ML20024B511
| ML20024B511 | |
| Person / Time | |
|---|---|
| Site: | Crane  |
| Issue date: | 05/07/1979 |
| From: | Labelle D BABCOCK & WILCOX CO. |
| To: | Roy D, Womack E BABCOCK & WILCOX CO. |
| References | |
| TASK-02, TASK-03, TASK-04, TASK-11, TASK-2, TASK-3, TASK-4, TASK-GB GPU-0286, GPU-286, NUDOCS 8307090057 | |
| Download: ML20024B511 (5) | |
Text
-
., i e. rM }.,it$. 3w pr.,;..
v..n.,..
,.:m u. :
%.*h.
11-., - - 9.y qQ3..;;.,wa. J.*%?,,*D
- ~ f --l. : - -
- r
,.o...
&i
'Jg.
..v'
-=Q~.*,*l..&*..~cj l,Q WW=.:t :r@p*W.?y;;t.r,.gg ;. :.4,c.,. -
gwp..d.m:s.:
~ v.-
n c m=.
.;;.,. :g. fz =- --y.. _ ;p_.
-~. =..._; j,.
.m:
.-...'.. c..c. c.:,
4..
- ~.mi g.~.ug g..
..z.
.,m
..m
~
-M 6.
2
.a.
.n-5r
$?~
N...?hj$?
W8,&
k N h A k h : h f M.a...$..hh t._>. m.__
y_.
_,5.;;7. %.
.w x.~s:.w.,.w.i* 5.t.Ww n_ s~sp.).
.m.,-gy.::g,._
c.,.
r-i C.
p.u
%.es.
ha 4
s s..
M
,e.
THE BABCOCK & WH.COX COMPANY E%%
POWER GENERATION GROUP ph
" ';,d To I
- n. a. ey, Mar.ager, tagineerist oepare=ent M. T. Bonaca
=-3 U mack. Manare_r. Plant Desten
- 1. O. Vosburgh (y
Cc:
t.
A.
"-Q'. Us F rer.
A-afvv's Ii!e ho.
D. V. f.sNelle. Mangeer. Saf ety i-or Ret.
K..
Ce s t.
Fd
^;
Date w.5d
%gq Subj.
v,v t me
~r 3 m,--..
m--. -, o., ---v. e t,...v b#r;.v.>
y",&q 4W:
J. :
i
.~w Massrs.
Attacted is a copy of scue recocuandations wh.ich were prepared by
- 70. :, ::r W
., - o.r;g l dge Vosburgh and Bonaca based upon ::7 request that they draw upon their kn
.y
';i h. y [
of osst work efforts, involvement m.11ent job of charactarizing f
I f eel they have done an both the problems and the direction that SW f aces as it verts toward rescore
_' "f:]w and C;.trability Report.
, 'L':..-
the saf ety of the tion of the NRC. utility. and general public cesfidence on
-w49 a
~ *s..ha~2 S&W F%a desis=.
- r, w
since I do not know the speci.fic directio:s is *.ich you are presently T. ~
.. f d plants, I pursuing the upgrade of system desigss for operation sad standarbonaca als:s e.
r e -,
d vanced to pass the orogram outl.ined by vosburgh andassigning Mr. Ec.aca pri=m b
R.~
--1:' 'e*-2 M
ferrther deve.lopme=c o.' detailed prcgra:s for nur apperval f or both operating review. I an at te;s time.
l nt c;*e Fit 5.~h $
asd standard plants le d*=g to justifiable ecx:puter cada capabi ity, eve fD
.y )*C,W asi i=cegrated systc=s desigt basis d
r ecc=senda-jgy formulation for the B&*.! reactor desic supported by saf ety and operability analysis, operational prece ure Tc; your additio:Al infor=ation, the T
~ %m tiens, and design change recc :nendati ins.
..g gl following activities are presently in progress:
b
,3.-
...{
Develo west of detailed event tree for LCX79cocaequences erstem designs a.4
- 91 A)
Competitor review of laGV evest quench tank, ste.a:a and M
ftTlf)
B) p*rtaininr. to pressurizer relie*,
a,Q' f" C L gg f ubveter eystes aren. ate Ptur and/or pressurizer M
Coirpetitor review of ev==ts whici:
W fg
.. hPlf0 Q C)
%fTREVtting Inc.h(* As Saf ety Analysis proceeds with its e' forts, I ca
[,t.
t
"'
- c7 ' **
4 ildup l i and I. :.scdig.. 74g of an interf ace structure with the Performar.:e Analysis, r.CCS Ana ys a, for confusion over As such, there is the pote="* al I theref ore requaat
' g w.;;:.3 Flant Integration units. cost recovery, and scri co= tral.11 and p;cvide directional
. f :..';y.
verk responsthilities, t
' 4- '@' 3 that you plea se, review our recomettatics.s by May i
~..
.. Q M, input quick.ly as to the,folleving:
if
-s Any geners.1 cc:sec=ts ou the suggestad approa
[5
... 4, v 1)
EE u.,
l sis WR vith, or perhaps duplicate, the ef f orts suc::csted by S lp 4 - ~' 5I'S 2) 5 fh./.dz@H
. ld',d in the wake of IMI-2 which nay establish a new precedest f or how
- 7.-
J) lled.
the proposed pre;zam should be coctdinated and contro
[
.;.2,7.;.[
h7790507 Ms h.:
- 05000289 p
HOL i
L-ru 4 W-'d:.Q Attaciment t
.., ?
4-j-.-
yi l
hlE SABC0CX & WiLCOX COMPANY 1g POWER GENERATION GROUP I,.
- seJ-To l
' Cl D. V. t aBelle. Manseer. Safety insivsis h.*
a From R. O. Yosburgh, Safety Analysis (2176) M T
0705) kf b M,
nos ses.s M. v. sonaca. Safeev Analvsis Cust.
File No.
or Ref.
j C
c.n. rte Subj.
Date Proerne and I.essons fren TMT-2 Incident May 1. 1979
)
l n s,....
k.s IhTRODUCTION A major incident has occurred at our DI ~ facility - B&W responded with a great deal of manpower and resources to aid our customer in assess-ing the situation and saf ely bringing the unit to a cold shutdevn. Currently, a large effort exists to respond to FRC required analyses for all B&W operating plants. However. several organizations are being or have been established to follev-up ca the 21-2 incident and it is new appropriate for Safety Acalysis to make its reco=endations concerning:
(a) follow-on analytical progracs for User's Croup (b) in-house code mcdificatiens/ development (c) equip =ent nedifications required for safety (d) Philosophical approach to saf aty Our opinion is that the following recomendaticas are of c:tre:c 4
.i= porta:ce and sh uld receive.Er,gineering Depast c.c.t_attentio:tto nrovide the
(~
i=petus necessary for their i=plementation.
.g j,.
s Proposed Program For User's Crous Isole=entation This program is a ecuplete, ccuprehensive safety evaluation of the 177 FA plants based on event tree evaluations. and not strictly on Licensing require-ments. The program is proposed in phases to allow an orderly accu =ulation and evaluation of data.
Phase 1:. Event Tree Simulation E
Lead Unit: Safety Analysis /ECCS Each class of accidents should be modeled with the possible paths that the transient can take. The event trees must in-clude such itens as ICS action. ICS failure, probable operator actions and single failures. For each event tree branch, recoacanded operator actions will be identified. as well as hypothetical actions that the operator could be led to take due to the inforzation provided in that situacion by the con-trol instruments. The combination of relevant paranecers for 7
i l
each key branch will be translated in a diagnostic assessment and an adequate operator action recocmendation which can be programed in the on-line computer for operator guidance (see l
l
~
Frogracs ard Irma w Incideat May 1. 1979 h
Page Tuo t[
4 Phase 3). The relationship to bounding SAE analysis can also be shown. Interface with the customer /A-E's for as-built secondary system design will be necessary.
t Phase 2: Analvtica1' Evaluation of Events t
Lead Unit Safety Analysis /ECCS The analytical follow-up to Phase 1 is to evaluate all branches W
of the event trees to assure bounding SAR analysis are in place and to identify those areas where operator error and/or ICS failure will propagate to a limiting situation. This analysis should carry the plant to cold shutdown under such conditions as natural circulation. This capability must lie in the Safety Analysis area (see code nodification section below).,
It is possible that personnel from Control Analysis and Plant Integration should be requested to temporarily transfer to Safety Analysis during this phase of the program.
Phase 3: Systes Parameter identification /Destra Reccenendations Lead Unit: Safety Analysis /ECCS Once a complete safety evaluatics has been performed, system paranaters important to the analysis can be identified. This identification can then be worked into a sche =e of simple checks and systen readings the operators can perform to diagnose I
and nitigste consequences with no room for error such as icd to the.TMI-2 %ssijlent Such a sche e could be progrs= sed for in-
- - f.
control room confuter ' aid bUbig *G~ print 4ut suggest~e3""CTist*
c and probable situations (or the operator. This work phase would require strong interface ties with the customers and with our Nuclear Service and Training groups.
Computer Code ?'odifications and Development The current versions of CADDS and TRAF 2 are adequate for Licensing type l
analyses, but as was painfully obvious f rez Vll-1 incident, they are not adequate for more complex or " realistic" situations. The following is an outline of reconnendations for code nodifications or develop =ent.. These mods will be requested forns11y via procedures; however, departnerit level priority it asked by this neco in order to fulfill the progran objectives outlined above. From l
current NRC analysis requests, these progran =ods will be required regardless l
of future User's Group work.
Tor CADDS Ta) expand code to a two loop model (b) incorporate the ability to model natural circulation with l
one or two generators operational l
(c) expand the property searches to faclude two-phase sixtures j
(d) incorporate current D::B correlations with hot channel l
capability l
For TEAP 2-(a) complete the OTSC and pressurizer models needed for over-l
.~
I Progra=s ans Lesson:
a Tt.
Incident May 1.1979 i
Page Three
(.G y
A heating and stagnant flev conditions (b) model RC punp power into prcper nodes (c) develop or help i=ple=ent more stable reduced code capability for long tern analyses i
~
Equie,ent Modif teatien Reco :nendations One of the nost useful pieces of infor=ation that vill co..e fren the event tree analysis vill be those paths that lead to severe results with high probability of occurring. By identification of these paths the type and degree of quality for equipuent vill be very graphically displayed for user infor:ation. Further judgment on equipnent codificatient will avait the event tree type analysis, for hasty decision on equip =ent modificatices can lead to proble=s in other areas when inplemented to solve only one probles.
Approach to Safety Analysis The most i= po rtan t, and saddest. lesson learned from IMI-2 is the general complete lack of knowledge concerning reactor safety. Ibe following points are made not to point fingers at anyone person or group. but are offered as candatory and insediate renedies to a situation that nust be solved if BW is to stay in the NSS business. Thece points are (1) Safety evaluatio: =ist be not considered a 1.icensing tool but done for plant safety. The licensing vill automatically follow.
- o longer can we ask. "'a~nat is required for the SARf" but we =ust ask.
"... is this design saf e!" Our definition of safe is to be based on adherence to the prebability of occurrence versus the results if it happens.
~,
(2) Decisions regarding plant safety, or eq0ipuent quality and functio *nal specifications, cust be =ade within Safety Analysis (and, ECCS). These tw units alone posseas the technical ability to make such decisions.
This neans involve =ent in conceptual design stages instead of serely
" blessing" a finished product. This is true for such ite=s as long-tern cooling. as well as RPS etc.
(3) Plant operating cccurrences uust be revieve.1 by Safety Analysis for safety evalution, and not af ter a " screening" process by which soce persen judges whether or not a safety issue lis involved. Any occur-rence that is not a nor=al plant operation cust be reviewed by Safety Analysis.
(4) All safety evaluations uust be done by Safety Analysis (escept LOCA).
Tuis pertains particularly to over-pressuritation events and events involving eperation or failure of the ICS.
If this means perfecting the analog capability within Safety Analysis. then this should be done ic:=ediately. Ccusistency of safety judgments =ust be centralized and strengthened. The degree of decisions c.tde affecting plant safety outside of Safety Analysis during the IMI-2 incident was appalling, although recognized as necessary under the current code of operation. I!ovever, the lack of a central " safety" grcup was largely responsible for a tresendous amount of duplication and vasced effort during IMI-2.
I~
~~~
i f
i Frograss and Lesso:a f ren
-2 d.I "N
Incident j
May 1, 1979 i'
- g, i
Fase Four
.. C (
.n==--.
Conclusions i*
The key to assuring that the probability of another TMI-2 event does g
One, a strong User's Group progran =ust be not occur lies in two items.
implemented as outlined in this meno. Secondly, a strong central Safety l
Analysis group must energe, precluding the frag =ented, self-styled safety
'erparts' that currently exist throughout the Engineering organization.
ROY MY3/km e
h o
~ 3 --
e.
.g f '
~
+
.