ML20141K365
Text
{{#Wiki_filter:/ e * \ 8 OW1
,, ( k i
n s-October 5,1M7 'y"k Q; ... t ..e 3 N. To I Dr. J. E. Metes Mr. M. J. Palladino Pr. S. H. Dush Dr. W. R. Stratton Dr. S. M. naamuer Dr. C. W. S e al e,... v.,.a. e c..,.. -_ Froc M. C. Caska, Aa's'is cant' to' itmec. Secretary ActFS 1 1-
Subject:
CffSTF.R Cf.K!I l. Dr. McEme has requested that the attached evaluettom by Mr. Epler I of the ister Creek instnmeatstion efetu be distethted to the Oyster Crook, Subcommittee. There is a questtom as to d ether Mr. Epler's ===nts have bosa edespetely t4uia Lato escount in the f tnal desty,,n of t.he Oyster Creek las trumentation system. Attschaenti Kr. Epler's comments dated May 26, 1964. I l l i , i 4 OmCE > 3C83... . ... ... . . .. . . . . . . . . . . . . . ... . . . . . . . . . . . . . . - - - . - -- - suRNA'dt > . D.bj. . .. ... . . . . . . . - - -. .. -- . - . . . . . . . . . . . - 04rt > 10.7.5.d ? _ nwm Acc.ats ta" *4b u s .omorn me.as cern e m4.
/ / ,s M I,
9203190065 910807 -2py /f,. c, PDR FOIA l'tl* DEKOK91-282 PDR
~ . 3. ;, ; a wa i r .W w A ,Lg.iv lad y ;,h.g. .u u, e..r.
t
- - \ /
,s
( U/ -w y; +d' w l ,, .w., u. ' g OAK RIDGE NATIONA'L L.AEORATORYqd .
+ ,
)
oet=Avto av {, 7 UNION CARBIDC CORPORATION '
-S ItdCL(An Oln110h M'3f .
g ~ ' ' 6 f; m.,,<,....
- .s ,
_ _ , .At $10 % TO.ANtt$tt Staal , ,_,<,
,S 8,ff 4,8 ',, t May 26, 1964 -'
~~
y,w o E. W. Nevson -e' D1ke University 9ahat, II . C . .
'ect Sir:
... uccordance with year request I attended the GE presentation to DI23 -
- . 02enday,fhy 12. The following su=:tarizes my evaluation of the systes.-
M ed on material presented at the aceting plus further study of the docume.r.ts gfver us rt the meeting. -
h
. Cyster creek system is an improvement over the Dresden system, ir. that
.t consists of two complete independent channels each consisting of two S .uo channels placed in coincidence. In principle this is scand providing tLa.t er.ch of the independent channels is itself sound. The reliability Of the pair vill be half that of each ncober, which is entirely accept-able. The success of the system vill rest entirely on the canner in
'/c.'.ch the details are executed. i.e. it is vital that the subehannels be s:Je independent. Further study of the drawing presented by GE dis-c'.aes, on the face of the schc=atic, numerous violations of this principle indicatirc that the desi5ner has given no thought to thic aspect of the l
7 oble= cnd that it vill- be corrected only on the insistance of LiR. i l R ECEIVED L.S. A10MIC D,iR3Y CCf/MISSID j ' i p.cas -q l e wIl d ,I ; .' 4's U.t P." , rx '
. . . . . . [$ .
)!fl l.915.5 1 5?! Ihl I N M \;d Y b "
~".
g q.en ~w)y, nt I .
$[pht{ ' g j.{%,/ C,
# ._ f
-l A.
l ) ,- ' s - i
.A , e h
, , c
______ : , ::,::rD,,%.,
, .w ,TfS, p~ \
. :.nw
= - ;;u ~ _ . . .
Sc b e mc--
,f
'77] e. Di'e.sden
'?f$f@f,ks i k s-
' .,W @ 4%
- -g, :;. ,, j.;i r.
,4 : y ~
14 h. . - C 40H"el l . ' fio siulL ~.. >
- . it.d
. . J:dk/
f o u.,e P / S-rfll<s % ,I _ 4 f ,m. y y _ 6?A BfD , y Ihy4b lep.A A _ L L- ,1 g
~
- u. B 3k- , 6PB bPc- 3 .
b~f 2 0FC
.. c'-- . p pg .
(
.. q 6 yyast' 3 =
s pp.
-h hBPB PV h0f5 0?5
, y p, f, ]d __
;-- B P c BP$f 1 6 f '-
Of OPE'~iE 5ts J 4......
.1 f
~ ~
f CI. anne } } Coi Is ClunniI 2 Cot l.s
'lt $ c m"1- p 3 coa
, e ge $
. e e O .,.p
'F. .
, w su-Q.{ ! v-
~q, Zs ',yl '
a~
~
, tw
.* 4 a- s ,
O
. q
- ,. j ; gl.', - - .
O
y .~. "
.-a :4~,,*/p p A 3
'O8d . , , _
. l
{1 '
,~.
. %ggjg 3 ' -? . g y 1 1
~
-1 The Dresden Scheme .' YS1$j ;; j y ;; ; , *; y y
'1. Channel 1 coils plus channel 2 coils ;:ust be de-ener61v.ed to y e produce scram, e-,9',43, 8 1
- ~ j
- 2. A or B or C can produce a half scram. "
- ' ' 3: ?
- D or E or F can produce a half scram.
- 3. One channel only can be bypassed at one time jn either string. + SZ O The Dresden scher.:e contains two outstanding; defects:
- 1. An electrical fault which bypasses all three inputs in either channel 1 or channel 2 vill parklyze the system.-
- 2. The bypassin5 equipment places an excessive amount of conitoring equipnent of minor _importance directly in the safety string thus increasing the exposure to electrical faults.
As an ext :ple note that the contact BPA appears at the top of the string
- and again at the bottom. These are two conta'ets on she same w icy or svitch which may.be no more than one inch apart which if connectnd togeth-er _vould rake the safety system inoperative. ,
A 4 y 9 O i k 2 N,1h!$.k" - m:x
-b f
- y~$:. ~
^
, a v
e Y. . O O 5
m - - - -
;$b'$@l7
~
^
' . ' ' l. -
y, :: , . .
.~;
v;cyxig.n -// - rs, Jca,1 sc hem e m.
. : 9.;;. % % y ~ %+ s..d} ,- k' ~
~ ,., . .cgu}g . s ; < (. 4,; .
aggl sq l' S16P,=w ;
. .: y . ,
. ,: f, w e is ff ' 4: ChannoII y ,,,)( f Fx ;: x u t' ^u,;,
'L powir
{-
^\9 Ssf th,es 1, .,c .
1 K.g. [ A
]
A z g P v 7 s. [-
--t "r- DF' L-6 - L B mL.E
'::" 5 i i
L - p
-e L. c I
.s. . p .
- j- -
W O ' jo U.S 14 7 {[jY0 --a g3 f . t t f
-l P5] "% - - -
-) -
y :H 1 U" # F ,.J X31 .L n,, 3 u7.J -L.hu, ' lc w - . I
-~'ov . -
m g'
] j.
1 i -l I 41 (I1 T_ - 6 % 63. Si 6; G 2. 6 a, sY pl. j, t A3 A t As pt As A't C honnrl 9- A
- 0 0"'f 5 Chpnei i Rod C'**f5 -
1 I g, Sc ro n)+
'jt Secom :
t
~
S 4 a s 4 ,- = % : . e s , - g k
' _l " '
E'4 , m WyN: -
% fr >
I O ' t
.J -
. ~ i i: .
~- , , a - J .* , ,
t- '
,4, q " ~ g. ; ,
e
,N ,.,
' ^ '
'syf-gjjr ?"S &
( h '
)
kg, g, ,- . . 3 M. j , DfyQ ' The Jersey Schene c.4.q,g) so,, .
- 1.
- Both channel 1 and channel 2 mst operate correct 2y in ordcr[ y
- to produce a reren. ,,
y cy 3
- 2. Each input channel can produce a half scram in each of fcntr .- l" ,
groups of rods, i.e. channel A can produce e half scram 1.n rod groupt,A1, A2, A3, & A4. A.i ! f'
- 3. Bypasses and inputs for other process variables are not shown. . -J
-~
The Jersey schene is c:uch improved over the Dresden schede and $s funf meatally sound. Each of tva channels cust operate correctly to prode:c a scram, hei./ever each channel itself consir ts of two parts - (sub channes'L) cither of which can produce o half scram. Thus if the sub channels are independent, no single fault can prevent a scram. If on the other band the ao sub channels are not independent a eingle fact which reuins undetected can set up a siriation such that the next fault vil3 make the systen inoperative. It is therefore imperative that the two sub channels bc.made physically
$ndependent. The GE scheratic attached shows many instaves where unusually great expmure exists, which unnecessarily increases the probability of failure.
- 1. EF1-1 and El-2 are adancent contacts on relay HF1 A fault in the relay or viring which can tie these centacts together, ,
vin tie together the sub channels. The test which is perfomed by tripping, for example Pico Amteter fl, vill not disclose this fault. This illustrates a general situation which is true of all contacts for both process and nuclear vanables as well as for 731, a single push button for manual scram and to MRl-1 ard MRl-2 (nanual reset) . The bypta . system, when developed, vill add many c.dditional pointG of exposure and should be Asolated frcs tile safety circof t.
- 2. K1-1 contact is adjacant to the lin contacts 10.-2, Kle3, 8.
Kl-4. - A failure v.'.11ch connects Kl-1 te KL-2 Vill tie th:L scram relsty to the bus and thereby fail the sub channel. This ex-N su e exists for all relays Kl through K8. This fault cannot be tietected by tha tnt Es proposed. Although tsc system is basicany sound, it is wiscund as presented it
+
that no thou62t has bien given to channel irsoJation in the schematics gg" When vill 'bethisintxw. system
.:ed is in re4 the :ced tolayout, viring hardware, many espec'. allyeMitional teminations, exposurethus points !4 A[-
greatly increasing the probabi.lity of failure,
' j%
f,' , m, . . . , 9 m f # U 5 g.
+ , , _,
7 h* , i f e 4 k,ee , he pqy '
.[. t sG l. ~
4
+
. J ow . .
.y
( ( l 6 1
- ) t1 l h is ny reccendation that measures be t2.cn to obtain ec=plete 'I cub channel isolation and that a test'oe deviced so as to test cach sub channel separately so as to discover the first failure as it occurs rcher than the second faihtre ac now designeA
//
V i;"U , E. P. EP,Ner
'sPE :,ip Attach >
cc: M. M.16:$n , R. F. L'11ccx / M. C. Gaske O. Fri::cli g a h 4 1 I l
*W f
~. 6 --,,
a , \
' w s, J,s E
' s 0
1 - ' Ili[i 1: 51 til b l 'h 3 ( 1 i I i
. g I I <% 98 r : li: ) (1 i 4 i Qi. s i 1 2-y- :
g~ p." gm,.... -. j 1 V p-
-9 a< .
, i c' g i
')
i, # 4 ,: s .= - 7s f t I- s' f .4 .. t, L % a .$ 2 L m [,
-n : m h y 7
- rS e
g us a, ijl., ." b - . .
~
- ' N N W
e ,,
* -i t-+-
>+ is
)
I e2,7" g
,h !UI 4 -4 s E 4 1
jgg I -
- e. M ,T %
b-. , . m f+9,'!g tg t
'5 g
3 y 1 {-
,o 1 J. ) L..
v-- _ y b 44>-0 .5 g4 s Ga . 1 3 4- . 1 i L-F* F.+-
- : Tin i
i s. m . j. 1.1 w' , . I i ' i+
+-
I
.i 4 4{ - ,
,2 4..:_c t 'l*' - ~~~~ ~~~
s
...,-..;_ Tr . _ _ , ra e) 3, a J ,q e, m *-{'"~~~hIN d *- 5
(( i
' d' hi };
-~
s
*f vi
;g % v a
< ~ ff 7 +u n 8131 t i ?
a- 4 kk Us 5# 1
,g
@,93 h
-6 +e a.r.n
-y ;. f I t < : 5t5 en
,4p
\*
% 3 ,, e.
O'*f J5 4
- r:
%i 2, " .&,,- + .-4 '
i
- g. s g f --,f- -i >-.,, i r .
_ 4F
!* y i'
; p 6 4 , l'S
-- n b t.
y f HH s u.- -
? t 4- - '
l I h
~i N wi Q. a\
e il ._ s t ' k
~ ir {
I,s'I si ' ;} j 8 h 1,< He. s 3 .
.. . ))[" y (7 .. _ . .
) f %I FD i
$ a i s i i"s:qFill.ut~<!! ,1:!._ N !l d ~f~Tg{lFM wp+t '
, p gg b :
A -
/
3' i < i . i e o . . ., c . i , , s j** a $,
.*}}