ML20137E486

Safety Evaluation Accepting Revised Response to NRC Bulletion 88-08, Thermal Stresses in Piping Connected to Reactor Coolant Sys, for Plant,Units 1 & 2

ML20137E486
Person / Time
Site:	South Texas
Issue date:	03/24/1997
From:	NRC (Affiliation Not Assigned)
To:
Shared Package
ML20137E454	List: ML20137E451 ML20137E486
References
IEB-88-008, IEB-88-8, NUDOCS 9703270250
Download: ML20137E486 (4)

Text

_

p r.ro j

p *t* UNITED STATES E

NUCLEAR REGULATORY COMMISSION E WASHINGTON, D.C. 30006 4 001 i

I s ...../

SAFETY EVALUATION BY THE OFFICE OF NUCLEAR REACTOR REGULATION SUPPLEMENTAL RESPONSE TO NRC BULLETIN 88-08 " THERMAL STRESSES l IN PIPING CONNECTED TO REACTOR COOLANT LOOP SYSTEMS"

FACILITY OPERATING LICENSE NOS. NPF-76 AND NPF-80

HOUSTON LIGHTING & POWER CONPANY j CITY PUBLIC SERVICE BOARD OF SAN ANTONIO CENTRAL POWER AND LIGHT COMPANY i CITY OF AUSTIN. TEXAS i DOCKET NOS. 50-498 AND 50-499 l SOUTH TEXAS PROJECT. UNITS 1 AND 2 BACKGROUND l- In Reference 1, the Nuclear Regulatory Commission (NRC) provided an evaluation j of the revised response by Houston Light and Power (HL&P) to NRC Bulletin j 88-08. This response was authored by Westinghouse Electric Corporation (W) as

! contractor to HL&P (Reference 2). M based its response on an analytical

! methodology developed under a program sponsored by the Electric Power Research

! Institute (EPRI) to investigate Thermal Stratification, Cycling, and Striping l (TASCS) (Reference 3). In Reference 4, HL&P submitted its response (prepared -

1 by W) to the NRC staff safety evaluation.

' EVALUATION i

The staff has reviewed the HL&P response in Reference 4, and has concluded that I i the comments made in its evaluation, as stated in Reference 1, remain valid and l current.

! A major weakness previously identified is the lack of correspondence between the j' TASCS methodology and the observed failures at Farley and Tihange. The TASCS ,

! methodology assumes that the temperature in the unisolable pipe is the same as the I reactor coolant loop (RCL) cold leg temperature, with and without leakage, within  ;

j' - the turbulent penetration length. It also assumes that the leakage through the

~

i check valve in the connected piping heats up a negligible amount over the length i of the horizontal pipe from the check valve to the cold leg, and calculates the

)

steady state thermal stress distribution based on this temperature difference. A cyclic frequency is then postulated and the fatigue life determined.

i Reference 3 reported limited results of a proprietary high temperature simulation i test of Farley, in which the temperature along the bottom of the pipe with leakage ,

j was measured. Figure 5.3-8 of the TASCS report shows that the axial time average 1

! 97o3270250 97o324 PDR ADOCK 05000498 1 P PDR

. . * =, - -.r e .- .--r , .-r< ~v --3s- e -s =

l l- l i

temperature of the leakage at the bottom of the pipe increases as the flow approaches the elbow. This figure shows the bottom inside wall temperature history at the five-diameter distance measured from the header nozzle, which is l the location of the weld joining the elbow to the horizontal segment. It also I

shows the bottom inside wall temperature history at the six-diameter distance, which appears to be similar to that at Farley, shown in Figure 5.3-2. The mean temperature at the location of the weld is about 500*F, with very small time

! variation. No top temperature at the'same location was stated, but the top measurement at Farley indicates almost the same value. The calculated thermal stress, which is based on the top-to-bottom difference of the mean temperatures, is therefore very small at this location. Yet, the crack at-Farley occurred at this location and not at the six-diameter distance, which shows a considerably greater transient temperature variation. No explanation for this paradox has been provided. Although no temperature measurements have been reported at Tihange, it can be surmised that the cracks in the elbow, as shown in Bulletin 88-08, Supplement 2, also occurred where temperature differences were small.

The response states again, that in the application of TASCS methodology to the South Texas normal charging and alternate charging lines it was determined that:

"At other locations within the unisolable piping, where turbulent penetration cycling is possible, the alternatip stresses were determined to be below the endurance limit stress based on 10 cycles." The response concluded on this basis that no cracking would occur in these lines. However, the "other" locations referenced in HL&P's response are those between the maximum penetration length and ,

the RCL nozzle, and one of these corresponds to the locations in Farley and l Tihange where the cracks were found.

The HL&P response also states that the Farley and Tihange failures were evaluated using the TASCS methodology, and that "the TASCS methodology does predict failure i at the observed crack locations." No such calculations were provided to support i this response. The response also provides "a summary of the results of the TASCS i methodology for the Farley/Tihange cracking. The fatigue and fatigue crack growth  !

analysis resulted in a prediction of failure of 3 to 6 calendar years, which was  ;

conservative with respect to the Farley plant age." This result was stated in '

, Reference 7 and is therefore not a result of the TASCS methodology.

In Reference 1, the staff stated in Evaluation Item 2 that the velocity component I which was measured in the Low Temperature Turbulent Penetration Test program was not specified. The HL&P response states that: "In the Low Temperature Penetration Test program, velocities were measured in the axial and the tangential direction; only the axial velocity was significant." This contradicts the data shown in References 5 and 6. The best fit curve of Figure 5.3-3 of the TASCS report for the 3" piping is almost the same as similar figures for all tests shown in these references. These figures are labeled " Maximum tangential velocities for-f all tests" and represent the decay of maximum (tangential) velocity with increasing distance from the tee. Very little information on the axial velocities is provided. The curves in Figure 5.3-3 and 5.3-4 are therefore tangential velocity curves. The equation of the upper bound curve in Figure 5.3-3 has the same slope as the best fit curve, and was used to determine the equation for the extent of the turbulent penetration. Therefore, the statement that "only the axial velocity was significant" is incorrect. In fact, the axial velocity does not enter into the calculations anywhere.

In the response HL&P also stated that the TASCS methodology was not intended to simulate an actual plant. Rather, it is intended to be used to generate a conservative set of thermal loadings. However, any such methodology must be able to reflect known events which are known to have occurred under. actual operating conditions. The events listed in NRC Bulletin 88-08 may be considered as benchmark problems by which to judge the adequacy and reliability of this i methodology. , l l Finally, thermal stratification cycling is a transient heat transfer process. The i L TASCS methodology is based on steady state heat transfer methods, which do not )

reflect transient thermal stress conditions.  ;

1

3.0 CONCLUSION

Based on a review of the additional material submitted in Reference 4, the staff concludes that the findings stated in Reference I remain valid, namely: ,

l

1. The mechanism of turbulent penetration was not fully investigated under the '

1 TASCS program. Its relevance to the fatigue failures described in Bulletin 88-08 has not been clearly established.

i

2. The TASCS methodology does not address the observed fatigue failures l described in Bulletin 88-08. The thermal loading conditions for.the failures described in Bulletin 88-08 remain undetermined.

i The staff also concludes that some elements of the TASCS methodology are l acceptable for application to the normal charging and the alternate charging lines at South Texas, Units 1 and 2, provided HL&P commits to the following provisions: ,

l

1. The thermal stresses due to isolation valve leakage are based on temperature distributions determined from transient heat transfer calculations using the l largest temperature difference between the turbulence source and in-leakage or out-leakage in the unisolable sections. The temperature cyclic history i should be such as to provide the greatest stress range.  ;

2. Fatigue failure is postulated at the nearest weld to the turbulence source in the case of a horizontal branch line attached to the RCL, or at the first ,

elbow in a branch line with a vertical segment attached to the RCL.

3. The American Society of Mechanical Engineers (ASME)Section III Class 1 fatigue analysis is based on the rigorous combination of design basis transients and the thermal stresses due to isolation valve leakage.

l

4. The time interval from start of check valve leakage to crack initiation is  !

based on the ASME Section III, Class I fatigue analysis and a cycling j frequency of one cycle per minute is assumed, as recommended in Reference 3.

l 5. Monitoring provisions to determine inadvertent internal leakage through the  ;

isolation valve are established and implemented. If such leakage is 4 detected, the time interval determined in step 4 establishes the allowable

! time interval to implement repairs. I r

j .

l

)

i l l

! REFERENCES

1. Letter of February 23, 1996, from T. W. Alexion, USNRC, to W. T. Cottle, Houston Light and Power (HL&P).

2. WCAP-12598, Supplement 1, "NRC Bulletin 88-08, Evaluation of Auxiliary Piping for South Texas Project Units 1 and 2," Westinghouse Electric Corporation (W), November 1993.

3. EPRI TR-103581, " Thermal Stratificatibn, Cycling, and Striping (TASCS),"

prepared by Westinghouse Electric' Corporation for the Electric Power Research Institute, Palo Alto, California, March 1994. (Licensable l material, proprietary and confidential).

4. Letter of July 15, 1996, from S. E. Thomas, HL&P, to the USNRC Document ,

Room, with attached M letter dated June 19, 1996, from M. A. Sinwell, W, to '

W. T. Cottle, HL&P, with enclosure " Response to NRC Safety Evaluation of l WCAP-12598, Supplement 1, and EPRI Report TR-103581." l

5. Kim, J. H., A. F.

Deardorff and R. M. Roidt,

" Thermal Stratification in Nuclear Reactor Piping System," presented at the International Conference on Nuclear Engineering, November 1991, Japan.

)

6. Kim, J. H., R. M. Roidt and A. F.

Deardorff,

" Thermal Stratification and Reactor Piping Integrity," Nuclear Engineering and Design 139 (1993). North l l Holland. ,

1

7. Westinghouse Report WCAP-11786, "J. M. Farley Unit 2 Engineering Evaluation i of the Weld Joint Crack in the 6" SI/RHR Piping," April 1988, proprietary.

Principal Contributor: Mark Hartzman Date: March 24, 1997 l

I 1

,