ML20205H050
| ML20205H050 | |
| Person / Time | |
|---|---|
| Issue date: | 08/13/1986 |
| From: | Olson J NRC OFFICE OF INSPECTION & ENFORCEMENT (IE) |
| To: | Johnston W, Kane W NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION I) |
| Shared Package | |
| ML20205H052 | List: |
| References | |
| FOIA-86-891 NUDOCS 8608190697 | |
| Download: ML20205H050 (3) | |
Text
_ &; r 99 acey g '~ UNITED STATES 4 3 g NUCLEAR REGULATORY COMMISSION t j WASHINGTON. D. C. 20sSS \\,...../ AUG 131986
- MEMORANDUM FOR: William F. Kane, RI William V. Johnston, RI FROM:
Jon Olson, Contractor to Performance Indicators Section Events Analysis Branch Division of Emergency Preparedness and Engineering Response Office of Inspection and Enforcement
SUBJECT:
UNPLANNED SHUTDOWNS This memo outlines information pertinent to several of the unresolved issues surrounding the unplanned shutdown indicator. One issue that is not addressed is the relationship of the indicator to the " truth table." It will require several more days to get this information together because:
- 1) the " truth table" is undergoing revision, and 2) the machines here are currently dedicated to getting the data in shape to analyze.
As soon as possible, however, we will get you this information for your review. Issue #1: What is the overlap between automatic scrams and unplanned shutdowns? Are the two indicators redundant? Using Greybook data, there appears to be considerable, but by no means total over lap. For the fifty plants in the current analysis, the following break-downs are noted. 1984 1985 Number % of Total Number % of Total Manual Shutdowns 153 29% 142 27% Manual Scrams 71 13% 61 11% Automatic Scrams 313 58% 333 62% 537 100% 536 100% These numbers suggest to me that there is considerably more information in the unplanned shutdown data than in the automatic scram data alone. The question remains, however, whether this information is useful for the current effort. Issue #2: 00 the automatic scram data and the unplanned shutdown data tell you the same things about' plant performance? % y, ' @Q{ &O8l@ofq 7%
1 There are two ways to approach this issue:
- 1) determining whether the two indicators'are related to plant performance (the truth table or its derivative) in the same way, or 2) determining whether you get the same rank ordering of plants using the two indicators.
As mentioned above, we are not yet able to correlate the unplanned shutdown measure to a revised truth table. We can look at the correlations between automatic scrams and the other components of unplanned shutdowns, however. The following correlations answer the question of whether the plants that have a lot of automatic scrams are also the plants that have a lot of manual' scrams and/or manual shutdowns. Pearson Product Moment Coefficients 1984 1985 Manual Automatic Manual Automatic Scrams Scrams Scrams Scrams Manual Shutdowns .10 .20 .02 .22 Manual Scrams .18 .09 These correlations indicate that the three types of shutdowns are virtually independent of each other, at least at the yearly level. Plants with a lot of automatic scrams are no more likely to have a lot of manual scrams and manual shutdowns than are plants with few automatic scrams. The general indicator of unplanned shutdowns, therefore, appears to be multi-dimensional. When we relate unplanned shutdowns to the " truth table," I will try to relate both total shutdowns and each of the components so that we get a clearer understanding of the multi-dimensionality of this indicator. It would be helpful, however, to have your input on why there would be such low correla-tions among the three components of unplanned shutdowns. Issue #3: What is the relationship between shutdown information collected by the residents and shutdown information reported by the licensee? In order to explore the possibility that published sources of information on unplanned shutdowns give incomplete information as to their number and nature, residents were requested to provide data on unplanned shutdowns for the first quarter of 1986. This section outlines a comparison of the resident generated data to the Greybook data for the same period. Data were available from residents for 44 plants. For these plants, the residents reported sixty-one unplanned shutdowns. For the same plants, seventy-one unplanned shutdowns were listed in the Greybook. There was an overlap of fifty-nine shutdowns. The residents reported several shutdowns that were from low levels of power (e.g., 3%) and that were not reported in the Greybook. The Greybook provided data on shutdowns that were not mentioned by the residents for one or more reasons: (1) the residents forgot.to report it (2) the residents combined several shutdowns into a more complex outage
as
- (3) the shutdown in question did not involve the reactor going sub-dritical and did not, therefore, reflect the TI instructions to the residents.
For the fifty-nine shutdowns listed by both sources, there was general agreement on timing, nature, and causes. With very few exceptions, the residents's narratives and the Greybook narrative identified the same components / systems and sequence of events. While power levels are not given, per se, in the Greybook, they may be easily inferred from the power graph provided. An area where agreement was less perfect, however, concerned causes of the events. There was very good agreement on cause for forty-eight of the fifty-nine shutdowns (81%). For the other eleven, there was a pronounced tendency for the Greybook to report shutdowns as due to equipment failure and for the residents to trace the cause back further to design, installation, or human-error. Thus, relying on Greybook data in the future may result in a systematic under-recording of design, installation and human-error problems. A decision must be made whether the additional richness of detail provided by the resi-dents is sufficiently important to justify the additional data collection burden on the residents. Origleef W W Jon Olson, Contractor to Performance Indicators Section Events Analysis Branch Division of Emergency Preparedness and Engineering Response Office of Inspection and Enforcement Distribution: DCS EAB R/F DEPER R/F EJordan SA hwartz Allison J01 son RSingh
- 3). s _,
IS:DEPER:IE FIUf6EPER:IE J0lson:md RSingh 8/ 3/86 8//]' 86 f / _}}