ML20215N328
| ML20215N328 | |
| Person / Time | |
|---|---|
| Site: | Summer  |
| Issue date: | 10/30/1986 |
| From: | Hopkins J Office of Nuclear Reactor Regulation |
| To: | Nauman D SOUTH CAROLINA ELECTRIC & GAS CO. |
| References | |
| NUDOCS 8611050244 | |
| Download: ML20215N328 (6) | |
Text
'
i
[
o UNITED STATES
!"3,,g[g NUCLEAR REGULATORY COMMISSION M,5 /
E WASHINGTON, D. C. 20555
((
October 30, 1986 Docket No. 50-395 Mr. D. A. Nauman Vice President, Nuclear Operations South Carolina Electric & Gas Company P.O. Box 764 (Mail Code 167)
Columbia, South Carolina 29218
Dear Mr. Nauman:
SUBJECT:
NPRDS TRENDS AND PATTERNS ANALYSIS OF MAIN FEEDWATER REGULATING VALVE AND VALVE OPERATOR i
The NRC Office of Analysis and Operational Data (AE00) has just completed the i
first in a series of studies to analyze the most safety significant components using data from NPRDS. This first study identified a pattern of failure of main feedwater regulating valves and valve operators and the plants that had unusually frequent problems with these components based on NPRDS data from January 1984 through October 1985. We want to bring some of these results to your attention and to ask you assistance in addressing these findings.
To ensure that the components studied in our NPRDS Trends and Patterns Analysis Program were those of the greatest safety significance, plant safety functions were ranked in order of these risk importance for PWRs and BWRs, respectively.
For each of these functions, the NPRDS key components for that function were then ranked in order of their importance for the selected function. Tha most safety significant function for PWRs was the primary coolant heat removal. The main feedwater regulating valve and valve operator were two of the most important components for this function.
Our analysis indicated that the valve design features most often related to frequent failures and probleras with these components were forged type valve bodies, mechanical (differential pressure to open, spring force to closel type operators, and operators with a geared subtype. Our study also indicated that there were possibly other factors related to failure such as valve inlet size, and valve maximum operating temperature.
The reported rate of failure of these valves and valve operators decreases as the valve inlet size increases from 8 inches to 18 inches and increases as the reported maximum operating terrperature increases.
The study identified the plants which reported frequent problems and failures of these components.
Interestingly, the design features related to failure (forged body type, mechanical operators and geared operators) exist primarily at these units.
Thus, the question remains whether these failures and problems are caused by the factors identified in our study or by other conditions at the units.
8611050244 861030 PDR ADOCK 05000395 P
a October 30, 1986
,,_ p r Mr. D. A. Nauman Enclosed for your information is part of the report on the analysis of the main feedwater regulating valve / valve operator. We request that you review these results and respond with your analysis and conclusions based on your experience.
Since this is the first in a series of studies, we would also appreciate your coments on our approach of disseminating results directly to you for review and discussion.
We appreciate your attention to this matter. We would like your response 60 days from receipt of this letter.
The reporting and/or recordkeeping requirements of this letter affect fewer than ten reepondents; therefore, OMB clearance is not required under' P.L.96-511.
Sincerely, i
i
/s/
Jon B. Hopkins, Project Manager i
PWR Project Directorate #2 I
Division of PWR Licensing-A Office of Nuclear Reactor Regulation i
Enclosure:
As stated cc w/ enclosure:
See next page DISTRIBUTION Docket, File NRC PDR L PDR Gray File PAD #2 Rdg T. Novak OGC-Beth D. Ailler P. Tam E. Jordan ACRS (10)
J. Partlow Tech Branch B. Grimes N.~Thomp on B. Brady-AE0D b
)JHopkins:bl LRubenstein
'l 86
/
10/3c/86 10/3t/86 I
k P
Mr. D. A. Nauman South Carolina Electric & Gas Company Virgil C. Summer Nuclear Station CC*
Mr. William A. Williams, Jr.
Technical Assistant - Nuclear Operations Santee Cooper P.O. Box 764 (Mail Code 167)
Columbia, South Carolina 29218 J. B. Knotts, Jr., Esq.
Bishop, Liberman, Cook, Purcell and Reynolds 1200 17th Street, N.W.
Washington, D. C.
20036 Resident Inspector /Suniner NPS c/o U.S. Nuclear Regulatory Commission Route 1, Box 64 Jenkinsville, South Carolina 29065 Regional Administrator, Region II U.S. Nuclear Regulatory Commission, 101 Marietta Street, N.W., Suite 2900 Atlanta,. Georgia 30323 Chairman, Fairfield County Council P.O. Box 293 Winnsboro, South Carolina 29180 Attorney General Box 11549
~ Columbia, South Carolina 29211 Mr. Heyward G. Shealy, Chief Bureau of Radiological Health South Carolina Department of Health and Environmental. Control 2600 Bull Street Columbia, South Carolina 29201
~
's ENCLOSURE As of January 1986, the NPRDS data base had population and failure data for 121 main feedwater flow control valves (including their associated operators). The data are from 42 PWR units.
The primary statistical analysis in this study focuses on component lifetimes (i.e., times between failures).
In an effort to have as complete a data set as possible, lifetimes preceding 1984 t' hat ended in a failure in 1984 or were cut off at the end of the study were included. There were two groups cf valves, those with few failures and those with many failures, resulting in two groups of lifetimes.
In the first group, lifetimes of 10 years are common; in the other group approximately six months is the most comon. Thus, there are wide bounds on'some of the results of this study. Overall, the mean time to failure was 82 months.
The lifetimes were studied with regard to many possible factors that might influence their lengths. These factors, or variables, include most of the engineering infomation from the NPRDS component population records. The most-significant variable is probably the plant unit (although there are too many units for fomal statistical tests).
In any case, the plant unit provides a starting point to pursue insights from other factors.
With regard to other possible variables, the most striking cases of factors associated with short main feedwater regulating valve lifetimes are forged type valve bodies, mechanical (differential pressure to open, spring force to close)~ type operators, and operators with a geared subtype. The mean times to failure with the lifetimes grouped according to these variables are 8.74, 7.66, and 4.77 months, respectively. However, there is little in the narrative of the NPRDS failure reports to indicate why valves with these conditions have such high failure rates.
in addition, these conditions exist primarily at the units that reported low mean times.to failure.
-2 Valve inlet size, maximum operating temperature, and maximum operating pressure were identified by statistical methods for closer review and are quantitative variables. One of the statistical methods evaluated whether there could be a linear trend in the logarithm of the occurrence rate as these variables change. A trend was identified for inlet size and temperature but not for pressure. Specifically, the occurrence rate was found to decrease as the valve inlet size increased from 8 inches to 18 inches and to increase as the reported maximum operating temperature increased from the 400"F range through the 450'F, 500"F, and 900 F ranges. As with the valve body and operator type variables, these trends are associated with the plants reporting the most problems. Valve maximum operating pressure, on the other hand, was identified as being significant in that the lifetimes differed for different ranges of this variable, but there was no trend.
The variables, valve type and valve operator actuation speed were not identified as significant from the study component lifetim:s.
Data quality concerns precluded the analysis of internal and external environments, operating mode, and valve force rating.
Valve material and valve function / application were not analyzed because nearly all the valves are listed as carbon steel flow control valves and the function of all the valves is flow control.
A statistical method that was used for the component lifetimes producted smooth curves approximating the time-varying occurrence rate for each subgroup of the data. These curves in most cases were decreasing in tine; f.e., in most cases they indicate that the chance of a failure in the next month for a component is smaller if it has been several months since the last failure.
I An analysis focused on net failures in calendar time rather than lifetimes between failures of individual components. The results are similar to those already described in that failures are more closely spaced for some subgroups of the data than for others.
In addition, however, this methods permits one to observe cases of fluctuating occurrence rates in calendar time.
For example, 71 (66%) of 107 failures during the study period occurred in 1984 Thus, 5.9 failure reports
s
. l l
l were filed per month in 1984, while 3.6 were filed per month for main feedwater regulating valves in 1985. This pattern exists in spite of the fact that there j
were more plants reporting in 1985. This again points to action at the units.
Other cases of time periods with different failure rates were identified, but l
no general pattern has energed.
j A study of the failure reports is a first step in further identifyino and solving main feedwater flow control valve problems and thereby improving the availability and response of the main feedwater system.
Since many of the results that have been discussed relate to the plant units and stations that had the most reported problems, the records of those plants were reviewed. A variety of typical valve and valve operator problems occurred at those plants. Together, they represent 78 of the 107 failures covered in this study. As with the data as a whole, 64%
of the failures were reported for the valve operators. Another point of interest concerning those plants is that for those with multiple units at a site, all of the units at the site reported problems.
In conclusion, the following are some issues observed in the course cf this study that might merit further investigation:
- 1) geared subtype valve operators,
- 2) forged valve bodies, 3) mechanical valve operators, and 4) the decrease in problems reported from 1984 to 1985.
Finally, several data quality concerns have been raised from the viewpoint of improving the data base for future studies of this type. The issue that has had the greatest in. pact on this study is a concern about whether many of the lono lifetimes are valid.
,