ML13330B538
| ML13330B538 | |
| Person / Time | |
|---|---|
| Site: | San Onofre  |
| Issue date: | 07/12/1991 |
| From: | Ray H SOUTHERN CALIFORNIA EDISON CO. |
| To: | Martin J NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION V) |
| References | |
| NUDOCS 9107300282 | |
| Download: ML13330B538 (3) | |
Text
Southern California Edison Compa h 23 PARKER STREET IRVINE, CALIFORNIA 92718 HAROLD B. RAY TELEPHONE SENIOR VICE PRESIDENT July 12, 1991 714-458-4400 Mr. John B. Martin, Administrator U. S. Nuclear Regulatory Commission, Region V 1450 Maria Lane, Suite 210 Walnut Creek, CA 94596-5368
Dear Mr. Martin:
Subject:
Docket Nos. 50-206, 50-361 and 50-362 Erosion/Corrosion of Piping Systems San Onofre Nuclear Generating Station The purpose of this letter is to provide a status report concerning our ongoing assessment of the subject matter. My letter to you dated February 14, 1991 summarized our prior discussion of erosion/corrosion (E/C) of piping systems and indicated that Southern California Edison Company (SCE) would shortly issue an informational Licensee Event Report (LER).
The LER was to include our assessment of the reasons why through-wall leaks had developed in steam plant piping at San Onofre Nuclear Generating Station (SONGS) and had not been prevented by monitoring for the effects of E/C.
Revision 0 of our LER 90-011 was submitted by letter dated February 20, 1991.
It reported our assessment that:
"Internal geometric discontinuities (e.g., fitup misalignment, backing rings, and unanticipated angle velocities) are causing accelerated E/C which is not predicted by our program. It is further observed that accelerated E/C due to internal geometric discontinuities is also not predicted by other E/C programs commonly in use within the industry."
This assessment was based on the following:
- 1.
Localized wall thinning had occurred that was substantially in excess of generalized thinning which had either occurred or been predicted.
- 2.
The localized thinning was associated with internal geometric discontinuities and changes in flow direction which were not explicitly included in the E/C models.
0 ik JI~
7I:
Mr. John B. Martin July 12, 1991
- 3.
Available descriptions for models used in E/C programs did not appear to address the geometric effects which had been identified at SONGS.
The E/C monitoring program previously in use at SONGS used a different analysis methodology than the CHEC code, which is commonly in use for this purpose in the industry. Accordingly, we recently performed a telephone survey which indicated that a number of plants that use the CHEC code have also experienced localized wall thinning leading to leaks,.similar to SONGS.
Although we believed that localized wall thinning due to internal geometric discontinuities and changes in flow direction may not be predicted by existing E/C programs, there remained a question as to whether such thinning would nevertheless be consistently identified and located before leaks or other unsafe conditions develop, through combined use of the program and systematic engineering evaluation and exercise of judgement as intended by EPRI.
In that regard, we presented a summary of our ongoing evaluation at a meeting in the NRC Region V offices on February 22, 1991. As a result of that meeting, and a.subsequent management meeting on March 1, 1991 which is summarized in a letter from Mr. Zimmerman to me dated April 2, 1991, we undertook the following:
o To communicate with EPRI concerning the results to date of our assessment and seek to achieve a common understanding concerning the intended use of the CHEC code in the overall program of E/C monitoring.
o To determine whether, if the CHEC code had been used at SONGS in accordance with then-existing EPRI guidance, the specific leaks which we experienced would have been avoided.
The principal result of the work thus far has been to emphasize the extent to which engineering judgement is expected to be used, in combination with either the CHEC code or its potential successor code CHECMATE, in an iterative inspection and evaluation process to monitor for E/C effects. Limited guidance concerning this is provided in the program literature.
In addition to discussion with EPRI, we responded to inquiry from NRR concerning the results of our assessment, as reported in the LER. Thus far, these discussions have emphasized the importance of engineering judgement in an effective E/C monitoring program and reduced the relative importance in our view of specific predictions by the codes themselves.
We have expended very substantial resources in this effort to date. The work has expanded beyond what we intended as a result of the February 22 meeting referenced above, to include
Mr. John B. Martin 3
July 12, 1991 use of the CHECMATE code at SONGS and verification that the CHEC and CHECMATE codes provide essentially the same results. As a consequence, we have not yet reached a conclusion concerning whether, if the CHEC code had been implemented in accordance with then-existing EPRI guidance, the specific leaks experienced at SONGS would have been avoided. The necessary work in this regard will be completed by the end of next month. A revision to the informational LER will be submitted by the end of September 1991.
More importantly, however, in discussions with EPRI we have not been able to compare our understanding of localized E/C phenomena with the assumptions which are included in the codes in use. Accordingly, we will apply our understanding of these phenomena in exercising the substantial engineering judgement required in E/C monitoring at SONGS, but we will not be able to generalize concerning whether or not equivalent results would consistently be obtained by straightforward application of the codes and associated guidance alone.
It should be noted that E/C monitoring programs progressively benefit from sharing of industry experience and evolving guidance through the CHEC/CHECMATE Users Group. This is an important, ongoing source of input to the engineering judgement which is a fundamental part of E/C monitoring methodology. SCE is a longstanding participant in the group, and has provided input at the April 1991 meeting concerning our specific experiences and assessments.
Finally, as noted above, CHECMATE is being used by SCE at SONGS as an improvement on the use of the CHEC code. Data obtained during the upcoming refueling outage of San Onofre Unit 2 will be used in an evaluation of CHECMATE which should be completed by the end of 1991.
(In addition to CHEC and CHECMATE, a data processing companion code, CHEC-NDE, has now become available. Use of this code may improve the rigor of results previously obtained using CHEC alone.)
If you have any questions or comments concerning the above, or if you would like additional information at this time, please let me know.
Sincerely, HBR:bam cc:
Mr. L. E. Kokajko, USNRC Project Manager, SONGS 2 and 3 Mr. C. W. Caldwell, USNRC Senior Resident Inspector, SONGS Mr. D. F. Kirsch, Chief Reactor Safety Branch, Region V