ML20054K641
| ML20054K641 | |
| Person / Time | |
|---|---|
| Issue date: | 06/30/1982 |
| From: | Disalvo R, Gery A, Pittman J NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| To: | |
| References | |
| NUREG-0872, NUREG-872, NUDOCS 8207060004 | |
| Download: ML20054K641 (96) | |
Text
l NUREG-0872 A Feasibility Study of Using Licensee Event Reports for a Statistical Assessment of the Effect of Overtime and Shift Work on Operator Error U.S. Nuclear Regulatory Commission Office of Nuclear Regulatory Research R. DiSalvo, A. Gery, J. Pittman f.... c,,
s...h-d G
7 4 820630 0872 R PDR
NOTICE Availability of Reference Materials Cited in NRC Publications Most documents cited in N RC publications will be available from one of the following sources:
- 1. The NRC Public Document Room,1717 H Street, N.W.
Washington, DC 20555
- 2. The NRC/GPO Sales Program, U.S. Nuclear Regulatory Commission, Washington, DC 20555
- 3. The National Technical Information Service, Springfield, VA 22161 Although the listing that follows represents the majority of documents cited in NRC publications, it is not intended to be exhaustive.
Referenced documents available for inspection and copying 'or a fee from the NRC Public Docu-ment Room include NRC correspondence and internal NRC memoranda; NRC Office of Inspection and Enforcement bulletins, circulars, information notices, inspection and investigation notices; Licensee Event Reports; vendor reports and correspondence; Commission papers;and applicant and licensee documents and correspondence.
The following documents in the NUREG series are available for purchase from the NRC/GPO Sales Program: formal NRC staff and contractor reports, NRC-sponsored conference proceedings, and NRC booklets and brochures. Also available are Regulatory Guides, NRC regulations in the Code of -
Federal Regulations, and Nuclear Regulatory Commission Issuances.
Documents available from the National Technical Information Service include NUREG series reports and technical reports prepared by other federal agencies and reports prepared by the Atomic Energy Commission. forerunner agency to the Nuclear Regulatory Commission.
Documents available from public and special technical libraries include all open literature items, such as books, journal and periodical articles, and transactions. Federal Register notices, federal and state legislation, and congressional reports can usually be obtained from these libraries.
Documents such as theses, dissertations, foreign reports and translations, and non NRC conference proceedings are available for purchase from the organization sponsoring the publication cited.
Single copies of NRC draf t reports are available free upon written request to the Division of Tech-nical Information and Document Control, U.S. Nuclear Regulatory Commissinn, Washington, DC 20555.
Copies of industry codes and standards used in a substantive manner in the NRC regulatory process are maintained at the NRC Library, 7920 Norfolk Avenue, Bethesda, Maryland, and are available there for reference use by the public. Codes and standards are usually copyrighted and may be purchased from the originating organization or, if they are American National Standards, from the American National Standards institute 1430 Broadway, New York, NY 10018.
$5.50 GPO Pontent wov pocer
NUREG-0872 RX
_ = = _
A Feasibility Study of Using Licensee Event Reports for a Statistical Assessment of the Effect of Overtime and Shift Work on Operator Error
$ ate Iu'Nshe une 1 R. DiSalvo, A. Gery, J. Pittman Division of Facility Operations Office of Nuclear Regulatory Research U.S. Nuclear Regulatory Commission Washington, D.C. 20555 o
e ** *' 4,
4 g
i
TABLE OF CONTENTS Section Page No.
1.0 PURPOSE.........................................
1 2.0 RESULTS 1
3.0 DATA REVIEWED...........
2 4.0 ANALYSIS....
3 4.1 LER Compilations 3
4.2 Statistical Section....
3 3
4.2.1 The Hypotheses 4.2.2 The Statistical Tests....................
12 4.2.2.1 Primary Procedure..
13 4.2.2.2 Secondary Procedure.
13 13 5.0 RESULTS 5.1 The Statistical Hypotheses 13 5.2 Statistically Significant Trends 15
- 5. 3 Uncertainties and Limitations 20 20
6.0 CONCLUSION
S APPENDIX A - THE LER DATA AND CELL CONFIGURATIONS A.1 The LER Data A.2 Cell Configurations A.3 Raw Data i
TABLE OF CONTENTS (CONTINUED)
Section Page No.
APPENDIX B - STATISTICAL METHODS B.1 The Factors Upon Which the Conditional Validity of the Statistical Results Rest B.2 Terminology and Notation B.3 The Poisson Random Variable B.4 Significance Levels B.5 The Statistical Tests B.6 H and H Under the Poisson Assumption u
y APPENDIX C - DETAILS OF STATISTICAL RESULTS FOR 7 CELL CONFIGURATIONS STATISTICAL BIBLIOGRAPHY REFERENCE BIBLIOGRAPHY LIST OF TABLES Table Page No.
3.1 Results from the LER Screening Process..................
2 4.1.1 Sample Error Distribution Across Days of the Week By Individual Years...
11 4.1.2 Sample Error Distribution Across Days of the Week Arranged Into Three Subsets 11 5.1.1 Results of the Primary Procedure for H 16 u
5.1.2 Data Display and Results of the Primary Procedure for H 17 y........
,1 ii g
LIST OF TABLES (CONTINUED)
Table Page No.
B.5-1 Results of the 2-sided Likelihood Ratio Test for Cell Configurations 1-7...................................
B.5-11 & 12 C-2 C-1 Cell Configuration #1 C-2 Cell Configuration #2.....................................
C-3 C-3 Cell Configuration #3......................................
C-4 C-4 Cell Configuration #4.....................................
C-5 C-5 Cell Configuration #5....................................
C-6 C-6 Cell Configuration #6....................................
C-7 C-7 Cell Configuration #7..................................
C-8 LIST OF FIGURES Figure Page No.
4.1.1 1978 Histogram of Time-identified LERs...................
4 4.1.2 1979 Histogram of Time-identified LERs.......
5 4.1.3 1980 Histogram of Time-identified LERs...................
6 4.1.4 1981 Histogram of Time-identified LERs 7
4.1.5 1978, 1979 & 1980 Histogram of Time-identified LERs 8
4.1.6 1979, 1980 & 1981 Histogram of Time-identified LERs..
9 4.1.7 1978, 1979, 1980 & 1981 Histogram of Time-identified LERs 10 4.2.1 Illustration of the 7 Cell Configurations 14 5.1.1 Ceasistency of "CC #7 Pattern "Over Days of the Week.
18 5.1. 2 Consistency of "CC #7 Pattern" Over Time..
19 UAY",
6.1.1 Distribution of Errors in the Material from 1912 to 1931 l,'j in 24 Hours of the Day and Night (Bjerner, et al., 1955)
.?
Plotted on the 1978, 1979, 1980 & 1981 Histogram of Time 21
- 3 Identified LERs (Figure 4.1.7) e.
91 111
- l. 0 PURPOSE Since the TMI incident considerable interest has been focused upon the human factor aspects of nuclear power plant operation.
One of the concerns hasbeentheeffectsofshiftwork(rotationandlength)andovertimeogoperator performance.
In its request for human factors research (RR-NRR-81-2)g
, the Office of Nuclear Reactor Regulation requested research be performed to identify the effects of shift work and overtime on operator performance.
The staff of the Office of Nuclear Regulatory Research performed the study here as an initial response to that request.
The purpose of this reported preliminary research is two-fold:
1.
To determine the adequacy of using the Licensee Event Reports (LERs) to evaluate:
a.
The effects of shift work and overtime on human performance in nuclear power plants.
b.
The statistical nature of human error in nuclear power plants.
2.
Given that the LERs are adequate for 1.b above, to draw statistical conclusions about human performance in nuclear power plants.
2.0,RESULTS It was determined that the LERs are useful in evaluating the statistical properties of human error in nuclear power plants given that certain conditions regarding the LERs hold, (see appendix B.1).
However, we determined that the LERs are not adequate for assisting in the evaluation of the effects of shift work and overtime on human performance.
A statistical analysis of the data in the LER files from January 1978 through June 1981 indicates that the distribution of human error times for licensed operators is not uniform throughout the day.
In the sample, errors occurred more often in some time intervals and less often in others than would be expected under the hypothesis of uniformity and given equal opportunity for error.
The analysis also indicates that the distribution of the number of errors per day is not uniform throughout the week.
Errors occured most often on Wednesday and least often on Sunday, with the frequency histogram over the week taking on a rough bell-like shape.
Sufficient data (quantitative and qualitative) are not available in the LERs for determining the reason for these variations.
The limited descriptive data supplied in the LERs did not enable us to distinguish the effects of shift work or overtime from other potential causes of degraded performance.
We can calculate some of the statistical properties of human error in nuclear power plants.
However, we cannot be certain of their validity because we do not know if we have met the requirements described in Appendix Bl.
[1] Memorandum from Harold R. Denton, NRC, to Robert B. Minogue, "NRR Research Needs for Human Factors Safety," March 27, 1981.
T 1
3.0 DATA REVIEWED l
The data which were reviewed for this analysis were the LERs identified as personnel error for the years 1978 through 1980 and the LERs for 1981 which were stored in the NRC LER computer file as of June 30, 1981.
The set of LERs identified as personnel error is further divided into more specific subsets according to the type of personnel involved.
The subset categories of personnel errors are identified as:
Licensed operators and senior operators, nonlicensed operations personnel, maintenance and repair personnel, radiation protection personnel, construction personnel, contractor and consultant personnel, and a category entitled "other."
Since the purpose of this analysis was to measure the effects of shift work and overtime on human performance for plant personnel the subsets of errors attributed to construction personnel, and contractor, and consultant personnel were deleted from the data base. We planned initially to analyze three different classes of plant personnel, (1) licensed operator and senior oper-ators, (2) maintenance and repair personnel, and (3) a group consisting of nonlicensed operations personnel, radiation protection personnel and the
" category" called "other."
The final effort, however, was restricted to licensed operators and senior operators only.
Excluded, nonlicensed personnel included the auxiliary operators.
The operator error LERs for the three and one half year period were carefully reviewed to determine those errors for which the occurence times were recorded.
Table 3.1 provides the results of this screening process.
As can be seen in Table 3.1 only approximately 14% of the licensed personnel LERs reported the actual time of occurrence.
Actually, some of these 84 LERs did not have reported times, but rather were described with sufficient supporting information to allow us to estimate error time by judgement.
Appendix A gives a more detailed description of the data.
l TABLE 3.1 RESULTS FROM THE LER SCREENING PROCESS YEAR TOTAL TOTAL TOTAL LICENSED LER PERSONNEL LICENSED PERSONNEL ERROR PERSONNEL ERRORS ERROR SHOWING TIME 1978 3,168 532 214 28 1979 3,162 626 151 20 1980 3,706 584 185 26 1981 (Available on 6/30) 1,901 159 49 10 TOTAL 11,937 1,901 599 84 2
4.0 ANALYSIS One incentive for this study was to establish the degree of validity in the commonly expressed opinion that the errors are more likely to occur in the pre-dawn than at other times of the day.
The technical approach used here was to hypothesize that the number of errors occur uniformly throughout the day and week and then examine the data for any statistically significant deviations from that hypothesis.
To make this determination the number of errors was plotted as a function of the time of day for each year's data.
Then, several statistical methods were employed to determine if there were intervals during the day in which the number of errors differed significantly from expectations.
4.1 LER COMPILATIONS Figures 4.1.1 through 4.1.4 are histograms of the LERs for each year from 1978 to 1981.
Figure 4.1.5 is the histogram of the errors for the years 1978, 1979 and 1980 and Figure 4.1.6 is the histogram of the errors for the years 1979, 1980 and 1981.
Figure 4.1.7 is a histogram for all four years.
The purpose of examining the subsets of data given in Figures 4.1.5 through 4.1.7 was to help in determining if the conclusions based substantially on pre-TMI data continue to hold when the analysis is based on most'y post-TMI data.
The errors occurring by days of the week for each year is shown in Table 4.1.1.
Table 4.1.2 shows the data arranged in three subsets; the years 1978, 1979 and 1980 summed together; the years 1979, 1980 and 1981 summed together; and all four years summed together.
4.2 STATISTICAL SECTION 4.2.1 THE HYP0 THESES The statistical section of this study addresses part (2) of the central purpose:
To draw statistical conclusions about licensed personnel errors
- in nuclear power plants that will be conditionally valid given the representative-ness of the data and the validity of aggregating the data from all plants.
This is a very general and broad task.
We made it more manageable, specific, and well-suited to the LERs by reducing it to the following tasks:
testing two statistical hypotheses and indicating some statistically significant trends.
These hypotheses and trends are considered equally applicable to all plants.
Hypothesis #1, (H ), is that the distribution of error occurrence times is uniform over the 24" hour day.
Hypothesis #2, (H ), is that the distribution ofthenumberoferrorsperdayisidenticalforeabhdayoftheweek.
H is equivalent to the hypothesis that the probability of an error occurring wYthin a given time interval is proportional to the length of the interval.
This implies that the expected fraction of total daily errors occurring within any time interval is equal to the fraction of 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> formed by that interval.
H is equivalent to the hypothesis that errors are equally likely for each day 1
or the week.
Like the case with H implies that the expected fraction of total weekly errors occurring withYn, Hakydayis1/7.
It is important to keep in mind the distinction between the two types of random 'cariables of concern
- Hereafter referred to as just " errors".
3
l 28 ERRORS LEGEND l
Reported occurrence time
@ Time interval within which error must have occurred, O
based on LER data and judgment Histogram bars drawn only for "I" errors
@ 's are separate from "I's" O@
w O
Number 6-l C
l of Errors 5
4 to 3
2 e
c.-
isi!!(
1
- l..==== ;
~"
.s.'<
5l lf !5 -5lg li!ElIn:..
-!l~lllif...
iifl g
- jilE l$3173 g
g g
g
- l l
l jg5 0
1 2
3 4
5 6
7 8
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Time of Day (Hours) i i
Figure 4.1.1 1978 Histogram of time-identified LERs
,l d,e r
rucc 4
o 2
eva 3
h 2
tsu r
2 s
mt o
2 n
r r
r o e e
m r
1 reg '
2 d
l "s e
h u
cj mi hd fo 0
i r l 2
t wn y m j
e a
ln o j
9 c
nia n
r i
ht of e
a j
1 t
r id ne w
wta ru R
c laE r
p a
8 r
c a
1 o
vL d
r e
d en s s t
e n o r
t i d b r
7 s
a e R
r 1
o e e a
E p
ms ms L
e ia a'
6 R
Tb r
1 d
WH@
g D
e o
i N
t j
if is j
5 t
E 1
n Gl j
e E
d L
}
4 j
i -
1 e
j
)
S j
s m
r j
R j
u it 3
o 1
f H
O
(
o 2
y m
R a
a 1
D r
R g
f o
1 o
E t
e s
1 m
iH 0
i 2
0 T
9 1
79 1
9 2
1 4
8 e
rug 7
iF
- } 6 5
4 3
, 2 1
0 6
5 4
3 2
s r
r e o brmE r
u f No w
l l
ll
l l
LEGEND l
Reported occurrence time
@ Time interval within which error must have occurred.
based on LER data and judgrrent Histogram bars drawn only for "l" errors
@ 's are separate from "l's" 26 ERRORS a
O b
Number 6-of Errors 5
4 cn 3
e m
2 1
m!sisisisi!!!
=~
= ~ ~
.=
is!!jlli,,!illil !!:iit llfih!l[il_
i l.
n ill"3!ijfisij!!!mi!l!i!!!
iliilik "Il
!!!i!!!
Illi HliII 7
i i
i O
2 3
4 5
6 7
8 9
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Time of Day (Hours)
Figure 4.1.3 1980 Histogram of time-identified LERs l
i
1 l
d,e r
ru cco eva h
4 ts 2
u s
mt o 3
r n
r l
2 r
r e
o e
m rreg '
2 d
l "s I
e hu 2
cj m
i o'
r l i
hd f
1 t
wn e
a m
I 2
n i a ly c
n n ro ht of e
a 0
t r
id ne 2
w r
c iR wta u
r a
c E
r a
9 o
vL d p r
1 e
d en r
s s
t t
in o e
a e d
b r
8 ro e e a
1 e ims ms p
a a'
R Tb r
f.i o@
7 g
l s
D 1
R
+
N t
is E
E L
H 6
Gl ll 1
d E
ie L
f 5
i t
1 ne d
4 i
1
)
e S
s m
e r
ml u
i R
3 o
t l
1 H
f O
(
o y
m R
l 2
a 1
D a
r R
f g
1 o
o E
O i
1 t
e m
is 0
i' i
H 1
i 0
T 1
1 89 n"
i 9
1 4
1 4
8 e
rug I
7 iF I
6 5
. i!
.. z.I
. c[
4
- l 3
2 i
1 i 0 6
5 4
3 2
1 s
r r
e o brmE r
u f No y
i il' jll l)
,i l
M l
2 2
4 4
l 3
3 2
2 2
2 l
2 2
7 1
1 l
2 2
0 0
I 2
2 7
9 9
l 1
- @T 1
..,;a s
8
- i R
l 1
- l 8
=
1 E
=
L
=
d 7
=
7 l
1 l
1 ie f
i tn 6
6 e
l 1
1 id
-e 5
xi 5
m i
1 pl 1
1 i
t 4
f
- 1 4
o
)
S l
1
!:l 1
s m
y..
r R
3.
uo a
O l
3
-~
3 r
H g
1 1
(
o e
y t
R R
l 2
m 2
a is D
H 1
MI i
1 T
f 0
E o
8 1
1 pf e
9 1
1 1
m 4
i E
t 7
l 0
- i 0
T
+l 1
9 1
79 1
- 1 l
9 pl-9 879 1
I 8
- a. ][
8 5
4 7
wl 7
4 i
1 e
M l
6 6
g r
- i u
oj iF M
1 5
a}
5 nl
- }
4
/ ol 4
^
l 3
a :l 3
- l M
l 2
- -f 2
- f ai O:
l 1
1
.I 0
0 8
6 4
2 t
ly ly r
r o
s o u nt o o u d
t o h
h n sd sd r
sr en a ot oten r
r al) e or n B
al) r n r
a b r r
a ei noiv ei noi v r
r or ph e mEr ph r r
e oni r
t u
onit t
t r
Fi n
Fi n
f E
I pwi No( pwi
)0 l
l1
I I
4 4
2 2
I 3
3 2
2 I
2 2
2 2
=,
1 1
I 2
2 0
0 1
2 2
s l
9 9
n l1 1
1 s
R
- gf I
8 8
E Ol 1
i[;
1 i S L
d l
7 7
ie 1
f 1
i tn 6
6 e
I 1
id 1
-e
- i 5
m f
5 al 1
it E
1 s
f j.::
o S
l 4
ul 4
)
1 1
m s
ru a
R l
3 3
o rg O
1 1
H o
(
t e
y s
R l
2 m
2 a
i H
1 i
3 1
D R
T
@O l
1 1
f 8
-. ?
o 9
E 1
$;l 1
1 1
e m
6
- i 0
iT 5
I 0
wl 0
1 j;
1 89 1
O :
I 9
~.l 9
i 979 1
q1 l
I 8
- l 8
6 1
I 7
i 7
4 sl p:
e I
6 iel 6
ig r
q u
p-F
- i I
5
<l 5
s W
I 4
- l 4
.;i l
3
- l 3
- l l
2 vj 2 i
- f l
1 i
1 s.
l I 0 l
O
~
8 6
4 2
t ly y
r l r o
u s
o u nt o d
t o
h h
nsd sd r
en s r a ot e n r
ot r
al) e or n B
a
)l r
n r
a r
a n
b r ei oiv ei n
r oiv r
or ph mE r
r e
r ph e
r o t
nit u
onit t r
F f
iwi No( pwi p
n Fi n
E(
e
lt l
o 1
2 2
4 4
T l
t 3
e 3
2 2
2 2
I 2
2 e
s3;s 1
1 I
2 2
I3 0
t 2
02 T
I 9
if 9
s 1
8{
R 1
E L
I 8
s, i 1
i:M[
8 1
d ie f
7 i
7 t
i 1
1 ne d
i 6
6 i.
1 e
1 T
\\
m i]
it n
I 5
if 5
i 1
- !.!i 1
f
_sii m
s o
- g S{
4
)
r 4
i.
]!
I 1
1 a
i'.
s r
g u
o I
3 3
o t
ii 1
H is 1
e
(
H m
y I
2 i 2
a 1
1 1
D 8
1 T
9 o
f 1
nv F I
1 i'
1 e
o i
1 8 {i 1
e m
0
- f.
i 8
I 0
$j 0
T 9
1
" i::
i!
1 1
gi i
ii.
9 c
I 9
7
$s {:
9 9
i II i!
5 1
~
8 8
I
'l 8
Ol 8
7 tI i
i.I 9
1 Sj O.
I 7
i.::_
7 7
_4:
4 l
1 55 i!
e H
I 6
$l!'
6 r
.m.
s u
s Ii i8i ig
_:::-p H
l 5
8j 5
F s
- ii K
^Y 4
p
- 5 ss i1 E,
4 i
nv f
3 f-;.
- I i:
Si i
h;.: ' ::j 3
hg. s.
hi
- = ',,,
o f
2 i:8j 2
i)i F
i'!
l i
I 1
I 0i fh 1
Ei K
0 0
I!
~
8 6
4 t
ly ly r
r o o u o u s
nt o d
t o h
h nsd sd r
r aoten s
en r
ot r
al) e o n
B r
al) r n
r a
b r
a ei noi v einoiv r
r or phr mE r
r ph r e
e r onit u
onit t
t r
Fi n
f Fi E( pwi No( pwin g
Table 4.1.1 Sample error distribution across days of the week by individual years S
M T
W Th F
S Total 1978 3
5 3
7 3
2 4
27 1979 1
2 4
5 4
3 1
20 1980 1
2 6
6 4
4 3
26 1981 0
0 4
2 2
1 1
10 Total 5
9 17 20 13 10 9
83 Expected totals 11.86 11.86 11.86 11.86 11.86 11.86 11.86 83*
ander H g Table 4.1.2 Sample error distribution across days of the week arranged into three subsets Expected Day Data daily subset 5
M T
W Th F
S error count Total 78,79,&80 5
9 13 18 11 9
8 10.43 73 78,80,&81 2
4 14 13 10 8
5 8
56 78,79,80,&81 5
9 17 20 13 10 9
11.86 83*
AAn error contained in the 1978 data could have occurred on either Tuesday or Wednesday, and hence, was discarded.
This resulted in the total of errors distributed across the days of the week to be 83 instead of 84.
11
here.
H involves error occurrence times (continuous random variables), and H invo1Ves the number of errors per day, (discrete random variables).
y For dealing with these hypotheses, we make two basic statistical assumptions:
i) The distribution of human error occurrence times over the 24-hour day is identical for all days and plants, and that these error times are independent of the error times of all other days.
ii) The distribution of the number of errors per day depends upon the day of the week, but not the particular week itself, or the particular power plant.
Also assumed is that the number of errors per day is independent of the total on any other day.
For example, the number of errors occurring on any Monday is independer.tly, identically distributed as the total for any other Monday, and is independent of the totals for all other days.
The same holds for the other six days of the week.
l The independence aspect of these two assumptions is reasonable in light of the nature of human error in nuclear power plants.
The plants have many operators working over a few shifts to a large extent on independent tasks.
Also, most errors are relatively small ones whose threats are almost always neutralized by a sophisticated safety backup system, rendering them relatively inconsequential.
Hence, they are often not reported at large or even learned of for days or even weeks.
Hence the total population of operators may often be unaware of the errors of the previous days or weeks.
Thus, their own care and attentiveness won't be immediately affected by earlier error occur-rences, as they normally might be if these errors were immediately made known.
Such an awareness and reaction response would appear to be the major mechanism for dependence between daily error counts or occurrence times.
Hence, its absence, or relatively minor presence, adds to the plausibility of independence.
4.2.2 THE STATISTICAL TESTS Combinations of three statistical tests are applied to the hypotheses:
2
-The Goodness of Fit Test (X =G0FT)
-The Conditional Binomial Test (CBT)
-The Likelihood Ratio Test (LRT)
These statistical tests are employed in two combined approaches to test each 2
of the hypotheses.
In the first approach the X _G0FT is applied under assump-2 tions i and ii only.
The multinomial assumption of the X,G0FT is implied by each of these hypotheses.
The second approach consists of making a Poisson assumption in addition to assumptions i and ii and then applying the Conditional Binomial Test and the Likelihood Ratio Test, (applied to H only).
These two tests are the most appropriate tests which can be applied Onder the second approach.
The Poisson assumption is widely accepted as one of the best for modeling rare count data.
(See Appendix B-6).
The nature of the combination will be described in the next section.
The advantage of combining these two approaches lies in the fact that a set of results that hold with and without the Poisson assumption is more credible than a set of results which hold either only with or only without the Poisson assumption.
12
These tests and procedures to apply them are described in more detail in the Appendix B.S.
They are described briefly here.
4.2.2.1 PRIMARY PROCEDURE This study takes a conservative stand and will reject H only if all the tests show significance.
In this way, if H is rejected, thO rejection will hold both with and without the Poisson assufrIption.
The basis for applying each test to H consists )f dividing the day into disjoint time intervals, (and thus creating N " cell configuration," see Figure 4.2.1), counting the number of errors in the sample that fall into each cell, and then calculating the "p-value."
The "p-value" is the likelihood of observing a sample cell count equally or more extreme than that of the given sample, under H and the assumptions of u
the particular test in use.
This routine is repeated for 3 subsets of the data and for 7 different cell configurations; (21 repetitions). We created 7 different cell configurations, some by random choice independent of the data and some as suggested by patterns in historical data.
The purpose of data subsets and the choice of cell configurations will be discussed in the appendix.
p-values foe a(i.e., conclude the observed distribution is nonuniform) if the We reject H ll tests and the largest data subset are significant, (p<0.01*),
for at least one of the seven cell configurations.
For the second hypothesis, H, only the X _G0FT and the Likelihood Ratio 2
Test (LRT), are applicable.
Therbisonlyonecellconfiguration,itscells being the days of the week, and the week taking the place of the day in the y (if the p-values for above definition of " cell configuration." We reject H p<0.05)*
both tests and the largest data subset are significant, 4.2.2.2 SECONDARY PROCEDURE Regarding H.
after analyzing the 3 data subsets on all 7 cells configurations, we isolated the cell configuration which revealed the greatest deviation from uniformity.
Then, to obtain further insight into the consis-tency of this deviation over the years, we applied the statistical analyses to that same cell configuration for each of the single year data subsets 1978, 1979, 1980, and 1981.
Furthermore, the statistical analyses were also applied to this cell configuration with the entire data set for each day of the week individually, to check for consistency over the days of the week.
No secondary analysis was done regarding H.
g 5.0 RESULTS 5.1 The Statistical Hypotheses.
3 cellconfiguration#7(CC#7)lOadtokignificantp-valuesforalltesEs,only Primary Procedure - Both H and H were rejected.
In rejecting H and all the data subsets.
In fact, it was the only cell configuration that lead to significant p-values for the largest data subset, and hence H could not have been rejected without it.
However, the p-values for CC #6 Eame very close to significance, ranging from above 0.01 to 0.04.
The only other source of
- See Appendix B.4 for a discussion on the choice of significance level.
13
TIME I
I I
I I
I I
I I
I I
I I
I I
I I
I I
I I
I I
l l
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 CELL CONFIGURATION For each cell configuration, each pattern represents exactly one cell.
.i.',.
((((((((((((((((
- 1
{J-4;;
3 cells 7
8 16
' '.' (-
~
..,f '.*. ;
- f 2 cells 9
16
- 3
[
'-L,.'
2 cells
),:,C
- y.?.[.
~ ~.; J 11 16
.,[
U 2 cells
[. ',, *-[
}
..[,
.[
I-J s
13 16
- 5 4 cells
/
6 12 18
- 6
.l [.~.'... : l._
N!flllllllll) fllllllllllllllll) 3 ccIls 8
13 16
~
' ff [) l _
ffffffflllllll)
, _ O j ~ 'L;'
! ' l ', -
3 cells 10 13 16 Figure 4.2.1 Illustration of the 7 cell configurations 14
significant p-values was CC #4, on all 3 tests, for subset 79-80-81 only.
(Table 5.1.'.).
In rejecting H, both the X*=G0FT and the LRT lead to significant p-values
~
y for 2 of the 3 data subsets:
79-80-81 and 78-79-80-81.
(Table 5.l.2).
Significance in the largest subset was crucial for rejecting H, since we y
required tt'is for rejection in the design.
The major results are that the distribution of error occurrence times is.
not uniforta over the 24-hour day, and-that errors are not. equally likely for all days o the week.
It is reassuring to note that these results hold with and withou' the Poisson assumption.
Seconilary Procedure - The pattern revealed by CC #7 is suf.ficiently _
consistent that it remains visible, (using the entire data set), for each of' six days o f the week individually, all except Sunday, and it remains well pronounced in 3 of these:
Monday, Wednesday a'nd Saturday.
(Figure 5.1.1).
The patter 1 is so well pronounced for Wednesday and Saturday that it is statistically significant at the.06 level' via the LRT.
(The X,G0FT is not applicable since the cell means are too low).
Thus, H is almost rejected by two individual days alone.
Recall that rejection on tNe basis of individual days was not necessary to reject H 'u The consistency of the "CC #7 pattern" ove the' days of the week lends credence to part of our first basic statistical assumption in section 4.2.1:
i.e., that the distribution of error occurrence times over the day is identical for all days.
Some consistency over the years is also evident.
(Figure 5.1.2).
When the data are left aggregated over the days of the week, the pattern.. remains visible for each of the 1978, 1979, 1980<, a1d 1981 data sets' individually.
It is well pronounced for 1978, 1979, and 1980, and so striking for 1978 and 1979 that it is statistically significant for 1978 and almost, so for 1979, (p =.06),
2
~
(Again, the X,G0FT is not applicable because some or all of the via the LRT.
cell means are too low).
5.2 STATISTICALLY SIGNIFICANT TRENDS
't In rejecting H we observed that for the time -intervals of (0-10)-and (13-16) the proportTo,n of errors was higher than'would be expected'under H '
u Thus the trend in which errors are more likely during these intervals is statistically significant.
Since H was rejected for all 3 data subsets, this trend, (the pattern of CC #7), is sEatistically significant when based on either 5/12, 2/7, or only 1/10 pre-TMI data.
In rejecting H, we observed that errors occurred more frequently on Tuesday, Wednesday, and Thursday than would be expected under,H.
Thus the trend in which errors are more likely on these days is statistibally significant.
l s
\\
15
Table 5.1.1 Results of the primary procedure for H u
Cell Data subset configuration 78,79,&80 79,80,&81 78,79,80,&81
- 1 NS NS NS
- 2 NS NS NS
- 3 NS NS NS p<.01
- 4 NS p<.01 NS p<.007
- 5 NS NS NS
- 6 NS NS NS
- 7 p<.005 p<.007 p<.005 p<.001 p<.01 p<.003 So we reject H n the basis of C.C. #7.
u NS - not significant ND 2
Significant p-values x -G0FT listed in this order LRT C.B.T. (if applicable) 16
2 1
Table 5.1.2 Data display and results of the primary procedure for H y Sample error distribution across days of the week i
Expected Day Data daily subset S
M T
W Th F
S error count Total 78,79,&80 5
9 13 18 11 9
8 10.43 73 78,80,C81 2
4 14 13 10 8
5 8
56 78,79,80,&81 5
9 17 20 13 10 9
11.86 83*
Statistical results Data LRT 2
i subset X -G0FT (c =c =...=c *1) 2 3 7
l 78,79,&80 X$ = 9.94
.15<p<.10 X{ =9.66
.15<p<.10 79,80,&81 X$ = 15.75 p<.02 xl = 16.96 p<.01 78,79,80,&81 xl = 13.56 p<.05 x[ = 13.59 p<.05 i
50 we reject H g
- An error contained in the 1978 data could have occurred on i
either Tuesday or Wednesday, and hence, was discarded.
This resulted in the total of errors distributed across the days of the week to be 83 instead of 84.
I 4
i I
17
33.7 Using the entire data
[" Observed number of errors set only('78, '79, '80, Et '81h Expected number of errors 2
21 M X - statistic for testing 44 uniformity with K + 1 cells.
LRT - (Based on the likelihood ratio test and the Poisson 33.3 I
2 A cell deviation in direction N
from the CC #7 pattern X -GOFT - (Based on the chi-square goodness of fit test.)
15 10 I All da s o the week v
u 2
,13.19, p <.003 (LRT) 2]12.61, p <.005 (x -GOFT) 2 0
HIGH 10 13 16 LOW 24 HIGH 3 m 2.29 1
m 2.08 u
1_.-
0.625 u 2
g x, = 1.09 0.55 < p < 0.65 (LRT)
HIGH LOW jJ4.13 hh3.75 u 2%.125 V 2
1 y = 5.11 0.05 < p < 0.09 (LRT)
Monday illGH HIGH LOW 2
6.88
_Sg 4bl Er3 x, = 2.39 0.10 < p < 0.15 (LRT)
U 2
HIGH LOW 9.17 8.33 t
enI\\.
u 3 m 2.5 v
2 Wadnesday y, = 6.02 p < 0.05 (LRT)
HIGH HIGH LOW
~
4 g 5.5 y 1 m1.5 u
2 Thursday y# = 1.35 0.50 < p < 0.55 (LRT)
HIGH LOW 3d 4.17 u 1 m.25 u
2 1
p y = 1.39 0.45 < p < 0.50 (LRT)
HIGH LOW l
l 3 J 4.13 S Q 3.75 y 3 % 1.13 u
2 ur ay y* = 5.93 p < 0.06 (LRT)
HIGH HIGH LOW Figure 5.1.1 Consistency of "CC #7 Pattern" over days of the week 18
4 Observed number of errors X * - statistic for testing h
Expected number of errors 21 uniformity with K + 1 cells.
44 @
LRT - (Based on the likelihood ratio test and the Poisson
.)
Y#,,6 assumption.)
g X -GOFT - (Based on the chi-
' A cell deviation in direction square goodness of
- from the CC #7 pattern fit test.)
g {-
15 s i.
.N\\
1978, 1979. 1980, u
and 1981 m
M w."""I Mr""""""....:.................I X2, 13.19, p <.003 (LRT) 2 12.61, p <.005 (X GOFT)
Combined 0
HIGH 13 16 LOW 24 HIGH 12.4 21 y
!.:::p 5g :
[
11.25 1 # 3.38 o
o X,2 14.72 p s.001 (LRT) 1978 Alone
- ~ '
=
HIGH LOW 9.17 SQ 9{h:wsq8.33 l
o 2.5 o
X, = 5.83 p *.06 (LRT) 1979 Alone 2
HIGH HIGH LOW g 11.9 12 i
10.83
. h h 3.25 v
2 1980 Alone X = 3.42 0.15 < p < 0.20 (LRT) 2 HIGH HIGH LOW 363.21 i
X l = 1.39 p a 0.50 (LRT) 1981 Alone
-)l-HIGH LOW j
Figure 5.1.2 Consistency of "CC #7 Pattern" over time j
19 1
5.3 UNCERTAINTIES AND LIMITATIONS There were several sources for uncertainty and limitation recognized during this analyses.
They are described herein.
5.3.1 Any effect on personnel errors resulting from overtime could not be determined for several reasons.
First, the LER does not provide information about the amount of time an operator has been on duty prior to committing an error.
Second, the LER provides no information about the hours of rest which were available to the operator since his prior duty cycle.
These data are necessary because the effects of overtime may be manifest in either of two ways:
(1) work hours longer the operator's normal work shift may result in fatigue and a possible increase in operator error; (2) an operator may suffer from fatigue and its effects on his next normal work shift following a shift involving overtime due to inadequate rest between work cycles.
5.3.2 While the effects of shift work may result in a non-uniform distribution of errors, it is not the only factor which may produce this effect.
A higher than expected number of human errors during a certain period may be the result of an increase in work load or an increased opportunity for error during that period.
This study could not determine such potential variations in human error rate.
It is the practice at some plants that testing and calibration is to be performed during the shift from midnight to 8 am, or its equivalent.
This time is selected because there is generally less demand on the grid during these hours; if a plant trip inadvertently occurs as a result of the tests the impact on grid is less than should it occur at or during a higher demand period.
During the day shift other maintenance, heavy demand loads, etc. may increase the operator's work load and thus increase the opportunity for error.
6.0 CONCLUSION
S The operator error occurrence times in nuclear power plant operations display a nonuniform distribution over the day.
To our knowledge this study i
produces the first such statistical evidence based on publicly available data.
Errors were found to be more frequent during the intervals midnight to 10:00 a.m. and 1:00 to 4:00 p.m. than would be expected.
While it cannot be concluded that these observations result from circadian desynchronosis, a greater opportunity for error or some other cguse, the peak error periods are similar g
was shown in NUREG/CR-1765 23to those reported by Bjermgr et al Figure 6.1.1 shows the Bjermer plot as overlaid on the accumulated human error vs. time histogram (Figure 4.1.7).
The effects of overtime or shif t length cannot be determined from the LER reports in their present form.
There is nothing to indicate in the existing format the hours an operator had been on duty when the human error occurred or
[i]B. Bjerner, A. Holm and A. Swensson, "Dirunal Variation in Mental Performance:
A Study of Three-Shift Workers," British Journal of Industrial Medicine 12, 103-110 (1955).
[2]H. Price, P. Wallace, M. Bauman, and M. Smith, Bio Technology, Incorporated,
" Review of Staff Requirements for Near-Term Operating License Facilities,"
USNRC Report NUREG/CR-1764, October 1980.
20
l 1
l t
1 84 ERRORS Error Bands (For errors not pinpointed to Wtthin an hourly l
n l
enterval)
I v
i K H K >l KH l-c4
}-c l KH lo!
I A
I I
V B
n I
^
l V
i V
I t
t I
I t
i f
f i
I t
I f
I I
f I
l I
t f
i 1
0 1
2 3
4 5
6 7
8 9
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Time Number 8
of Errors IFor errors muum 4
pinpointed to S
4000 within an hourly 0
~
interval) f f
J.i: :
- i:'-
o y
'i:i:
U_
Y h" -
4 Jg !e!!W
' ~ ~
6%
g !
2000yE p'
g g
- im e,,
!!]ijhii!
[!!$:!!!s3!5isi8!
2
- !jigMii;
' sim".
i!G# tis s) 3 l
s l
2
$3..1 m
i$:.
" '~
~*
,,, ?
kj T
l w-
, 1====.
s s
l M!!!?f:iss!{p;! iiii, Qs X.jlT:ir!!!Ipi:!F Qi!!!s!sj?i ::I:)? Ij1:i >!!!ji!:i
, g
.gs'+
< lffill. :!jiy
@$H 0
1 2
3 4
5 6
7 8
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 l
Time of Day (Hours)
Figure 6.1.1 Distribution of errors in the material from 1912 to 1931 in 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> of the day and night. (Bjerner, et al.,1955) plotted upon the 1978,1979.1980. Er 1981 histogram of time identified LERs, (Figure 4.1.71.
l the amount of overtime worked in the previous 24-hour, 48-hour, or week interval, or whether the operator is in the beginning or end of the day, or several day, shift duty cycle.
The impact on operator error occurrence time resulting from increased opportunity (more control manipulation) or an increase in work load cannot be measured in an analysis of this nature.
A statistical study of the LER pro-vides valuable insights into the effects which may be observed but it does not permit drawing conclusions as to the causes which produce these effects.
Researching plant logs to increase the total number of LERs for which the error occurrence times are known would at most increase our confidence in the statis-tical results.
It would not increase our knowledge of the impact on error rates of shift rotation, shif t length, cr overtime use.
Hence, it is felt that further efforts to increase the number of usable data points are not justified.
To review the shift log books in attempting to identify cause factors which may have established a more error likely situation would appear to be more produc-tive.
However, the costs are judged to be high and the results uncertain.
It appears that a series of planned and controlled experiments in which the vari-ables affecting operator performance could be limited and measured would be more cost effective than further LER review.
The LER reporting system for human errors needs major revision.
The current human error reporting system is too dependent upon a system developed for hardware systems and components.
To be suitable for the reporting of human error an entirely different approach (format and style) may be required than that which now exists.
It is possible that the criteria for reporting human events should be different from hardware related events and in the extreme a different reporting system may need to be established.
A reporting system similar to the FAA/ NASA Aviation Safety Reporting System is a possibility.
Some requirements for a human error reporting system could include:
(a) Operator hours on duty at occurrence.
(b) Hours rest since previous duty cycle.
(c) How many days since start of the current shift.
(d) What are the operator's present duty hours, previous assigned duty hours, duty hours on the next shift assignment.
(e) Existing environmental conditions.
(f) When the event actually occurred (not only when it was detected).
In summary, we need to know more about the plant conditions and the state of the operator involved in human errors, and not just the fatt that errors occurred.
22
0 APPENDIX A THE LER DATA AND CELL CONFIGURATIONS
APPENDIX A.1 THE LER DATA The data for this study was selected from the LERs of 1978, 1979, 1980, and January through June of 1981.
This data were gathered from all U.S. nuclear power plants.
The focus was successively narroweo from consideration of all 11,937 errors, to the 1901 personnel errors, then to the 599 licensed personnel errors, and then finally to the 84 licensed personnel errors for which the time of occurrence was identified.
The data were analyzed as a whole and by subsets, such as only the data from 1978, 1979, and 1980 combined.
This was done to get some idea of the con-sistency of the conclusions over the 3.5 years during which the data were gathered.
The following data subsets were analyzed for all 7 cell configurations:
1978, 1979, & 1980
- together, 1979, 1980, & 1981 together, and 1978, 1979, 1980, & 1981 together The advantages of using these 3 data subsets are as follows:
All 3 contain enough data points to make the x2-G0FT applicable.
In addition, they help in determining if the conclusions based substantially on pre-TMI data continue to hold when the analysis is based on mostly post-TMI data.
To see this, observe that the 78-79-80, 78-79-80-81, and 79-80-81 data subsets c e, respectively, almost equally, moderately, and slightly representative of the pre-TMI environment; (5/12, 2/7, and 1/10 of these data subsets cover pre-TMI days respectively).
Such a determination is worth pursuing because the shock of the accident probably significantly affected the attitudes and performances of the power plant operators, and hence may have had some significant effect on the time distribution of errors.
For 22 of the 84 time-identified errors, the exact time of occurrence could not be determined from the LERs.
But we identified an " error interval" or
" error bound" within which we were sure the time had to fall.
For each cell configuration, if an error interval was not contained within a single cell, it was omitted from the analysis of that cell configuration.
Histograms of error occurrence times (see Section 4.1) were formed by representing the day as a linear axis of 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> and then placing above each hour a rectangle whose height indicated the number of errors in the sample that occurred within that hour.
Very few errors occurred exactly on the hour.
When an error occurred exactly on one of the endpoints of a cell of any cell configuration, it was counted as occurring within the cell which had the largest number of errors in excess of its expected number under H '
u
. - ~ _ _ _
Note that error #20 of 1980 occurred at 2400 hours0.0278 days <br />0.667 hours <br />0.00397 weeks <br />9.132e-4 months <br /> on a Friday.
By the above j
remarks, for the analysis of cell configuration #7 (CC#7), this error should i
have been counted as occurring within the (0-10) cell of the following i
Saturday.
However, by an oversight, it was counted as within the [(10-13)
U (16-24)] cell.
If it had been counted as within the (0-10) cell, the pattern of CC#7 would have been even more well pronounced, yielding statisti-cal calculations even more supportive of the current results, (i.e., test statistics consistent with the current results, but with even lower p-values).
4 i
4 1
A.1-2 1
APPENDIX A.2 CELL CONFIGURATIONS In testing H a number of distinct cell configurations (C.C.) were chosen, each of whicN, led to two or three distinct statistical tests; (two if the C.C.
had 3 or more cells and three if it'had 2 cells).
As dascribed in the primary procedure, H will be rejected only if all two (or three) tests show 0.01 level of significance for at least one C.C.,
for the largest data subset.
The criteria for choosing the C.C.s was to maximize the probability of detecting a deviation from uniformity, (i.e., the probability of rejecting H when H is false), subject to the time limitations of this project and the EonstraiNt that the significance levels on each test remain fixed at 0.01.
The total number of errors was sufficiently small that no C.C. would contain Eore than 4 cells.
It was judged that using only one C.C. would be inadequate, since the tests based on just one C.C. would fail to detect severe non-uniformity in the particular case that the cells of the C.C. were overlapped equally by the intervals of high and low error likelihood that must exist in the non-uniform case.
Clearly, the detection probability increases with the number of distinct C.C.s used, since there is a greater chance that one of the C.C.s will have cells in near alignment with the intervals of high and low error likelihood.
In light of this, seven different C.C.s were created, some by random choice independent of the data and some as suggested by patterns in historical data.
These are the 7 configurations, (see Figure 4.2.1):
[( 0 -
8), ( 8 - 16 ), ( 16 - 24 )]
[( 9 - 16), ( 9 - 16 )C)
[(11 - 13), (11 - 13 )C)
[(13 - 16), (13 - 16 )C)
[( 0 - 6), ( 6 - 12), ( 12 - 18 ), (18 - 24 )]
[( 0 -
8), (13 - 16), (( 8 - 13 ) U (16 - 24 ))]
[( 0 - 10), (13 - 16), ((10 - 13 ) U (16 - 24 ))]
Terminology: (x y) represents the complement of ( x y ) with respect to the 24-hour day, i.e.,
the set of all daily times not in ( x y ).
( x y ) U ( w - z ) represents a single cell as a union of the intervals
( x y ) and ( w - z ).
a APPENDIX A.3 RAW DATA
LER data 1978 Licensed Operators 214 Total events 28 Time identified errors Time Day Date 1
[2330-0030]
Tues.*
1/24/78 2
[0055-0155]
Weds.
8/16/78 3
[0635-0735]
Weds.
8/16/78 4
0351 Tues.
1/24/78 5
0935 Weds.
2/15/78 6
0215 Fri.
7/14/78 7
1000 Sun.
7/23/78 8
0552 Weds.
12/13/78 9
2000 Weds.
10/25/78 10
[0100-0200]
Sat.
7/15/78 11
[0100-0500]
Sat.
7/15/78 12
[0600-0605]
Mon.
12/18/78 13
[0130-0230]
Weds.
2/15/78 14 1020 Sun.
4/23/78 15 0952 Weds.
9/13/78 16 0254 Sat.
11/4/78 17 1720 Thurs.
12/28/78 18
[0000-0800]
Sat.
9/16/78 19 0940 Tues.
9/19/78-20 1345 Mon.
11/13/78 21 0157 Mon.
7/17/78 22
[0100-0140]
Mon.
8/14/78 23 0251 Tues.
8/22/78 24 0030 Fri.
12/15/78 25 0301 Thurs.
5/4/78 26 0045 Mon.
1/30/78 27 1115 Sunday 7/23/78 28 1935 Thurs.
10/19/78
- or Weds.
A.3-2
LER data 1979 Licensed Operators 151 Total events 20 Time identified errors Time Day Date 1
1446 Sun.
8/19/79 2
1108 Thurs.
9/20/79 3
0830 Tues.
11/27/79 4
0640 Thurs.
5/17/79 5
0600 Mon.
9/24/79 6
1515 Thurs.
12/13/79 7
[1500-1520]
Tues.
12/4/79 8
[0022-0105]
Thurs.
7/26/79 9
1345 Tues.
12/18/79 10
[0532-0632]
Weds.
2/28/79 11 0010 Tues.
3/20/79 12 1945 Mon.'
9/24/79 13 1403 Sat.
5/5/79 14 2303 Fri.
2/16/79 15 2125 Weds.
12/26/79 16 0640 Fri.
3/30/79 17 0400 Weds.
3/28/79 18 0400 Fri.
1/12/79 19 1328 Weds.
5/23/79 20 1900 Weds.
8/15/79 A.3-3
LER data 1980 Licensed Operators 185 Total events 26 Time identified errors Time Day Date 1
.[1403-1500]
Mon.
11/17/80 2
[0430-0500]
Weds.
2/13/80 3
[1500-1512]
Sat.
2/2/80 4
0920 Weds.
7/23/80 5
2300 Thurs.
12/25/80 6
[1045-1145]
Tues.
4/1/80 7
2311 Sun.
12/14/80 8
1300 Weds.
1/16/80 9
0440 Weds.
4/23/80 10 1357 Sat.
6/7/80 11
[0930-1000]
Weds.
8/20/80 12 0910 Tues.
10/14/80 13 1600 Tues.
10/14/80 14 1912 Sat.
2/23/80 i
15 0206 Fri.
10/31/80 16
[0355-0455]
Mon.
5/19/80 17 0200 Thurs.
6/19/80 18 0708 Fri.
7/18/80 19 1349 Fri.
8/15/80 20 2400 Fri.
2/29/80 21 0345 Thurs.
3/6/80 22 0840 Thurs.
3/13/80 23 0816 Weds.
8/6/80 24 1730 Tues.
11/25/80 25
[2100-2200]
Tues.
1/8/80 26 1030 Tues.
9/30/80 A.3-4
LER data 1981 (Up to 6/30)
Licensed Operators 49 Total events 10 Time identified errors Time Day Date 1
2000 Weds.
4/22/81 2
1600 Tues.
1/16/81 3
0330 Sat.
4/18/81 4
[0825-2025]
Tues.
3/24/81 5
0430 Thurs.
1/8/81 6
2300 Tues.
3/17/81 7
[0853-1253]
Thurs.
4/23/81 8
1703 Fri.
4/3/81 9
[0818-1018]
Tues.
3/3/81 10 1325 Weds.
3/25/81 A.3-5
Sample error distribution across days of the week by individual years S
M T
W Th F
S Total 1978 3
5 3
7 3
2 4
27 1979 1
2 4
5 4
3 1
20 1980 1
2 6
6 4
4 3
26 1981 0
0 4
2 2
1 1
10 Total 5
9 17 20 13 10 9
83 Expected totals 11.86 11.86 11.86 11.86 11.86 11.86 11.86 83 under H y 4
A.3-6
Page A.3-8 through A.3-28 contain the LERs which describe the 84 errors contained in this report. Three LERs reported the same error as being repeated at a later time. Hence, the following tabulation contains only 81 LERs. The information contained in each item is arranged in the following format.
FACILITY / SYSTEM /
DOCKET NO./
COMPONENT / COMPONENT SUBCODE/
LER NO./
EVENT DATE/
CAUSE/CAUSE SUBCODE/
CONTROL NO./ REPORT DATE/
EVENT DESCRIPTIDH/
COMPONENT MANUFACTURER NSSS REPORT TYPE CAUSE DESCRIPTION tab
ECWIN I. HATCH-2 050003GG 091678 WHILE REVIEWING PREVIOUS SHIFT OPERATIONS. DURING WHICH PLANT CONDITION SYSTEM CODE NOT APPLICABLE 78-032/03L-0 092978 CHANGED FRCM SHUTDOWN TO STARTUP. IT WAS DISCOVERED ON 9-16-78 AT 1000 C COMPONENT CODE NOT AF PLIC ABLE 022498 30-DAY ST THAT A CONTINU0US FIRE WATCH AS RE0UIRED IN TECH SPEC 3.7.A HAD NOT B SUSCOMPONENT NOT APPLICABLE GE PERSONNEL ERROR EEN ESTABLISHED AT AN UNSEALED CONDUIT PENETRATING FLOOR & CEILING BARRI ERS OF RPS-MG SET ROOM. PENETRATIONS WERE FOR A 1 1/2 INCH CONDUIT CNLY.
LICENSED 8 SENIOR OPERATORS 8 A 4 HR. PERIODIC FIRE WATCH WAS IN EFFECT. 8 SMOKE DETECTION WAS AVAI ITEM NOT APPLICABLE LABLE DIRECTLY ABOVE CEILING PENETRATION 8 DIRECTLY BELOW FLOCR PENETRAT UPON COMMENCEMENT OF STARTUP AFTER A PERIOD IN COLD SHUTDOWN THE SUPERVI l
SOR ON DUTY FAILED TO CHANGE THE FIRE WATCH FROM PERIODIC TO CONTINUOUS.
WHEN THE EVENT WAS RECOGNIZED A CONTINUOUS FIRE WATCH WAS IMMEDIATELY E XTABLISHED, AND SHIFT FOREMEN SUPERVISORS WERE MADE AWARE OF THE CONDITI ONS TO PREVENT RECURRENCE.
DAVIS-BESSE-1 05000346 032079 3/14/79 AT 1140 ALL GROUP 8 AXIAL POWER SHAPING RODS PLACED IN ASYMMETRI REACTIVITY CONTROL SYSTEMS 79-042/03L-0 041879 C BYPASS MAKING ROD FAULT CIRCUITRY GROUP 8 INOPERABLE.
T.S. 4.1.3.2.1 INSTRUMENTATION + CONTROLS 025587 30-DAY AND 4.1.3.3 REQUIRE INCREASE IN SURVEILLANCE REQUIREMENTS FDOM ONCE EVER NO SUBCOMPONENT PROVIDED B+W Y 12 HR. TO ONCE EVERY 4 HR.
AT 0010 ON 3/20/79 REACTOR OPERATOR FAILED PERSONNEL ERROR TO PERFORM VERIFICATION. NO DANGER TO HEALTH AND SAFETY OF PUBLIC OR U LICENSED & SENIOR OPERATORS HIT PERSONNEL. SURVEILLANCE REQUIREMENTS WERE SATISFIED AT 0115 ON 3/20 ITEM NOT APPLICABLE
/79 1 HR. AND S MIN. LATE.
NP-33-79-44.
+
PERSONNEL ERROR.
REACTOR OPERATOR FAILED TO PERFORM SURVEILLANCE REQUIR EMENT WITHIN SPECIFIED TIME INTERVAL.
INCIDENT WAS REVIEWED WITH OPERAT OR INVOLVED. MODIFICATION TO CONTROL ROOM STATUS BOARD WILL BE MADE TD ASSIST OPERATORS WHEN INCREASED SURVEILLANCE REQUIREMENTS ARE REQUIRED.
w do D. C. C00K-1 05000315 072679 DURING MODERATOR COEFFICIENT TESTING STARTING AT 0022 HOURS. REACTOR PO REACTOR CORE 79-040/03L-0 082279 WER DECREASED FROM 79 PERCENT TO 58 PERCENT IN 43 MINUTES.
THE SURVEILL COMPONENT CODE NOT APPLICABLE 026583 30-DAY ANCE REQUIREMENT OF T.S. TABLE 4.4-4 ITEM 4.B (ISOTOPIC ANALYSIS FOR ICD SUBCOMPONENT NOT APPLICABLE WEST INE) WAS NOT PERFORMED WITHIN THE REQUIRED TIME LIMITS.
~
PERSONNEL ERROR LICENSED 8 SENIOR OPERATORS ITEM NOT APPLICABLE CHEMICAL SECTION PERSONNEL WERE NOT NOTIFIED OF THE POWER DECREASE BY OP ERATIONS PERSONNEL. THE EQUIPMENT OPERATOR AND UNIT SUPERVISOR BOTH TH0 UGHT THE OTHER HAD NOTIFIED CHEMICAL PERSONNEL OF THE POWER CHANGE. AR CUTINE SAMPLE TAKEN AT 0900 INDICATED ONLY A VERY SLIGHT INCREASE IN IOD INE ACTIVITIES. OPERATORS WERE REINSTRUCTED.
MILLSTONE-1 05000245 121578 ON DECEMBER 15, 1978. AT APPROXIMATELY 0230 HOURS WHILE RETURNING THE UN CNTNMNT COMBUS GAS CONTROL SYS 78-030/03L-0 011579 IT TO FULL POWER IT WAS DISCOVERED THAT ALTHOUGH THE DRYWELL OXYGEN CONC COMPONENT CODE NOT APPLICABLE 023355 30-DAY ENTRATION WAS WITHIN THE 5 PERCENT SPEC. WITHIN 24 HOURS AFTER "RUN" MOD SUBCOMPONENT NOT APPLICABLE GE E. THE SUPPRESSION CHAMBER OXYGEN CONCENTRATION WAS S 6 PERCENT.
IT WAS PERSONNEL ERROR NOT FELT THAT THE 0.6 PERCENT EXCESS OXYGEN FOR 2 HOURS CONSTITUTED A 5 LICENSED 8 SENIOR OPERATORS IGNIFICANT REDUCTION IN THE MARGIN OF SAFETY.
ITEM NOT APPLICABLE IT APPEARS THAT THE NITROGEN PURGE RATE ON THE CONTAINMENT WAS TOO SLOW IN ORDER TO MEET THE OXYGEN CONCENTRATION SPECIFICATION SPEC. WITHIN THE 24 HOUR LIMIT.
THE PPRGE RATE WAS INCREASED. THE OPERATOR WAS REMINDE D OF THE IMPORTANCE CF CAREFUL PARAMETER MONITORING.
SALEM-1 05000272 013078 AT 0045 HOURS. TRANSFER OF NO. 11 SGFP TO AUTOMATIC CAUSED SUDDEN INCREA CONDNSATE + FEEDWTR SYS + CONT 78-006/03L-0 022478 SE IN S/G LEVELS.
INSURGE OF FEEDWATER CAUSED TAVE REDUCTION AND PRESSU COMPONENT CODE NOT APPLICABLE 023233 30-DAY RIZER PRESSURE REDUCTION. PRES $URIZER PRESSURE DECREASED TO 2165 PSIA W SUBCOMPONENT NOT APPLICABLE WEST HICH IS LESS THAN THE 2220 PSIA ALLOWED BY TECH SPEC 3.2.5.
PRESSURIZER PERSONNEL ERROR HEATERS RECOVERED PRESSURE IN 4 MINUTES. THIS IS THE STH OCCURRENCE OF
~
LICENSED'8 SENIOR OPERATORS THIS TYPE.
ITEM NOT APPLICABLE THE CAUSE OF THIS OCCURRENCE WAS PERSONNEL ERROR IN THAT THE OPERATOR FA ILED TO MATCH THE MASTER DEMAND TO THE PUMP SPEED PRIOR TO PLACING THE M ASTER CONTROLLER TO AUTOMATIC. THE OPERATOR WAS REINSTRUCTED BY THE SHI FT SUPERVISOR. ALL LICENSED OPERATORS WILL REVIEW THIS LER IN THE REQUA LIFICATION PROGRAM.
ARKANSAS-2 05000368 081678 DURING MODE 5 OPERATION. PRIOR TO INITIAL CRITICALITY 8 SDC LOOP 1 OUT 0 COOLANT RECIRC SYS + CONTROLS 78-001/01T-0 083078 F SERVICE. THE LOOP 2 LPSI PUMP TRIPPED AT 0055 HOURS.
THE REACTOR CCRE COMPONENT CODE NOT APPLICABLE 022337 2-WEEK WAS WITHOUT COOLANT FLOW FOR 2.8 HOURS UNTIL LOOP 1 WAS RETURNED TO SER SUBCOMPONENT NOT APPLICABLE COMB VICE. VIOLATING T.S.
3.4.1. AT 0635 HOURS. LPSI PUMP A WAS STOPPED FCR A PERSONNEL ERROR HYDROSTATIC TEST OF THE RECIRC LINE RESULTING IN NO CORE FLOW FOR 3.9 LICENSED 8 SENIOR OPERATORS HOURS, VIOLATING T.S.3.4.1.
FLOW WAS AVAILABLE. IF NECESSARY, VIA RCP'S.
ITEM NOT APPLICABLE REACTOR COOLANT FLOW WAS NOT RE-ESTABLISHED IN < 1 HOUR DUE TO OPERATION S PERSONNEL NOT COGNIZANT OF THE REQUIREMENT DURING MODE 5 OPERATION. TE CHNICAL SPECIFICATIONS REQUIREMENTS FOR MODE 5 WAS REVIEWED BY PER3ONNEL INVOLVED.
Y JOSEPH M.
FARLEY-1 05000348 081478 AT 0140. DURING A UNIT STARTUP, TAVG DECREASED TO 537.8F WHILE CHANGING CONDNSATE + FEEDWTR SYS + CONT 78-056/03L-0 091178 OVER FROM THE AUXILIARY FEED SYSTEM TO THE MAIN FEED SYSTEM.
T/S 3.1.1.
INSTRUMENTATION + CONTROLS 022234 30-DAY 5 REQUIRES THAT THE RCS LOWEST OPERATING TAVG BE GREATER THAN OR EQUAL T CONTROLLER WEST 0 541F.
ALL FEED TO THE S/G'S WAS SECURED UNTIL THE TAVG INCREASED TO W PERSONNEL ERROR ITHIN THE T/S LIMITS AT 0146.
ALL T/S 3.1.1.5 ACTION REQUIREMENTS WERE LICENSED 8 SENIOR OPERATORS MET.
THE HEALTH AND SAFETY OF THE PUBLIC WERE NOT AFFECTED BY THIS CCCU HAGAN CORPORATION RRENCE.
REACTOR OPERATOR RESET THE FEEDWATER ISOLATION SIGNAL WITHOUT ADJUSTING THE FEED REGULATING VALVE M/A STATIONS TO THEIR ZERO DEMAND POSITION. WH EN THE SIGNAL WAS RESET THE FEED REGULATING VALVES OPENED APPROXIMATELY 10% RESULTING IN OVERFEEDING THE S/G'S AND COOLING DOWN THE RCS. THE VAL VES WERE IMMEDIATELY SHUT.
C00PER-1 05000298 071578 DURING REACTOR PCWER DECREASE FOR CONTROL ROD SEQUENCE EXCHANGE. SCHEDUL SYSTEM CODE NOT APPLICABLE 78-025/03L-O 072678 ED LOAD DROP DEVIATION DUE TO SYSTEM DEMAND, REACTOR POWER WAS DECREASED COMPONENT CODE NOT APPLICABLE 022102 30-DAY FROM 1646 MWT TO 1071 MWT,
(~ 2 4 ?. ). BETWEEN THE HOURS OF 0100 AND 0200.
SUBCOMPONENT NOT APPLICABLE GE THE REACTOR COOLANT SAMPLE AS REQUIRED BY T.S. 4.6.B.1.A.4 WAS NOT OBT PERSONNEL ERROR AINED UNTIL 0950 HOURS.
THERE WAS NO ADVERSE EFFECT UPON THE PUBLIC HEA
~
LICENSED 8 SENIOR OPERATORS LTH AND SAFETY.
ITEM NOT APPLICABLE
> 20% POWER CHANGE IN ONE HOUR WAS NOT NOTED BY OPERATIONS PERSONNEL AND CHEMISTRY SAMPLE WAS NOT OBTAINED WITHIN FOUR HOUR PERIOD AS REQUIRED B Y T.S.
4.6.B.1.A.4.
THIS OCCURRENCE AND REQUIREMENTS OF T.S. HAVE BEEN DISCUSSED WITH PERSONNEL INVOLVED.
i
D. C. C00K-1 05000315 021578 UNIT WAS OPERATING NORMALLY IN MODE 1.
THE 143 EMERGENCY DIESEL GENERAT SYSTEM CODE NOT APPLICABLE 78-016-03L-0 031478 OR WAS REMOVED FROM SERVICE AT 0130 HOURS ON 2/15/78 FOR MAINTENANCE. T COMPONENT CODE NOT APPLICABLE 020572 30-DAY HE BREAKER ALIGNMENT VERIFICATION AS REQUIRED BY APPENDIX A TECH. SPEC.
SUBCOMPONENT NOT APPLICABLE WEST 4.8.1.1A WAS NOT PERFORMED WITHIN ONE HOUR.
THE CORRECT BREAKER ALIGNME PERSONNEL ERROR NT WAS VERIFIED AT 1045 HOURS ON 2/15/78.
THE PLANT REMAINED AT STEADY LICENSED 4 SENIOR OPERATORS STATE POWER AND BOTH OFF-SITE POWER SUPPLIES WERE AVAILABLE BUT NOT CALL ITEM NOT APPLICABLE ED UPON.
THERE WAS NO EFFECT ON THE HEALTH AND SAFETY OF THE PUBLIC.
THE EVENT WAS CAUSED BY THE UNIT SUPERVISOR FAILING TO REQUIRE HAVING TH E BREAKER ALICNMENT VERIFIED. THE SHIFT OPERATING ENGINEER REINSTRUCTED THE UNIT SUPERVISOR IN THE REQUIREMENTS FOR REMOVING VITAL EQUIPMENT FR0 M SERVICE WITH RESPECT TO PERFORMING SURVEILLANCE AS DIRECTED BY THE TEC H.
SPECS, AND OPERATING DEPARTMENT INSTRUCTIONS.
JOSEPH M.
FARLEY-1 05000348 071778 AT 0157 DURING STARTUP. TAVG DECREASED TO 532F WHILE TYING GENERATOR ON-SYSTEM CODE NOT APPLICABLE 78-049/03L-0 081478 LINE WITHOUT INCREASING REACTOR POWER TO 12-14% AS REQUIRED BY PROCEDURE COMPONENT CODE NOT APPLICABLE 022049 30-DAY
. T/S 3.1.1.5. REQUIRES THE RCS LOWEST OPERATING LOOP TEMPERATURE TAVG B i
SUBCOMPONENT NOT APPLICABLE WEST E GREATER THAN OR EQUAL TO 541F.
TAVG RESTORED WITHIN T/S LIMITS AT 020 i
~ PERSONNEL ERROR 5.
ALL T/S 3.1.1.5 ACTION REQUIREMENTS WERE MET.
HEALTH AND SAFETY OF LICENSED & SENIOR OPERATORS GENERAL PUBLIC NOT AFFECTED.
ITEM NOT A?PLICABLE INCOMPLETE COMMUNICATIONS BETWEEN REACTOR PLANT AND SECONDARY PLANT OPER ATORS.
IN FUTURE A SPECIFIC INDIVIDUAL WILL BE DESIGNATED TO ASSURE PR OCEDURAL REQUIREMENTS ARE MET.
INCIDENT TO BE REVIEWED BY ALL LICENSED PERSONNEL. SIMILAR INCIDENT REPORTED IN LER 78-042/03L - O.
Y POINT BEACH-2 05000301 061980 AT 0200 HOURS ON 6/19/80 WITH UNIT 2 AT 100% POWER. TEST IT-04 WAS DONE EMERG CORE COOLING SYS + CONT 80-007/01T-0 070380 INCLUDING STROKING OF "B" RHR PUMP SUCTION VALVE.
EMERGENCY SAFEGUARDS PUMPS 031566 2-WEEK TRAIN EQUIPMENT WOULD HAVE BEEN IHOPERABLE ON A LOSS OF AC POWER BECAUSE CENTRIFUGAL WEST EMERGENCY DIESEL GENERATOR. 3D. WAS 005 FOR MAINTENANCE. BOTH RHR SYST PERSONNEL ERROR EMS WERE MDMENTARILY INOPERABLE HAD A LOSS OF AC OCCURRED. REPORTABLE U LICENSED 4 SENIOR OPERATORS NDER TECHNICAL SPECIFICATION 15.6.9.A.2.
SIMILAR LER 80-006.
ITEM NOT APPLICABLE PERSONNEL ERROR: SHORT-TERM ACTICN; PROHIBIT PERIODIC TESTING UNTIL EME RGENCY DIESEL GENERATOR. 3D. IS RETURNED TO SERVICE. LONG-TERM ACTION ADDITIONAL ADMINISTRATIVE POLICIES AND PROCEDURE CHANGES ARE BEING REVIE WED.
NORTH ANNA-2 05000339 103180 ON OCTOBER 31 1980. WITH THE UNIT AT 91% POWER. PT-26.3 (AXIAL POWER DI REACTOR CORE 80-089/03L-0 111980 STRIBUTION MONITORING) WAS NOT PERFORMED AT 0206 HOURS AS REQUIRED. SIN COMPONENT CODE NOT APPLICABLE 033021 30-DAY CE THE PERIODIC TEST WAS PERFORMED WHEN THE ERROR WAS DISCOVERED AND ALL SUBCOMPONENT NOT APPLICABLE WEST DATA WAS SATISFACTORY. THE HEALTH AND SAFETY OF THE PUBLIC WERE NOT AFF PERSONNEL ERROR ECTED.
THIS ITEMS IS REPORTABLE PURSUANT TO TECHNICAL SPECIFICATION 6.9 LICENSED & SENIOR OPERATORS
.1.9.C.
ITEM NOT APPLICABLE THE PERIODIC TEST 2-PT-26.3 WAS MISSED DUE TO PERSONNEL ERROR. WHEN DIS COVERED. THE PERIODIC TEST WAS PERFORMED WITH SATISFACTORY RESULTS.
)
I
)
CALVERT CLIFFS-1 05000317 071478 WHILE OPERATING AV STEADY STATE POWER THE VEMPERAVURE DISE ACROSS VME MA CIRCULATING WATER SYS + CON 78-035/04T-0 072178 IN CONDENSER $ BEGAN RAPIDLY INCREASING. AT 0215 A.M.
THE OPERATOR BEGAN CCMPONENT CODE NOT APPLICABLE 021822 2-WEEK DECREASING POWER TO CONTROL DELTA T.
AT 0300 THE TEMPERATURE RISE ACR0
$UBCOMP0HENT NOT APPLICABLE COMB SS THE MAIN CONDENSERS EXCEEDED THE MAXIMUM ALLCWED AT 10.1 F PER ENVIRO PERSONNEL ERROR NMENTAL SPEC.
2.1.1.
POWER WAS FURTHER RELUCED AND CONDENSER DELTA T WA
~
LICENSED 4 SENIOR OPERATORS S RETURNED TO SPECIFICATIONS BY 0400.
LER'S 77-97 (U-1) AND 77-36 (U-2)
ITEM NOT APPLICABLE DESCRIBE SIMILAR EVENTS.
THE CAUSE WAS DUE TO THE OPERATOR'S FAILURE TO DECREASE POWER ENOUGH TO COMPENSATE FOR THE RAPID INCREASE IN MAIN CONDENSER DELTA T CAUSED BY FO ULING OF THE CONDENSERS BY BAY DEBRIS.
THE CORRECTIVE ACTION WAS TO DEC REASE PCWER UNTIL THE DELTA T DECREASED TO 10.0 AND 10 CLEAN THE CONDENS ER WATER BOXES.
t I
JOSEPH M.
FARLEY-1 05000348 032278 AT 0251. DURING A REACTOR SHUTDOWN. TAVG DECREASED TO 538F. TECH SPEC 3.
REACTOR CORE 78-058/03L-0 092078 1.1.5 REQUIRES THAT THE RCS LCWEST CPERATING LOOP TEMPERATURE TAVG BE CR COMPONENT CODE NOT APPLICABLE 022375 30-DAY EATER THAN OR EQUAL TO 541F. TAVG WAS RESTORED 10 WITHIN TECH SPEC LIMIT SUBCOMPONENT NOT APPLICABLE WEST S AT 0255 ON 8/22/78. TECH SPEC 3.1.1.5 ACTION STATEMENT REQUIREMENTS WE PERSONNEL ERROR RE MET. THE HEALTH AND SAFETY OF THE GENERAL PUBLIC WERE NDT AFFECTED BY l
LICENSED A SENIOR OPERATORS THIS OCCURRENCE. SEE LER 78-056/03L-0 ITEM NOT APPLICABLE REQUCTION IN TAVG OCCURRED DURING TRANSFER FROM MAIN FEED TO AUX FEED SY S. REACTOR POWER INCREASED TO BRING TAVG WITHIN T/S LIMIT. INCIDENT DISC USSED WITN SHIFT PERSONNEL TO EMPHASIZE LIKELIHOOD OF TAVG REDUCTIONS DU RING FEEDWATER MANIPULATIONS AT VERY LOW POWER LEVELS.
w
.L.
~~
DAVIS-BESSE-1 05000346 110478 AT 02:54:03 OPERATIONS PERSONNEL INTENDING TO START REACTOR COOLANT PUMP AC ONSITE POWER SYS + CONTROLS 78-125/03L-0 011979 1-1-1, WHOSE ESSENTIAL BUS FEED BREAKER UNDERVOLTAGE RELAYS WERE DISCON CIRCUIT CLOSERS / INTERRUPTERS 023425 30-DAY NECTED FOR START, STARTED RCP 1-1-2 INSTEAD. ESSENTIAL BUS D1 CE-ENERGI SWITCH (OTHER THAN SENSOR)
B+W ZED AT 02:54:08.
EMERGENCY DIESEL GENERATOR 1-2 STARTED AND ESSENTIAL B
- PERSONNEL ERROR US Di RE-ENERGIZED AT 02:54:18.
DISABLING ESSENTIAL BUS UNDERVOLTAGE RE LICENSED 3 SENIOR OPERATORS LAYS IS CONTRARY TO T.S.
3.8.2.1 ESSENTIAL BUS D1 DE-ENERGIZED TEN SEC ITEM NOT APPLICABLE ONDS.
UNIT WAS SUSCRITICAL.
NP-33-78-148.
PERSONNEL ERROR AND DESIGN DEFICIENCIES. ESSENTIAL BUS D1 TRANSFERRED B ACK TO NORMAL POWER SUPPLY AND EDG 1-2 RETURNED TO NORMAL STANDBY STATUS
. OPERATIONS PER$0NNEL WILL TRANSFER ESSENTIAL BUS POWER SUPPLY TO BUS TIE TRANSFORMER WHILE STARTING RCPS OR CWPS. FCRS PREPARED TO INSTALL M0 MENTARY BYPASSING OF UNDERVOLTAGE RELAYS AND MODIFY T.S.
PILGRIM-1 05000293 050478 DN MAY 4, 1978. AT 0301 HOURS. THE MANUAL SCRAM ON RPS CHANNEL "A" WAS R SYSTEM CODE NOT APPLICABLE 78-025/03L-0 052378 ESET TO ALLOW FOR PERFORMANCE OF FRONT PANEL CHECKS AND SCRAM TESTING OF COMPONENT CODE NOT APPLICABLE 021459 30-DAY CONTROL ROD DRIVES.
THE MANUAL SCRAM HAD BEEN APPLIED BECAUSE OF INSUF SUBCOMPONENT NOT APPLICABLE GE FICIENT IBM CHANNELS IN THE "A" RPS TRIP SYSTEM IN ACCORDANCE WITH SECTI PERSONNEL ERROR ON 3.1 TABLE 3.1.1 AND NOTE 1.A.
THE UNIT WAS SHUT DOWN FOR REPAIRS TO LICENSED A SENIOR OPERATORS INBOARD MSIV A0-203-ID WHEN THIS EVENT OC URRED.
ITEM NOT APPLICABLE THIS EVENT OCCURRED BECAUSE OPERATIONS PERSONNEL WERE CONDUCTING TESTS R EQUIRED PRIOR TO STARTUP UNDER LESS THAN OPTIMUM CONDITIDHS. THOSE PERS ONNEL RESPONSIBLE HAVE BEEN CAUTIONED THAT ADHERENCE TO THE REQUIREMENTS OF TECHNICAL SPECIFICATIONS SUPERSEDES ALL OPERATIONAL OBJECTIVES.
DAVIS-BESSE-1 05000346 041381 (NP-33-81-26) AT 0330 HRS IT WAS NEC TO RAPIDLY TAKE 34S Id BUS "J" OUT-RESIDUAL HEAT REMOV SYS + CONT 81-024/03X-2 071381 0F-SERVICE DUE TO 2 BURNING POTENTIAL DEVICES ON THAT BUS. 7.T WAS ISOLAT PUMPS 037012 OTHER ED PER SP 1107.01 HCWEVER. LOADS ON 13.8 KV BUS "A" POWEP SUPPLY FROM S OTHER B+W TARTUP TRANSFORMER "01"(WHICH IS FED BY "J" BUS) HAD NDY BEEN TRANSFERRE
~
PERSONNEL ERROR D TO "02".
RESULT WAS LOSS OF DECAY HEAT PUMP 1-1 4 ENTRY INTO ACTION ST LICENSED & SENIOR OPERATORS ATEMENT OF T/S 3.4.1 NO DANGER TO HEALTH / SAFETY OF PUBLIC/ STATION PERSO OTHER NNEL. RCS PRESSURE / TEMP VARIED ONLY SLIGHTLY. SEE LER'S 80-54, 58, 8 60.
CAUSE WAS PERSONNEL 8 PROCEDURAL ERROR. IN AN ATTEMPT TO RAPIDLY ISOLATE BUS, OP FAILED TO PROVIDE AN ALTERNATE POWER SUPPLY FOR DH PUMP 1-1 PRI OR TO ISOLATING "J" BU
S. PROCEDURE
ONLY PROVIDES A SHORT NOTE OF WARNING
. DH PUMP 1-1 WAS RESTARTED 1 MIN & 28 SEC AFTER IT WAS STCPPED. MEMO 81
-1013 WAS WRITTEN TO INFORM OPS 8 M-4S53 WAS WRITTEN FOR PROC SP 1107.01 SEQUDYAH-1 05000327 030680 THE CONTENTS OF THE TASK DECONTAMINATION COLLECTOR TANK WAS DISCHARGED V LIQ RADIDACT WSTE MANACMNT SYS 80-002/01T-0 031980 IA THE LIQUID RADWASTE EFFLUENT LINE TO THE COOLING TOWER BLCWDOWN WITH CCMPONENT CODE NOT APPLICABLE 030571 2-WEEK THE FLMf ELEMENTS IN BOTH LINES INOPERABLE AND WITHOUT FLOW ESTIMATES B SUBCOMPONENT NOT APPLICABLE WEST EING FERFORMED. EVENT OCCURRED AT 0345 CST (T.S. LCO 3.3.3.9). THERE WA PERSONNEL ERROR S NO DANGER TO THE HEALTH AND SAFETY OF THE PUBLIC, PREVIOUS OCCURRENCE LICENSED & SENICR OPERATORS S - NONE.
ITEM NOT APPLICABLE MISUNDERSTANDING OF TECHNICAL SPECIFICATIONS. OPERATIONS PERSONNEL DID NOT UNDERSTAND NEED FOR FLOW MONITORING IF NO RADI0 ACTIVITY WAS PRESENT.
PLANT PERSONNEL INSTRUCTED IN NEED TO COMPLY WITH TECHNICAL SPECIFICATI ONS REGARDLESS OF LACK OF POSSIBLE RADI0 ACTIVITY.
3D
'wL PILGRIM-1 05000293 051980 ON MAY 19, 1980 AT APPROXIMATELY 03$5 HOURS A REACTOR SCRAM WAS INITIATE SYSTEM CODE NOT APPLICABLE 80-026/03L-0 070230 D BY THE REACTOR LOM WATER LEVEL SWITCHES OF THE RPS SYSTEM.
CONTRARY T COMPONENT CODE NOT APPLICABLE 031546 30-DAY 0 STATION PROCEDURE NUMBER 2.2.17 ENTITLED " COMMUNICATIONS SYSTEMS" SECT SUBCOMPONENT NOT APPLICABLE GE ION 2.H.,
THE NRC WAS NOT NOTIFIED WITHIN ONE HOUR OF THE ABOVE EVEt4T.
PERSONNEL ERROR MANAGEMENT REVIEW OF STATION OPERATION IDENTIFIED THIS OVERSIGHT AND DET LICENSED & SENIOR OPERATORS ERMINED ITS REPORTABILITY IN ORC MEETING 80-86 CONDUCTED ON 6-3-80.
ITEM NOT APPLICABLE THE REQUIREMENTS OF PROCEDURE 2.2.17 WERE EMPHASIZED TO THE OPERATIONS P ERSONNEL AND REVIEW OF THIS PROCEDURE IS INCLUDED IN THE OPERATOR REQULI FICATION PROGRAM.
PERSONNEL ERROR C5000334 012478 AT 0343 HOURS, DURING NORMAL CPERATION, A REACTOR TRIP OCCURRED DUE TO A LICENSED 3 SENICR OPERATORS 78-011/99X-0 032878 TURBINE TRIP.
THE AUXILIARY FEEDUATER PUMPS AUTO STARTED ON LOW-LOW ST BEAVER VALLEY-1 020796 OTHER EAM GENERATOR LEVEL.
AT 0551 HOURS. A SAFETY INJECTION OCCURRED AS A RE EMERG CORE COOLING SYS + CONT WEST SULT OF LCW PRESSURIZER PRESSURE AND LEVEL.
ALL SAFETY EQUIPMENT FUNCTI COMPONENT CODE NOT APPLICABLE ONED SATISFACTORILY AMD, AFTER VERIFYING THE 5.I. WAS INADVERTENT, THE S SUBCOMPONENT NOT APPLICA3LE IGNAL WAS RESET.
SAFETY IHJECTIONS TO DATE: 8 OPERATIONAL, 2 PRE-OPERA ITEM NOT APPLICABLE TIONAL.
THE PRIMARY CAUSE OF THE SI WAS FAILURE OF THE OPERATOR TO ISOLATE STEAM TO THE MSRS IMMEDIATELY AFTER THE REACTOR TRIP.
SECONDARY CAUSES WERE I!F.RGDUCTION OF COLD AUXILIARY FEE 0tiATER AND A LEAKING PRES $URIZER S* RAY VALVE.
THE OPERATOR INVOLVED WAS REPRIMANDED. THE PRESSURIZER SPR4Y V ALVE MILL BE REPLACED. AUTCMATIC MSR ISOLATION WILL BE IMPLEMENTED.
~__
THREE MILE ISLAND-2 05000320 032879 BEGINNING ABOUT 0400 AN ACCIDENT OCCURRED CAUSING CORE DAMAGE AND RADICA FEEDWATER SYSTEMS + CONTROLS 79-012/01T-0 032979 CTIVE RELEASE.
LOST FEEDWATER; TURBINE TRIPPED; PRESSURIZER RELIEF VALV VALVES 027456 2-WEEK E OPENED; RX TRIPPED. AUX FW VALVES WERE CLOSED 8 MIN.
PORV STAYED OPE OTHER B+W N; RCS PRESSURE DECREASED; ECCS STARTED. WAS THEN THROTTLED ON RISING PR PERSONNEL ERROR ESSURIZER LEVEL.
AT 1-2 HR RCS PUMPS TURNED OFF.
OPEN PORV ALLOWED RC LICENSED 8 SENIOR OPERATORS DRAIN TANK TO OPEN TO RX BLDG. SUMP.
SUMP PUMPED TO AUX BLCG.; RADIDACT l
DRESSER INDUSTRIES, INC.
IVITY RELEASED.
I SEVERAL MAIN FACTORS INCLUDING PERSONNEL ERROR DESIGN DFrICIENCIES. CCM PONENT FAILURE. VALVES VALED 00T.
PORV FAILED TO CLOSE.
INCCRRECT PRE SSURIZER LEVEL INDICATION. RAD LEAKS FROM LETDOWN SYSTEM TO AUX BLDG.
TRIPPING RCS PUMPS LEG TO FUEL DAMAGE. SEVERAL LICENSEE /NRC/ VENDOR ACTIO NS UNDERWAY. PRESIDENTIAL COMMISSION; OTHER STUDY GROUPS APPOINTED.
ZIDH-1 05000295 011279 DURING NORMAL OPERATION INST. INVERTER 114 TRIPPED AT 0400 HRS. 1 COULD REACTOR TRIP SYSTEMS 79-004/03X-1 030979 NOT BE RESET.
INVERTER WAS NOT OPERABLE ON EITHER ITS AC OR DC FEEDS.
GENERATORS 025150 OTHER DG
'O' WAS 005 FOR REPAIR SO A CONTROLLED SHUTDOWN WAS INITIATED AT 0750 INVERTER WEST HRS. ON U-1 AS REQUIRED BY T.S. 3.15.2.H.
HOWEVER. UNIT WAS NOT AT HOT PERSCHNEL ERROR SHUTDOWN WITHIN FOUR HOURS.
REDUNDANT CCMPONENTS WERE AVAILABLE TO PER LICENSED 4 SENIOR OPERATORS FORM ANY REQUIRED SAFETY FUNCTION SO HEALTH AND SAFETY OF THE PUBLIC WER WESTINGHOUSE ELECTRIC CORP.
E NOT AFFECTED.
INVERTER 114 DC BREAKER WAS FOUND TO HAVE A LOOSE CONTACT AND WAS REPAIR ED AT ONCE; THE 2.5 KVA INVERTER SECTION WAS FOUND DEFECTIVE AND WILL BE REPAIRED WHEN PARTS ARRIVE.
FAILURE TO BEGIN SHUTDONN EARLIER WAS DUE i
TO AN OVERSIGHT BY LICENSED SHIFT PERSONNEL. SHIFT PERSONNEL HAVE BEEN INSTRUCTED IN PROPER USE OF T/S.
NO FURTHER ACTION IS REQUIRED.
Y C
BRUNSWICK-1 05000325 021380 DURING A NORMAL SURVEILLANCE TOUR OF THE REACTOR BUILDING AT 1600 HOURS.
I REACTOR CONTAINMENT SYSTEMS 8 0- 018 / 0 3 L -0 031280 THE AUXILIARY OPERATOR NOTICED THAT THE "B" LOOP OF CORE SPRAY WAS STIL VALVES 030409 30-DAY L DRAINING AFTER BEING PLACED UNDER A CLEARANCE AT 0430 HOURS FOR MAINTE GLOBE GE NANCE.
IT WAS DETERMINED THAT THIS WATER WAS COMING FROM TORUS AND THAT PERSONNEL ERROR TORUS LEVEL WAS IN ITS NORMAL OPERATING BAND.
AFPROXIMATELY 5500 GALLO LICENSED 8 SENIOR OPERATORS NS WERE DRAINED FROM THE TORUS. TECHNICAL SPECIFICATIONS 3.6.1.1 6.9.1 ITEM NOT APPLICABLE
.9C.
THE FO-10B VALVE. A 3" MANUAL ISOLATION VALVE IN THE MINIMUM FLOW LINE.
l WAS FOUND OPEN WITH THE PUMP VENTS AND DRAIN VALVES. THREE 3/4" LINES AN D ONE 1" LINE. ALSO OPEN. PROVIDING A GRAVITY FLOW PATH FROM THE TORUS T 0 THE CORE SPRAY SUMP.
THE FO-10B VALVE WAS SHUT PER THE CLEARANCE INST RUCTIONS.
FT. ST. VRAIN-1 05000267 010881 WHILE EVALUATING PCRV COOLING WATER SYSTEM DATA IN PREPARATION OF A REPO COOLANT RECIRC SYS + CONTROLS 81-007/01T-0 020381 RTABLE OCCURRENCE. IT WAS DETERMINED THAT A VIOLATION OF AN LCD HAD OCCU COMPONENT CODE NOT APPLICABLE 036201 2-WEEK RRED RATHER THAN OPERATION IN A DEGRADED MODE AS WAS ORIGINALLY BELIEVED SUBCOMPONENT NOT APPLICABLE GGA
. LOOP 2 INLET & OUTLET AVG. TEMP WAS < 100 F. FROM 0430 HRS ON 1/8/81 PERSONNEL ERROR UNTIL 1000 HRS ON 1/10/81. CONTRARY TO LCO 4.2.15(E) 8 REPORTABLE PER T/
LICENSED & SENIOR OPERATORS S AC 7.5.2(B)2. PREVIOUS OPERATION < 100 F WAS REPORTED IN 79-07 & 77-28 ITEM NOT APPLICABLE NO CONSEQUENCES OF THIS EVENT & NO EFFECT ON PUBLIC HEALTH 8 SAFETY.
CAUSE OF THE EVENT WAS FAILURE TO MAINTAIN SUFFICIENT HEAT INPUT TO THE COOLING SYSTEM FROM THE SURGE TANK HEATING COILS. LOW HEAT INPUT ALLOWED l
COOLING WATER SYSTEM TEMP TO DECREASE BELOW LCO LIMIT WHILE REACTOR WAS SHUTDOWN. ALL LICENSED OPERATORS WILL BE REINSTRUCTED IN PCRV COOLING W ATER SYSTEM OPERATION. EVAL OF EFtECTS ON TOP HEAD LINER MAT'L REQUESTED
DAVIS-9 ESSE-1 05000346 042300 (HP-33-80-48) ON APRIL 23, 19G0 AT 0440 HOURS, THE TECHNICAL SPECIFICATI 0FFSITE POWER SYSTEMS + CONTRL 80-039/03X-2 072881 ON 3.8.1.2 SURVEILLANCE REQUIREMENT L ATE DATE PASSED FOR THE OFF-SITE AC COMPONENT CODE NOT APPLICABLE 031084 OTHER SOURCES LINED UP AND AVAILABLE WEEKLY TEST. ST 5080.01 THEkE WAS NO DA SUBCOMPONENT NOT APPLICABLE B+W NGER TO THE HEALTH AND SAFETY OF THE PUBLIC OR TO STATION PERSONNEL. THE PERSONNEL ERROR TEST WAS SUCCES$ FULLY PERFORMED AT 1030 HOURS THE SAME DAY AND WAS LATE LICENSED 3 SENIOR OPERATORS BY ONLY FIVE HOURS AND FIFTY MINUTES. SIMILAR EVENTS WERE REPORTED IN N ITEM NOT APPLICABLE P-33-80-37 (LER 80-31) AND NP-33-77-84.
THE CAUSE WAS AN OVERSIGHT BY THE SHIFT SUPERVISOR. THE TEST DID SHOW UP ON THE COMPUTER PRINTED TEST SCHEDULE. GIVEN THE WORK LOAD OF THE $NIFT SUPERVISOR DURING AN OUTAGE, ADDITIONAL REMINDERS OF IMMINENT TESTS ARE NEEDED AND WILL BE SUPPLIED BY SECTION PERSONNEL.
DAVIS-BESSE-1 05000346 022879 AT 0532 STARTUP TRANSFORMER 41 TAKEN OUT OF SERVICE FOR ADJUSTMENTS. TH OFFSITE POWER SYSTEMS + CONTRL 79-023/01T-0 031279 IS IMMEDIATELY ENTERED UNIT INTO ACTION STATEMENT (A) 0 F T. S. 3. 8.1.1 WH COMPONENT CODE NOT APPLICABLE 025394 2-WEEK ICH REQUIRES SURVEILLANCE TESTS BE PERFORMED WITHIN ONE HOUR.
AT 0730 P SUBCOMPONENT NOT APPLICABLE B+W ERSONNEL REALIZED SURVEILLANCE TESTS SHOULD HAVE BEEN PERFORMED AND TEST PERSONNEL ERROR ING WAS COMPLETED BY 0744.
ALTERNATE OFFSITE AC ELECTRICAL POWER SOURCE LICENSED 3 SENIOR OPERATORS WAS IN SERVICE AND SUBSEQUENT SURVEILLANCE TESTS SHOWED BOTH EMERGENCY ITEM NOT APPLICABLE DIESEL GENERATORS OPERABLE.
NP-32-79-01.
PERSONNEL ERROR. OPERATIONS PERSONNEL MISINTERPRETED ACTIDH STATEMENT (
A) 0F T.S. 3.8.1.1 AND FAILED TO PERFORM REQUIRED SURVEILLANCE.
Y_
CALVERT CLIFFS-1 05000317 121378 DURING A ROUTINE TRIP RECOVERY THE REACTOR WAS MADE CRITICAL AT 0552.
A COOL SYS FOR REAC AUX + CONT 78-054/01T-1 122778 T 0610 THE SHIFT SUPERVISOR RECALLED THAT 11 COMPONENT COOLING HEAT EXCH COMPONENT CODE NOT APPLICABLE 023025 2-WEEK ANGER WAS OUT OF SERVICE AND 11 SALTWATER SUB-SYSTEM WAS PARTIALLY DRAIN SUBCOMPONENT NOT APPLICABLE COMB ED, THEREBY INVOKING ACTION STATEMENTS OF T.S.
3.7.3.1. AND T.S. 3.7.5.1 PERSONNEL ERROR
. THE REACTOR WAS SHUTDOWN AT 0615.
MAXIMUM POWER LEVEL ACHIEVED DURIN LICENSED 4 SENIOR OPERATORS G THIS EVENT WAS.0001%. THIS IS NOT A REPETITIVE OCCURRENCE.
ITEM NOT APPLICABLE THIS EVENT WAS CAUSED BY FAILURE OF LICENSED PERSONNEL TO VERIFY THAT TH E COMPONENT COOLING AND SALTWATER SYSTEMS WERE NOT IN A DEGRADED CONDITI ON PRIOR TO TAKING THE REACTOR CRITICAL. PERSONNEL REINSTRUCTED AND MAD E CHANGE TO SCRAM RECOVERY PROCEDURE TO PROVIDE ACTION STATEMENT CHECK L IST.
CALVERT CLIFFS-1 05000317 092479 AT 0600, THE REACTOR EXCEEDED THE LICENSE LIMIT OF 2700 MWTH AS PRINTED SYSTEM CODE NOT APPLICABLE 79-050/03L-0 101179 OUT ON THE COMPUTER LOG TYPEWRITER WITH A POWER OF 2702 MWTH.
THE OPERA COMPONENT CODE NOT APPLICABLE 027169 30-DAY TOR TOOK IMMEDIATE ACTION AND REDUCED REACTOR POWER TO LESS THAN 2700 MW SUBCOMPONENT NOT APPLICABLE COMB TH BY 0605.
THIS IS NOT A REPETITIVE OCCURRENCE.
~
PERSONNEL ERROR LICENSED 4 SENIOR OPERATORS ITEM NOT APPLICABLE UPON INVESTIGATION. IT WAS DETERMINED THAT THE OPERATOR, REALIZING THAT THE INSTANTANEOUS REACTOR POWER WAS HIGH, BORATED TO REDUCE POWER; HOWEV ER, THIS ACTION WAS NOT ADEQUATE. AFTER REALIZING THE AVERAGE HOURLY PO WER EXCEEDED 2700 MWTH. THE OPERATOR IMMEDIATELY REDUCED POWER BY INSERT ING CONTROL ELEMENT ASSEMBLIES.
s
CRYSTAL RIVER-3 05000302 121878 AT 0600 DURING ROUTINE OPERATION OF AHF-14A AND AHF-14C.AHF-14C OVERLOAD AIR COND HEAT, COOL, VENT SYSTEM 78-065/01T-0 121978 ED, WAS SECURED TO ATTEMPT TO BALANCE AIR FLOW AND TRIPPED DURING RESTAR BLOWERS 023162 2-WEEK T.
NO EFFECT ON PUBLIC HEALTH AND SAFETY.
AHF-14C PROMPTLY RETURNED TO SUBCOMPONENT NOT APPLICABLE B+W SERVICE AT 0610.
EVENT CONTRARY TO T.S. 3.7.8.1.
5 PREVIOUS OCCURRENC PERSONNEL ERROR ES.
FIRST THIS TYPE.
LICENSFD & SENIOR OPERATORS JOY MANUFACTURING CO.
ATTRIBUTED TO NOT ALLOWING SUFFICIENT TIME FOR OVERLOAD CONDITION TO RES
(
ET BEFORE ATTEMPTING TO RESTART AHF-14C.
OPERATIONS PERSONNEL INSTRUCTE j
D OF FAN OPERATING CHARACTERISTICS.
I BRUNSWICK-2 05000324 051979 DURING A REVIEW OF STARTUP DATA FOLLCWING A REACTOR STARTUP, IT WAS NOTE REACTOR CONTAINMENT SYSTEMS 79-039/03L-0 061879 D THAT PT 2.6.5, PRIMARY CONTAINMENT LEAK RATE TEST ON THE AIRLOCK, WAS COMPONENT CODE NOT APPLICABLE 026123 30-DAY NOT COMPLETED WITHIN THE REQUIRED 72 HOURS AFTER REACHING CONDITION 3.
SUBCOMPONENT NOT APPLICABLE GE CONDITION 3 WAS REACHED AT 0640 ON MAY 14, 1979 AND THE LEAK RATE TEST W PERSONNEL ERROR AS NOT DONE UNTIL MAY 19, 1979, FOR A TOTAL ELAPSED TIME OF APPROXIMATEL LICENSED 8 SENIOR OPERATORS Y 133 HOURS.
(TECHNICAL iPECIFICATIONS 3.6.1.3, 6.9.1.9C)
ITEM NOT APPLICABLE PT 2.6.5 WAS COMPLETED SATISFACTORILY ON THE AIRLOCK.
DURING THIS PERIO D OF OPERATION, AN ADMINISTRATIVE OVERSIGHT PREVENTED THE DETECTION OF P T 2.6.5 NOT HAVING BEEN PERFORMED. THE PLANT STARTUP PROCEDURE GP-1 HAS l
BEEN REVISED SUCH THAT PT 2.6.5 MUST BE COMPLETED BEFORE THE REACTOR IS I'
TAKEN CRITICAL.
y>
w us POINT BEACH-1 05000266 033079 ON 3/30/79, AT APPROX. 0640 HRS, FOLLOWING A COMPUTER SHUTDOWN, AN OPERA SAFETY RELATED DISPLAY INSTR 79-005/01T-0 041279 TOR INITIALIZED THE COMPUTER INCORRECTLY RESULTING IN LOSS OF REACTOR CO INSTRUMENTATION + CONTROLS 025586 2-WEEK RE AXIAL FLUX DIFFERENTIAL MONITORING PROGRAM & ASSOCIATED ALARM. UNIT COMPUTATION MODULE WEST OPERATED FOR NEXT 9 HRS WITHOUT MANUALLY LOGGING FLUX DIFFERENTIAL AS RE
~
PERSONNEL ERROR QUIRED BY T.S.
15.3.10.B.2.F ALTHOUGH NORMAL VISUAL SURVEILLANCE WAS MAI LICENSED 8 SENIOR OPERATORS NTAINED. THIS EVENT IS SIMILAR TO UNIT 2 LER 78-006/01T-0.
WESTINGHOUSE ELECTRIC CORP.
INVESTIGATIONS SHOWED THAT THE OPERATOR DID NOT RECOGNIZE THAT THE COMPU TER DID NOT START THE MONITORING PROGRAM.
THE COMPUTER WAS REINITIALIZE D AT 1540 HOURS ON 3-30-79.
THE RE-INITIALIZATION PROCEDURE HAS BEEN CH ANGED TO PROVIDE THE OPERATOR WITH A P3SITIVE VERIFICATION THAT THE PROG RAM IS RUNNING.
SALEM-2 05000311 071880 DURING ROUTINE OPERATION, IT WAS DETERMINED THE AUXILIARY FEEDWATER STOR CONDENSATE STORAGE FACILITIES 80-017/03L-0 081280 AGE TANK LEVEL FELL BELOW THE MINIMUM OPERATING LEVEL OF 200,000 GALLONS ACCUMULATORS 031965 30-DAY AT 0708 HOURS AND ACTION STATEMENT 3.7.1.3 WAS ENTERED (DOCKET 50-272, SUBCOMPONENT NOT APPLICABLE WEST
- 78-40).
~
PERSONNEL ERROR LICENS2D 8 SENIOR OPERATORS OTHER THE INLET VALVE WAS THROTTLED TOO FAR WHILE MAKING UP WATER TO THE STEAM GENERATORS FROM THE TANK, THEREBY CAUSING THE LEVEL TO DECREASE TOO FAR THE LEVEL WAS RESTORED AND THE ACTION STATEMENT WAS TERMINATED AT 073 1 HOURS.
SEQUOYAH-1 05000327 030830 UNIT 1 OPERATING IN MODE 2 AT 3.5 PERCENT POWER. UNIT 2 PREOPERATIONAL EMERG GENERATOR SYS + CONTROLS 80-129/03L-0 090280 TEST TVA-13B1 IN PROGRESS. AT 1240 HOURS DG 1A-A FAILED TO START FROM A ENGINES. INTERNAL CCMBUSTION 032236 30-DAY SIMULATED BLACK 0UT SIGNAL ON 2A-A SHUTDOUN BOARD.
THIS EVENT RESULTED SUBCOMPONENT NOT APPLICABLE WEST IN UNIT 1 ENTERING TECHNICAL SPECIFICATION ACTION 3.8.1.1.
THERE WAS NO PERSONNEL ERROR DANGER TO THE HEALTH AND SAFETY OF TNE PUBLIC, PREVIOUS OCCURRENCES -
LICENSED & SENIOR OPERATORS NCNE.
OTHER DGIA-A FAILED TO START DUE TO 43T (L) TEST SWITCH BEING IN THE TEST POSI TION.
THE SWITCH WAS PLACED IN THE NORMAL POSITION ALLOWING DG 1A-A TO START.
DURING PERFORMANCE OF SI-7 ON 8/6/30 AT 0816, 43T (L) WAS PLACED IN TEST 4 WAS NOT RETURNED TO NORMAL AT CONCL OF SURVEILL. INTRUC ARE BEING REV TO INCLUDE SIGN OFF SPACES FOR VERIF OF DG STANDBY READINESS.
SALEM-1 05000272 030381 WHILE INCREASING POWER LEVEL DURING STARTUP, IT WAS DISCOVERED THAT THE CONDENSATE STORAGE FACILITIES 81-024/03L-0 040281 LEVEL IN THE AUXILIARY FEEDWATER STORAGE TANK WAS BELOW THE MINIMUM LEVE ACCUMULATORS 036601 30-DAY L (200.000) GALLONS) REQUIRED BY TECHNICAL SPECIFICATION 3.7.1.3.
ACTIO SU3 COMP 0HENT NOT APPLICABLE WEST N STATEMENT 3.7.1.3 WAS ENTERED AT 1018 HOURS. PREVIOUS OCCURRENCES: 7
~
PERSONNEL ERROR 8-40, 80-17, AND 80-27.
LICENSED 4 SENIOR OPERATORS OTHER THE CONTROL OPERATOR HAD GIVEN INSTRUCTIONS TO OPEN THE SUPPLY VALVE, BU T NOT IN SUFFICIENT TIME TO PREVENT THE TEMPORARY LOSS OF REQUIRED LEVEL IN THE TANK.
THE CONTROL OPERATOR WAS COUNSELED FOR HIS ERROR IN JUDGM.
ENT.
YM INDIAN POINT-3 05000236 032481 WHILE AT HOT SHUTDOWN AT 2330 HOURS, IT WAS OBSERVED THAT SERVICE WATER STATION SERV WATER SYS + CONT 81-004/03L-0 042381 PUMP NO. 36 HAD BEEN TAKEN OUT-OF-SERVICE AT 0825 HOURS THAT DAY FOR MAI PUMPS 036738 30-DAY NTENANCE.
SERVICE WATER PUMPS 35 AND 36 HAD BEEM SUPPLYING THE NON-ESSE CENTRIFUGAL WEST NTIAL HEADER.
TECH. SPEC. 3.3.F.1 REQUIRES 2 SERVICE WATER PUMPS TO BE PERSONNEL ERROR OPERABLE ON THE NON-ESSENTIAL HEADER AT ALL TIMES ABOVE COLD SHUTDOWN, A LICENSED 8. SENIOR OPERATORS ND SPEC. 3.3.F.2 REQUIRES C00LDOUN TO COMMENCE WITHIN 12 HOURS IF THE CO LAYNE-BOWLER, INC NDITIONS OF 3.3.F.1 CANNOT BE MET.
IMMEDIATELY UPON IDENTIFICATION OF THIS SITUATION, DESCENT TO COLD SHUTD OWN WAS INITIATED. THE HOLD-OFF ON SERVICE WATER PUMP NO. 36 WAS CLEARE D AT 2400 HOURS. AND THE PUMP WAS DECLARED OPERABLE, THEREBY SATISFYING THE REQUIREMENTS OF TECHNICAL SPECIFICATION 3.3.F.1.
NO SIMILAR EVENTS HAVE BEEN REPORTED TO DATE.
BEAVER VALLEY-1 05000334 112779 AT 0637 HOURS. NO. 1 EMERGENCY DIESEL GENERATOR WAS REMOVED FROM SERVICE EMERG CORE COOLING SYS + CONT 79-046/01T-0 121179 FOR PREVENTIVE MAINTENANCE. GENERATOR PROVIDES EMERGENCY POWER TO CH-11 VALVES 027663 2-WEEK 5B. ONE OF TWO PARALLEL MOTOR-OPERATED VALVES ON SUCTION LINE FROM REFUE GATE WEST LING WATER STORAGE TANK TO HHSI PUMPS. AT APPROX. 0830 HOURS, REDUNDANT PERSONNEL ERROR NOV, CH-115D, WAS PLACED ON CLEARANCE IN CLOSED POSITION TO CORRECT A M0 LICENSED 5 SENIOR OPERATORS TOR OPERATOR PROBLEM. ERROR WAS DISCOVERED AT 1015 HOURS. VALVE WAS REST ITEM NOT APPLICABLE ORED AT 1029 HOURS PROVIDING THE REQUIRED FLOW PATH.
FLOW PATH TO HHSI PUMPS WOULD HAVE BEEN AVAILABLE WITH OFFSITE POWER AVA ILABLE. IN EVENT OF COINCIDENT HHSI AND LOSS OF OFFSITE POWER. HHSI WOUL D ONLY HAVE BEEN AVAILABLE UNTIL VOLUME CONTROL TANK EMPTIED. INCIDENT W AS CAUSED WHEN CLEARANCE APPROVAL WAS GRANTED IN MIDST OF NORMAL EARLY M ORNING DEMANDS OH OPERATOR TIME.
I
SEQUDYAH-1 05000321 031480 AT 0840 CST ON 5/13/80, ESSENTIAL RAW COOLING WATER RADIATION MONITORS 0 OTHR INST SYS NOT PECD FR SFTY 80-006/01T-0 032780 N TRAIN A WERE IHOPERABLE. AT 0300 CST 3/14/80, IT WAS DETERMINED THAT COMPONENT CODE "i JPPLICABLE 030573 2-WEEK LCO 3.3.3.9, TABLE 3.3-12 REQUIRES SAMPLING (ST 414) IF NO RADIATION MON SUBCGMPONENT %V 't*LICABLE WEST ITOR$ ARE CPERABLE. SI 414 WAS IMMEDIATELY INITIATED. ALL EFFLUENT REL PERSONNEL ERROR EASES OCCURRING DURING THE PERIOD OF ERCN RADIATION MONITOR INOPERABLILI LICENSED 8 SENIP --itATORS TY WERE MONITORED BY COOLING TCWER BLCWDOWN RADIATION MONITOR. NO UNMON ITEM NOT APPLIC/ u :
ITORED RELEASES WERE MADE.
TECHNICAL SPECIFICATION 3.3.3.9, TABLE 3.3-12 STATES ONLY ONE RADIATION MONITOR IS REQUIRED. RADIATION MONITORS WERE OPE *ABLE ON B TRAIN ERCW B UT THIS DID NOT SAFTISFY T.S.
ONE RADIATION MONITOR PER ERCW EFFLUENT P ATH IS REQUIRED. ALL NECESSARY PERSONNEL REINSTRUCTED AS TO CORRECT INT ERPRETATION OF T.S.
JOSEPH M.
FARLEY-1 05000348 042381 AT 1253 ON 4/23/81, ACCUMULATOR 1B WAS DECLARED INOPERABLE DUE TO HIGH N EMERG CORE COOLING SYS + CONT 81-028/03L-0 052681 ITROGEN PRESSURE.
TECH. SPEC. 3.5.1 REGUIRES THE 1B ACCUMULATOR TO BE D ACCUMULATORS 037079 30-DAY PERABLE.
TECH. SPEC. 3.5.1 ACTION REQUIREMENTS WERE MET.
THE HEALTH AN-SUBCOMPONENT NOT APPLICABLE WEST D SAFETY OF THE PUBLIC WERE NOT AFFECTED BY THIS EVENT.
Pf(SONNEL ERROR LICENSED 8 SENIOR OPERATORS e
-JESTINGHOUSE ELECTRIC CORP.
OCCURRENCE WAS DUE TO ACCUMULATOR BEING INITIAiLY PRESSURIZED TO VALUE T 00 CLOSE TO ALARM SETPOINT OF 649 PSIG. AS CONTAINMENT TEMPERATURE INCRE ASED, ACCUMULATOR PRESSURE INCREASED TO ALARM SETPOINT. ACCUMULATOR 1B I MMED VENTED TO 640 PSIG 1 RETURNED TD NORMAL'AT-1255 4/23/81. INDIVIDUAL S INSTRUC TO MONITOR ACCUMU/ PRES 3 MGRE CLOSELY, ESP DURING PLANT PEATUPS s
$b se H.
B. POBINSON-2 05000261 101480 OH 2 OCCASIONS, AT 0910 8 1600 HOURS, WHILE IN A HOT SHUTDOWN SUPCRITICA REACTOR CONTAINMENT SYSTEMS 80-024/01X-1 120280 L CONDITION, THE CONTAINMENT PERSONNEL AIRLOCKS WERE OPENED SIMULTANEOUS PENETRATIONS,PRIMRY CONTAINMNT 032869 OTHER LY. 1) MAINTENANCE WAS BEING PERFORMED ON AIRLOCK DOOR INTERLOCKS WHEN A PERSONNEL ACCESS WEST SMALL FIRE WAS REPO*TED INSIDE CONTAIHMENT. OPERATIONS PERSONNIL OPENED PERSOHNEL ERROR EXTERIOR DOOR WHIL E INNER DOOR WAS NOT SEAL ED.
- 2) INTERLOCK FAILED TO LICENSED 8 SENIOR OPERATORS FUNCTION PROPERLY,' ALLOWING' BOTH DOORS TO B E OPENFD. EVENTS CONTRARY TO CHICAGO BRIDGE 3 IRON COMPANY TECH. SPEC. 3.6.1.A; REPORTABLE PER TECH. SPEC.
6.9.2.A.2.
AIRLOCK INTERLOCKS WERE NOT IN SERVICE AT TIME OF FIRST EVENT DUE TO MAI HTENANCE. IN HASTE TO EXTINGUISH FIRE, OPERATCR OPENED DOOR WITHOUT FIRS T CHECKING WITH MAINT. PERSONNEL WORKING ON DOORS. AIRLOCK INTERLOCKS FA ILED TO FUtlCTION PROPERLY DURING SECOND EVENT WHILE PERSONNEL WERE EXITI NG CONTAINMENT.. VENDOR WILL INSPECT
- RECCUMEND: IMPROVEMENTS TO AIRLOCK.
CALVERT CLIFFS-2 05000318 072380 AT 0920 EMERGENCY POWER SUPPLY WAS REMOVED FROM #21 CHARGING PUMP CAUSIN REACTIVITY CONTROL SYSTEMS 80-032/01T-0 080680 G PUMP TO BECOME IHOPERABLE PER T.S.
1.6.
AT SAME TIME TWO REDUNDANT CH PUMPS 031932 2-WEEK G PPS WERE INOPERABLE DUE TO MAINTENANCE ALLOWED BY T.S.
3.1.2.4.
AT 11 RECIPROCATING COMB 56 823 CHG PP RETURNED TO OPERABLE STATUS.
TWO OFFSITE POWER SOURCES AN PERSONNEL ERROR D EMERGENCY POWER SUPPLY FOR #23 CHG PP REMAINED OPERABLE. NOT REPETITI LICENSED 8 SENIOR OPERATORS VE.
ARMCO METAL PRODUCTS CORPORATI OPERATOR'S FAILURE TO VERIFY WHICH SAFETY-kELATED EQUIPMENT OUT OF SERVI CE PRIOR TO REMOVING EMERGENCY POWER SUPPLIES FROM SERVICE.
TO PREVENT RECURRENCE APPROPRIATE INSTRUCTIONS WILL BE CHANGED TO REQUIRE OPERATOR TD-DOCUMENT HE HAS VERIFIED OPERABILITY OF REDUNDANT EQUIPMENT PRIOR TO PLACING SAFETY-RELATED EQUIPMENT GUT OF SERVICE.
/
o FT. ST. VRAIN-1 05000267 082080 FROM 1000 TO 1115 HOURS ON AUGUST 20, 1980, WITH THE PLANT OPERATING AT COOLANT RECIRC SYS + CONTROLS 80-046/03L-0 091980 APPROXIMATELY 25% THERMAL POWER, PCRV (PRESTRESSED CONCRETE REACTOR VESS INSTRUMFNTATION + CONTROLS 032421 30-DAY EL) COOLING WATER OUTLET TEMPERATURE EXCEEDED THE 120 F LIMIT OF LCO 4.2 CONTROLLER GGA 15 B).
DEGRADED MODE OF LCO 4.2.15. REPORTABLE PER FCtT ST. VRAIN TEC" PERSCSNEL ERROR NICAL SPECIFICATION AC 7.5.2(B)2.
NO ACCOMPANYING OCCU2RENCE OR PROBAbt LICENSED 3 SENIOR OPERATORS E CONSEQUENCES. NO EFFECT ON PUBLIC HEALTH AND SAFETY.
FOXBORO CO.,
THE CAUSE OF INCREASED TEMPERATURE WAS ISOLATION OF COOLING WATER DUE TO OPE RATION OF THE WRONG CONTROLLERS. UPON DISCOVERY OF THE OUT OF LIMIT TEM PERATURES. THE CONTROLLERS WERE RETURNED TO THE PROPER POSITION. TEMPER ATURES RETURNED TO ACCEPTABLE LIMITS BY 1115 HOURS AND NO FURTHER DEGRAD ED MODE OPERATION WAS OBSERVED.
BIG ROCK POINT 05000155 021578 DURING MAINTENANCE ACTIVITY ON THE REACTOR DEPRESSURIZING SYSTEM CONTROL OTHER ENGNRD SAFETY FEATR SYS 78-008/01T-0 022878 CIRCUITRY, THE FIRE PUMP CONTROL SWITCHES UERE INADVERTENTLY PLACED IN CIRCUIT CLOSERS / INTERRUPTERS 020545 2-WEEK THE INHIBIT POSITION WITH BOTH FIRE PUMPS HDT RUNNING BETWEEN 0935 AND 1 SWITCH (OTHER THAN SENSOR)
GE 315 HOURS: THUS THE SYSTEM WAS NOT AUTOMATICALLY OPERABLE AS REQUIRED BY PERSONNEL ERROR T/S 3.1.5A. THE FIRE PUMP FUNCTION COULD HAVE BEEN MADE FULLY OPERABLE LICENSED 4 SENIOR OPERATORS BY OPEnaTOR ACTION IN THE CONTROL ROOM IF CONDITIONS HAD WARRANTED USE O ITEM HOT APPLICABLE F THE SYSTEM. INCIDENT NOT REPETITIVE.
THE FIRE PUMP CONTROL SWITCHES ON THE RDS CONSOLE HAVE BEEN MARKED WITH SPECIFIC INSTRUCTIONS FOR USE OF THE INHIBIT POSITION SINCE THE ERROR TH AT OCCURRED MIGHT NOT OTHERWISE BE POSITIVELY CONTROLLED FOR ALL POSTULA TED TROUBLESHOOTING ACTIVITIES. THE INCIDENT WILL BE REVIEWED WITH LICEN SED PERSOHNEL.
Y EDWIN I. HATCH-2 05000366 092178 WHILE STARTING UP BUT STILL SUB-CRITICAL, IT WAS DISCOVERED THAT THE DRY SYSTEtt CODE NOT APPLICABLE 78-039/03L-0 100378 WELL AIRLOCK DOOR SEALS SURVEILLANCE REQUIRED EVERY 72 HOURS WHILE MAKIN COMPONENT CODE NOT APPLICABLE 022491 30-DAY G MULTIPLE ENTRIES WAS NOT PERFORMED BY 0940 DN 9/19/78 AS REQUIRED.
SUBCOMP0HEHT NOT APPLICABLE GE DERSONNEL ERROR LICENSED 8 SENIOR OPERATORS ITEM NOT APPLICABLE SURVEILLANCE WAS PERFORMED UPON DISCOVERY AND COMPLETED AT 1700 9/21/78.
PERSONNEL WERE INSTRUCTED TO WATCH THIS REQUIREMENT CAREFULLY SO THAT IT DOES NOT GET MISSED AGAIN.
DAVIS-BESSE-1 05000346 091478 DG 1-1 DECLARED IHOPER.- 09/13/78 AT 0952 FOR PREVENTIVE MAINTENANCE. SU EMERG GENERATOR SYS + CONTROLS 78-096/03L-0 101178 RVEILLANCE TESTING REQUIRED BY T.S.
3.8.1.1 NOT PERFORMED UNTIL 09/14/78 ENGINES, INTERNAL COMBUSTION 022692 30-DAY AT 0134.
ON DISCOVERY OF OMISSION SHIFT FOREMAN IMMEDIATELY ARRANGED T SUBCOMP0HENT NOT APPLICABLE B+W 0 PERFORM REQUIRED TESTING. NO DANGER TO HEALTH AND SAFETY OF PUBLIC OR PERSONNEL ERROR UNIT PERSOHHEL. EMERGENCY DIESEL GENERATOR 1-2 ABLE TO START AND SUPPL LICENSED 4 SENIOR OPERATORS Y ELECTRICAL POWER TO ESSENTIAL BUSSES.
HP-33-78-113.
ITEM NOT APPLICABLE PERSONNEL ERROR.
THE SHIFT FOREMAN FAILED TO RECOGNIZE SURVEILLANCE TES TING REQUIRED. SHIFT FOREMAN IMMEDIATELY COMPLETED APPROPRIATE SECTIONS OF SURVEILLANCE TESTS ST 5080.01 AND ST 5081.01.
THIS LER WILL BE DIST RIBUTED TO OPERATIONS PERSONNEL AS ADDITIONAL CORRECTIVE ACTION.
I CALVERT CLIFFS-1 05000317 072378 WHILE RETURNING TO FULL POWER FOLLOWING AN CUTAGE THE CIRCULATING WATER CIRCULATING WATER SYS + CON 78-036/04X-0 072478 INJECTION TEMPERATURE APPARENTLY TOOK A SUDDEN DROP.
AT 10:00 A.M.
THE COMPONENT CODE NOT APPLICABLE 021823 OTHER TEMPERATURE RISE ACROSS THE MAIN CONDENSERS EXCEEDED THE MAXIMUM ALLOWED SUBCOMPONENT HOT APPLICA3LE COMB AT 10.3 DEGREES PER ETS 2.1.1.
POWER WAS IMMEDIATELY REDUCED AND CONDE PERSONNEL ERROR NSER DELTA T WAS RETURNED TO SPECIFICATION BY 11: 00 A.M.
LER'S 78-35 (U LICENSED 8 SENIOR OPERATORS
-1) AND 77-86 (U-2) DESCRIBE SIMILAR EVENTC.
ITEM NOT APPLICABLE THE CAUSE WAS APPARENTLY DUE TO A RAPID DRCf IN THE CIRCULATING WATER IN l
JECTION TEMPERATURE. THE COMPUTER PROGRAM iMAT CALCULATES CONDENSER DEL TA T DOES NOT HAVE PROPER RTD DEL AY TIMES TO DAMPEN THE EFFECTS OF RAPID WATER TEMPEkATURE CHANGES. THE DELAY TIMES WILL BE MEASURED AND PROGRA MMED INTO THE COMPQTER DURING THE HEXT SCHEDULED POWER REDUCTION.
DAVIS-BESSE-1 05000346 042278 REVIEWING REACTOR OP LOG 4/22/78 DISCOVERED GROUP 5 CONTROL RODS 100% 20 REACTIVITY CONTROL SYSTEMS 78-040/03L-0 051878 NE REFERENCE LIGHT LAST CHECKED 4/21/78 AT 2220.
SINCE GROUP 5 IN ASYMM INSTRUMENTATIDH + CONTROLS 021214 30-DAY ETRIC BYPASS. TS 3.1.3.3, ACTION A.2.B REQUIRES GROUP 5 POSITION BE VERI OTHER B+W FIED AT LEAST ONCE EVERY 12 HOURS. NO DANGER TO HEALTH AND SAFETY OF PU PERSONNEL ERROR BLIC OR UNIT PERSONNEL. CONTROL RODS OF GROUP 5 NOT IN AN ASYMMETRIC CO LICENSED 8 SENIOR OPERATORS NDITION.
NP-33-78-50.
ITEM NOT APPLICABLE PERSONNEL ERROR.
OPERATOR INADVERTENTLY FAILED TO PERFORM REQUIRED OBSE RVATION OF ZONE REFERENCE LIGHT.
OPERATOR DISPATCHED TO READ ZONE REFER ENCE LIGHT AT 2117 HOURS ON 4/22/78.
AN INVESTIGATION OF AIDS TO REMIND PERSONNEL OF REQUIREMENTS IS UNDERWAY.
Y; THREE MILE ISLAND-2 05000320 093080 ON SEPT. 30, 1930 AT 1030 HOURS THE FEEDER BREAKERS FOR BUSSES 2 28 AND AC ONSITE POWER SYS + CONTROLS 80-046/01L-0 103080 2-48 WERE REMOVED FROM SERVICE FOR MAINTENANCE. THROUGH THE SHIFT FOREM CIRCUIT CLOSERS / INTERRUPTERS 032784 30-DAY AN'S ERROR THE TIME CLOCK WAS NOT INITIATED FOR REMOVAL OF THE TS REQ BC CIRCUIT BREAKER B+W SSES. CONSEQUENTLY, OUT OF SPEC. CONDITION EXCEEDED APPLICABLE ACTION PE
- PERSONNEL LRROR RIOD BEFORE SITUATION WAS CORRECTED. THIS WAS A VIOL OF TECHNICAL SPECIF LICENSED 8 SENIOR OPERATORS ICATION 3.8.1.2 8 IS CONSIDERED REPORTABLE UNDER SECTION 6.9.1.8(B).
TH GENERAL ELECTRIC CO.
IS HAD NO EFFECT ON PLANT. ITS OPER, OR HEALTH 1 SAFETY OF FUBLIC.
EVENT RESULTED FROM SHIFT FOREMAN NOT CHECKING THE T.S. AS APPROPRIATE, AND CONSEQUENTLY, NOT INITIATING TIME CLOCK.
ONCE IDENTIFIED, BUS FEEDE R BREAKERS WERE REPLACED 8 BUSSES RESTORED TO AN OPERABLE STATUS. TO BOL STER FOREMAN'S KNOWLEDGE, IN TRAINING CYCLE STARTING OCT. 8, 1980, OPER REQUAL PROGRAM WILL INCLUDE A CLASS IN WHICH ALL TS REQ EQUIP IS IDENT.
CRYSTAL RIVER-5 05000302 040180 AT 1145 IT WAS DISCOVERED THAT AN INADVERTENT ACTUATION OF THE FIRE SERV AIR COND. HEAT COOL, VENT SYSTEM 80-016/01T-0 041680 ICE DELUGE SYSTEM RESULTED IN DAMAGE TO AUXILIARY BUILDING VENTILATION E OTHER COMPONENTS 030865 2-WEEK XHAUST FILTER AHFL-2A IN EXCESS OF TWO-THOUSAND ($2,000) DOLLARS.
THIS SUBCOMP0HENT NOT APPLICABLE B+W CREATED AN EVENT REPORTABLE UNDER 10 CFR 20.403.B(4).
NO RISK TO THE HE PERSONNEL ERROR ALTH OR SAFETY OF PLANT OR PUBLIC. AUXILIARY BUILDING VENTILATION WAS S LICENSED 8 SENIOR OPERATORS HIFTED TO AHFL-2B AT 1300.
THIS IS THE FIRST OCCURRENCE OF THIS TYPE RE NUCLEAR CONSULT. SERV. INC.
PORTED.
THE CAUSE OF THIS EVENT IS ATTRIBUTED TO PERSONNEL ERROR.
THE INADVERTE NT ACTUATION OCCURRED WHILE ATTEMPTING TO RESET A FIRE SERVICE PANEL ALA RM.
AHFL-2A MEDIA WAS REPLACED ON 2 APRIL 1980.
POST REPLACEMENT TESTI HG IS SCHEDULED AND WILL BE COMPLETED PRIOR TO RETURNING AHFL-2A TO SERV ICE.
BEAVER VALLEY-1 05000334 092079 PLANT WAS IN OPERATICHAL MODE 3 WZTN RCS TEMP. LESS THAN THE LO-Lo AVERA ONSITE POWER SYSTEM + CONTROL 79-040/99X-0 121979 GE TEMP. VALUE OF 543 DEGREES. AT 1108 HRS, THE #4 VITAL BUS INVERTER W GENERATORS 028013 OTHER AS TRANSFERRED FROM ITS AUXILIARY POWER SOURCE TO THE INVERTER. HOWEVER INVERTER WEST THE INVERTER WAS NOT OPERABLE AND, WHEN THE VITAL BUS WAS TRANSFERRED
~
PERSONNEL ERROR TO THE DEENERGIZED INVERTER, IT RESULTED IN A HIGH STEAM FLOW SIGNAL COI LICENSED 4 SENIOR OPERATORS NCIDENT WITH THE LO-LO TEMPERATURE AND A SAFETY INJECTION OCCURRED. SAF STATIC PRODUCTS ETY INJECTIONS TO DATE:
13 OPERATIONAL, 2 PREOPERATIONAL.
THE INCIDENT RESULTED FROM CPERATOR ERROR IN TRANSFERRING THE NO. 4 VITA L BUS TO THE DEENERGIZED INVERTER. THE INVOLVED PERSONNEL WERE REPRIMAN DED AND REMINDED OF THE NECESSITY TO ADHERE TO WRITTEN PROCEDURES.
i THREE MILE ISLAND-1 05000289 072378 DISCOVERED AT 1452 A EMERGENCY DIESEL GENERATOR AIR VALVE HAD BEEN TAGGE EMERG GENERATOR SYS + CONTROLS 78-022/03L-0 081178 D SHUT AT 1115 DURING ISC SYS MAINTENANCE RENDERING A DIESEL INOPERABLE.
VALVES 022054 30-DAY CONTRARY TO TS 3.7.2.C. REDUNDANT DIESEL WAS NOT IMMEDIATELY TESTED.
OTHER B+W REPORTABLE PER TS 6.9.2.B.(3).
B DIESEL WAS SUCCESSFULLY TESTED. A DIE PERSONNEL ERROR SEL WAS RETURNED TO SERVICE UPON COMPLETION OF ITS REQUIRED SURVEILLANCE LICENSED 8 SENIOR OPERATORS ITEM NOT APPLICABLE
. NO THREAT TO HEALTH AND SAFETY OF THE PUBLIC.
CAUSE FOR TAGGING OUT A DIESEL WITHOUT TESTING B DIESEL WAS PERSONNEL ER ROR.
ALL PERSONNEL HAVE BEEN INFORMED OF THEIR IMPROPER ACTIDHS. ALL 0 PERATIONS DEPARTMENT PERSONNEL HAVE BEEN REBRIEFED ON IMPORTANCE OF PRDP ER TESTING OF REDUNDANT ES SYSTEMS.
w DAVIS-BESSE-1 05000346 011680 ON 1/16/80 AT 1300 HRS, INCREASED SURVEILLANCE FREQUENCY REQUIREMENT OF REACTIVITY CONTROL SYSTEMS 80-007/03L-0 021280 T.S. 3.1.3.1.1 WAS MISSED. INOP OF ASYMMETRIC FOR FAULT CIRCUITRY FOR G INSTRUMENTATION + CONTROLS 030144 30-DAY ROUP 5 ROD 11 REQUIRED A ZONE REFERENCE POSITIDH VERIFICATION WHICH WAS INDICATOR B+W DUE AT 1300 HRS BUT WAS NOT MADE UNTIL 1440 HRS.
WITH ALLOWABLE FREQUEN PERSONNEL ERROR CY STRETCH APPLIED, READING WAS 40 MIN LATE.
THERE WAS NO DANGER TO THE LICENSED 8 SENIOR OPERATORS HEALTH AND SAFETY OF THE PUBLIC OR STATION PERSONNEL. THE READING SHOW ITEM NOT APPLICABLE ED THE ROD WAS STILL AT ITS DESIRED 100% WITHDRAWN POSITION (NP-33-80-09)
THE INCREASED SURVEILLANCE FREQUENCY REQUIREMENT WAS OVERLOOKED BY THE O PERATOR DUE TO THE LACK OF AN ADEQUATE REMINDER SYSTEM.
THE OPERATORS W ILL BE PROVIDED WITH A MULTIPLE TIMER SYSTEM WHICH WILL PROVIDE WARNINGS FOR THESE NON-ROUTINE INCREASED SURVEILLANCE REQUIREMENTS AND PREVENT A RECURRENCE.
SEQUOYAH-1 05000327 032581 UNIT 1 IN MODE 1 AT 98% RTP. ON 3/24/81 AT 0436(C), 2B-B DIESEL GENERATO GAS RADIDACT WSTE MANAGMNT SYS 81-032/01T-0 040681 R WAS REMOVED FROM SERVICE FOR INSPECTION TO FULFILL SURVEILLANCE REQUIR INSTRUMENTATION + CONTROLS 036600 2-WEEK EMENT 4.8.1.1.2.C.f.
AT 1325(C), ABGTS "A" TRAIN WAS REMOVED FROM SERV T OTHER WEST 0 REROUTE AIR LINE FOR FCD-30-146B.
ON 3/25/81 AT 0300(C), IT WAS DETER PERSONNEL ERROR MINED OPERABILITY REQUIREMENTS FOR ABGTS COULD NOT BE MET WITH BOTH 25-8 LICENSED & SENIOR OPERATORS DIESEL GENERATOR 8 ABGTS "A" TRAIN INOPERABLE. THIS RESULTED IN BOTH AB OTHER GTS TRAINS BEING INOPERABLE 8 ENTRY INTO LCO 3.0.3. PREVIOUS DCCS-NONE.
MISINTERPRETATION OF LCO 3.0.5 BY ASST SHIFT ENGINEER RESULTED IN ABGTS "A" TRAIN BEING REMOVED FROM SERVICE WHILE 2B-B DIESEL GENERATOR WAS INO
THE RQMTS OF LCO 3.0.5 HAVE BEEN DISCUSSED WITH ALL SHI FT ENGINEERS, ASST SHIFT ENGINEERS, UNIT OPS, 4 SHIFT TECH ADVISORS.
ZION-1 05000295 052379 AT 132G HRS. ON 5/23/79. WHILE PERFORMING SAFEGUARDS LOGIC TESTING AT 99 EMERG CORE COOLING SYS + CONV 79-042/99X-0 082079
% POWER, AN INADVERTANT RX. TRIP AND SAFETY INJECTION WAS INITIATED ON U COMPONENT CODE NOT APPLICABLE 026756 OTHER NIT 1.
TECH. SPEC. 3.3.2.F.3 REQUIRES A 90 DAY REPORT. VISUAL INSPECTI SUBCOMPONENT NOT APPLICABLE WEST ONS INDICATED THAT HO DAMAGE OCCURRED AS A RESULT OF THE SAFETY INJECTIO PERSONNEL ERROR H.
THE HEALTH AND SAFETY OF THE PUBLIC WERE NOT AFFECTED.
LICENSED 8 SENIOR OPERATORS ITEM NOT APPLICABLE AN OPERATOR ERROR IN TESTING TRAIN "A" SAFEGUARDS LOGIC CIRCUITS SIMULAT ED HIGH STEAM FLOW SIGNALS CONCURRENT WITH LCW STEAM PRESSURE. THIS COI NCIDENCE INITIATED MAIN STEAM ISOLATION. WHEN 1B MAIN STEAM ISOL. VLV.
FAILED TO CLOSE. STEAMLINE DIFF PRESSURE REACHED 100PSID AND SI INITIATE D.
OPERATOR WAS REINSTRUCTED.
D.
C. C00K-2 05000316 111479 DURING REFUELING OPERATIONS BORON INJECTION FLOW PATH REQUIRED BY TECH.
CHEM, VOL CONT + LIQ POISH SYS 79-033/01T-0 112879 SPEC. 3.1.2.1 WAS INADVERTENTLY MADE IHOPERABLE TO PERFORM SURVEILLANCE COMPONENT CODE NOT APPLICABLE 027428 2-WEEK TESTING. THE IHOPERABILITY OF THE FLOW PATH WAS NOT IMMEDIATELY RECOGNI SUBCOMPONENT NOT APPLICABLE WEST ZED AND FUEL MOVEMENTS WITHIN THE CORE WERE CONTINUED. AN ALTERNATE BOR PERSCNNEL ERROR ATIDH FLOW PATH FROM THE RWST TO THE RCS THROUGH THE RHR SYSTEM WAS AVAI LICENSED 8 SENIOR OPERATORS LABLE AT ALL TIMES.
ITEM NOT APPLICABLE CHARGING LINE, THE NORMAL BORON INJECTION FLOW PATH TO THE RCS WAS OUT 0 F SERVICE FOR SYSTEM MODIFICATION. PATH FROM CHARGING PUMPS THRCUGH BIT WAS BEING MAINTAINED AS ALTERNATE FLOW PATH. LINE WAS MADE INDPERABLE AT 1345 HOURS NOV. 14 1 THE ERROR DISCOVERED 9.25 HOURS LATER. FUEL MOVEME NT IMMED HALTED 8 FLOW PATH RETURNED TO SERV. PERSONNEL WERE INSTRUCTED.
Y N
~~
EDWIN I. HATCH-2 05000366 111378 WITH THE MODE SWITCH IN RUN AND REACTOR POWER AT 1464 MWT, 2C DIESEL GEN EMERG GENERATOR SYS + CONTROLS 78-069/03L-0 112778 ERATOR WAS BEING RUN TO PROVE OPERABILITY FOLLOWING MAINTENANCE. DURING COMPONENT CODE NOT APPLICABLE 023635 30-DAY THE TIME THE OPERABILITY TEST WAS IN PROGRESS. THE EIGHT HOUR TIME LIMIT SUBCOMPONENT NOT APPLICABLE GE ON SURVEILLANCE ON THE OTHER DIESELS PER TECH. SPEC. 4.8.1.1.1.A RAN OU
~
PERSONNEL ERROR T AT 1345.
THE OPERABILITY TEST ON 2C DIESEL WAS COMPLETED AT 1416.
LICENSED 8 SENIOR OPERATORS ITEM NOT APPLICABLE SINCE THE OPERABILITY TEST ON 2C DIESEL WAS IN PROGRESS. THE OPERATING P ERSONNEL DID NOT DEEM IT NECESSARY TO SECURE THIS TEST TO RUN THE OPERAB ILITY ON THE REMAINING DIESELS.
SALEM-2 05000311 081580 DURING ROUTINE OPERATION, IT WAS DETERMINED THE AUXILIARY FEEDWATER STOR CONDENSATE STORAGE FACILITIES 80-027/03L-0 091180 AGE TANK LEVEL FELL BELOW THE MINIMUM OPERATING LEVEL OF 200,000 GALLONS ACCUMULATORS 032395 30-DAY AT 1349 HOURS AND ACTION STATEMENT 3.7.1.3 WAS ENTERED (DOCKET 50-272.
SUBCOMPONENT NOT APPLICABLE WEST 878-40).
PERSONNEL ERROR LICENSED 8 SENIOR OPERATORS OTHER THE INLET VALVE WAS THROTTLED TOO FAR WHILE MAKING UP WATER TO THE STEAM GENERATORS FROM THE TANK, THEREBY CAUSING THE LEVEL TO DECREASE TOO FAR THE LEVEL WAS RESTORED AND THE ACTION STATEMENT WAS TERMIFATED AT 142
.1 HOURS. A MEMORANDUM MAS SENT TO ALL LICENSED CPERATORS TO HEIGHTEN TH EIR AWARENESS OF THE HARROW BAND OF OPERATION PERMITTED BY TECH. SPECS.
DRESDEN-2 05000237 060780 WHILE PERFCRMING DRYWELL-TORUS PUMP 9ACK DP INSTRUMENT SURVEILLANCE DURIN REACTOP CONTAINMENT SYSTEMS 80-021/03L-0 070280 G NORMAL RX OPERATION, DRYWELL TO TORUS DP DROPPED BELOW 1 PSI FOR 2 HOU INSTRUMENTA110N + CONTROLS 031525 30-DAY R$ IN VIOLATION OF T.S.
3.7.A.7.A.
THE CCHSEQUENCES OF THE EVENT WERE N OTHER GE EGLIGIBLE BECAUSE OF THE SHORT DURATION AND BECAUSE THE DP DROPPED ONLY PERSONNEL ERROR
.99 PSI.
SIMILAR EVENTS:
50-237-76-7, 50-237-76-8.
LICENSED 8 SENIOR OPERATORS ITEM NOT APPLICABLE THE 1 PSI DP REQUIRED BY THE TECHNICAL SPECIFICATIONS WAS INADVERTENTLY ALLOWED TO DROP TO.99 PSI AT 1357 BUT WAS RESTORED TO 1.01 PSI AT 1600 HOURS.
APPROPRIATE PERSONNEL WERE INFORMED OF THE REQUIREMENT TO MAINT AIN 1 PSI DP THROUGHOUT THE TEST.
NO ADDITIONAL CORRECTIVE ACTION REQU IRED.
BEAVER VALLEY-1 05000334 111780 AT 1403 HOURS, THE STATION OPERATING SUPERVISOR (505) WAS NOTIFIED OF A EMERG CORE COOLING SYS + CONT 80-094/99X-1 030281 STEAM LINE LOW PRESSURE BISTABLE [BS-MS-485] INTERMITTENTLY TRIPPING AND INSTRUMENTATION + CONTROLS 033745 OTHER RESETTING. WHEN THE SOS ACCOMPANIED THE OPERATOR TO THE PROCESS RACKS POWER SUPPLY WEST TO TRIP THE BISTABLE, THEY DPERATED THE WRONG CHANNEL IBS-MS-484), IN VI PERSONNEL ERROR OLATIDH OF TECH. SPEC. 3.5.2 AND 3.5.3, ECCS SYSTEMS.
THE PUBLIC HEALTH LICENSED 2 SENIOR OPERATORS 2 SAFETY WERE NOT JEOPARDIZED SINCE ALL ACTIONS OCCURRED IN THE MOST CD ITEM NOT APPLICABLE NSERVATIVE MANNER. SAFETY INJECTIONS TO DATE: 15 OPERATIONAL 4 2 PRE-OP.
THIS INCIDENT OCCURRED BECAUSE SOS DID NOT LOOK AT STATUS PANEL HIMSELF, BUT TOOK WORD OF ANOTHER PERSON INVOLVED IN PLANT OPERATIONS. REQUIRED READING INCIDENT RPT DISTRIBUTION IS BEING USED TO INDICATE NEED TO RECH ECK IMPORTANT EVOLUTIONS SUCH AS TRIPPING BISTABLES. MALFUNCTIONING SIGN AL CONVERTER CARD REPLACED OH 11/19/80 WAS CAUSE FOR FLUTTERING BISTABLE
'w H.
B.
ROBINSON-2 05000261 050579 ON MAY 5, 1979, DURING FUEL SHUFFLE OPERATION, DETECTOR VOLTAGE WAS LOST AC ONSITE POWER SYS + CONTROLS 79-014/03L-0 060479 ON N-32 SOURCE RANGE DETECTOR AT 1403 HOURS.
THIS WAS DUE TO INSTRUMEN ELECTRICAL CONDUCTORS 026054 30-DAY T BUSS 2 BEING SWITCHED FROM MCC-8 TO INVERTER "A".
INVERTER "A" HAD BE SUBCOMPONENT NOT APPLICABLE WEST EN RETURNED TO SERVICE FOLLOWING REQUIRED MAINTENANCE AND SWITCHOVER FR0
?ERSONNEL ERROR M MCC-8 RETURNED INSTRUMENT BUSS 2 TO ITS HORMAL POWER SUPPLY.
FUEL MOV LICENSED 3 SENIOR OPERATORS EMENT WAS STOPPED AND VOLTAGE WAS RESTORED TO THE DETECTOR AT 1405.
THI WESTINGHOUSE ELECTRIC CORP.
S IS CONTRARY TO TECH. SPEC. 3.8.1.D 8 REPORTABLE UNDER 6.9.B.2.
DETECTOR VOLTAGE WAS LOST ON N-32 DUE TO THE POWER SUPPLY SWITCHOVER OF INSTRUMENT BUSS 2 FROM MCC-8 TO INVERTER "A".
VOLTAGE WAS IMMEDIATELY R ESTORED. NO COMPONENT FAILURE OCCURRED. OPERATIONS PERSONNEL WILL BE I NSTRUCTED OF CONSEQUENCES OF THIS BUSS SWITCHOVER.
I i
BEAVER VALLEY-1 05000334 081979 AT 1446 HOURS, THE REACTOR WAS MANUALLY TRIPPED FOLLOWING THE TRIP OF TH TURBINE-GENERATORS + CONTRCLS 79-028/01T-0 083179 E IB STEAM GENERATOR FEED PUMP DUE TO LOW NPSH.
THE OPERATORS PERFORMED CIRCUIT CLOSERS / INTERRUPTERS 026984 2-WEEK THE IMMEDIATE MANUAL ACTIONS FOLLOWING A REACTOR TRIP AND LOSS OF NORMA CIRCUIT BREAKER WEST L FEEDWATER. AN OPERATOR. THINKING THAT THE GENERATOR OUTPUT BREAKERS W PERSONNEL ERROR ERE TAKING TOO LONG TO TRIP, MANUALLY TRIPPED THEM PRIOR TO THE AUTOMATI LICENSED 8 SENIOR OPERATORS C TRIP WHICH NORMALLY OCCURS 30 SECONDS AFTER THE TRIP.
FREQUENCY DROPP ITEM NOT APPLICABLE ED ON THE 1A 8 1B KV BUSES AND ALL THREE RCPS TRIPPED ON UNDERFREQUENCY.
THE INCIDENT RESULTED FROM OPERATOR ERROR IN TRIPPING THE MAIN GENERATOR OUTPUT BREAKER PRIOR TO THE AUTOMATIC TRIP.
THE IC RCP WAS STARTED FOU R MINUTES AFTER THE TRIP AND FORCED FLOW WAS RETURNED TO THE RCS.
THE I NVOLVED OPERATOR WAS REPRIMANDED FOR HIS ACTION AND ALL OPERATIONS PERSO NNEL WILL REVIEW THE INCIDENT CIRCUMSTANC'S.
.1
CALVERT CLIFFS-1 05000317 020280 AT 1512 START FAILURE ALARM RENDERING DIESELS INOPERABLE WERE RECEIVED F EMERG GENERATOR SYS + CONTROLS 80-010/03L-0 020880 OR 11, 12 8 21 DIESEL GENERATORS FOLLOWING SHUTDOWN AFTER AN INADVERTENT ENGINES INTERNAL COMBUSTION 030217 30-DAY UNIT 1 SIAS (T.S.
3.8.1.2.
UNIT 1:
T.S.
3.8.1.3.D. UNIT 2).
ALARM WAS SUBCOMPONENT NOT APPLICABLE COMB RESET AND ALL DIESELS TESTED AND RETURNED TO SERVICE BY 1527. OFFSITE P0 PERSONNEL ERROR WER SYSTEM REMAINED OPERABLE. UNI ( 2 WAS AT FULL POWER AND UNIT 1 IN MO LICENSED 8 SENIOR OPERATORS DE 5 WITH NO REACTIVITY CHANGES BEING MADE.
THIS IS NOT A REPETITIVE OC FAIRBANKS MORSE CURRENCE.
TECHICIAN PERFORMING A TEST CAUSED AN INADVERTENT SIAS STARTING ALL DIES ELS.
SIAS FAILED TO FULLY RESET BEFORE DIESELS WERE STOPPED, CAUSING ST ART FAILURE ALARM.
STANDING ORDER ISSUED REQUIRING OPERATORS TO VERIFY NO ESFAS MODULES ARE TRIPPED PRIOR TO STOPPING DIESELS AFTER A FALSE SIA S.
CALVERT CLIFFS-2 05000318 120479 AT 1520 WHILE PERFORMING A POST REFUELING LOW POWER PHYSICS TEST, TAVE D MAIN STEAM SUPPLY SYS + CONT 79-047/03L-0 010480 R0PPED BELOW 515 DEG. F. CONTRARY TO THE REQUIREMENTS OF T.S.
3.1.1.5.
COMPONENT CCDE NOT APPLICABLE 027816 30-DAY THE OPERATOR IMMEDIATELY TRIPPED THE REACTOR AND PLACED IT IN HOT STANDB SUBCOMPONENT NOT APPLICABLE COMB Y,
SINCE IN HIS JUDGMENT TAVE COULD NOT BE RESTORED TO GREATER THAN 515 PERSONNEL ERROR DEO. F. WITHIN 15 MINUTES. LER 76-11 DESCRIBES A SIMILAR EVENT.
LICENSED 8 SENIOR OPERATORS ITEM NOT APPLICABLE CAUSE WAS DUE TO OPERATOR OVERFEEDING STEAM GENERATOR AND ERRATIC CONTRO L OF TURBINE BYPASS VALVES BECAUSE OF LOW STEAM FLOW.
THE COMBINED EVEN TS CAUSED TAVE TO RAPIDLY DECREASE. SINCE REACTOR OPERATION WITH A PTC IS LIMITED TO BRIEF PERIODS, NO CORRECTIVE ACTIDH IS DEEMED NECESSARY.
w CALVERT CLIFFS-1 05000317 121379 BECAUSE OF LOW OIL PRESSURE #12 CONTROL ROOM AIR CONDITIONING COMPRESSOR CONT ROOM HABITBLTY SYS + CONT 79-077/03L-0 011080 TRIPPED AND WAS OUT OF SERVICE FROM 1515 ON DECEMBER 13, 1979, UNTIL 200 HEAT EXCHANGERS 027942 30-DAY 0 ON DECEMBER 14, 1979.
THE AIR CONDITIONING COMPRESSOR WAS REPAIRED AN OTHER COMB D THE OPERABLE NUMBER RETURNED TO TWO AS REQUIRED BY T.S. 3.7.6.1.
A SI PERSONNEL ERROR MILAR EVENT WAS REPORTED IN LER 78-27.
LICENSED & SENIOR OPERATORS TRANE CD THE COMPRESSOR CRANKCASE DIL WAS DISPLACED TO THE EVAPORATOR COIL DUE TO OPERATION AT LOW AMBIENT TEMPERATURE. THE OIL WAS RECOVERED AND THE UN IT RESTARTED. AN FCR HAS BEEN INITIATED TO INSTALL A HOT GAS BYPASS VALV E TO PREVENT THIS CONDITION FROM RECURRING.
DAVIS-BESSE-1 05000346 010681 (HP-33-81-02) ON 1/6/81 AT 1600 HRS, A CONTROL ROOM OPERATOR INTENDING T AC ONSITE POWER SYS + CONTROLS 81-003/03L-0 020381 0 MATCH A SWITCH FLAG POSITION WITH BREAKER INDICATION LIGHT, INADVERTEN CIRCUIT CLOSERS / INTERRUPTERS 036030 30-DAY
.TLY TRIPPED BREAKER HXO2B WHICH DE-ENERGIZED VITAL BUS B, 1 0F 2 13.8 KV SWITCH (OTHER THAN SENSOR)
B+W BUSSES. THIS REDUCED ELEC DISTRIBUTION TO LESS THAN THAT REQ BY T.S.
3 PERSONNEL ERROR
.8.1.1, 3.8.2.1 8 3.8.2.3.
THERE WAS NO DANGER TO THE HEALTH 8 SAFETY 0 LICENSED 8 SENIOR OPERATORS F THE PUBLIC OR TO STATION PERSONNEL. THE BREAKERS WERE ALL RESTORED WI ITEM NOT APPLICABLE THIN 15 MINUTES. THIS IS THE FIRST REPORT OF THIS TYPE PERSONNEL ERROR.
CAUSE WAS PERSONNEL ERROR BY OPERATOR.
IN ATTEMPTING TO MATCH SWITCH FL AG COLOR WITH BREAKER INDICATION, HE ACCIDENTALLY TURNED SWITCH THE WRON G DIRECTION EVEN THOUGH HE WAS AWARE OF ITS FUNCTION. RESTORATION OF BR EAKERS WAS ATTAINED WITHIN 15 MIN.
OPERATIONS PERSONNEL BEING MADE AWAR E OF OCCURRENCE TO EMPHASIZE NEED FOR ATTENTIVENESS DURING OPERATIONS.
PILGRIM-1 050002C3 6603G1 ON APRIL 3, 1981 AT APPROXIMATELY 1703 HOUR 3,A REACTOR OPERATOR WHILE PE EMERG CORE C00L2HG SYS + CONT G1-012/03L-@ 050181 RFORMXNG ROUTIHE CHECKS CN THE LOW PRESSURE ECCS SYSTEM INADVERTENTLY CL CCMPONEH1 CODE NOT APPLICABLE 036916 30-DAY OSED 4KV BREAKER A507 WHICH FEEDS CORE SPRAY PUMP P215A.
OPERATOR IMMED SUSCOMPONENT NOT APPLICABLE GE IATELY TRIPPED BREAKER. HOWEVER, OPERATING PERSONNEL COULD NOT RECHARGE
~ PERSONNEL ERROR BREAKERS CLOSE MECHANISM SO THEY IMMEDIATELY ISSUED MAINTEHANCE REQUEST LICENSED 4 SENIOR OPERATORS 3 INITIATED REQUIRED SURVEILLANCE TESTING. PROBLEM RESOLVED. SURVEILLANC ITEM NOT APPLICABLE E TERMINATED & CORE SPRAY SYS RETURNED TO NORMAL BY 2100 HRS SAME DAY.
l THE ACTION TAKEN BY THE OPERATOR APPROXIMATED THE AUTOMATIC SEQUENCE WHI CH OCCURS DURING A " TRIP FREE" BREAKER OPERATION. HOWEVER, THE CLOSE/TR IP SEQUEHCE TIMING WAS DIFFERENT RESULTING IN A CHARGING MOTOR CAM INCOM PLETE CYCLE WHICH REQUIRED MANUAL INTERVENTION TO CORRECT. THIS IS AN I SOLATED EVENT.
DAVIS-BESSE-1 05000346 122878 1720 UNIT ENTERED MODE 4.
AT 1930 NOTED CONTAINMENT SPRAY PUMP MOTOR BR CNTNMNT HEAT REMOV SYS + CONT 78-123/01T-0 011079 EAKERS HAD NOT BEEN RACKED IN.
THIS WAS IN EXCESS OF ACTION STATEMENT 0 CIRCUIT CLOSERS / INTERRUPTERS 023567 2-WEEK F T.S.
3.6.2.1 AND PLACED UNIT IN LCO 3.0.3 UNLESS CORRECTIVE MEASURES A CIRCUIT BREAKER B+W RE COMPLETED. AT 2025 BREAKERS WERE RACKED IN.
NO DANGER TO HEALTH AND PERSONNEL ERROR SAFETY OF PUBLIC OR UNIT PERSONNEL. BREAKERS WERE RACKED IN BEFORE CRI LICENSED 8 SENIOR OPERATORS TICALITY WAS ACH 5VED.
NP-32-78-13.
ITEM NOT APPLICABLE PERSONNEL ERROR.
OPERATOR MISINFORMED REGARDING REQUIREMENTS OF CONTAIN MENT STRAY PUMPS AND FAILED TO FOLLOW WRITTEN PROCECURE. BREAKERS WERE RACKED IN BY MAINTENANCE PERSONNEL. ALL PERSONNEL INVOLVED HAVE BEEN AD VISED OF THE IMPORTANCE OF PROCEDURAL COMPLIANCE.
?
Y2 SEQUOYAH-1 05000327 112680 UNIT IN MODE 3 WITH RCS TEMPERATURE 548 DEGREES F AND PRESSURE 2235 PSI CNTNMNT ISOLATION SYS + CONT 80-190/03L-0 122480 G.
ON 11/24/80 AT 1730 CST, 20 CONTAINMENT PURGE ISOLATION VALVES WERE COMPONENT CODE NOT APPLICABLE 033456 30-DAY STROKED PER SURVEILLANCE REQUIREMENT 4.6.3.1.3.
THE VALVES WERE NOT LEA SUSCOMPONENT NOT APPLICABLE WEST K RATE TESTED WITHIN 24 HOURS FOLLOWING THE STROKE TEST AS REQUIRED BY S
~
PERSONNEL ERROR URVEILLANCE REQUIREMENT 4.6.3.1.4.
THERE WAS H0 EFFECT UPON PUBLIC HEAL LICENSED & SENIOR OPERATORS TH OR SAFETY. PREVIOUS OCCURRENCE - NONE.
ITEM NOT APPLICABLE FOLLOWING STROKE TEST, SHIFT OPERATIONS PERSONNEL SUBMITTED A MAINTENANC E REQUEST TO PERFORM REQUIRED LEAK RATE TESTS. RESULTS SECTION DID NOT R ECEIVE MAINT. REQUEST WITH 24-HR TIME LIMIT. VALVES WERE SATISFACTORILY TESTED WITHIN 41 HRS OF STROKE TEST. OPS PERSONNEL HAVE BEEN REINSTRUCTE D TO CONTACT RESULTS PERSONNEL DIR. FOR TESTS REQ'D WITHIN SPEC TIME LIM ZION-1 05000295 081679 OH 8/15/79 AT 1900 HR$ WHILE PERFORMING SAFEGUARD TEST PT-10A, CONTAINME CNTNMNT ISOLATION SYS + CONT 79-063/03L-0 083079 NT ISOL. VLV 1A0V-PR21B, FAILED TO CLOSE. REDUNDANT VLV PR-21A, WAS 0 VALVES 026751 30-DAY PERABLE BUT WAS NOT IMMEDIATELY SECURED IN CLOSED POSITIDH (VIOLATING CO PLUG WEST NT. INTEGRITY TECH. SPEC. 3.9.5).
VALVE WAS SECURED 12 HOURS LATER BY N
- PERSONNEL ERROR EXT SHIFT. PUBLIC HEALTH AND SAFETY WAS NOT AFFECTED.
LICENSED & SENIOR OPERATORS MASONEILAN INTERNATIONAL, INC.
FAILURE TO PROMPTLY SECURE 1FCV-PR21A IN THE CLOSED POSITION WAS DUE TO PERSONNEL ERROR.
OPERATORS HAVE BEEN IHSTRUCTED TO TAKE PROMPT ACTION A ND PROVIDE CLEAR NOTIFICATION OF FAILURES TO SHIFT SUPERVISOR.
1FCV-PR2 1A WAS SECURED BY NEXT SHIFT AND TAKEN OUT OF SERVICE.
1FCV-PR21B FAILU RE IS V'DEF INVESTIGATION.
NORTH ANNA-1 0500033G 022380 AT 1912 CN FEBRUARY 23, 1980. AN ACTUATIDH OF THE EMERGENCY CORE COOLING EMERG CORE COOLING SYS + CONT 80-028/01T-0 030780 SYSTEM OCCURRED DUE TO A HIGH STEAMLINE DIFFERENTIAL PRESSURE.
THIS IS VALVES 030512 2-WEEK A REPORTABLE CCCURRENCE AS PER T.S.
6.9.1.8.I AND T.S.
6.9.5 9.B.
SINC BUTTERFLY WEST E THIS WAS AN INADVERTENT ACTUATION (EQUIPMENT PERFORMED THE INTENDED IN
~
PERSONNEL ERROR JECTION FUNCTION) AND THE TRANSIENT WAS NOT SEVERE.
THE HEALTH AND SAFE LICENSED 4 SENIOR OPERATORS TY OF THE GENERAL PUBLIC WERE NOT AFFECTED.
ROCKWELL MANUFACTURING CD.
CAUSE OF HIGH STEAMLINE DIFFERENTIAL PRESSURE 1 SUBSEQUENT SI WAS VALVE HRV-MS101C BEING OPENED WHILE NRV-MS101A
- B REMAINED CLOSED 1 STEAM FLO W EXISTED. REASON HRV-MS101C CPENED WAS TO ALLCW NON-RETURN NOISE DATA T 0 BE OBTAINEC :1 CONNECTION W/PREVIOUSLY DETECTED LOOSE PARTS INDICATION
$ IN S/G. CORRECTIVE ACTION-RETURN PLANT TO STABLE CONDITIONS BY OPER.
THREE MILE ISLAND-2 05000320 101978 PERFORMING SURVEILLANCE PROCEDURE 2301-3D1 ON 10/19/78 DETERMINED THAT L REAC COOL PRES BOUN LEAK DETEC 78-062/01T-0 110178 CO ACTION B FOR TECH SPEC 3.4.6.2 NOT INVOKED WHEN SURVEILLANCE PROCEDUR COMP 0HENT CODE NOT APPLICABLE 023080 2-WEEK E 2301-3D1 DATA OBTAINED AT 1935 ON 10/16/78 SHOWED UNIDENTIFIED LEAKAGE SUBCOMPONENT NOT APPLICABLE B+u GREATER THAN 1 GPM (2.6 GPM ACTUAL).
ALL LEAKAGE FROM RCS IS PROCESSED PERSONNEL ERROR THROUGH RADWASTE TREATMENT SYSTEM AND EVENT DID NOT AFFECT PUBLIC HEALT LICENSED 8 SENIOR OPERATORS H AND SAFETY.
ITEM NOT APPLICABLE MISINTERPRETATION OF T.S.
3.4.6.2 AND 4.4.6.2 LED TO PERFORMANCE FREQUEN CY FOR SURVEILLANCE ABOVE THAT REQUIPcD BY THE T.S.
NOT CLEAR TO PERSON NEL WHICH SET OF DATA CAME WITHIN T..
nEQUIREMENT AND WHEN TIME REQUIRE MENTS OF ACTION STATEMENT WERE APPLIC dLi UNIDENTIFIED LEAKAGE SUBSEQU ENTLY REDUCED TO ALLOWABLE.
w DAVIS-BESSE-1 05000346 092479 AT 1530 FOUND BORON INJECTION FLOWPATH TEST HAD HOT BEEN PERFORMED TO ME CHEM, VOL CONT + LIQ POISH SYS 79-094/03L-0 192279 ET SURVEILLANCE REQUIREMENT 4.1.2.2.A.
UNIT ENTERED ACTIDH STATEMENT OF COMP 0HENT CODE NOT APPLICABLE 027333 30-DAY T.S.
3.1.2.2 AT 1945 9/24/79 WHICH IS FINAL TEST DUE DATE WITH MAXIMUM SUBCOMP0HEHT NOT APPLICABLE B+W EXTTNSIDH. FLOWPATH FROM BORATED WATER STORAGE TAHK WAS OPERABLE.
(NP-
~
PERSCNNEL ERROR 33-79-109)/
LICENSED 8 SENIOR OPERATORS ITEM NOT APPLICABLE PERSONNEL ERROR.
BREAKDOWN IN COMMUNICATIONS OCCURRED BETWEEN TRAINING AND TESTING SHIFT AND OPERATING SHIFT.
TEST WAS SUCCESSFULLY PERFORMED AT 1700 HOURS ON 9/25/79.
CHANGE MADE IN SCHEDULING OF TEST TO ENSURE T RAINING AND TESTING SHIFT IS MORE AWARE OF THIS TEST.
SIGNOFF TO PERFOR M TEST HAS BEEN ADDED TO MONTHLY EQUIPMENT LOG.
CALVERT CLIFFS-2 05000318 102578 DURING LPPT, WITH THE REACTOR IN MODE 2. THE CONTAINMENT AIR LOCK INNER REACTOR CONTAIHMENT SYSTEMS 78-038/01T-1 111378 DOOR WAS OPENED AT 1600 DN 10/27/78 TO REPAIR A FAILED OPERATING MECHANI MECHANICAL FUNCTION UNITS 022740 2-WEEK SM.
CONTRARY TO THE REQUIREMENTS OF T.S. 3.6.1.3 THE INNER DOOR REMAINE LUBCOMP0HENT NOT APPLICABLE COMB D OPEH UNTIL 0930 OH 10/31/78 AT WHICH TIME REPAIRS WERE COMPLETED AND A PERSOHHEL ERROR IR LOCK AGAIN BECAME OPERABLE. CONTRARY TO REQUIREMENTS OF T.S. 3.0.4 T LICENSED & SENIOR OPERATORS HE REACTOR WAS MADE CRITICAL AT 2000 ON 10/28/78 FOLLOWING A TRIP AT 190 CBI NUCLEAR COMPANY 5 ON 10/27/78.
THE OUTER AIR LOCK DOOR REMAINED SHUT AND LOCKED.
THIS EVENT WAS INITIATED BY A BROKEH CAM FOLLOWER, MCGILL MFG. CO. PART
- CF-1-S.
THE BROKEN CAM FOLLOWER IS BEING INVESTIGATED BY MAINTENANCE A ND METALLURGICAL PERSONNEL. INNER DOOR LEFT OPEN DUE TO LACK OF COMMUNI CATION BETWEEN MAINTENANCE AND SUPERVISORY PERSONNEL. OPERATIONS PERSON HEL HAVE BEEH COUNSELED REGARDING THIS EVENT.
DAVIS-BESSE-1 05000346 042381 (HP-33-81-27) DURING A REVIEW OF SAFETY TAGGING LOG PRIOR TO ENTERING M0 CHEM, VOL CONT + LIQ POISH SYS 81-025/03X-1 061881 DE 3 A SHIFT SUPERVISOR DISCOVERED THAT MAKEUP (MU) PUMP 2 WAS STILL TA PUMPS 037013 OTHER GGED OUT FROM 4/11/81. UNIT HAD ENTERED MODE 4 WITHOUT REQUIRED 2 MU PUM CENTRIFUGAL B+W PS OP. THERE WAS NO DANGER TO PUBLIC OR STATION PERSONNEL. ACTION STATEM PERSONNEL ERROR ENT FOR T/S 3.1.2.4 REQUIRES A SHUTDOWN TO HOT STAND 3Y WITHIN 72 HOURS I LICENSED & SENIOR OPERATORS F LESS THAN 2 MU PUMPS ARE OP WHEN IN MODES 1-4 2ND MU PUMP WAS RESTORE BINGHAM PUMP C0 D WITHIN 7 HRS OF ITS REQUIRED OPERABILITY, & UNIT WAS IN HOT SHUTD0'N.
A THOROUGH REVIEW OF SAFETY TAGGING LOG 8 AN ABSOLUTE VERIFICATION OF T/
S RQMTS WERE NOT DONE PRIOR TO CHANGING MODES. UPON DISCOVERY, IMM ACTIO N WAS TAKEN, 8 2ND MU PUMP WAS RESTORED BY 0445 HRS ON 4/23/81. INCIDENT WAS REVIEWED W/ PERSONNEL INVOLVED. A MEMO WAS WRITTEN TO DPERATIONS PER SONNEL TO REINFORCE NECESSITY FOR PROPER PROCEDURAL,VERIF_ICATIONS, SURRY-1 05000280 010880 FROM 2200 TO 2300 ON JANUARY 8 THE CONDENSER COOLING WATER OUTLET TEMPER STATION SERV WATER SYS + CONT 80-001/04T-0 011880 ATURE RATE OF CHANGE EXCEEDED 3 DEG. F/HR, THE TS LIMIT.
THE RATE OF CH HEAT EXCHANGERS 030006 2-WEEK ANGE INCREASED TO 7 DEG. F/HR DURING UNIT 1 STARTUP. UNIT 1 WAS HELD AT CONDENSER WEST 50% POWER UNTIL THE TEMPERATURE RATE OF CHANGE RETURNED TO WITHIN LIMIT PERSONNEL ERROR LICENSED 4 SENIOR OPERATORS ITEM NOT APPLICABLE AT THE TIME, TWO OF FOUR CONDENSER WATER BOXES WERE NOT AVAILABLE TO DIS SIPATE HEAT AS THEY WERE BEING INSPECTED FOR LEAKS.
A SEARCH OF THE RIV ER ENVIRONMENT YIELDED NO VISIBLE DETRIMENTAL EFFECTS.
Y W
JOSEPH M.
FARLEY-1 05000348 122679 AT 2125 1C CHARGING PUMP WAS SHUT DOWN TO PERFORM MAINTENANCE. AT 0500 CHEM, VOL CONT + LIQ POISH SYS 79-068/03L-0 012580 SHIFT FOREMAN HOTED 1A AND 1B CHARGING PUMPS ELECTRICALLY AND MECHANICAL PUMPS 028084 30-DAY LY ALIGNED TO "A" TRAIN.
AT 1025 ALIGNMENT OF IB CHARGING PUMP TO "B"
T OTHER WEST RAIN, ELECTRICALLY AND MECHANICALLY, WAS COMPLETED.
T.S. 6.9.1.9.C REQU PERSOHHEL ERROR IRES REPORTING INADEQUACIES IN IMPLEMENTATION OF ADMINISTRATIVE OR PROCE LICENSED 8 SENIOR OPERATORS DURAL CONTROLS WHICH THREATEN TO CAUSE REDUCTION OF REDUNDANCY IN REACTO ITEM NOT APPLICABLE R PROTECTION OR ESF SYSTEMS.
EQUIPMENT STATUS CONTROL AND MAINTENANCE AUTHORIZATION WAS INCORRECTLY I MPLEMENTED. WHILE REMOVING IC CHARGING PUMP FOR MAINTENANCE, 1B CHARGIN G PUMP WAS INCORRECTLY LEFT ALIGHED TO "A" TRAIN, LEAVING NO CHARGING PU MP IN "B" TRAIN.
PERSONNEL REINSTRUCTED THAT WHEN EQUIPMENT REMOVED FR0 M SERVICE HD LCO RESULTS OR LCO IS IDENTIFIED.
JOSEPH M.
FARLEY-1 05000348 031781 AT 2335 ON 3/17/81, IT WAS DETERMINED THAT DIESEL GENERATOR 1-2A HAD BEE EMERG GENERATOR SYS + CONTROLS 81-009/03L-0 041381 N INOPERABLE FOR A 35 MINUTE PERIOD FOLLOWING PLANT ENTRY INTO MODE 4.
T ENGINES. INTERNAL COMBUSTION 036661 30-DAY HE D.G. WAS INCAPABLE OF AUTO-STARTING (I.E. D.G.
1-2A MODE SELECTOR SWI SUBCOMP0HENT NOT APPLICABLE WEST TCH WAS IN MODE 3).
TECH. SPEC.
3.8.1.1.
IN PART, REQUIRES D.G.
1-2A TO PERSOHHEL ERROR BE uPERABLE. UPON DISCOVERY OF THE CONDITION, lHE MODE SELECTOR SWITCH LICENSED 8 SENIOR OPERATORS WAS PLACED IN MODE 1 AND TECH. SPEC. 3.8.1.1 ACTION STATEMENT REQUIREME OTHER HTS WERE MET.
HEALTH / SAFETY OF THE PUBLIC WERE NOT AFFECTED.
CAUSED BY PERSONNEL ERROR. D/G 1-2A MODE SELECTOR SWITCH HAD BEEN PLACED IN MODE 3 AT 1855 OH 3/17/81 IN ORDER TO PREV AUTO-START WHILE MAINT BE ING PERFORMED. ALTHOUGH OPERATOR HAD BEEN INSTRUCTED TO PLACE SELECTOR S WITCH IN MODE 1 PRIOR TO ENTERING MODE 4, THIS WAS NOT DONE. METHOD FOR IMPL PROPER ADMIN CONTROLS FOR THIS TYPE IHCIDENT WILL BE DISCUSSED.
=-
CALVERT CLIFFS-2 05000318 122680 AT 1945. WHILE INVESTIGATING ABNORMAL CONTAINMENT AIR PARTICULATE AND GA AIRBCRNE RADIDACT MONITOR SYS 80-059/03L-0 012681 SEOUS READINGS. FOUND CONTAINMENT ATMOSPHERE SAMPLE ISOLATION VALVE SHUT VALVE OPERATORS 033707 30-DAY (TECHNICAL SPECIFICATION 3.4.6.1).
VALVE WAS REOPENED AND GASEQUS READ PNEUMATIC / DIAPHRAGM / CYLINDER COMB INGS RETURNED TO NORMAL; HOWEVER. APD REMAINED IHOPERABLE (REFERENCE (U-PERSONNEL ERROR
- 2) LER 80-60).
GRAB SAMPLES TAKEN BOTH PRIOR TO AND DURING EVENT.
CTMT LICENSED 8 SENIOR OPERATORS SUMP LEVEL ALARM OPERABLE. NOT REPETITIVE.
MASONEILAN INTERNATIONAL. INC.
OPERATOR INADVERTENTLY SH'JT SAMPLE ISOL ATION VALVE AFTER DRAINING CTNT S UMP AT APPROXIMATELY 2300 ON 12-25-80.
ALL LICENSED OPERATORS TO BE MAD E AWARE OF THIS EVENT.
ALSO. DURING CUR ENT OUTAGE. ONE OF THE TWO CTNT SUMP VALVE HAND SWITCHES WILL BE MADE SPRING RETURN TO CLOSE (REFERENCE FFCR 80-85).
JOSEPH M.
FARLEY-1 05000348 021679 AT 2303 CONDUCTING PL ANT S/U WI)'i REACTOR AND STEAM GENERATOR CONTROLS I MAIN STEAM SYSTEMS + CONTROLS 79-009/03L-0 031479 N MANUAL AND STEAM DUMPS IN AUTO ON PRESSURE CONTROLLER EXCESSIVE OPENIN VALVES 025432 30-DAY G OF STEAM DUMPS OCCURRED CAUSING REDUCTION IN TAVG TO 538 F.
T.S.
3.1 OTHER WEST 1.5 REQUIRES RCS TEMP. TO SE 541 F.
OR GREATER.
STEAM DUMPS TURNED OFF PERSONNEL ERROR AND TAVG RESTORED ABOVE 541 F AT 2308.
ACTION STATEMENT REQUIREMENTS OF LICENSED 8 SENIOR OPERATORS T.S.
3.1.1.5 MET.
HEALTH / SAFETY OF PUBLIC NOT AFFECTED.
FISHER CONTROLS CO.
TAVG DECREASE OCCURRED BECAUSE EXCESSIVE VALVE OPENING RESULTED FROM INC ORRECT SETPOINT ON STEAM DUMP PRESSURE CONTROLLER. OPERATIONS PERSONNEL WILL BE INSTRUCTED REGARDING POTENTIAL FOR RE;CCURRENCE WITH PRESSURE C ONTROLLER SETPOINTS LESS THAN 1005 PSIG.
DESIGN CHANGE REQUESTED TO PRE VENT CONTROLLER WINDUP.
Y O
CRYSTAL RIVER-3 05000302 121480 DURING TROUBLESHOOTING OPERATIONS WHILE ISOLATING A LEAK IN FIRE SERVICE FIRE PROTECTION SYS + CONT 80-051/01T-0 122680 SYS., FSV-107 WAS SHUT, ISOLATING FIRE SERV. FROM AUX. BLDG. DELUGE. SP VALVES 033356 2-WEEK RINKLER 8 HOSE STATIONS CONTRARY TO T/S 3.7.11.2 8 3.7.11.4.
VALVE HAD GATE B+W BEEN SHUT AT 2311 ON 12/14/80 BUT NO BACKUP FIRE SUPPRESSION SYS. WAS AV
~
PERSONNEL ERROR AILABLE OR ESTABLISHED FOR AFFECTED AREA.
NO EFFECT UPON HEALTH OR SAFE LICENSED 4 SENIOR OPERATORS TY OF GEN. PUBL.
FIRST OCCURRENCE FOR FSV-107, FIRST EVENT REP. CONTRAR M AND H GATE VALVES Y TO T/S 3.7.11.4 3 SECOND EVENT REP. CONTRARY TO T/S 3.7.11.2.
CAUSE OF EVENT IS ATTRIBUTED TO FAILURE TG RECOGNIZE DESIGN FUNCTION OF FSV-107.
WHEN SYS. WAS DISCOV. TO BE ISOLATED, FSV-107 WAS OPENED 8 OPE RABILITY WAS RESTORED AT 0105.
CAUSE OF LEAK WAS DETERMINED TO BE A PAR TIALLY OPEN HYDRANT.
IT WAS CLOSED.
PERS. INVOLVED HAVE CRITIQUED EVEN T 8 WILL PRESENT LE! SONS LEARNED TRAINING TO ALL OPER. SHIFTS BY 1/15/81 PERSONNEL ERROR 05000317 021778 (78-13/3L)
DURING REACTOR COOLANT SYSTEM (RCS) C00LDOWN ON JANUARY 23, CAUSE SUBCCDE NOT PROVIDED 78 03L 031078 1978, THE 100 DEGREE PER HOUR MAXIMUM C00LDORN RATE SPECIFIED DY TECHNIC CALVERT CLIFFS-1 020638 30-DAY AL SPECIFICATION 3.4.9.1 WAS EXCEEDED BY 3 DEGREE F.
AT 2330 RCS TEMPEe REACTOR VES. + APPURTENANCES COMB ATURE WAS 473 DEGREE F.
BY 0030 ON JANUARY 24, 1978, RCS TEMPERATURE HA COMPONENT CODE NOT APPLICALLE D DECREASED TO 370 DEGREE F.
SUBCOMPONENT NOT APPLICABLE ITEM NOT APPLICABLE THIS EVENT WAS CAUSED BY THE OPERATORS FAILURE TO MAINTAIN THE RCS COOLD 044 RATE WITHIN T;HE PRESCRIBED LIMITS.
SEQUDYAH-1 05000327 030330 DURING INITIAL FUEL LOADING, THE LICENSEE FAILED TO ESTIMATE FLOWRATE FR GAS RADICACT WSTE MANAGMNT SYS 80-001/01T-0 031430 OM THE SERVICE BUILDING EXHAUST ONCE PER FOUR HOURS FOR A PERIOD OF MARC OTHER CCMPCHENTS 030S17 2-WEEK H
1, 1950, AT 0000 CST 70 MARCH 3, 1980, AT 1000 CST (T.S. LCO 3.3.3.10)
SUCCCMPONENT NOT APPLICABLE WEST THERE WAS NO DANGER TC THE HEALTH AND SAFETY OF THE PUBLIC. PREVIOUS PERSONNEL ERROR
. OCCURRENCES - HONE.
LICENSED 3 SENIOR OPERATORS ITEM NOT APPLICABLE MISUNDERSTANDING OF TECHNICAL SPECIFICATION REQUIREMENTS. ESTIMATES OF FLOWRATES WERE IMMEDIATELY CHANGED FROM EIGHT HOUR FREQUENCY TO FOUR HOU R FREQUENCY. PLANT INSTRUCTIONS REVISED TO SHOW CORRECT FREQUENCY.
Y M
APPENDIX B STATISTICAL METHODS
I APPENDIX 8.1 THE FACTORS UPON WHICH THE CONDITIONAL VALIDITY OF THE STATISTICAL RESULTS REST This addrasses part 1(b) of the central purpose.
Even if the two statistical assumptions of this section are valid, the validity of the results still depends upon two other factors.
One is the validity of aggregating the data-of all nuclear power plants, drawing conclusions based on this aggregate data, and then assuming that these conclusions hold for each plant individually.
This would require that all the power plants are well represented in the LERs.
Secondly, the LER data, and in particular the time identified errors, must be representative of the true error occurrence time distribution.
In statistical terms, this is the requirement that the LERs provide a random sample of error times.
This would not hold, for example, if the detection and recording rate differed substantially amongst errors that occurred at different times.
In particular, if early morning errors were twice as likely to be recorded in the LERs as were late afternoon errors, then representativeness would be violated.
APPENDIX B.2 TERMIN0 LOGY AND NOTATION This terminology and notation will be used frequently throughout the technical appendices, often without being redefined.
Therefore, it is recommended that they be studied first, and then reviewed when necessary.
An equality of definition a
Approximately equal As The random variable A is distributed as...
sE Not E; logical negation Implies
+
If and only if
++
Set membership dim H Dimension of the subspace H x!
x(x-1)(x-1) 3 2
- "x factorial" m
Ra aa a,
- the product.
j y2 j=1
ild Independent, identically distributed.
N(u, v)
A normal random variable with mean u and variance v.
P(E,
- , E,)
Pr {E, and E and and E,}, for events E,
, E,.
1 y
2 1
Pr Probability under the hypothesis H.
g D
Converges in distribution to
+
"( )A The alternative hypothesis to H( )
X/Y = y The random variable X given that Y=y The Standard Normal cumulative distribution function Max P (Z)
The maximum of the function P evaluated at 1 O
O OcH over all points e in the set H.
L(j)
The length in hours of cell j of a given cell configuration.
C.C.
Cell configuration In some instances, the distinction between a random variable and its realiza-tions, (the actual observations), will be made by using upper and lower case letters, respectively.
However, the meaning should be clear from the context.
Everything below, (to which the above comments apply), is a random variable.
B.2-2
The number of weeks in the data subset being m
analyzed.
Clearly this will vary between data subsets of different lengths.
Hence, m will be the sample size for the likelihood ratio test.
t The total number of errors in the data subset.
It is to be implicitly understood that t depends both upon the data subset and the cell configura-tion being analyzed.
It depends upon the cell configuration because, as discussed in Appen-dix A.1, an error time could be omitted from the analysis of a given C.C. if it does not fall within a gingle cell.
t will be the sample size for the X -G0FT and the Conditional Binomial Test.
Variables The following assumes a cell configuration of K > 1 cells.
For i = 1, 2,..., 7 (Except wnere their use as dummy variables is j = 1, 2,..., K, clear), the following are defined:
Y The number of errors falling into cell j on day i gj 7
X.s I
Y..
The total number of errors occurring within J
j=1 1]
cell j over the days of the week.
X,)
The iid sample of size m of X) generated by a data X) a (Xji, j
subset of m weeks.
X s (X,..., X )
The vector of samples X).
3 k
m Observe that IX is the total number of errors falling into cell j 1=1 )j for an entire data subset of m weeks.
m m
Xa(IXjj,...,IXki)
The vector of cell totals for an entire data 1=1 1=1 subset of m weeks.
D The number of errors occurring on day i.
j B.2-3
0 s (D 9
93,..., D,)
The iid sample of size m generated by a data j
subset of m weeks.
D a (ps,.... D )
The vector of samples D.
7 9
m Observe that I D..
is the total number of errors occurring on day i j=1 '3 for an entire data subset of m weeks.
m m
-Ds(ID I D.)
The vector of daily totals for an entire da'ta II,...,
3 j=1 j=1 subset of m weeks.
B.2-4
APPENDIX B.3 THE POISSON RANDOM VARIABLE A discrete random variable X is said to be Poisson with parameter A, (X
- Poisson (A)), if A > 0 and the probability mass function of X is given by:
x
-A A *
, x = 0, 1, 2,...
-P(x) = Prob {X = x} =
0
, otherwise To indicate that P(x) depends on the parameter A, we add A as a subscript, and write P (x), instead of P(x).
g It is well known that if X
- Poisson (A), then:
E(X) = VAR (X) = A i
The Reproductive Property Of The Poisson Random Variable:
If X,..., X
- Poisson (A ),...
Poisson (A ) independently, issob(A +... + A )
I 1 +..."+ X Then X s
3 n
n
l APPENDIX B.4 SIGNIFICANCE LEVELS Recall that a statistical test is called "significant" if its p-value is less than some previously chosen significance level.
Recall also that the rejec-tion of H and H require significance in either two or three tests.
Within the standNrd ranhe of generally chosen significance levels, (0.001 to 0.05),
or choices of 0.01 and 0.05 are arbitrary, as is ultimately the choice of the significance level for any testing situation.
To this extent, or acceptance or rejection of H and H are also arbitrary.
u y
Hence the additional advantage of the p-values, (or bounds for them), listed in this report:
they aid the decision maker with an objective, quantitative basis for forming his own conclusions as to the acceptability of H and H.
7 They also help suggest alternative hypotheses for further testing.u B.4-1
APPENDIX B.5 THE STATISTICAL TESTS The x (Chi-square) Goodness-Of-Fit Test.
This test assumes that the number 2
of errors per cell is described by a multinomial distribution.
(See Appen-dix B.7).
For a given cell configuration and a given number of total errors, the hypothesis to be tested will dictate the expected number of errors per The x -G0FT will measure the discrepancy between the observed and 2
cell.
expected error counts in such a way as to indicate the approximate likelihood of observing a discrepancy equally or more extreme, under the hypothesis being tested.
Once again, this likelihood is the p-value.
A widely accepted requirement for accuracy is that the expected cell count be greater than or equal to five in each cell.
For our largest data subset this is the case, and most of the expected cell counts are well above five.
For large n, (the total number of errors), the X approximation is good, and it 2
improves as n approaches infinity.
The Conditio.nal Binomial and Likelihood Ratio Tests.
In general, a wide variety of parametric assumptions can lead to these tests.
In this study, for testing H, both of them are based upon the assumption that, for any given cell configu-r3 tion, there are independent Poisson random variables describing the number of errors occurring within the cells of that configuration.
(See Appendix B.3 for a discussion of the Poisson Random Variable).
This amounts to a separate assump-tion for each cell configuration.
These assumptions are not necessarily incon-sistent.
The assumption is the same for testing H;, except, again, that the cells are the days of the week.
It was shown in Appendix B.6 that testing H and H, under the Poisson assumption reduces to testing for the proportionaliEy of Poisson means.
The following tests each do this in a different way.
The Conditional Binomial Tesh ) This applies only to case of a cell configura-tion with two cells.
Lehman' has shown that this is a Uniformly Most Pc,werful (UMP) unbiased test.
The Likelihood Ratio Test, (LRT).
This test can be usefully applied in most parametric testing situations.
In general, the LRT is not an UMP test.
How-ever, it is quite intuitively appealing and in our testing situation has useful asymptotic properties.
The General Test sampTesoTsizem,kuchthkk,ea...,ZchinbTv,idualsNp,leissT)d.
...,Z
..., Z be K independent Let Z = (Z,..., Z ) = (Z I
K i Suppose that Z_ has known joint density or likelihood function p, where the subscript 0 is g
a parameter indicating that p depends on 0.
AssumG that 0 is an k-dimensional vector.
Let H be the set of all possible values of 0, and let H be a proper subset of H.
Consider the following hypothesis testing situati8n:
H : 0 H
vs.
H 0
H-H-g g
3 g
[']E. L. Lehman, Testing Statistical Hypotheses, John Wiley and Sons, Inc.
New York, 1959.
B.5-1
Define the likelihood ratio (LR):
Max 0 H 0 (E)
P LR = Max 0 H o (E)
P Intuitively, we would like to reject H when LR is too small.
This is indeed whatth{2}RTdoes.
The question of hoO small is answered by applying Wilk's Theorem Under H and given certain regularity conditions on P, -2 log (LR) converges o
i in distfibution, (as m + m), to a X2 random variable with d degrees of freedom, where d = dim H - dim Hg Next observe that small values of LR correspond to large values of -2 log (LR).
So we reject H at the a-significance level exactly when:
l g
-2logLR>Xj,1_,
whereX$.1-a istheupper(1-a)percentilepointofaXjdistribution.
The Specific Test We now derive the specific LR for testing h will be a specific case of b.
(See Appendix B.6).
The LR p
for testing H' is.
In terms of the notation justintroduchMwe can rewrite H as:
" PPM (A'***'A)
"o vs.
H (A ' **** A )
" ~ "o l
k PPMA l
k H = {(a ' ' "k) "i > 0 i = 1,..., K}
l "o * ("lg* ' "k "1 > 0, a$ = c;a, cj > 0, y
Recall that the relevant data is:
X = (X,..., f ) * (X11,..., X,,..., Xki' ' Km) 1 k
y Let p(x a ) be the likelihood function of X;) evaluated at xjj jj j
p(x; a ) be the joint likelihood function of X_; evaluated at x;...
j p(x (a,..., a )) be the joint likelihood function of X evaluated at y
k x...,
WHEN the Poisson means A,..., A ssume the respective values ay,...,a' y
K K
b S.S. Wilkes, Mathematical Statistics, pp. 394-427, John Wiley and Sons, Inc.
New York, 1962.
B.5-2
i
- tecall that under the Poisson assumption:
for i = 1, 2,..., K X.
- Poisson (A )
IJ j
j = 1, 2,..., m
- ij '"i a;
e (1) so we have p(x a)
=
jj j
xq!
By independence, the joint likelihood function for X; is the product of the likelihoods in (1).
m 1 x '..J j=1
-ma j a.
e 1
P (x; a )
(2)
=
j II (xjj !)
j=1 Since the components X,..., X are all independent, the joint likelihood functionforXisthehroductofthelikelihoodsin(2):
m m
( I *1j) -ma
( I *kj) -ma J=1 j=1 l
k a
e a
y K
P (X; a,..., o )
y K
m m
Il (xyj !)
Il (xkj !)
,J:1 j=1 The specific LR statistic is now:
Max p (x (ay,..., a ))
k (a ' '**' "K)
"o l
LR =
Max p (x (a,..., a ))
y k
(a,..., o )
y K
8.5-3
We first derive the numerator, N:
i m
m k
(I X yj)
(I X Kj)
~* 1 "i J=1 j=1 i=1 1
"k N=
Max m
m
("l ' * * * * "k)
"o H
(X ) !)
(X
)
H y
Kj J=1 J:1 K
-m I a.
t t
t 1
y 2
k N=c Max a y a,
g o
e (a,..., a )
"o y
k where c = c (X,..., X )
- y k
m m
,H (Xyj ! )
H (XKj !)
J:1 J=1 m
is a constant in the a 's, and t. = I X..
g I
13 j=1 Recognizing the restrictions imposed on a2'***'"K by H yields:
g t
t t
-m (1 + c2 *
- CK) "1 y [c2 "l] 2
- [cK "13 K
N=c-Max a
y y>0 a
-m (1 + C2+
+C)"1 K
T
=c Max a
e (3) y y>0 a
nd c' = c - (c c
)
where T=t1+
+tK 2
K B.5-4
1 l
1 The function to the right of the " Max" is positive and continuous in a, so, l
if it has a maximum, it occurs exactly where its log is maximized.
So we calculate:
Max T log a - m (1 + c2 +
+ c ) "1 l
K 1>0 y
Since the function to be maximized here is a sum of concave functions, it is also concave.
Hence if it has a turning point &
then S will be the unique y
Sincethefunctioniscontinuous,wecansda,rchfortheturningpointby max.
setting the derivative to zero:
T log a - m (1 + c2 "
- CK) "I y
da 7
- m (1 + C2+
+C) 0
=
K 0
1 m (1 + c2+
+c)
K Substituting S for a in (3) gives:
y y
T
-m (1 + c2+
+c)
I K
m (1 + c2+
- +c)
T K
N -
c' [ m (1 + c2+
+c)3 K
T N = c' [ 1 ]T T
-T
[1+c
] *
(4)'
2+
+cK Repeating the initial steps of this reasoning for the denominator, D, gives:
K
-m I t
t i=1 a; y
g
=
c-Max a
o e
D y
g (a '
l
' "K)
B.5-5
by H gives:
Recognizing the restrictions imposed on a,
, og y
K
-m 2 a.I t
t i=1 1
K D =
c-Max a
a e
(5) y K
a,
,og>0 y
The function to the right of the " Max" is positive and continuous in a,
,o and so, if it has a maximun, it occurs exactly where the log is mdximized.g,So we calculate:
K 1 "i I I
Max
{t I 9"I+
+tK 9 "K 1
i=1 a,,aK>0 y
The function to be maximized here is also the sum of concave functions, and concavity), if it has a turning point (&,y,
& )f.
Hence, (again bySince this function ks cont
,y,
, S ) "III hence it too is concave in the variables a
,a then (&
K be the unique max.
,a we can searchfortheturningpointbysettingthepartialderivakivestoke,ro:
For j = 1, 2,
, K:'
K t.
=
-m 0
[t log ay+
+tK I 9 "K ~ ".1 "i ]
a.
y i=1 J
J t.l 5.
=
+
3 m
Substituting S for a in (5) gives:
j j
'".1 t
t t
i=1 (k)K t
1 1
c(-)
e D
=
8.5-6
K K
It
-m I t g
g c(f)i=1 t
t i=1 1
0 =
t
... t 1
K K
By the definitions of t,
, t y
g and T, T =,I t.
So the above becomes:
g i=1 c(f) 1 D
t t
e (6)
=
y K
Combining (4) and (6) yields:
T T
1 T
c,
- T m
1+C LR = 0 =
2+
+CK D
T t
t y
y K
-T L
t e
c g
3 g
Recalling the definition of c' (right after (3)), and cancelling now gives:
t t
2 K
cc c
T 2
K T
LR =
t t
t 1+c +
+c I
ct t
t 1
2 K
t (f)1([c 2
c K
T
)
-(f)
T LR
=
(7) y, 2
K 2,,,,, K Note that LR is independent of the sample size m.
B.5-7
Concerning the X cutoff point, the proper degrees of freedom is d = dim H -
2 dim H
=K-1.
g For application to Hppg:
m
, and c
=c
=
=c
=1 1 D take K=7, tg = J:1 jj 2
3 7
For our situation, the regularity conditions of Wilk's Theorem involve the validity of the interchange of multiple summations and the partial derviatives These condi-of the joint Poisson Likelihood function p (Xtions hold, and so Wilk's Theorem indee (a
providing an extremely useful statistic for testing H and H PPM PPM' As described in Appendix B.6, m is large, and hence the asymptotic approxima-tion of -2 log LR will be very good.
Example of the calculation of -2 log LR:
We calculate -2 log LR for cell configuration #6, based on data set C, (see Table B.5-1).
This C.C. has 3 cells, so k=3.
We are also given:
c2 = 3/8, ca = 13/8, and m
t s I xi. = 34 i
J j=1 m
t2 a I x2. = 15 T=ti+t2 + ta = 82 J
j=1 m
ta s 1 x. = 33 a3 j=1 2
So, -2 log LR = -2 log 1+c +ca 34 15 33 82 1
3/8 13/8 82
= -2 log R
15 33 1+3/8+13/8
~
(3)15 (13)33 (82)82
_ -2 1 9 (34)34 (15)l3 (33)33 (8)*6 (3)6#
B.5-8
33 log (13 + 82 log 82
= -2
-34 log (34) -15 log (15) -33 log (33)
-48 log (8) -67 log (3)
= 6.655 (Recall:
log = log e As described in Appendix 8.6 that the 84 time identified errors that were recorded over the n weeks are assumed to be representative of the true error occurrence time distribution.
It is reassuring to note that if the error times of these 599 were in the same proportion as the 84, then the LR statistic
(-2 log LR), would have an even lower p-value.
To show this we prove the following general result:
m
,I X
for i = 1, 2,
, K, as before, let LR (t,
, t )
M: We let tg = J=1 gj y
K be as in (7), and let the number of errors in the' sample with more errors the number of errors in the sample with fewer errors, y Then:
-2 log LR (at,
, at.g) = a P 2 log LR (t,
, t )3 y
y K
)
Proof:
at at "K
aT l c 2
c LR (a t,
, at ) * ( a
)
( at )
( a
)
(1+c
)
y K
1 2
K 2+
+cK at at at "l
"2 y
2 K
aT
-([c
([1
=(f)
Ch)
Ch)
(a)
)
)
2 at K aT
- ([c )
(1c+
4
)
c K
2 K
a (ty+
+t) aT 1 c 2
c K
T t
t t
k (h)
(a)
(
) (
)
-(
) (l+C
=
2 + + CK B.5-9
=(f) oT
[ LR (t,
, t ) 3 a
l K
a
= [ LR (t,
, t )3 y
K The result now follows easily by taking logs and multiplying by -2.
Example:
For CC #7, the vector of sample cell counts was (44, 15, 21) for a total of 80 time identified errors.
The Likelihood Ratio statistic, (using i
c
=.3,c 1.1), was:
2 3
-2 log LR (44, 15, 21) 13.19 (p <.003)
=
599 = 7.49.
Here,a=8gferrors:
If we had observed a proportional cell count in the entire set a (44, 15, 21) 7.49 (44, 15, 21)
(330, 112, 157), then the
=
=
Likelihood Ratio statistic would be:
X = a [13.19] = 7.49 [13.19] = 98.79 The p-value for this is astronomically small.
B.5-10
Table B.5-1 Results of the 2-sided likelihood ratio test for cell configurations 1 - 7 157, for 78,79,&80 M = 122, for 79,80,&81 175, for 78,79,80,&81 l
Data set A=78,79,&80 m
m m
m
-2 log LR B=79,80,&81 IX IX IX X
1k 2k 3k 4k Cell configurations C=78,79,80,&81 k=1 k=1 k=1 k=1 LR p-value
%x3
- 1 c
=c
- 1
.0106 9.087 p<.02 2
3 A
32 28 13 (0000-0800), 3
.1352 4.002
.25<p<.3 1
B 17 25 13 (0800-1600), 32 C
34 32 16
.0203 7.793 p<.03
- p (1600-2400), 33
%XI
- 2 c = 17/7
.69
.742
.3<p<.5 2
A 25 49 (0900-1600), 31 B
20 33
.4080 1.793
.15<p<.2
.7178
.663 4<p<.5 (0900-1600)c C
27 54
%X}
c
=1 5
2 A
15 58
.9982
.004
.95<p<.98 (1100-1600), 3
.4352 1.664
.15<p<.2 1
B 15 38
.999
.001
.95<p<.98 (1100-1600)c,x C
17 64 2
%XI
- 4 c = 21/3 = 7
.456 1.570
.20<p<.25 2
A 13 61 (1300-1600), 31 B
14 41
.034 6.770 p<.01 C
15 68
.348 2.110
.10<p<.20 (1300-1600)c'52
Table B.5-1 (continued)
Data set A=78,79,&80 m
m m
m
-2 log LR B=79,80,&81 IX X
X X
Cell configurations C=78,79,80,&81 k=1 k=1"k k=1 k=1 LR p-value 1k 3k 4k
%XS
- 5 c
=c
=c
=1 i
2 3
4 A
25 20 15 11
.0424 6.322
.05<p<.10 (0000-0600), 31 B
13 13 16 11
.6218
.950
.80<p<.85 (0600-1200), 32 C
27 21 18 13
.0725 5.249
.15<p<.18 (1200-1800), x m
3 (1600-2400), x_4
%X!
l l
- 6 c = 3/8; c = 13/8 2
3 A
32 13 28
.0247 7.403
.01<p<.04 l
(0000-0800), El B
17 14 24
.0307 6.969
.01<p<.04 (1300-1600), 32 C
34 15 33
.0359 6.655
.01<p<.04 (0800-1300)U(1600-2400), X 3
%X3
- 7 c = 3/10; c 1/1 2
3 A
42 13 18
.0009 13.990 p<.001 (0000-1000), El B
23 14 16
.0088 9.459 p<.01 (1300-1600), 32 C
44 15 21
.0014 13.187 p<.003
((1000-1300)U(1600-2400)), x3
l APPENDIX P 6 H AND H UNDER THE POISSON ASSUMPTION u
y Recall that two approaches are are being combined to test H and H and thatoneofthemistomakeaPoissonassumptioninadditio8tothb, basic assumptions (i) and (ii).
The Poisson assumption for H is that the number of errors falling into the cells of any cell configuration are independent Poisson random variables.
For H, it is that the error per day counts are independentPoissonrandomvariables,foreachdayoftheweek.
We have assumed that the distribution of the number of errors per day depends upon the day of the week.
Thus, for each cell configuration, the number of errors per cell will also depend upon the day of the week, as well as upon the particular cell itself.
To formally accommodate this we write:
H is:
Y..
- Poisson (Ajj)
The Poisson assumption for H is:
D s Poisson (a )
j (j = 1, 2,..., K and i = 1, 2,..., 7) jj > 0 and ay > 0 are the unknown means.
where A DEVELOPMENT OF THE HYPOTHESES OF PROPORTIONAL POISSON MEANS:
7 1 A (1)
For j = 1, 2,
, K, let A 3 = 1=1 jj.
7 Recall that X) =,1 Y ).
g i=1 So, by the reproductive property of the Poisson random variable we have:
X)
- Poisson (A)), (1 1 j 1 K) and these X s are independent since the Y s are independent in i and j.
gj The same ho ds for each sample of X; iid Poisson (A ), (1 1 j $ m)
X s
jj j
If H is true, it follows that, for each day i = 1, 2,
.., 7 of the week, the meany of the y s will be in the same proportion as the lengths of their jj B.6-1
corresponding cells.
This proportionality of Poisson means can be represented by the hypothesis HOPPM, (Daily Proportional Poisson Means):
"DPPM A
=C i2 2 il 3
3 1
For each day i = 1, 2,..., 7 of :
the week
^ik K ^il and A
>0 jj where C)
, (j = 2, 3,..., k)
=
Now, since each equation in H is true for each i = 1, 2,..., 7, we can sumoveriineachequationa0pYU110ws:
7 7
7 1 A (j = 2, 3,..., K)
C) i=1 jj C) All I
=
1 A
=
gy, i=1 i=1 4
(Using (1))
C) A, (j = 2, 3,... K)
A)
=
y and this gives the hypothesis Hppg, (Proportional Poisson Means):
" PPM A
- A "S'
NPPMA 2
2 l A
_C A
2 - 2 l A3=C A 3 1 not A =CK ^1 and Ay>0 and A,..., AK>0 y
Once again, the C.'s of H and H depend on the particular cell configurationoffocus.
N511elb$Msimpler considerations of H, and the Poisson assumption for it lead to the hypothesis H'PM, (Proportio6a1 Poisson Means over the Week):
B.6-2
Hppg 2 = l vs. H 'PMA a
p a, = a3 3
"1 a
not o
=a g
3 u, = a3 and og>0 and ay,...,u7>0 By all of the above remarks it follows almost immediately that:
UNDER THE POISSON ASSUMPTION:
"u "DPPM
" PPM H
and H
+
g PPM and by logical contrapositive:
' " PPM
' "u (x) s H'PM
'H I
For testing H, we will work with H X,
..,X will be the relevant random variabYes for this, and X, PPM.
Iwillbekherelevantsample.
In X
lightof(*)andtheremarksonhage..,Moutrejectioncriteria,wemust 15 2
reject H if testing H leads to a p-value less than.01 in the X G0FT, theCondYtionalBinomi$$" test,andtheLikelihoodRatiotestforthelargest subset and for at least one cell configuration.
On the other hand, we are not forced to accept H even when H le insignificantp-values (those>.01)forall7cellconfigura$$Uns.adsto (Let us call this Case B).
H is true for all 7 cell configurations" can in fact be false even when HThiscouldoccur,forexample,$k"therewereeq cell means within each of the 7 cell configurations, but non-uniformity within the cells of one of them.
Thus, in the event of Case B, above, we will say that although H may be false, we do not have sufficient evidence to reject it.
u FortestingH,wewillworkwithHh The reasoning and procedure are analogous,wi[hthespecificdifferebM.ces that the significance level is.05,
.., 0 are the relevant random variables, that D
..,D not.01, that D,
7 is the relevant sample, dnd that there is only one cell configuratidn,:
the one consisting of the days of the week.
So by using a Poisson assumption, we have reduced testing H and H to u
g testing H ndHhPM.
There are good tests for the latter two.
PPM Note that no UMP (Uniformly Most Powerful) test exists for testing H H'$$.
For any particular alternative hypothesis, there exists a bias $$ or p
test tb is more powerful against that particular alternative than any candidate B.6-3
for a UMP.
This biased test cannot be a UMP either, by consideration of another appropriate alternative hypothesis.
SAMPLE SIZE UNDER THE POISSON ASSUMPTION There is a lag in the LERs between the latest recording date and the availabil-ity date.
Hence we know that although the LERs for 1981 were available up until 6/30, the latest day recorded was surely earlier in June or May. We estimated this date to be May 6, since this was the last day for which an error was recorded in the 1981 LERs as of 6/30.
Thus the total number of days in our data subsets are:
78 80 :
365 + 365 + 366 = 1096 79 81 :
365 + 366 + 126 = 857 78 80 - 81 :
365 + 365 + 366 + 126 = 1222 By recalling the random variables that we are sampling:
X.
the number of errors occurring cell j over the days of the week
=
J (15 j $ k)
D
=
j the number of errors occurring on the ith day of the week, we see that we have one observation of X. and D. for every week of observation.
Since every 7 days corresponds to a sample poinl, the 3 data subsets have the following sample sizes:
10 6 :- 157 weeks 78 80 :
79 81
= 122 weeks 78 80 - 81.
1222 = 175 weeks l
So m = 157, 122, and 175 for the 3 data subsets.
In Appendix B.5, we see that these are our sample sizes for the Likelihood Ratio test.
It is important to avoid confusing m, the sample size under the Poisson assumption, with t, the number of errors in the sample.
Of the 599 errors that were recorded, 84 were time identified.
These are treated as representative:
as though the propor-tion of cell counts given by these 84 is representative of that given by the total 599.
B.6-4
t APPENDIX C DETAILS OF STATISTICAL RESULTS FOR 7 CELL CONFIGURATIONS
Table C-1 Cell configuration #1
[(0000-0800), (0800-1600), (1600-2400)]
Results Data X Goodness of Conditional binomial LRT -
set Data fit test test (2-sided)
(2 sided)
Cell 0-8 8-16 16-24 X] = 8.242 X] = 9.087 78,79,&80 Observed 32 28 13 Expected 24.33 24.33 24.33
.01<p<.02
.01<p<.02 TOTAL = 73 Cell 0-8 8-16 16-24 X] = 4.073 X] = 3.059 79,80,&81 Observed 17 25 13 Expected 18.33 18.33 18.33
.10<p<.20
.25<p<.30 TOTAL = 55 Cell 0-8 8-16 16-24 X[ = 7.12-X] = 7.793 78,79,80,&81 Observed 34 32 16 Expected 27.33 27.33 27.33
.01<p<.03
.01<p<.03 TOTAL = 82 C-2
i Table C-2 Cell configuration #2
[(0900-1600), (0900-1600)c)
/
Results 2
Data x Goodness of Conditional binomial LRT -
set Data fit test test (2-sided)
(2 sided)
Cell 9-16 (9-16)c x} =.84 x? =.742 78,79,&80 Observed 25 49 Expected 21.58 52.42
.3<p<.4 p=.456
.3<p<.5 TOTAL = 74 Cell 9-16 (9-16)c x = 1.88 x} = 1.793 79,80,&81 Observed 20 33 Expected 15.46 37.54
.15<p<.2 p=.222
.15<p<.2 TOTAL = 53 Cell 9-16 (9-16)c x} =.68 xi =.6632 78,79,80,&81 Observed 27 54 Expected 23.63 57.37
.4<p<.5 p=.484 4<p<.5 TOTAL = 81 C-3
l Table C-3 Cell configuration #3
[(1100-1600),(1100-1600)c3 l
Results Data X Goodness of Conditional binomial LRT -
set Data fit test test (2-sided)
(2-sided)
Cell (11-16) (11-16)c X{=.004 xi=.0036 78,79,&80 Observed 15 58 Expected 15.21 57.79
.95<p p=1
.95<p<.98 TOTAL = 73 l
Cell (11-16) (11-16)c X} = 1.79 X}=1.6639 79,80,&81 Observed 15 38 Expected 11.04 41.96
.15<p<.20 p=.242
.15<p<.2 TOTAL = 53 Cell (11-16) (11-16)c X{=.001 X{=.0012 78,79,80,&81 Observed 17 64 Expected 16.88 64.12
.98<p p=1
.95<p<.98 TOTAL = 81 I
C-4
Table C-4 Cell configuration #4
[(1300-1600), (1300-1600)c)
Results 2
Data x Goodness of Conditional binomial LRT -
set Data fit test test (2-sided)
(2-sided)
Cell (13-16) (13-16)c x} = 1.74 x}=1.57 78,79,&79 Observed 13 61 Expected 9.25 64.75
.15<p<.20 p=.26
.20<p<.25 TOTAL = 74 Cell (13-16) (13-16)c x}=8.42 x} = 6.77 79,80,&81 Observed 14 41 Expected 6.88 48.12 p<.01 p=.007 p<.01 TOTAL = 55 Cell (13-16) (13-16)c x} = 2.35 x} = 2.11 78,79,80,&81 Observed 15 68 Expected 10.38 72.62
.11< p<.16 p=.19
.10<p<.20 TOTAL = 83 C-5
Table C-5 Cell configuration #5
[(0000-0600),(0600-1200),(1200-1800),(1800-2400)]
l Results Conditional 2
Data x goodness binomial test LRT -
set Data of fit test (2-sided)
(2-sided)
Cell 0-6 6-12 12-18 18-24 x3=6.55 x3=6.3216 78,79,&80 Observed 25 20 15 11 Expected 17.75 17.75 17.75 17.75
.05<p<.10
.05<p<.10 TOTAL = 71 Cell 0-6 6-12 12-18 18-24 x3=.96 x3=.9502 79,80,&81 Observed 13 13 16 11 Expected 13.25 13.25 13.25 13.25
.80<p<.85
.80<p<.85 TOTAL = 53 Cell 0-6 6-12 12-18 18-24 x3 = 5.20 x3=5.2489 78,79,80,&81 Observed 27 21 18 13 Expected 19.75 19.75 19.75 19.75
.10<p<.20
.15<p<.18 TOTAL = 79 C-6
Table C-6 Cell configuration #6
[(0000-0800),(1300-1600)0]
((0800-1300)U(1600-2400))
Results Conditional Data x Goodness binomial test LRT -
set Data of fit test (2-sided)
(2-sided)
Cell 0-8 13-16 c x3 = 7.42 x3=7.40 78,79,&80 Observed 32 13 28 Expected 24.33 9.13 39.54
.01<p<.025
. 01< p<. 04 TOTAL = 73 Cell 0-8 13-16 c x3 = 8.59 x3 = 6.97 79,80,&81 Observed 17 14 24 Expected 18.33 6.88 29.79
.01<p<.02
.01<p<.04 TOTAL = 55 Cell 0-8 13-16 c x3 = 6.77 x3=6.66 78,79,80,&81 Observed 34 15 33 Expected 27.33 10.25 44.42
.01<p<.04
. 01< p<. 04 TOTAL = 82 C-7
Table C-7 Cell configuration #7
[(0000-1000), (1300-1600)g]
((1000-1300)U(1600-2400))
Results Conditional Data X Goodness binomial test LRT -
set Data of fit test (2-sided)
(2-sided)
Cell 0-8 13-16 c X{ = 13.19 X3 = 13.99 78,79,&80 Observed 42 13 18 Expected 30.42 9.12 33.46 p<.005 p<.001 TOTAL = 73 Cell 0-10 13-16 c X] = 11.06 X3 = 9.46 79,80,&81 Observed 23 14 16 Expected 22.08 6.63 24.29 p<.007 p<.01 TOTAL = 53 Cell 0-10 13-16 c X] = 12.61 X3 = 13.19 78,79,80,&81 Observed 44 15 21 Expected 33.33 10 36.67 p<.005 p<.003 TOTAL = 80 C-8
STATISTICAL BIBLIOGRAPHY j
1)
Lehman, E. L. Testing Statistical Hypotheses New York:
John Wiley & Sons, Inc., 1959 2)
Wilks, S. S. Mathematical Statistics New York:
John Wiley & Sons, Inc.,
1962.
Bi-1
REFERENCE BIBLIOGRAPHY The following publications and articles have been identified in the course of this research as sources containing further information on the effects of shift work and overtime on the human operator.
These articles have not been reviewed and their listing here does not signify either relevance, accept-ability or availability.
This list is provided as an aid for those who may wish to delve further into this subject.
Akerstedt, T.,
P. Patkai, and K Dahlgren, " Field Studies in Shiftwork:
II.
Temporal Patterns in Psychophysiological Activation in Permanent Night Workers, Ergonomics 20(6), 621-631 (1977).
Christie, M.
J., and E. M. McBrearty, " Psychophysiological Investigations of Post Lunch State in Male and Female Subjects," Ergonomics 22(3), 307-323 (1979).
- Folkard, S., and T. H. Monk, "Shiftwork and Performance," Human Factors Journal 21(4), 483-492 (1979).
- Folkard, S., T. H. Monk, and M. C. Loffan, "Short and Long-term Adjustment of Circadian Rhythms in ' Permanent' Night Nurses," Ergonomics 21, 785-799 (1978).
Horn, J. A., C. G. Brass, and A. N. Pettitt, " Circadian Performance Differences Between Morning and Evening Types," Ergonomics 23(1), 29-36 (1980).
Patkai, P., T. Akerstedt, and K. Pettersson, " Field Studies in Shiftwork:
I.
l Temporal Patterns in Psychophysiological Activation in Permanent Night Workers," Ergonomics 20(6), 611-619 (1977).
Price, H., P. Wallace, M. Bauman and M. Smith, Bio Technology, Inc., " Review of Staf fing Requirements for Near-Term Operating License Facilities, "USNRC Report NUREG/CR-1764, October 1980.
- Reinberg, A., et al, " Oral Temperature, Circadian Rhythm Amplitude, Ageing and Tolerance to Shift-work," Ergonomics 23(1), 55-64 (1980).
l Wilson, J. T. and K. M. Rose, "The Twelve Hour Shift in the Petroleum and Chemical Industries in the United States and Canada:
A Study of Current Experience," Industrial Research Unit, The Wharton School, University of Pennsylvania, Philadelphia, Pennsylvania, May 1978.
Bi-2
MRC FORM 335 U S NUCLE AR REGULATORY COMMISSloN NUREG - 0872 BIBLIOGRAPHIC DATA SHEET 4 rittE AND SusTiTLE tadd voe me no,,F preneni A Feasibility 2 Ite=e w -'*J u
Study of using Licensee Event Reports for a statistical assessment of the effect of overtime and shift work on a ReclPIENT S ACCESSION NO operator error.
- 5. DATE REPORT COMPLETED
- 7. A U T HO H is)
M ON TH l VE AR R. DiSalvo, A. Gery, J. Pittman May 1982 9 PE HF OHMING OHGANilATION N AME AND MAILING ADDRESS traciude I,p Codel DATE REPORT ISSUED
.ONis lveAa Division of Facility Operations Office of Nuclear Regulatory Research June 14p U. S. Nuclear Regulatory Connission 6 treee ueans Washington, D.C.
20555 8 treeve Nan *)
17 SPONSOHING OHG ANil ATION N AME AND tJ AILING ADORE SS (lactude lip Codel
- 10. PROJE CT/T ASK/ WORK UNIT NO Division of Facility Operations Of fice of Nuclear Regulatory Research s i CON T R A(,T NO U. S. Nuclear Regulatory Commission Washington, DC 20555 9E RtOO COVE RE O fenclusove dams) 13 T YPE OF HE POH T Final 14 tte*, 0/"'*/
183 SUPPLE ME N T ARY NOTE S 16 ARSTH ACT (100 words or ress; A study based upon the reported licensed operator errors from January 1981 to determine if a valid statistical determination could be made of the effects of shift work and overtime on operator error. The study concludes that the data reported in the Licensee Event Reports are inadequate to draw conclusions on the influence of overtime and shift work on operator error. The analysis did show that the errors are not uniform over the hours of the day or the days of the week; the causes of the non-uniformity could not be determined.
17 KE Y WORDS AB D DOCUME NT AN ALYSIS 17a DESCRIPTORS Licensed Operators Human Error Licenser Event Rdports Operater Error Overti.ne Shift Work 17h IDiN TIFIE RS OPE N ENDE D TE RVS 19 U TV CL S Th s reporr/
21 NO OF PAGES 18 AV AIL ABILITY ST ATE MEN T
"~
- Wals%C"'"
s unlimited N RC 8 ORM 335 L7 7 7I
UNITED STATES s ou st u ctass uait NUCLEAR REGULATORY COMMISSION
'037^G,3 l,'"5 O 3 g WASHINGTON, D.C. 20%5 ansa o c Ps nuit n. In OFFic AL susiNtss PtNALTY FOR Pkr/ ATE USE. 6100 f
17055 8877 1 ANPX US NRC ADM DIV 0F TIDC POL IC Y C PUBLICATIONS MGT RR PDR NUPEG COPY LA 212 WASHINGTON OC 20555