ML20235C936

Part 21 Rept Re Consolidated 3787WA Safety Valve Spring Failure for Plant.Investigation Indicated That Presence of Precracks Filled W/Nonmetallic Deposits Combined W/Low Impact Strength to Cause Preservice Failures

ML20235C936
Person / Time
Site:	Catawba
Issue date:	09/04/1985
From:	Walsh R DRESSER INDUSTRIES, INC.
To:
Shared Package
ML20235B661	List: ML20235B659 ML20235C936
References
REF-PT21-87 NUDOCS 8707090590
Download: ML20235C936 (20)

Text

{{#Wiki_filter:- _ _ _ _ _ - _ _ ) w y-. - DRESSER r# ' -' ed ' I INDUSTRIES 'h.ww.4 INDUSTRI AL VALVE OPER ATION S 0 5 0X 1430 O ALEXAN DRI A, LOutSIAN A 713 01 j Tst. sis /e4o s a no O twx sia.ev e svaa vsLsxiso e4az cAsLc oivio l j l Duke Power / Catawba Station Consolidated 378'MA Safety Valve Spring Failure Report September 4,1985 8707090590 070702 PDR ADOCK 05000413 S PDR 8 8 8 CONSOLIDATEC HANCOCK DEWRANCE DSESSER IN OW S TSIE Se SNC.

ABSTRACT: The fai lut e o f two sa fe t.y valve springs occurted Duke Pows.r's Catawba Nuclear Power Stati.on. An extensive investigdtion . indicated that the presence of precracks filled with i nonmetallic deposi.ts combined with low impact strength to cause the unusual preservi.ce failures. The identical location and singularity of the fractures further suggests t ha t the precracka were established during the spririgs' manuf ac ture and processing. BACKGROUND: During the routine Hydro-Uetting of the 3707WA Consolidated Safety Valves by Dresser Service Engineer Mr Larry Comtr., t wo valves were found to have their springs t>roken on orie of their end coils, P/N OA796CRVN. These.prings were sent to _ the Drenser Industrial Valve operations Lat> orator y in Alexandria Louisiana for analysis. The typical as-toceived condition of the cprings is shown in Figures 1, 2, & J. The similarity in location of fracture sur f ace is readi.l y apparent, although one failure was on the " top" of the sptinq (as installed in the valve) and the other

was on the " bottom" of the spting. Since there was no obvious c.ause of failure, and ber.ause the failures occuteed 4 1 1 in an area that in norma.lly censidored to be of very low stt eca;, half the specimens wete sen t to the c.ptinq manu fac t.uror, Al.CO uptinq Industties, Inc. for a tandom evaluation. ALCn Spring's shott tesponse is attached, F.iqure 4, and al.tributed failure to nonmetal Lic inclusions. INVESTIGATION: The chemistries of both springs were checked and found to be correct for the "CR" designation. Fiqures 5 & 6 documen t the resultc which confirmed the SlB60H allay composition. Hardness tests were aslo conduc ted. The hardness of both spt ings was in t.he 44--40Rc t ange which is normal for the alloy in this size range. Tensile and yield strengths also were found to be acceptable. The test report is attached aa Fiqure 7. Impact t est. sper imens were machined from a full diameter segment o f a spring. 1ho longitudinally oriented Charpy v-not.ch specimens q-tve impact values of 2.5 & 3.0 foot pounds in the as-received condition. A hydroqen releif at

400 Degrees F for 24 houta made no difference, and resulted in impact values of 2.5 & 5.0 foot pounds. An 800 Degroes F tempering for 24 hours increased the impact values to 7.0 & 8.0 foot puounds, tp.it decreased the hardneas to an unacceptable 40Rc. The char at t er and origin of the fracti..tres are shown in Figurec G & 9. The fractures are very brittle in nature, appeared to have originated in the middle portion of the flat, and did not start from a single point. 1he origination region in ohnwn by the arrows in the figures. The fact that failure did not irti tiate on the inside of the spring indicates that stress was not the nnly factor effecting failut e, since stresses are the highest. on the inside of the coils. Multiple fractute cites and a band of fracture initiation also point towards other factors. Microscopic examination of the ground npring flats detected unusual surface checking adjacent to the fracture sites themselves. This discovery is shown in Figure 10. Dye penetrant testing, Figure 11, further accentuated this obsetvation and showed t. hat this was not simply cutface gouging. A similar indication, although not as extensive, was found on the tetained portion of the other spring and is depicted in Fiquee 12. As an additional check, dye penett ant testing was perfomed on al.1 ground cufraces of the retained woments. 't is impott. ant to note that indications were ordy_

found in the ftacture regions of broken ground endc. A cross sectional mounted specimen was made of the region shown in Figurn J0 A photo-mont. age of a typical surface check in c.hown in Fiqure 11 The unusual wide and curving natut e clow to the spring's morface was in stark contrast to the narrow and angular Inwer portJon of t.he fracture. A microsttuctural evaluation was then :anduc ted. Figure 14 depictu the typical fine or ained tempered martensitic microstructure that was present. No surface decarburization was found, nor was any " white martensite" that can be formed during improper grinding. The next obvious step that was needed was to try to determine the age and origin of the " surface checking". To do this work the Dresser Scanning E]ectron Microscope (SEM) wit.h elemental microprobe analytical cababili ties, located at Dresser Security Division in Dollas, was used. Mr. Ron Gruener from the Valve Operations Laboratory conducted the investigation. Figure 15 shows how t he surface checking ct ack appeared in the SEM. An n00x magnification close-up and elemental ccan for the element zine is shown in Figures 16 & 17. The presence of high concentrations of zine in the surface cherking in important since the zinc indicaten that the

cracks, or at least the defects, were present prior to the zine phosphating of the speing, which occurs af t er rate and mag particle testing. Following this same fracture zone to even deeper leve.ls reveals some additional interesting points. Figure 18 shows the continuation of the fracture. Up untii the point shown by the arrow, the fracture is filled with a material that the SEM indicated was of L.he same composition as the base metal. This indicates oxides that are for med at temperatures above ambient or service conditions. Such oxides form during working or heat treating temperatures. Below the arrow, the fracture is unfilled and markedly angular in character. A 1200x magnification photo, Figure 19, documented this same frac ture front as connecting with a normally present manganese sulphide inclusion. The specimen was also broken out of it's mounting media and photo-documented. Area "A" in Figure 20 1s the zinc phosphated ground surface of the spring, and area "B" is the cross secional region that has been examined in Figures 13 through 19. DISCUSSION & CONCLUSIONS: The fact that both fractures displayed nid nonmetallic inclusions and cracks is readily apparent. However, the

I cause is.not quite so obvious. The following is considered to be the most logical explaination. 1. Dur.i n g the hot forming of springs, it. often becomes necessary to f latten the endo or bend the bot tom coils upwar d towards the first full coil. If this bending is done when the spring is too cool, the coolen outer surface could cr.tek or r etain suf ficent out foco tensile stresses that quench cracking would occur during heat treatment. The fact that the fractures were only observed in the area where this i bending stress would be expected to be at a maximum, is an t indication of the likely correctness of this senario. Additionally, if the nonmetallic inclusions were present in the mill delivered bar, they would be long striagers, and not the singular and perpendicular t.ype that were found.

2. During subsequent heat treatment, the defects (fractures) then proceeded to fill and grow with the high temperature oxidation products.

It may have been that the fractures were not easi.ly discernible after grinding. 3. The agressive acid cleaning portion of the zine phosphating process then succeeded in eating away at the outer regions of the oxide layer to leave the unusual ditch steucture shown in Figure 10.

4. Once installed in a valve, the high spring stress.es

combined with the wedged oxide structure and low notch toughness to result in the complete fractor en that have been observed. /'O Richard D. Walsh, Jr. P.E. Chief Engineer, Materiale. & Processes

, g.3 e 9 e s m t r.w,; fp. +! 5* 'R ~f 4 9 t wu sa c y my :.. c., '. } g,4 's n_.,9a ' &'y~; \\.g& s u ,_4 e u[k '. ~ ~ ,me ' ,y g.@ !h ;&~ *:g'. .v'" y g ..g g j ~ ..,t ~ [ y ' '.g SW ( { 4 s.. _ Ik. s b.f.' o Figure 1 Figure 2 Spring Segment Top View of Figure 1 Spring #1 J 'T,. ik-f%.. sr y5 $r ,,k-A 3., h j c ',.Q.'.T bth-Q g>R'.,g 'e i 1 n. g . ~.. y .y>. 4 l y-figure 3 '4e End Coil , ks...' i Sprmg #2 %r s. k 'h d ,7'* a 7, ,,,A,.

il Q Jt?i Ln a.~. ! N<,.W '1vtd.

s

8 LCO SPRING INDUSTRIES I N C. 23nostatsr a suctio ca cico sts..ittisois.o ii,. 3is.rsso 3.,3,2,2.s.oo

n. o. sex is..

August 5,1935 ] I:D-1 d u cau i 65 -c37 fir. Joe P. Smitn T-78 Industrial Valve Division { Dresser Industries 4 Dox 1h30 M- Z.LJ 4 k hv Alexandria, LA T1301

Subject:

P.O. 28889 h Aleo Pef. 23h6-1

Dear !!r. Smith:

Please refer to your letter of June 18, 1995 and follow up dated June 25,1905 Unfortunately, due to the extended time interval (over 7 years) we have not been able to locate our original order file for these springs. The spring se6ments sent back by P.on Gruener have been checked out by our manufacturin6 and 0.uality Control snd they advise the reason for failure is non-metallic inclusions. Very truly yours, ALCO SPRI'iG T" UPTRIES, IIIC. C. Harrison Custoner Service !!anager D g\\h U Y@cd s 5 gi Figure 4 I

a ,,. ~ l. . e s,; ..')n4;1 ab b'= 1. 4'L i

s...

i t..1 .$..,,p. - . _ ~ - 'y^f 6.' l'.EUt.'i.M I SijF'liR-- !gU d l l'

• • 2t&$$$$$$$*4*$3***

H.EUEX $ fhh E.4

• l u

8';P. .012% i ) MM

1. 03";

g ~e ; . !}. ,1,,. ., eir% FE 'C 't!. h.4 (Uh GQ ~

5. u 1

NI.125": i;v. CU NI .12P'% i E ..y fF. .-t, s- ' ch.i '. ' g'p' 4 i I"A .,.i

v. !

i f d V i 1. l n.v g.) l1 PB P9 Y y ZR dR i IIS t16 NV. OS'<!"; MO. O'P'2%

• • • • A***************

4444444334444g;g:.;.: l i r ). Figure 5 Figure 6 b' Qualitative chemical Analysis Qualitative Chemical Analysis .}l.,. SPrirg 41 Sprirg #2 I e 9 ~p e

t.

f, i. I e I h...'.. + .+ s-e k ie s

r:...._,,,,>,.. ..L,> %,gte Ce.c...TI ,Ie.. ... :. a - -n PLEmo! :: L0u t sI A..'A i r

• e l. -. -.,, cs t-Fe s.-* r..re.- T sleiv-

.,.. _... gg i . o, e., !~~4!9 r# %IFICA.TIGN 1.+0 ? ,a..,..

e.. n. 4..... a.

m _..___r<.. 2........ i i _.... ..,, a

.y -.

e:1 .... c.a. ,.4. 4 4 4.. 4 %. 7.

f. s... :.. t. j 1.3.r. a. g c ; q %:. r...,.

.;, =, 4 .5 .?. s.r. q, g 1 ,t.i ., s m a. .,.' a .i. r a.... .. s m. .g.. ...e e.....

e...

.'g W Figure 7

L l.h kl I r 'M ,I.;, t e . ') ) ' t (- ),. ) ,TJ,'9'; ( i i f ~.., - - Eigure 8 Fracture Surface Spring #1 Arrow Indicates Origin ( ' t y ,..; y .a a.,. l .S.. ;. .+*'o, i ? ! ^ \\

Ic

-. L _*?.,(,a p'.W.. ' n~n-).< j .c .x >fg

' ^.. ;..

.y +t * ' ^ ;., l f:.Q;N_. . 9 I l Figure 9 l l Fracture Surface Spring #2 Arrow Indicates Origin

Sk !~i '

c.. x, Q. DV, l?;."~'.

.. ~ ~. .s s Jg,x r - w.V '. v,,,.. &~oeg'M: s-Ps.y % Qa. - y' .r-pp a % - x - 4 z *1.. & w, 4 y ;t,.:::,, s.., s., w wr,., \\ ~ q, % 9:. v Q:... h

• 2 :,

v v g.~h-{p;,' rg W. ~ ' * ~ *. I6 D#$dM../y,*.CI[MU.~ ._ m n ~ n:., ~ ~ g. f.. ;.dlW.. ; f, '.'d ',_^ + x % y ~., ,', ~ ;,'. w,z,.. z Q'<., , h, *4, . 3,,.,$ n. g= r.' 2 '.h . 't 4.r, p Figure 10 Microscopic View of Ground Flat Surface Checking Visible w... ..-t... ..i q f ' . ; D,.,. %,' e [.. .: W. .c . -o - ,Y. A.:. "., ~ - :... n -.. r,. u,- ,. :.. (.. umi- .. n , 7,( . ' a ;y '.

~...

.,

~~ -w - Q, r :..a ~

• y. ~ ~4

~Jy ;,7% - s,.;)

p. -

• ~, -.

.ra ) Figure 11 Dye Penetrant View of Figure 10

- e. n, g',', m.,.. m/ ).;,: Le k','Q!."s'a.g.. - ' ' ' ' - ~~ 4Ggd kdQ af . v., - t,

• ]'( k/4 ' ( q.': g. i9 5, < y ~, s.,.

t i,, j - ~ v r y". y ..,. v .... s'sl-f.'. i,b.;, *~y"Q.% ? L,' Vi., \\'e' "., ' "Y 'h. >t. b. d

w..8s, e / y(a.,.7. I,.~~'){ a %%q.y f 'js. -)*; I M)l.,.r,? ? JT.

gV t ..n? t". !y-.

w e

y a,. fn.,7 <. + .s. . e a,y..... 7. g.. .,y,. >. ;.( e.

p...L:

. 4 a: t : f.',,*' *.. ;7.,. > ;, ' w '. p'. g. v:. , C. ~4. v ya. ';f 't{t:na. y; J> ;;tg,3. '.'v',-l, t, f g yg,,, ggi,,y.c'. ,4,9 4 ' 'f}ll[(4;ph,Jr y,

i.,

4 i Y .. L, n% ' a@, p." X.;ic., ? 3..,,' _ -...i, 1 m... Figure 12 Dye Penetrant Crack Shown With Arrow

"Y cQ.1,. s ~.**' ;T,t ? *f, %fs ' .~ ;; :,, ~. ' ;. ,y.y'*, ..e ,? Y. IET fkf ' '/ r s.I * ;I,; & f **. 4 gl ji (,! p/ L'. s p;6is s,

p. c: @g e,': %,

O y..:,,. ~ u};f.g4 g. w> ,..,M sl, 'Q O.b 9 b .,g ,, 8% ',, d #4,, A e .s .- {l h,t? e i 4 , ~ 2i %,s

• n t

e e 6 4, a g f. .*s., d,

• • ]

y ',.C

i i. *., '

(,2.~.' E' '. '\\ L ;c.i*,

• • • j p e

m. e e-y _- ( !.,;c., b

• i<:

,g.g a3' = . c C.'g4'o , f ?}l,1c. C O 9 4

• j1.7 Q:q, 8

+' j. i-l t i 0 I i 'n f ? W ..t' ie $w, :.: YY Figure 13 ..emp e

l l } Figure 14 Tempered Martensite Microstructure 300x Magnification Nital Etchant l' ., s

"

_;jj.

~ o I,.

• Ny

~ y T' @lm-Figure 15 I . SEM Photo of Surface Crack (Cross-Sectional View)

s I i, J i '~ Pvg Figure 16 SEM Photo of Surface Crack (Upper Center of Figure 15) .. m:.;. ;.,. :. - . :;E.- g - ;-.. ,f.;.

.,..','l

~~ ~

f ?f

-c. T ,M b .,r Qf % ^,.}$~

• y : L w-

. rg *. 7:.;,*p Figure 17 SEM Elemental Map for Zinc Bright Areas Indicate Zine Phosphating

Kt%:l"' . :g.,q[.y:vwlq y ,q y: z c -

,<.

g g y o'

.x -4 c ! *tt j 'y -. . o-ea. 'h. $f a%?pi s.o,..p m ;w :. .. w '*', $f,'f;iT i' Wh . f, ).). 5 lay V a i .r ...s 3eku x23e es es275 oso Figure 18 Continuation of Figure 15 .w:: v: r :n ...,,. ~,.. f '^ +_, L . f,!. fu ._,M~. ,; 7 ; M, e., :y' A I.. . e.

w. k,.

g e # .w ..,.y 2 .g f[.

4.,.

ps,. T . s / rE ,:S S% I ..e s a.

k..

e e..- <kEW:1" F, ;..i GQ., wp w.: ,,;z.g l-f*qd;p f.> 'g' -_,. fjt:a ; 3{,y i ey .; u w., 36s 3eku x12ee M se7.08275 oso Figure 19 Figure 18 Close-Up Showing Fracture / Inclusion Interaction

- myg ,,n ., ' '., d '? $# f f J...?l;;.; > o,.j .,' '?/ ~ f .D?fg }.y!!. k*f <~ , M.,, ' 'y- '.'l ~ r. 4 1 ;ry ~' l' :}'S ? - . ;/C/ F _ \\ - % -< '[~' ,', i - Figure 20 SEM Photo of " Surface Checking" Area (White Spots Are Polishing / Mounting Media) Fractures Visible in "A" Dept. Visible in "B" . -.. - -...}}