ML20244E141
Text
,
.A,
t 4-e d
. -,. u' ft.,; A e
.... a,
- . n.
- y..
e.
mr ~ '
3 ~"~M '- ~ l'- f -
s gy.h( '?m%fjbf y:.$.t c is,1,' 7[,,.
7.y q,
.pi,i w ama.. >
m.
a c:4 r.._.
M., m
@gm, e m.
. a.s.pu o. ww%. mar so, levo
-e m
..x
.. 4
.;.. ;-i w
- A m
y
+
v~
e.g
_{,,
,; R
'" q, h,K.,"*
c
- h.>.
.c
. n...
a;
~.n y c
- r. -
p.s,
.,. 3 A e.
v..
m.
.,. =
,,z[y ; n,:
s 2 -;
3 g,';
77 w,
Br',JJoeoph M. Eendrie Obstraan, Advisory comm?/,ittee, ii )
~
"E. i
+'
s on,Rosator Safeguards w_J W. S.-Atomie Energy Commissies:("9 g7j~,
Q', g. w,y,
- .. _..y
~
~ d:
J Washington,.D. C.
30645
.. g s
.a.
Dear Dei Eendries i
i.~ :
Eighteen copies of a letter.to Harold Deaton from
.U:
. _.._. R uW. Staehle,, dated.May118,' 1970,.
' ' recommesadations'fora' course ~of action to be taosaceraias., ken yf.y..cyg.
^-^
for the repair of safe ends es the resetor vessels of Dresden 2 and Nine Mile Point, are enclosed for the use of the Committee.
"C
.Am Biseerely,-
e.4m p,.w
_;.y s.
^
is s
p
.m wsow?' n;;.o...s.
a.
' Peter A.. Morris, Director
...,.,. ~.
.q_,,,,,
i Reteter.-
asingr mo
-.9 e, l NW y.n l. q y y
~
R!w.-
gangee,y,,,
~
,c As stated G
Distribution:
HLPrice, DR Compliance 1-=
DR reading C
DRL reading RLTedesco, DRL NO' EGCase, DRS
<.n 1;
e 9
omer >
...D.R..L.......
V sunawr >
..P.4Mor ri s/b.b..
04Tr >
_5/2.0/70 Forse ABC-SIS #Rev. 9-53) AECM 0240 e a novaemmang ewerimo orrecs e ssee.-ase-us
[
h1 8904240313 890413
[/
i
)..
m.
y i
- %,), 4 3,7.u g
- I R..W. Stachic.
.i k
o;, 9 4.-
,y l'
l 5060 Havens Corners Roal W@p}.fm l. a*......(
.6
.%s Gahanna, Ohio t
is
. j
. g 2.d'+
t.
,h Pkfb%
t reply:tos Dr. R. W. Staehlc., g The Ohio State Univ.
.f.f !& p,.
. !,F wm
- f. ;..,$.,,: - /.
.' ;.. ~..
3 a116 W.'19th Avenae, Dept.. of Met. - Engr.y my -
,p,
,g p,
[ Columbus, Ohio _43210
..p
^qJy
- r.
y- /. 9
- y wo Qq 18 Maye*1970 y% ;
.s, 2,. y ;p y p.a.'
- i. -
g f 'g f.'.@ @^..l. 7 4..Gfj! 't. n . ac - r. r. 'b Mr. Harold R. Denton, Chief. ~ ^ 9,2
- o l
' Technical Support Branch j p@ '1 U.' S.. Atcmic Energy Ccumission ' Division of Ccapliance. a, j.y . Washirgton, D. C. 20545 4,, pw M. 2
Subject:
Course of Action on Repairs for Dres-len -2 St Nine Mile-Point' u a j
Dear; lbrold:
1: @6 M. -1 gg l < On May 14 you requested my reccmnerdations for a course of action ' to be taken for the repair of safe ends on the reactor vessels of D; e Dresden -2 an$ Nine Mile Point. The purpose of this letter is to transmit g' b these reo:mmenlations together with their justification. In addition to M w. 4 @. Ik these reco.mundations I have included suggestions for measures to be N? I C. , g~ig..... taken to reduce the' likelihood of spurious cracking. ., :, e ~ ! < 9 E,c Primary Reccnnendations l )M% F . -fiff The recommendations here are contingent upon the kini of " failure" 4 a which the AEC wishes to avoid: ~ e w l.
- 1..For example, if it is desirsi to avoid the inciderce of any transgranular stress corrosion cracking, then.it is becessary to replace all stain 3ess steel pipin; with either a higher
$, M. ~ . nickel alloy, a ferritic stainless steel or a duplex alloy. l 5 h 2df In my opinion this is an unreasonable criterion in view of ?@ the econcmic implications. Itrever, it must be kept in mini that ircidences of transgranular cracking are always likely 9jh in austenitic stainless steel systems and extreme care must i w.- i t s w
- 4.;w 7
9 $wip* v i . g -n L s-yyt. -44*s Dh// o am s a g e 8 w w w' V * * ' * * .. _, g .x N Y*b p g,
- r*
4, s mW
ri ..g tu ph ~'- ' .~ f. be exercised to avoid this phenonenon. . ;.[ 3.. ": O - 2. If it is' desihed to avoid only the' type of cracking observed. N 4 ? M in the first failure of the safe end at Nine Mile Point,, _c 4 mf y then a different course of action is appropriate. It'is my [ , r.J,.. * '. operating asstunption, and I think it is the only reasonable' L
- p.,
one, that the cracking which ms observed could have lead to.. '? .;c,., a large' rupture during a thornni transiens. Ibwever, it'
- f. c
.is not appropriate here to cotment on the consequences of ^ 'g Your interpretation of this is auch areL. - n It is my contention that such' a failure can%',r..a,.. p ( the large rupture. R '6 expert than mine. 9' ' result andithat there is a reasonable posaih414 ts initial 1 'y, k"'. 1A stages would not be detected. Tims, it 'inot. ' sufficiently'{ ~, , f'., fee it.to be' detected before the major fe e occurs. F' f- '~. e. fy?%%, U.. J If the AEC considers' that such a failure must.be avoidedHh; T' & s? s'. ~ then I reccnneni that: c bdan"M ' t z e. ;- en - y 3 m. ..c )). 4 a.,; 7. f, - a. The intergrarular inode of sensitized 800 be considered ( ~ ~ as leading to the extensively penetrated cracking.of the- ? ^ + type observed at INP and IACBWR. 7
- a.,
i e. 1 .. b'. The AEC should require that all safe, ends be replaced _,_. WM :c ~ ' either with' solid so:tions or weld"overlayed inside' ~ ard outside. Mfi W 1 c. The AEC should require that a cordition for replacement . ifC - p
- d is that the licensee demonstrate (by test results) the adequacy of the replacanent material ^and prcccdure' l
. M. a t > .l,l through agreed upon tests. Incidentally, scme of these tests may be aircady canpleted. These tests"should~ EQ !.1(. ' demonstrate (l) that the new material will resist ^ SOC
- gi i
'N in' the fonn used'in the replacement ard' (2) that the ..., [T.' . configuration used will not be subject to prematuie ' ~ y ~ ;r 1.- Jy [.; 5 other am.v.iate ma:hanical requiranents..l shock, fatigue f 'J. ,f . failures as a result of therma 3 yb
- g.
- nd.. The'nost reasonable replacement raterial is probably' 0;f
~ Type 308L mld deposited materialisince this has had . 3Q. wide use as weld overlay material. Acceptable materials ) f M, would also irclude Incoloy-800 and Inconal-600 providing that other defects are not introduced.. The 308L hast j ~ ; 1.. ' the disadvantage relative to the other two materialail.. ' of a ready susceptibility to transgranular SCC in chloride NIk environments. The uncertainties in the latter two materials involve their behavior in the weld sensitized h condition with resps.ct to intergranular ECC in the water b(m E -0 envir naents. ' 2 .m M ~ b.e + i c n. 4. 4 ,y g-sm ~.. - - -.ef 4 a Q.J~ 3 _ n_ ~ ..r *' < : l
3-. N)1$cing the repairs the syste ' ust be hydrostLtically. i n. wa t.u.l. n s : l )7 f a d i N I believe that in its regulatory capacity the AEC has no choics )
- py) but to require repair of all safe erds. The pattern of oo:urrerce of failures contains elements of frequency ard unpre51ctability y
which cannot be denied. Regardless of the detailed rationalization W of the cracking ard attenpts by various parties to obfuscate the issues, the fact renains that this cracking is cccurring with grt it frequercy .o;. "y QNP, IACBE, Oyster Creck) in sensitized stainless steel, that' the J l crackiry is extensive, and that it was not predicted. Frau a j management point of view one nust corclude that, regardless of,the. %+d b stress analysis or any other int *4mtely smalirrating fixes'ori. $jbf5 ;, calculations, there is a high probability of failures occurrirg again; 1 (.7 further, there is the reasonable prokhility that one of the failures y "*Y be ** o9$n ended type-ypg 1 f4 Secordary Rc:ommerdaEions 1 ,W.. x 4 -l There are a number of other reconnendations which should be considerai at this time. While they are rot as imnaliately critical as the primary 2;t.a($}, l reccamordations above,. I believe that they should be seriously considered !l l and chould form the basis for action taken in the near future.. um I r; La ; 1. Consideration should be given to the development of improved i alloyc for service in reactor vessels ard pipirq. Those present.ly 3; l availablu montinue to exhibit problens. Two areas in which ^ work might k-urdertaken are the duplex alloys ard the high >j pp ' i purity ferritic alloys. t e.,.g 2. Consideration should be given to shot peening inside ard outside l ~ t surfaces of stainless steel. This would prevent the, initiation I .. cracking. While there are aheys questions about coverage of shot
- .g%
peening the probability of a:C is greatly lowered. - w, e.. ~
- 3. ' consideration shoulv. he given to an exterior coating which would inhibit SCC fran tha outside in.
4. The criteria for leaks of valves, fittirgs, etc. in operating f g< plants should be reviewed ard tightened. In my opinion there (' *.t - should be no dripping water of any,kird anywhere near the reactor or its canponents, j n., <,t } I a 'I, 1-I .g i t ,h .l. .c. - r 7, , my-)we s m, ' " ,
- 4 0. [ k.
,4 'I
g~.e
- 5. b MC +'^jld conduct a p enram to dete[ 'n the effe:t of y
l' crack extent (depth + circuMercatial distlhtion) on tic @y(. f
- node of mechanical failure under appropriate loading schates.
~ This is addressed to the question: "Ibw ad uMer what , Mjd : g corditions would the soc crack propagate under machanical loads- .,s g' expected during service." s t ,y y, , mg g, c (:,1 6, There is a reasonable incentive to determine the extent of the -Q transgramlar. SCC on the interml' surface observed by BNRL in y p exanining the interice surface of the second core spray nozzle. g n$[^ wh. Cmments an! Justification in Support of Primary Reconnendation - 7; 9 f- ' ;. y g Eh ~.. s 1. The first question in W mind is whether the AEC will consider 5 f p-? that a conplete break of any safe end is sufficiently serious %i .4[- - y3 l 'to pose a threat to safety. If there is no threat, then there is no point in promulgating any required repair ' action. ^ ? ,?-4. 2.- I believe that there is a reasonable possibility that the 'intergranular cracking of the type observed at NMP ard IACBWR, can lead to a ccnpleto severance or substantial yaw ing of the n Q"%p pipe. Such an assertion is basci on the fact that in both p' cases the cracks were penetrated 50-100%.of the wall over 1/3-N of the circumference. That such a defect can occur before ij Mi leaks are found was a great surprise and also a cause for F Th substantial concern. q 3 i .s 6$. 3. The critical questions here relative to the causative parameters .{ . g[.g*. ]. are the levels of oxygen and stress which are critical. For example there is the reasonable possibility that the cracking c 3a may have cccurred in the gas phase where radiolytically produced J p ', oxygen lad accumulated and which would be saturated with water. However, for various reasons I believe this to be an unreasonable
- 9.,%
Yh10.O assumption. For example, the deposition of copper in the j IACBWR case suggests that water was in the crevice for a protracted ~
- period of time.
I suggest that it is, practically speakirg, . unreasonable,to ascociate these failures exclusively with these gasified regions. 1 -s%,9 With respect to the variable of stress; I believe, along with @{Tf'dd With others, that a higher stress may be required to propagate the intergramlar cracks than in the case of transyranular cracks. ? N ~. However, I believe it is_ unreasonable to cut this effect too fine i with respect to deciding which safe end should be repairal. Despite "3j the fact that stress analyses are being con 3ucted on as-built pipes there is no clear certainty that these analyses are Mt sufficiently accurate with respect to what really exists. ?ik Q n.c I gy l b j w s @M ^= .:- r.: 7. y d
. 5-ie hj,. Further, there is no b sis for confiderce which suggests that I 'g@ these stresses will stay the same during operation. They may (g either increase or decrease. EYihi c 1 4 A possibly causative factor may be considered the fluoride ion. ,g Despite the paper by Ward, Mathis and myself shTing the virulence jE j of this contaminant in causing a:C of sensitized materialf there is no basis for suggesting that it was uniquely causative in %@g 7 .the wide range of cases observoS. It must be concluded for the g present that fluoride (as well as chloride) is undesirabl'e but l- ~ that it is not uniquely causative for cracking of the heavily ' dg , s #y sensitized material. M..,
- y(
g f
- 1 i
5. There is the additional issue concerning whethde the smaller- ..f cracks will propagate later. This matter is conjectural. l l' However, frcm the poi.nt of view of public safety it is reasonable A to assume that existing cracks will propagate, hit at a slower $p rate. This is the approach often used in the chemical irdustry 1 82 when scheduling the reordering of replacement m terials. Naturally, ; 2% - it is tappting to conclude otherwise depending on ones cconcmic i prejudices or on one's aversion for politicalinfighting'. I W($;Sg$.D l There were two instances at NMP where specimens were'Itaken ~ to assess the effects of stress: The flared region of the first ^ 4,,,s ' safe end ard the secod safe end. In both cases cracks were ohc m cd. Thene were not co dccp bat wire clearly of tle propagating -y g' type as opposed to the " disk geometry" associated with the external " strawberry" cxamined. Thus there was ro evidence for these additional cracks at the other two locations to be " running out (Q.w,gE of gas." The possibility that the cracks in the flared region ydi'A of the first safe eni were related positively to oxygen concentration s 4* must be rejected-in view of the narrowness of the crevice, i.e.. i ? the high surface to volume ratio for the solution would permit Wj I . the oxygen to deplete rapidly. Di
- ] F crack in the second R
The observatiw of the longitudinal core spray rozzle frcm the GE irr/estigation suggest that the [~f residual stresses may be unpccdictable and significant. 5' f QM 6. The strong incentive for repair at the present time is related also to the fact that repairs later, will be increasingly difficult g. ~ E6 Mf ~ owing to the accumulation of radioactivity. A failure several years herce will be increasingly difficult to repair ard will g1 { result in greater risks to personnel than performing the repair immediately. Khile this point is obvious, it should be considerci
- @fjA relative to the very high probability that. failures will cccur
- i. O unicss repairs are made.
x:he W Dh l i l )Y ~ a-, x= et,f,+
]&~
- p,.
' m',/,e ; ;. 7.z :I do'_not believa that a program of intcrium inspection 3 ] 3 i isLaufficient'to prevent the occurrence of a serious crack.c The initial cracks observed at WP occurrai within three nonths. -It is reasomble that a cracking system which L' starts' slowly ani rennin undetcetsi'early could, within, a short time, begin propagating at a very rapid rate. j ~ .w i ~. ~ mg E;. ^ b 8 There'is no basis for the fact that the reacta systera nust LK* 3 - ' he dynamic (i.e. heating, cooling, or otherwise contorting)
- for cracks to propagate.. The components can be essentially j
y. . still, and the.residtial stresses wou141 be sufficient forf h f f - cracks to initiate and propagate. y s;y. I J A % y y q g y p.g 7-j n - 9. ' The signifha of the transgranalar. cracks startirrf frat the 7 i jp
- V inside of the-safe erx1 of the second core spray nozzle at W P is not' clear. They were clearly caused by simultaneous
" - r presence of chloride-ions and oxygen at the 1-10 ppn _ level. )- However, it is not clear at what time chloride was present. - n, The natural question here is why was more extensive transgrarulary not observed in other parts of the systen, i.e. 'in the other A core spray nozzle. This problan could be reasonably associated L with only this second core spray line. However, this issue is not clear. There is an incentive for pursuing this matter, 4 further. _ 10. Sann of the protmtion agnire.t the prc= cation of trar,:ranuhr cracks fran the oatside despite the reasonability of occasional l-g> contamination probably results from the fact that the metal surfaces are dry and when cool and wet are not sufficiently t' when lot 1 warm for cracks to prvpagate. 5 t x ~ gd ( ' Concluding Remarks _ y,., Ei The stress corrosion problen is characterized by itsiircredibility. l ' After all, the material is " stainless" and does not %.igEpcot angy l occur once but,after. all,an.irdividual fallurep does not re
- ~
, j g&, g epidemic.* This subconscious belief has anesthetized both technical perspective of experts and the practical judgment of engineering managenent. I believe ? @$ it is time for the AEC to publish a decisive instruction that all sensitized f stainless steel nust be eliminated. Unless such action ~is taken sunmarily, the reactor irdustry jeopardizes itself in terms of a major failure. Thgt' such a failure would lie grevious not only to'the AEC but also tw country s ^ ^ already deficient electrical power goes without saying. T.i While the above paragraph as well as the entire letter is net constructed py in quantitative tcrns and is rhetorical in plac/it portrays the 14 seriousness with which the problen must be considered acted upon. ly*& It is never possible to inve anywhere near sufficiently quantitative information to l 1 ..WW}, make such engineering decisions.
- gfg
.I urge you to implanent t ese rcccnm tions without delay. h S
- erely, This is in jest tut represents, unfortumt R.
chle prevailing opinion. c g333 ,~ Consultant m e ___}}