation provided for review, we find acceptable the procedures and methods used in reanalyzing these problems.

The reanalysis included those pipe stress problems involving the reactor coolant pressure boundary and the supports associated with those problems. Since the reactor coolant pressure boundary is inside containment and all of the supports which must be modified will be modified prior to startup, there is no potential for a loss-of-coolant accident in the event of a DBE.

The licensee has stated that I&E Bulletin 79-04, "Velan Valve Weights,"

presents no problem to the reanalysis program.

The pipeline support designs for affected system piping was inspected by the licensee to verify the location, orientati3n, support clearar.ces, and support type. Any deviations that were identified are incorporated into piping reanalyses. These piping systems were also verified by the NRC Office of Inspection and Enforcement.

The pipe supports were reevaluated in cases where the original support design loading was exceeded as a result of piping reanalysis.

In such cases, the support reevaluation has included the consideration of base plate flexibility and a verification of actual field construction of the support. Where concrete expansion anchor bolts were used, their capacities, without compromising the originally committed safety margin, were also included in the reevaluation.

There are approximately 5100 supports in the 15 safety-related systems involved in the reanalysis; of these, 3600 supports have Deen reevaluated. Among the supports reevaluated, 1548 supports are in the inaccessible area and 198 of them were identified to need modification based on the criteria stated in the September 21, 1979 submittal. The licensee has committed to complete all these modification prior to startup.

1295 004

There are 297 supports associated with auxiliary feedwater system; of these a total of 50 supports have been identified to need modification based on the criteria stated in the September 21, 1979 submittal. The licansee has also committed to complete all these modifications prior to startup.

In addition, there are also 771 Supports outside containment which are associated with hign and low head safety injection, containment and reciruelation spray, and main steam systems up to the isolation valves. All these supports have been reeval uated. Modifications to 147 supports which have been identified as failing to meet the design criteria as stated in September 21, 1979 submittal will be completed prior to entering Modes 1 and 2.

Based on the results to dete, we expect other suppori:s outside containment may be found that will not have a minimum factor of safety of 2 to ultimate, whNh is used as a criteria for support operability. However, if support reanalysis indicates this we will require the licer.sae to inform the NRC of the results of reaaalysis within 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> and that the affected system be considered inoperable as specified in the facility Technical Specifications until the necessary modifications are implemented or a reanalysis assuming support failure is completed.

The licensee has examined nozzle loadings on 2 auxiliary feedwater pumps and 1 containment spray pump. The new forces and moments obtained from the reanalysis were Meluded in the reevaluation.

With regard to penetration loads, 7rvative hand calculation methods were used for simple configuro; ions. For larger penetrations a finite difference computer program based on linear thin shell theory is used.

Current results indicate no over-stressed conditions. Effort is being continued on the nozzles and penetrations that are included in the reanalysis.

The licensee has committed to reevaluate all the nozzles and penetrations within the inaccessible area and for the entire auxiliary feedwater system. prior to entering Modes 3 and 4.

Field modifications will also be completed prior to entering Modes 3 and 4 for penetrations that fail t< meet the criteria stated in the September 21, 1979 submittal and for nozzles which fail to meet manufacturer's acceptar.ce criteria.

1295 005

. The licensee has also committed to compleoe reevaluation, prior to e iaring Modes 3 and 4, of penetrations and nozzles associated et Nigh and low pressure safety injection, containment and recirculation spray, and main steam system up to the isolation valves. Field modifications will also be completed during the same period of time for penetrations that fail to have a safety factor of 2 to %1timate and for nozzles which fail to meet manufacturer's teceptance criteria. Furthemore, penetrations that fail to meet the criteria as stated in the September 21,19T submittal will then be modified prior to entering Modes 1 and 2.

Within 60 days of the date of plant startup reevaluations and field modifications, as appropriate, of the remaining penetrations and nozzles will be completed and the same operability require-ment which is applied to ;upports will also be applicable to penetrations and nozzles.

The design and analysis of the supports and attached equipment are in accordance with the criteria specified in the plant FSAR.

The pipe break criteria of the FSAR was reviewed in connection with the possible effect of changes of the high stress point resulting from the reanalyses. Reanalysis completed thus far has not shown any requirement for postulations of additional break locations per the FSAR criteria.

In cases where the FSAR criteria should be exceeded, new break locations will be postulated and protection provided as required. For inside containment, the same pipe break criteria for outside containment can be applied.

The piping systems and supports were designed to the allowable limits on ANSI B31.1. Components used in pipe supports (rods, U-bolts, clamps and bolts) were designed in accordance with ~

ANSI B31.1 and MSS Sp 58.. Bolting of structural components associated with pipe supports were designed in accordance with AISC. Wcided connections were designed in accordance with B31.1. The maximum loading conditions do not allow stress levels to exceed 24.0 KSI for fillet welds.

1295 006

, The safety-related piping systems supports and attached equipment, wnere the original analysis used an algeoraic intramodal summation technique, have been, or are to be reanalyzed with acceptable methods. The procedures used in the support reanalyses arid their results have been reviewed.against the criteria in the FSAR r.nd found acceptuole.

4) Conclusions The licensee has cemonstrated that PIPDYN II is the only sethod of analysis used for t ae facility's safety-related systems which coubines seismic loads algeoraically. Safety-related piping systeras analyzed with PIPDYN IT have been reanalyzed with an acceptacle method. Results of che reanalysis indicated that the pipe stres:. and equipment loads, after necessary modifications, will be acceptable when comparea with the FSAR allowables and the manufacturer's specified load criteria.

The reevaluation of pipe stress problems indicated that aodifications in 248 supports in the inaccessible area and auxiliary feedwater system were found to be necessary in order to bring the pipestresse; to within allowable. Tnese modifications are identified in Section 1, and the licensee will camplete them prior to plant startup. Reevaluation will also be completed prior to startup for supports associated with hign and low pressure safety injection, containment and recirculation spray, and main steam system up to the isolation valves. Any modifica-tions requireo for these systems will then be completec: either prior t.o startup, if a safety factor of at least 2 to ultimate does not exist, or prior to entering Modes 1 and 2, if the criteria stated in the Septembe.' 21, 1979 submittal is not met.

Evaluation of the supports and schedule for completion of necess&ry modifications in the balance of the plant will be completed within 60 days of the startup. Further, in those cases where reanalysis exceeas code allowable, the staff requires that the criteria used to determine whether a factor of safety of 2 to ultimate does exist by linear elastic analysis techniques or no more than twice the rated load for snubbers. Use of Welding Research Council Bulletin #107 for evaluation of local stresses due to integral attachment is acceptaole. Supports in accessable areas which exceed the factor of safety of 2 to ultimate will be considered as inoperable as defined in the Technical Specifications.

We reviewed the analysis techniques which are currently the bases for the facility's piping design. We have determined that the applica-tion of these techniques at Salem Unit I assures that safety-related systems will withstand the design basis earthquake. Although the reanalysis of supports outside containment is not completed, there 1295 007

. is reasonaole assurance that the facility can operate during the interim period until the reanalysis and any required modifir,ations are completed without endangering the health and safety of the public. This assurance is based on the following factors:

(1) All safety system piping outside containment which was originally seismically analyzed with the PIPDYN II program has been reevaluated and, subject to modifica-tion, is acceptable.

~(2) All piping ano supports of the affected safety systems inside containment have been reevaluated and were found either acceptable as presently designed or will be modified as identified in this SER prior to startup. All the nozzles and penetrations in the same systems will be completely reevaluated and modified, if necessary, in accordance with the licensee's commitment as stated on its October 11, 1979 letter.

(3) Confirmation of input data through "as-built" verification provides assurance that analytical results are correct and significant "as-5uilt" deficiencies repaired.

(4) The licensee has completed reevaluations and will implement necessary modifications prior to entering Modes 3 and 4 for the supports associated with auxiliary feedwater systems. The remaining reevaluations will also be completed prior to entering Modes 3 ana 4 for the supports associated with high and low head safety injection, containment, and recirculation spray, ario main steam system up to the isolation valves. Any necessary raodifications for supports in these systems will be completed prior to entering either Modes 3 or 4 or Modes 1 and 2 in act.ordance with ttie licensee's letter of October 11, 1979.

These systems and auxiliary feedwater system assure that ECCS systems and systems necessary for m'aintaining hot standby will be capable of withstanding a design basis earthquake.

(5) The licensee has ccmmitted to complete all the support reevaluation in accessaole areas outside containnent within 60 days of the date of plant startup.

(6) The probability of an earthquake excaeding the design basis earthquake during the 60 day period when the remaining support reevaluation is being completed is small and the licensee has committed to shut down the facility in the event of an earthquake wnich exceeds 0.01 g acceleration and inspect all piping, penetrations, supports and nozzles which have not been reanalyzed for both OBE and DBE.

1295 008'

9 (7) The NRC will require prompt notification of inoperabia supports within 14 hours1.62037e-4 days <br />0.00389 hours <br />2.314815e-5 weeks <br />5.327e-6 months <br /> and either resolution by reanalysis of the piping system assu:aing a failed support or modification of the af ected support, if reanalysis of a support indicates that a fact,r of safety of two to ultimate capacity does not exist (or sr ibber loading greater than twice rated capacity).

Based on the above, we conclude that the licensee has demonstrated why Salem Unit 1 can be operated for 60 days pending completion of reanalyses required by IE Bulletin 79-07.

5.

Actions Taken to Eliminate Feedwater Piping Cracks On June 19, 1979 in response to IE Bulletin 79-13, the licensee perfomed radiographic i";pections of the steam generator feedwater nozzle / piping

' fitting welds the adjacent area on all four steam generators. Crack-like indications were revealed by RT and UT in Nos.11,13 and 14 steam generator feedwater nozzle to pipe fitting weld areas. UT indications were noted in the same region of steam generator 12 but not shown by RT.

A meeting was held with the licensee on July 12, 1979 to discuss the following items regarding feedwater piping cracks:

1.

Nature and extent of the cracking 2.

Metallurgical evaluation of the cracking including identification of the mode failure 3.

Stress analyses 4.

Operating history 5.

Feedwater chemistry 6.

Corrective actions 7.

Safety implications The licensee's interim report on the feedwater line cracking is enclosed.

In accordance with the requirements of IE Bulletin 79-13, the licensee perfomed volumetric examinations of all the feedwater piping welds with the exception of one inside of. containment.

In addition, magnetic particle examinations were perfomed of the auxiliary feedwater to main feedwater piping connections. No reportable indications were revealed from the results of the inspections.

The results of metallurgical evaluations by the licensee and their contractor of samples from loops 1, 2, 3 and 4 revealed cracks of a maximum depth of 0.120 inches in the region of the counterbase in the fitting in loops 1 and 4.

Shallow ecacks (0.025 inches) similar in nature to those in loops 1 and 4 were fcand in loops 2 and 3.

The cracks were generally straight and not branchtd. Fractographic examination at how manification revealed beach marks. TEM examination, with great difficulty, indicated fatigue situations on the order of 1 to 3 micro-inches. The probable mode of failure was identified as corrosion assisted fatigue.

1295 009

. The licensee performed stress analyses in an effort to identify an anomalies which could cause the observed cracks. The analyses were:

1.

Structural analyses using a 30 finite element nodel of the feedwater line including the effects of thermal, deadweight, and pressure (does not include stratification conditions). The licensee reports that cs show the stresses are within the allowable code limits.

resr 2.

2D finite element fatigue analysis of the feedwater nozzle / elbow configurations. The licensee reports that the results show an acceptable usage factor using the allowable cycles for a peak stress range from the ASME design S/N curves.

3.

Frequency analyses of the feedwater line and steam generator. The licensee reports that the results of the analysis indicate that feedwater line/ steam generator resonances is possible but consider this unlikely based on testing perfomed at the similar facilities that have been instrumented.

The piping fitting were removed and replaced on all steam generators.

Any cracks identified by the liquid penetrant examination of the nozzle base ID or OD were removed and, if required, repaired. Repairs to fabrica-tion rel.?ted discontinuities in welds in the feedwater lines have been compl eted. The nozzle to fitting welds were fully radiographed and ultra-sonically inspected following completion of the welding and stress relieving operations. The licensee has committed to perform radiography and ultrasonic examination of the nozzle to fitting welds at the next refueling outage.

In addition the licensee has installed instrumentation to measure pressure, thermal and mechanical transients during startup operations.

We conclude that the actions taken and proposed augmented inspections and sufficient to in.Jre that the piping integrity will be maintained.

If the causes of cracking cannot be determined by the next refueling outage, we will then decide what further actions, if any, are necessary.

Environmental Consideration We have determined that the amendment does not authorize a change in effluent types or total amounts nor an increase in power level and will not result in any significant environmental impact. Having made this determination, we have further concluded that the amendment involves an action which is insigificant from the standpoint of entironmental impact and, pursuant to 10 CFR 551.5(d)(4), that an envirorcental impact statement or negative declaration and environmental impact appraisal need not be prepared in connection with the issuance of this amendment.

1295 010

Conclusion Technical Specification Review Since the Technical Specification changes presented by the licensee for Cycle 2 operation were developed as a result of analyses using approved methods and because none of the changes result in a significant decrease in the margin of safety, the staff finds the changes to be acceptable.

Potential Safety Problems Addressed by IE Bulletins Responses to IE Bulletins 79-02 and 79-14 provided by the licensee have been determined by the Office of Inspection and Enforcement to be acceptable.

Likewise, the responses to IE Bulletins 79-07 and 79-13 have been determined to be acceptably sufficient to permit restart of Salem Unit 2.

Although initiation of Cycle 2 is not conditioned by the requirenents of IE Bulletin 79-06A, the licensee has resconded in an acceptable manner to all short term requirements of this Bulletin.

We have determined that the possible presence of an eddy-current template plug in the reactor coolant system does not pose a safety problem during Cycle 2.

• We have concluded,. based on the consit'erations discussed above, that:

(1) because the amendment does not involve a significant increase in the probability or consequences of accidents previously considered and does not involve a significant hazards consideration, (2) there is reasonable assurance that the healta and safety of the public will not be endangered by operation in the propos_J manner, and (3) such activities will be conducted in compliance with the Commission's regulations and the

• issuance of this amendment will not be ir.imical to the common defense and security or to the health and safety of the public.

Date: October 30, 1979 I295 011

References Page 1 1.

Public Servica Electric and Gas Company (PSEG) letter (Librizzi) to NRC (Schwencer) dated March 2,1979.

2.

PSEG letter (Librizzi) to NRC (Schwencer) dated April 30, 1979.

3.

PSEG letter (Librizzi) to NRC (Schwencer) dated August 8,1979.

4.

PSEG letter (Schneider to NRC (Schwencer) dated August 9,1979.

5.

FSEG letter (Librizzi) to NRC (Grier) dated July 30, 1979 LER 79-44/03'.-l.

6.

PSEG letter (Librizzi) to NRC (Grier) dated July 30, 1979 LER 79-47/03L-1.

7.

IE Bulletin 79-02 " Pipe Support Base Plate Designs Using Concrete Expansion Anchor Bolts," dated March 5,1979.

8.

IF. Bulletin 79-06A " Review of Operational Errors and System Misalignments Identified During the TMI Incident," dated April 18, 1979.

9.

IE Bulletin 79-07 " Seismic Stress Analysis of Safety Related Piping,"

dated April 14, 1979.

10.

IE Bulletin 79-13 " Cracking in Feedwater System Piping," dated June 25, 1979.

11.

IE Eulletin 79-14 " Seismic Analyses for As-Built Safety-Related Systems," dated July 2,1979.

12. PSEG Letter (Librizzi) to NRC (Lear) dated October 7,1977.

13.

NRC July 17, 1979 Meeting Summary (W. Ross) Docket No. 50-272 dated July 25, 1979.

14. Westinghouse letter (Anderson) to NRC (Check) dated September 4,1979.

15. NRC letter (Grier) to PSEG (Schneider) dated August 28, 1979.

16. "SEG letter (Schneider) to NRC (Grier) dated September 24, 1979.

1295 012

References PaSe 2

17. PSEG lette" (Schneider) to NRC (Grier) dated April 25, 1979.

18. PSEG letter (Schneider) to NRC (Grier) dated May 11,1979.

19. PSEG letter (Schneider,' to NRC (Grier) dated July 13, 1979.

20. PSEG letter (Schneider) to NRC (Grier) dated August 14, 1979.

21. NRC letter (Grier) to PSEG (Schneider and Martin) dated August 28, 1979.

22. PSEG letter (Schneider) to NRC (Grier) dated September 21, 1979.

23. PSEG lettar (Librizzi) to NRC (Schwencer) dated October 11, 1979.

24. PSEG letter (Libr5zi) to NRC (Grier) dated June 28, 1979.

25. PSEG letter (Schneider) to NRC (Stello) dated June 14., 1979.

26. PSEG let.ter (Schneider) to NRC (Grier) ted September 24, 1979.

E7. PSEG letter (Librizzi) to NRC (Grier) dated May 9,1979.

28. PSEG letter (Librizzi) to NRC (Grier) dated October 18, 1979.

29. PSEG letter (Librizzi) to NRC (Schwencer) dated February 15, 1979.

30. NRC letter (Schwencer) to PSEG (Librizzi) dated June 6,1979.

31. NRC memo from D. Ross to D. Vassallo dated December 8,1977.

32. Salem FSAR

33. PSEG ' letter (Librizzi) to NRC (Schwencer) dated August 3,1979.

,34.

HRC memo from R. Tedesco to D. Vassallo dated March 28, 1979.

35. WCAP-9286 " Optimized Fuel Assembly Demonstration Program" dated July 1978.

36. WCAP-8385 " Power Distribution Control and Load Following Procedures" dated September 1974.

37. PSEG letter (Librizzi) to NRC (Schwcacer) dated October 18, 1979.

1295 013