ML19249D366
| ML19249D366 | |
| Person / Time | |
|---|---|
| Site: | Crane  |
| Issue date: | 04/13/1979 |
| From: | Lewis H CALIFORNIA, UNIV. OF, SANTA BARBARA, CA |
| To: | Hendrie J NRC COMMISSION (OCM) |
| Shared Package | |
| ML19249D365 | List: |
| References | |
| NUDOCS 7909240386 | |
| Download: ML19249D366 (2) | |
Text
u..
/ o.
m UNIVI.RSITY O.F CALIFORNIA. SANTA BARBARA
, C..,
,h
,.,3 \\,
l g4 i SA.NTA RARRARA - $A.NTA CRC 1 P ""
1I76 - DAW * mtuE
- LOS ANCII.I3
- Ap*ERSIDE
- SAN DIEGO
- SAN TRAftC25CO a.
T' M/
h:------
/
. u.
DITA.?_N'T OF* FHTS3CS M
- :=--
SANTA BARRARA CALIFORNIA 93106 5
=
April 13, 1979
==
=
Dr. Joseph M.
Hendrie, Chairman U.
S. Nuclear Regulatory Corraission Lf
.]
iE:s WasM gton, D. C.
20555 7, -l m p! Sn-D n;
3r,
[a.9 i
i
.a
Dear Joe:
,.d.Aj j Dj ' j ;' L a
Now that the situation at Three Mile Island seems to have stabi-b lized, I assume that everyone will be beginning the long process of learning and absorbing the lessons of the events, and I hope that you wi'1 not think it presumptuous if I share my first try with you.
The following are truly first impressions based on fragmentary information,
=
and are hardly original, and I do eagerly await the results of the in-evitable thorough and complete elucidation of the course of events.
In particular, I hope that you do not read into the following any ef-fort to second guess anybody, since I have neither the f acts nor the inclination to do so.
It seems to me that the central lesson is te =rovide a boost for the ccncept of flexible response, as was the case' 'at Browns Ferry.
E' The fa :s seem to be that there were surprises at Th'ree Mile Island, f.l...
and see little reason to expect that there will not be surprises if any future accidents go this f ar down the secuence.
Further, you knew that I believe that constructive human response is one of our
=
greates:
ncuan:ifiable lines of defense, and that was true a: Three
. :. l e Island.
The concep: of flexible response has obvicus implications
- e. ning c:. reactor operators, and for the provision of rather
- re instrumentation designed to provide information under upset con-diticns than seems now to be the case.
Everyone knows about water level indicators -- are there other examples?
In addition, it suggests increasing, rather than decreasing, the flexibility provided to re-ac cr operators, concurrently with an upgrading of their training and J.;
- ssting.
The aviation analogy is, as usual, obvicus, and one recalls G
tha: there are some two 'housand commercial aircraft in the United 3:ates and cnly seventy reactors.
New reacters cost a billica dollars
=
each.
Thus, in addition to the things tha: the NRC will centinue te
- :: red::e the probability cf accidents, this cencep: wculd imply
..=-
...e-
.c...., c expec: accidents anyway (that is, after all, what probabilis-tic risk analysis is all about)'
2:
expe : surprises (ACE - anticipated unanticipated events?)
2 ',
- cvide redunfant and prolific instrumentatica to supply
.n:ormat:.cn.n upset cent:.::.cns (even to cc cer cara e er ranges that wculd be censidered "of f scale" fer nermal
- =
a-'
..s)
.-..- -.... a -
-- gg,
.u.,.
= '.n '..c.,.. e e" cn, a.. =-
as ~-..~..c
.n -
-,e...
.= c a. ' - -
eac*c c"-------e.
a"'^-
1002 063 790924o 1
i...
- =
N.
N. Joseph M.
Hendrie April 13, 1979
.e.
It also seems to me that this accident was not necessarily a j
f ailure for probabilistic risk assessment.
Though the probability
- f was apparently calculated by the infamous square-root bounding
[_
technique, WASE-1400 did have in it the likelihood that the dis-
- g..
charge valves on the. auxiliary feed water system might allie in-advertently left closed, as a testing and maintenance error.
Ob-
- t. :
viously, one doesn't take the calculated probability too seriously, MR nor.does one trivially extrapolate the situation at Surry to that TlE at other plants, but it is true that WASE-1400 provided a listing IEE cf a large number of credible accident secuences.
It would be in-teresting to go through those, asking whether there are any proce-
=:
dural actions er inspection reallocations that can be taken to reduce i+EE their probabilities, or instrumentation enhancements that might help Eff te control the sequences without ambiguity.
I still believe in the i.? ~
~
Review Group recor nendation to take probabilistic risk assessment i.
seriously.
p Finally, the need to learn as much as possible from the Three
.E hile Island experience is cbviout, and does not require recetition Irom 'ne.
I hope, and assume, that there really will be a ' serious an_ cm3 ective ef fort to do so, uncontaminated by the need to clace
-~~
the bla:ne.
Best regaris, La'rold W.
Lewis jd, D
D
.J. x, S }tftr HWL:ek 9[F;fl@!!b! OI a tr au dul La t
=
e k
1002 064