ML20236B169
| ML20236B169 | |
| Person / Time | |
|---|---|
| Issue date: | 07/05/1987 |
| From: | Bennett Brady NRC |
| To: | Meachum T IDAHO NATIONAL ENGINEERING & ENVIRONMENTAL LABORATORY |
| Shared Package | |
| ML20236B036 | List: |
| References | |
| FOIA-87-465 NUDOCS 8710230374 | |
| Download: ML20236B169 (2) | |
Text
.
, ?
- u. u o \
~ Ju',y 5,1987
/
T0:' TERESA MEACHUM q
~FROM: DENNETT BRADY i
SUBJECT:
STATISTICAL ANALYSIS OF TURBINE AND MOTOR DRIVEN PUMP REPORT.
i( l
. Thank you 'for sending me the draft of .the statistical analysis section-j
\
and tables for the f6rthcoming pump report. I am retuining a copy of l the draft on which I have made'some suggestions and comments. _Your analysis, as always,was very thorough an'd tho'ug!(ful;.'t'his report l seems more readable
]
for the non-mathematical 11y' inclined than the previous reports. A few-general coments: l
- 1. At the beginnrigg of each variable section, right.after stating the ;
different values for the variable, state which were outiers and which were the' good performers and where that type of pump. is found. Then have"the' i i
}
paragraph on the test statistics. This was done in seme cases but not in 1 p!' $11.
- 2. For each of the statistical methods, be consistent in stating goo'd 1 then bad or bad then good perfomers. If bad then good, have the Page (Wcussion mention those with failure rates above-baseline, those with fluctuating rates and those below baseline.
- 3. Can .thb Weibull models where they were performed, be used to identify.
(by using the Wilbull mean or median) which are.the better performers't . l 4 ., Where two different models confiem or don't confirm the same odlliers, I yom might: point this out. '
5 [1 hMr.k the statement of decreasing failure rates is clearer in some.of 4
p/i- the earlier. reports. - !
ix),
\
jfoprietary information Attach 8dl
.poy.p yg 9, j 8710230374 871020 l at s) <" PDR F01A ~
3 4 GORDON87-465 PDR 7 N/.1 h a,
n,- - - _ - - -
iNW A- ,
7
+ , ,
. '1i
- -a;
. a
- m. . g . .~
.i
- 6. . I-am not quite: sure why 'in the tests _'of significance the values'.for.lwhich 9
there was little data were excluded in sone'.of the tests.- . . ,; /
i .
- 7. Is there a discussion of what we mean by failure'.- somewhere.in'the report 1 and what;NPRDS means by catastrophic, degraded ~and incipient?.
1
- . , q
'. :l d
1
.1
.i
. 4 j
- ]
1 1
y a
i i
-l _
.q
')
.1 cc J. Crooks R. Dennig G.L. Plumlee 1
_ _ _ _