ML19268C034

Rept on Investigation of Weld Imperfections in ASME SA-312 Double Welded Austenitic Stainless Steel Pipe for Compliance W/Nrc IE Bulletin 79-03

ML19268C034
Person / Time
Site:	Wolf Creek, Callaway, Sterling
Issue date:	08/09/1979
From:	Cahn A, Koepke W, Stanley C BECHTEL GROUP, INC.
To:
Shared Package
ML19268C033	List: ML19268C032 ML19268C034
References
NUDOCS 7912190210
Download: ML19268C034 (115)

Text

-

(

a 8/09/79 t

'l BECHTEL NATIONAL, INCORPORATED

(

MATERIALS AND QUALITY SERVICES DEPARTMENT

^

RESEARCH AND ENGINEERING SAN PRANCISCO, CALITORNIA JUNE, 1979 Project Number:

Generic Titler Report on Investigation of Veld Imperfections in ASME SA-312 Double Welded Austenitic Stainless Steel Pipe for Compliance With NRC I & E Bulletin 79-03 Prepared Fort A. L. Cahn, Manager of Engineering Thermal Power Organization Prepared By:

,_C. Stanley and Iff foepke

/2

/g f

Materials and Q5ality Serv!ces P(partment

\\

Research and Engineering Reviewed By:

f,5 g/

C. R. Sefimid t N

Heta11urgical Enginearing Group Manager Materials and Quality Services Departe.ent Research and Engineering Approved By:

1. W R. C. Bertossa Assistant Manager Materials and Quality Services Department kesearch and Engineering Approved By:

W. R. Smith, Sr.

Manager Materials and Quality Services Department Research and Engineering

(

Log No. B100-57 g

79u190 t'

e s

C,s..-... b 4*gDL;

_a.

.s o.= ~

..w...n-Section Fern 1h. w, O.

....eT >a.T.rC...

i.,

w.s ALSTRACT..............................................

111 I.

,,. 0D,JC.13,.1 1

..A p.

.c. +. n.. tat..v.

2 III.

CR:CLUSIO!;3...........................................

3 IV.

PI COM:G'!3 AT I c h' 3.......................................

4 t,.

I, \\.. S.Ic.,4.. v.

A

.J..,>.

,u..

1.n.,.i 5

J A... aa VI.

A? ? I!; D I C E S 1 a nd 2....................................

25 i

I '...'. M.' " 'LP.~,~'$."...'"_.'5

.3 T.2.9. tr.. 9

7U?r9 P.ee

_;vre 1 - Cross-Sec tien of a Prod uction Pipe Veld...........

6

!;ur: 2-Ul t ra s t.".i c D:m in t io n s ke tch<t s...................

10 Ti ;,:r Photos o? Lur t Tested Pipa 13

...A...a

.a.

Table 1 Picte Supplier's Certified Cher.ical Anal:

10.......

9 Table 2 Mechanical Prnparties of Plate P.aae M.stal and Pused W 12 Met 21..................................

9 Table 3 Su=ary of NDZ I;esults on Fvur Labor s. tory Tes:

Velds.............................................

11 Tcbic 4 P.? o a of GT Perfor:ed c t Pla.c Tc.:t Welds Vith Know.i CLP.........................................

11 Tabic $

Machanical Propertian and Crosc-Sections of labori ory 7e5E U?idJ 12 Table 6 Me:h: nice.1 Proper 'as.and Cros -Sections of Laboratory ' lect

~.ds.............................

13 Table 7 Cheaical Analysis and Muhanical Properties on Youngr. town Pipa U:ad in Burst Tes:s...............

15 Tabulation of Eurst ler.;; Data.....................

17 Table 8 Table 9 - Mechanicc1 Propertica and Cross-Sections of Eurst Test Welds........................................

19 Tabla 10 -- Evaluation of AE.".2 SA-312 Double Walded Pip Fulleen Power Prcducts............................

23 Tabic 11 - Details of RT, UT and Etching E::c::inations "MorEt Cacc" Pi C 24

?

i 11

, 3 e T.'k. _%I.. ' ' '..

.$.%. _2 1 invas:C p.: ion ia und tri ns en to ch:-rnc: ari: 1 wC.d defa:. Pro to velop a pr-rdu:- f or vo'. net r t al'.) examining..lds in A5M: 't-212 e

.2513 -11;ed Austenitic 5: in'.ess Steel pipe. T.s progran v..

!.evelop ed

ar 17 with ta016 E Lulle

ia 79-03, of M;r:h 12,197?, en itled a

.a.NituJ1:.. 'Jeld uef ects in.it.; S *-312 Typ e 304 Stainlh i ': vel Pi; -

acl s l'.'.u i s::ured b; Toun pto ca '491 din 3 and Engineering Cw:in;*."

- veld i r arfectier.a of centern were deternirmd to be a result of

.:nterline lack c2 wold pene: ration (CLP) and ccte porosity. Severel

, ire weld; rnd test welds with C'.P vare exaninsd.

CLP could nat b?

da r eitt:d by either c1*.rasonlu (U ) or. iiograchy (R?) en.c.:ine.10...i wMa th"

,'o undused f t:w s ere in intin2 2 can:cet. 1.can wit.

d14 crete scp, ccaventional nethodu of ET exaaination could not detetr. LLP.

A specir.1 UT :echnique, l..

ver, was ceveloped which cc.uld it.dicate the prr.sence c# OL?, but or.ly when a discrete gap existed betv.an tha unf used

1:ss.

T' techanical praparties cf tes welds with intar:1cn.ly produced C.? indir.2:ed that the yield r>:rength of the veld w:uid use:

the AS!2 rateriale specifict ion rot,uir:nants even with es au.

'? JJ in percent CLP. All of the.e:tr.nical pt eperty requirene.nts war; vel:a wi:h up to 26 parcent CL2. Three burst testa voro p :fc.aed en

5;w 93

:13:.s containing r.p;.ro::itz:ely 13, 40 and 55 per:2 n C1.F. Tha pipe did to: burst until the interr.nl prest.ute retched saveca. tien thi ce d e-re :uir.1 hy::.:t : ;resau 2, evan fer the tip2 uith $5 ;.rcnn CLP.

A servay oi 71 veld c.ross-sections in over 500 f eet of pr:Jue: ion pipa vos perfor: )J to determine th+ nsxinua amount of CLP.

In e.ll of tna 71 e rc ru-7.ec :1:nc er:.:ined; nonu stawad CLP in e:: cess of 26 percent.

An cuisids ccasulting fi:

perf are d f racture unchn ics studica ard deter-tined that the criti: 11 defec: si:o was approni. 4taly 60 tercent ei the vall thickness at anbi ar.t teurerature usin; ASME III alloutbl3 at:23sas and ninir.u. echanical propertius. This report presents a detalle description of the work which v a perfortsd, cs well as the con:1caions drtvn fr.e. the data developed.

111

7.. p..,.. m.,.

.s In Septe'utr, IM s. def at**. vare discovered in len;itudin:1 pelf e of 3.'.-312 f > a 33'. r u s t e n '. t i., t t.' nl e ra st.el doible velded p*p-duria; c.

.dior.ry nit (ict > e; c.au tisc. ci c : rcu:.. -r : :::1:1 pipe :.ub;.a.;b1.y wr.lds.

'i n e pi p re catert;.1 van u nuf:ct. m. by Yar.;s ts.in ' elding cnd Engineerin.;

Co ps.y vains formed plate th at. ras auto genour, da:ble velded (tran bo: t side,) usir.: tue gr., tun *,. ten a:: s aldin,q (CTA*n') pre:ns:. Ultra m.; (';T) e:r i_..lan of t'ae.c.c.g t ud; n:.1 pip.- wei e s *. u pa. f o t ; e.! b; U1.ra L::

3, Inc., a m;bcontra::.c to Yce;;tm Th.:

op

:1:cu rf 2rential Lld a ' ' a cr.d: by Pulletn Power Pruduc*.s, e piping cutasswably f abricator.

d.e ;ip in; subn.+.2 %11n w e re 'ain.; f airi ?.ted for tha Pa.'.o Ve rd e Nu tlear Generat ing Station.

Further UI exacinations b:' Pull an and Ult ra Lt.bs, I t.. raulted in the rejectten of 74 :f tha: 171 completed and parti!.lly f ab rit tt ec pipin; sub.Lse9 mica.

Pullenn reported that longitudinal ueld defecta ex: ceding the UT eccept-nces :tc.ndtrds of Nce-25f 2 were discovere; in c11 si;. heats of Youngstown SA 312 pipe which had b-in purchased for the Palo Verde contract.

In sc.a inat:.nces, 2.T m perfor.;ed ta ve: fy the resuhr. of t..c Lt exa.-

t inn ti o n';.

Bned up;n the re t.ul:s of a limited UT at.d T.T cv11'ti.ti. n, the incier tica. were detha:iacd to contist of porosity and centarlina lack of

+

.2 t et a ton (Ci?). A sintlar percentage of rejection vca alsa reported

.malle r quac.tity of pipe subasaemblica f abricated by la?.inan for the c n t.

San L af ra 2 & 3 ::uc1ce.r Cuoratins, Stati n.

The uti:l ty ctr.r.ac.ies f or both projncts reported the a findinac to the Nucle:r aegulator/ Ctenission (:3C). Those re,> orts resu'.te:! in th2 iss'.c.nco of the Nuclear P.c3nlatory Cocmissica I & E Eulletin ra. 79-03.

Bnst:;lly, the !P.C Lulletin. required the inst $ tution of a pro;;ra:.t of v d.

.;1c emainct irn for tiu loa;1tudinsl walda in A5h2 SA-31'l pipe w :in;tured by Yo aptou:.

Since techtel is ar.nociated with the design and constructior ref several nucler.c power plcats, a decision was cade to c.pptcar. this proble: on a generic basis, rathe: thaa on a project-by pro;ect m sis. Aa investi-gation was instituted by Bethtel's Materi:1s t.r.J Quality S :rvicas Depcrt-(?SQS) to cb;rntteri:e the vald defects nn', to davelop aa ac:c; table tent nondestructive examination (ND*S) proccJure for detecting their proscnce by v.Autetric examination.

i 8

env u..

'. thv

~~*

'he invest!:stion.as ir-fated by gatherin; ir.forn..ine frma the pipe atnrtal tor.ufacturs ar d th-r pipe ruhusenly fabria. tor cnd by U: a t.;!

c.< : :... tion of pt.te,r p.us ::ntaining kaat.n CLi'.

L.the UT or si

.e t e :t e d C!? in a pi: r weld with 23 per:. ant k:.dua CL? whea it uns exa -

~

'ne; by Iachtel. Ad titi:nal productiot. pipe velds and sp tially tede jlste ar4 pipe toi: walds were elso ennained by RT and UI.

Although CLP cas cut."

%.-d by < talle;.::p.:1c ensaia: Lion, the CL:' ccui, not N tete:t-? ay RT c:le.2 re vaa a p:. butvan th i tn enf

-d f acca.

CL *'

^

ca.11d na he detected by conv2ation21 r. hear vava UT techniqu2s evaa with s 3up between the two unfuse1 fcce2.

Tours va.re cc:.dect.ed whi th confirmed that both the intinste cant :t of the pipe fccus and the orientation of the tefused ares were tr. rensnns f or the inability of UT to detect CLP.

A specially devolc?ad UT enaninr tisa procedura

.n de t ect CL?, but caly when a cis.-. rate gcp exista betweer. the unfu.ied facas.

Jelded plar.as vNra prepared to sist:.te the velds in the profuction pipe cor. ainia; CLP.

The CL? ranged fro 14 percent to 47 perennt.

Ze ails t e r. t e tad coaventianc. UI, specit.1 UT and P.! cuaaination; were p ?riarmed on ths.ae celts.

.11 of the :cf.anic 1 property rs;d renents of.\\S.T.

TA-311 va re =e with CLP up to 26 ptrcent. Tha yiel; strer.,th require-centa vera.xceedt.d with CL2 cp to 47 percont, the elen;ation require-

=ents vete ae with CL2 up tn 22 percen. and ths ulti.cte tensile strength requir.r ents were net with up to 25 p < rcent CL?.

Three b ;t testr vere perf ar-M on pipes with actual CLP equivalent to 15, 40..ad 55 percent of thi "c.11 thicke Mr.

Each pipe van 12 inthrs in dir. net er by 3/E inch wall by 46 inchua loa.:.

S;Yror.tatic preart a was exerted within the test assa blies until the pipas burst at pronsures of

$3 J, 37CJ, and 3000 pai e,:spe:.tively.

The c o.'u-ra r.u.'. red tinI:.u.a hy dr o--

static tc#

pressure ia 852 p21.

An evaluation vaa =ade of cpprontmately 520 feet of welded SA-312 pro:uction pipe to deteruine the cani un amouat of CLP. The evaluation consistei of polichin;, etchin;; cnl examining veld cross-sactions on all accessibic,pf.pe ends And in t.o envitiec unde by grinding. A total of 71 cross sections were c::acinec, 25 of

• .aa Shr ing cc:e degree of CL?.

The greatest CLP found was 2o percent of tha vall thicLnuca.

An outsido engineering cenculting fire (Aptech) was ensa.;od to perf ern a f ractur' an: lysis s tudy to deterr.ine the ef f net of CLP on the me h-anical properties of ASMI SA-312 pipe. The fre.cture analysis was done to deterning the eaxinua through-wsil height of an infinitely long CLP condition that can exist in a r!.p'. without caur,ing a failure under allow-able design stresses.

Pipe stresces were limited to AS':I Section 111 Class 1 allowable stress intensities. See Appendix 1 f or details.

o

• ~ }.'

_u..u_ute-

.-3 (e o m s=

End;ographf: (RT) extuinntions cannat detect cent rline 1 :k of-penetratie: ( C'.? ) la der.as weldad AS:. SA p.ie

.-7 ne two unf t.*:

l pin te edge s ra in in:1cate c:ntact.

b'h a t.

st :3 di. crete ;2p exisen '.:s wn :he ua user. f a u. 3, ra dic.;r sphic exacinatien :En direc CL?. T:.a size.. gap n ce mary in orcer to detect C;2 by radiogr:phy has not been deteceiaed.

2.

Conventional ultre.cni (U!) extuination :ct:.r.1;u.3 c er.n a de:ee:

C!2 La double we'.ed /R.s SA pipe either uhea tha twa enius2d plat eds.es are in inticate contact, or when ther: is a discre:c g4p between the:.

It ves feund in this investig::1ca the: conven:ioa21 thcar vavo UT uay indicste an accep:2cle weld in accord :nt with AS>l Section I'!, eve.n when :hs CLP exceeds 53 percent of tha vall thickness.

3.

A specially developed ultra,onic ensmiaation procedure can datect CLP, but Dr.ly vaen th we is a dicer:te gap bet cen the two unfus2d fcces.

F.owever, i. : a: ef tha pipra en2ain.d c. di:;ct.te gap did not exis:, even th ugh arazinations of thu end. of the pipe se::ioas and catallogrcph1: exttinations of cross-sections confirmed : hat CL? cid exist.

4 M?:h:nical pts. ettias were deterr.ined on plate veld utth inter-tior J1) pr a:ur_. CLP, Yi el.: st*".G:h3 te: the requir uants of A3X?. SA-312 even though the.* elds contair.2d as cuch as 47 percent CLF. As the CLP increa ms beycr.2 25 pa rc ent, the ul:iuste strength drops belav the ninizun requirements.

'.: elds with as uu~h as 32 percent CI.? had suf ficiant due:ility t - teet the ASMI Ss-11.?

min i *.. tiongc.: ion requiranents.

5.

Pipe, wi:h known CL? of 15, 40 and 55 perr nt withste ad internal bydrasta:1c pressuras of 5300, 3700, and 3000 pai prior to burst-ing.

The.sa results are several tines the. code-rcquirad hydro-static tes: pressure of 6L1 psi.

6.

The naxitu: dimension of CLP foend in enc =inaticas of 71 sections of approxima:cly 520 feet of produe:1on pipe was 26 percea: of the vall thi:*4 ness.

7.

The fracture analysis study determined that pipes in the diaceter range of 10 to 24 inches and with a vall thickness of 0.25 to 0.50 inches een have as ca:h as 60 percent CL? and not fail under the allowabic design stresses.

8.

The burst tests of the veldad pipe with CLP and tha results of the Aptech study indicate thLt ASME SA-312 pipe esa withstr.nd several cines the code required hydrostc. tic pressuru before fail-ure vccurs.

vr.e n.r.:..,4.. 9.v.

. ~

m.

.a follo :na recc.:a.'. ations are ofiered ba.Ted upon the r*c its of tha l

e

. t.sti.;s t.a n ::enue : s d tht.; fe i-resp:nse to r?.c 1 6

2. ZuLle-in 7?-03.

r?- operatin; pl :ts end pla. ts celer coas:rer.tian r(th in:t tlled pip *e or pipe which han baea fr.oricat2d in:a spoola:

1.1 Pipin; systens subject to draign heap stres.ses of leta than 85 ;4 ren t o f all;...'.. le s ::assa s n e... r.r-b.

a..

ined.

1.2 For pipi:; systenu cubject to design hoop stresses of 65 percea: of cedo allow. tale streaaes, or crestar, it is recon..: ended that the.alds, on n reados saajlin3 bcsis, be lo: ally grouni to a depth of 3/4 of cha udl dich.~n a. !

e::hri fo: visual er.a:Innion to deteni.es sch2;..ar CL?

exia:s.

It is re:.oemmisd that. welas vi:h up ta 30 per:ent CLP be ac:nyttd, bcaw upon t.tc.t3 chich shaa tha:. welda with 32 p-reent CLP cae; the AD:2 SA-312 Tr30' yisld ?.t: 19;th and du.-*ilicy requirnneats. Fini len3ths 7.ith expossa cada tay b.:

'.ifia.d by aa atch e::c.2.ina: ion of :!.?. waldo at e::h ena cf tha pipe inn;ths.

2.

For r.:rai?,ht langths of pipe hat have n.. baan f abricated into su a aasenblies, tr.e weld on acca end of t'..c pip-shacid be around, o n

."1 an:. <:aa. f.nud f or CLP.

Pipu with CLP gren..- than 30 percent sh.1 he rejected.

3.

It is recc-:: ended that the NRC Eulletin 79-03 be reviacd to:

3.1 Elininate the require 2xnt for volunst ric r. ndustru::1ve exan f.:.at.10 - of th= lor.,,itu:'.iaal v:ld s.

3.2 hequir e.:'.sinatica of pipe vald a in ne ordau:n with icta 1.1, 1.2 or 2 above as applicable.

_4_

s g :e-3,, _,, y. c-. s. m...y um.a. c -r r -. g, *

.;g -,,_

1.

.Y.:m.. 7_. ; s..t..y.,...

Li' t r 1.'_..a.. M. '

• " m a_n. ".

-ad L~.tel s.<dini,/.-

.llcr;.:.:nl e ;i.ae-vi:.ed Yon ptor 'aldin ; t-d Engint e :

.3 C:.:pr.y on - til 19, 19 7 '.e, t o ob.ain inf oraat; r e g a r di..;

nelr veliing and NUE practina s as well e.s.slditien il inic.natica to

-sai8t in Iren 1ficativa*of the questionabl* pipe.

'Jbe Yvm ;os t wel:..e pr;cedura specificatice had <.dient.ud a vid.. ran,a

VAG bles enc that in + c lec: Loa of r p ;;if i.c wel;.in;;....ine se r :ir.;.;

(vn;il recently) nude by tbt veldins operitor.

Ir. app ropr l.te a '.

selection of weldi:.; pa:. 2e*. cts could have r?sulted in CL? cad /or perosi:y it t ld.

UT v?a perf orn ed in Youngstuvn's chop by their SCE cr.b-

. actor Citra I.cb;, It.:.

Jo.:. atown t-., rar. urn:tured double V i'e!

J

i in t

:ardana siJa /.573

-31. for spproxt:r.:M y tver, years and A m 51.-312 e:.s_iad by UT sia:n i?7!..

The inf orcation ava'lable e.t Yu aratown did nut permit thn placemant of a time frr.me on J..a production period of the pip.: iavolv?d in thin investigatien er partit the isolation of pr

• nc :; 3n lo. vaich tr.y have questianaSie wclds. D win'. the visit, Y ou r, : t tv. r. i. r.ve L e e.b r.cl r. pip e w 31d r.1 pl ) with a typical CI' condition.

The actple ucs taken frca pipe ta:* rial, Ee.t 25337, Pipe T-2 thich had beca returned by Iull:.ar i; Yvan; ::9n.

Lt. 'lographic examination shwad that tha inngitudinal wall sean of this pipe ne:: ion con:c.ined 20 perennt CLP.Sub me.uen: I nnd 7.7 c a !.:2:. ion s perior,d by E n;htei were unnble to cetu:: :n;. lack et pene;rction. This said it shcwn in Fis :a 1.

2.

Pu '.h t' n N r P - &

• s Pullnaa Power Prod cca vaa visited by a Lect;el velding/.?r,llurgical c r. w. n. -

. Apri:

'4, IPO. to oa:ain f arth e inf orus. on eding; the entracterir. tion the wl.d d e f e c t s.

Radio,ruphs of sone len gitu.1:

pipt celda nad beca man by Pulitan in additior. to the seerca ult.rcso.'.'

exu ainction perf o:nad jointly by Pullman and Ultra Labc.

In sont casei volca we:e accepted by UT, yet they would ht.ve been rejec:23 by RI.

Radiographic film of son

• of the pipe showed:

Centerlina Icek of weld penetratics Porosity - isola:ed volds rand n3 in size f rom 1/16 to 3/16 inch i

in diameter Porosi:y - scaller volds in line alon; the veld, sca+what connected Porosity - teall, 1/15 inch, connsaterd by lack of ye.e: ration Seven radiographic filus on Ican f rc 1 Pullman which show these veld indications verc reviewed. Positive prints were ende and are now on file.

At Pull:en, two velda vere sectioned and examined without et:hing. One weld croxs-section was castined which contained a void 1/16 inch in din:+t: by 3/16 inch in lenrh.

Bechtel la:cr entined tni: cro:.u-sec:lon. -The second cross-section was not examined by techtc1; hoarenr, It was reported that CLP was present. Fullcan had perforced chcaical analyses of the veld cnd tht base cetal e.nd had confirnad hat the veld had been nede without filler t2tal, as required by the ta. trir.1 speci-ficatice, ASMI 5.s.-312.

i e

e t

si.ei...,

c ** y:l.. *.

1

,,,, ' ' " ' ~ -

o',

..,,,,.;. i... = i..

.;. z.-

s ~:

• e 3

.r-

• ' s d

g%

. \\...

'b,.

\\

\\

e.

3 lg e, P

.. % w 3

.p... at p.,

. b,,, f f ** !.' * *... -

2

.... -,,. m.., 4.t'.: r 7, v._, g M*

..'o

,,. '

• 9 : a.,%

as...

.,3 4

s..y.. ej*f.%.g #,3,3l A 5

f.

I f

~-

J s,

\\(

o 7

. ewyf.

-g e v.:.~., *. =. T,* **" 1 N *

• V ' # ' * *1

.y,

-,, e 3' ;

4;g.

.~ c

.v.-

y'. t c.. ->,c\\...... s.

,\\

. ~s

s. ~. r :., ',,.-

.s:., s

j, '$*c,

, h I e> d. 7. '<...f' i J,..,

a 1,

~'

~.

m, o

s

's,

L .11 4 '.::.. 1..

}l..!.',.,*

• . % ? ~, -

• ,, ^, 'i '

'nu -

s

• • . I l

- ',*;',....,' n.. : -:.* 1

.-.. x.*

.. T.,.

a. s. '.'.', *q :., ;,}. %^

,% ?,,. ? a "^'.1 b.

y..T

,.... "... * * -, j * *, 'Sj+ b " "', -

q

...:. 9 9.

l s

y -..:

• .'***l

.e..

~ '.""-*

e 0 b,[

• w...,^'

i

.~=s.

ab*

r.

v.

.s j[..'g +) * '

,y e.r,"l( k #.s ':.', W.. [.Q}

._s, s-

_ 4 7.n y.w%:.j.E.D.

L.-lL1

a;

.y u Y s2,

.~L

,.m,-

.-~#.-

, -y a,.,...

u ;. y4 y,,., c

c.. f, m.,.,..,

r..

.c s-

~.:.c.,

.....i w k :'. ;;s.o;>,,; V.

y 'n,

,s

... m

. n '. V.

. W %. :.9:, ; * ';

m.

.t

., : ~.:. L. = a. e... v.

a.,. -

.s.,., - ~-

. J -.

y e...

,m.

t m..

..t 7.-.

. 3..,. 'u 2...f-

, r. ; i..r..#.,7

.r p

1 g,,.,..

as~,...... w:.:.w..s.,....

us. w +.,e..-

~-..

u.

... n...-..,...,.,.....

. > ~ ~.

FICUP? 1 Cross Section of a Produc.tien Pip-Wld (Tub 2, heat 5337)

Ootair.ed ro: '.' or a tm, ( P r e v i o..u..s.. v. J..o_l d..."M i.u.!, '. -...O_r B e c h t c h M *) t t' ).

... '-T h e I _S M*.D.i e r e '. 4..'. a v t ;..: t. 3 -~..

.-..L. a c k O '

M*

w r._m_.. a t M '~ h i. ' C r.c.,.;..._.3. e c t b.

i..

.A.n.o r_o x 4. 4 t d_. _v 2 4

+

Weld P u :n c tic. a (G'..F ) L'a s S o r D e t r:...t t_d Bv E. *. e.._, ?.. M r '_s n._.".i c O r P.rc i n ::.: phi..:_ N.

im. io_n..- Th? h n_n i ?.e Pr m : 1 u '.d H i s

..d 1,a..,.-,. -... - -.,.. _....a..,.-1...

g;s_

e

.s.

Tonsile St: ength 81,000 psi Yield Strength 46,400 psi Eloegn f.on 25 perccat

p_.i.c. '..:.1 d 0. l e c_.

c-r.m

• 1 nus, o: pi, a about I? im h:2 long ch? b h a d 1..> t, retur ' d t >

rr.

n a by Pull -

vere el tu d fron Yom ; mim for e: o t :.:< t ! n 1,-

'hiel.

.'J 1 ti n - ;e.,;1tu?'nal vel.la 5.d b..e e,' c e t ed b/ U.I '. r n L 'u r.tru t'l rc ::ani.a '.on.

$: s,qcca: UT eu.ain: ;oa of tb v-thrc: pip.za l'ech t.el inllen ted tha t tb. w:1 In n ot t he r e quire' ea t s of.* OJE III

-2 *' 5 7..

A subser,urne nett.ing on Ly 23,.1 M./, o f repr e r.e n t e !.iv ic o f

.:r.r.:.:avn, L'ltr a Lna n:1d k:h:e1 in Son Trancinco ec.nc'c?r ? th* t th

1.. r 0.

• x p uf t nd ay citt L n g, was t.unpoet.ad chat 'b pipt nay av e 1 n a acce g 1'.;1 c in 3c co r ' ne n u t

• h.*J.:;i. I '.I, lD ' 252.

I:rw v e,

t c13 nrgler.1 m dnation confi re d that t.co of the t hree i! pes ccn-

. a ln e d CL? ', hich va c no t de r.ec t rl bf UT.

The pi; e sanpl : vere:

IM 39312, 7th l',

10 fn:t ; r.305 vall Tull reaecr tion. aid W st ?'J,0T/, Tube 5 12 fu-S 373 all 10 p ec n t C '. ?

E. u 5 :/ f l~,, 'i o S 2 10 inch,tz. s.250 vall 10-15 percent CLP

_S. A e c in i Tm t

• ,'s.1 d n n

w optcin: test v '.dc nu9:ered T., T, C and !!, vere prod'.ced f m a 4 'e n it,u o r t '.a *r-:: pi; e.

L y tiera velded vi t h UI/N vit1.uit if tier

.etn1 f:s

'.oth sihr v] h tl.e ve! Jins p trer.ete ra adjusted to ;;f v? very-n;, de;;nes of CL?. T1 -se pipa secti:;nn tre 3/u inch in thiebe n.

ad G coruined 35 perce. it CLP, t!. Id 1133 perce, 1:1d T $0 pe- : n *.

id.. d

'l. :D perm at.

Tha v al6e:' p.lates ver<... lutinn caaenl ed a t 19.Wp Iwc r f oll o :f by v/.te r <3 c hin,;. t o adt ula te t he t r e a t e.

givt n o r r;

Mu;::iva pipe.

'

• ter, six additi. al tielc's ut a ncde; three usinr, 3/8 ine' p'.ato and

'.a r usin,1/4 itch pitte vit:i varyic,; t'c;;r a s s CLl',

'i., u.1d s n c 'e

, 3/C inch platu, A-2, E-1 cnd C-1; and in 1/4 f neh plat, /.-2, t-2

' C ^ u r: 5.olution aanenle41 af t er velding.

Tc.e ebnical e:v.lyria of

' mne ple.ta v.aterhin, as chotea by the oteel produccc's c ;tifici test epcrts, are t.hown in Tabic 1.

Th2 rechanicci propertie:. of the l' ace:

' nt e ec:orial and the f use.d un 3 netal are cho in in Table ?.

'The 1%'

mult a la: these ic.ur la'mato::y t.es t pl a t e t. cre chovn in Table ~1 A

>tal ed t cin tr nsile te sts t.1/e con iucted; t u.) t o w.5 on sp: Peas 1:L ench trotw.t. c2 CLP cn shova dn Tables 5 ond 6.

Then, velda.ere 1so em..ained by s.pveial Ur procedures and by !!T (~'able 4).

'e remnt.v;ea of CLP ver e do c=..ed f rom the czer 'c vid h of CLP on

s frretur face divJded by th3 orf ginni pInte t'aichness.. All of the ansile npetinena failed in the veld.

.sti of the tensile specimens

'!ncluding the ones with 47 percent CLP) c>:hibited yf eld arrenath. Which

.cet Jtd the requirements of ASMC SA-312 Type 30!. auctend tic ntain!*us a el. The ultimate tensile 'strengtha and ductiltries decreact.d cs the ree:itabc of CLP f nerecsed. Tablen 3 and 6. include phote;;raphu of the

-id crace.uctions nnd a tabul:. tion of the occhanical propertiec the ter.t velds.

i In.2dition to tba tensile tor.u ef wel?qd plat a, t.o c'l-vald :r: 1 t e r s '. J e n.:ple,vera r &.

Th.

e.,-ci.

., vere 24.8a.

. in ;. lata r.:. C, '

tn e

.? J.; p:ccess

'. but il E e c.e::1.

Tha ;

xt,

-2; i m, ti re :

n.:a of

..:enitic ntsinl m: steel na w ?c us e-.

f o r th a t n;U.

t e r. t c c. ribed t'

3 CT! ' fu;t:.n p::. es vera run ca both sid2 a c tha plat:

^

achiev+ c aph.te fusion throu;;h th' catire :" l2.c a '. len,th and i L : s r .

J. rip cr.

c! the :a ;ile spe:1:er.

~ h e r n '.ta:.: :.. e d.a nical prope r'c wen tcmr at t to tha typical proper:ias of f32 3;4 au s :. r.

• A.- s

c. i n '. c :.., 3 :. -;.

aa resul:3 of th r.e astr.r.n tcbul :cl a ~t cc::.ar d with tr. b.aa cetal preperties in Tr.bl. 2.

U.' c:a.ic enninatix 33 sera performed both befora and.fter salution t.r.n e a l:(

and th: cute 02 its we 2 obtaiaed.

All DA'.: perc>en.au icas than "6

..: rc :..: cre cc 2pt M 2 'a acco rdance "':5 'n-2532.

The decuila r-;xedi.y, tha inve::' s tion cf ths ultrasonic n:.mtica proc t hra are 2

A?m 'i.- 2. Conver.:ioaa', che r: usve UT i,

not -2 b '. e o pick up th2 C;.P in the four sempl.3, G, H, F and E.

Threu ;h the un> of s 1cial UT t echniq u e. i, the Inch of wild peast 2tio van daacted f.: abau: the first ine.h of :: s lonCiud'.ual *.ald in each aumple, bu: it v'.2 not d.:ncta l ir the re:.u.n:!er of h" vald. Ta2 sp sci:1 UT technique it.iol' tad :he ta of a 70 de,;., a trar.:d.. :n r whici dirc :ed an t.lt ra.; u ed bc.a.2t appra:d ualy 90 dr ;;,ro :a to.he r h. r e c 2 th r.s CL?.

nadimrsphy canfi:-:3d tha prew w i of CLP f or tha ful. 1:c,:5 of the wold in the c..:ple uf.:h 60 part::nt L.P, bu: i'.T dete:nd C1' for only the first inch of th.: ue10.: with less.:

N.:ca n of CL?.

...a r u c.

. c

ha e #...;1: ::d ar are chwn la.c.bla 3 and FI; ' a 2 dupm a ta

'J e3 >.t tantions which ara de;cribad 12: thin z.e e :1 m..

It was postulued that the intimate contact of the t to unfused i tsi uc.a1 fnra,

• nich vu e used by uld cheir' /.39, ray per-!t uninterrupal trand d.u ion of sound c coas th s unf ua.ed cren c..d than the.:nfuaad c.ren c:m a a dc:cetad ly.. '. 70 conf'.ru this prcine, en additis 11 veld was rede with a 0.015 inch gap bet:.een th: joint curfan n cd with 35 percent of the will thicV.aasa uafuaad.

Both radiography aa! ths specint UT c.ncmina: dona date:ted the CL? for the full len3th of this m:1r Con tentica ri :.hur ww a U2.;hou d caly "sene inMc<. tion.:;"

hn >< r, none of -hace it.lications enreeded the acceptnece 11::lts of l'iK 532.

(An ad....1:nci G db. of gain uan required to obtain a railection cr.ual to 103 peraat of DV; per ASPI III requirements.) Thin test confirmed that:

4.1 If the unfuoad facas of t'au crea of CLP ara in in:i= ate contact, the CC' cannot be detected by c1'.her FT or by UT vhen usinp, either conventional or special UI techniques.

4.2 During UT c::aninstion by conventional techniques (as requir:d by AMI Ill), if the sound is reflected fro;a a CLP condition, the rajority of :he sound does r.o: return to th3 UT tranadun; ard is not displayed on the screen as n UT ind'. tion.

The The crienta:. ion cf the deree: appears to ;>rever - che detection of CL? by convention:.1 UT techniques.

g

11.o 1

.s a_

-.. '_ _c.4........t a.... _1. _3._......_ _. _ o f -..

e.-

3_:m

. 1

.C-1:n_

P

.S

.S_i_

.C_r 1.:.1

.C._u__M.a _C o_

.!.I.

.0043 1/ ' in:h 71r. e.03& 1.1' O.27.020.60 IS.17 8.5'.

14

.1' 22.003 19032 3/S inch pl

t..057 1.74 0.25.011.35 13.47 8.20.21.23.13.073 g. u.....,.

.Mec'.

l trow.rtiec, o f... '.... % e. e n,._1_ '.d F u a d W i._f % r. 1 Elo.;-

-.._i s.t l_e_

Y._i.a..l._d_ _.. _a_ _ : a

_PA T-

'nch Plate, Ls.: 90043 hec Fvts.'.

E2,000 pai 36,000 psi 6 0'.

71%

1.v h Plato, 1: 4 a t D C O '.3 /J..' - '/ 2 t + 5 tn1 0.,300 pai 3 b, 5',0 p -.

63 60%

inch ?l>:n, 1. cat ' ??32 h. N: al C.,003 pci 33,403 poi

71..

677, inch Pla:u, Eect 19032 /.11 'n'a.c-Matal 51,700 pa;.

39,3[0 psi 52" 667.

i 9

. ~.1

2. ( r-r Q

/

,. e

,.s

.f _...._. %. m

\\ ~ e*

\\

~~

,4

/..,--.-

,e

-.m.

l ', N T AN i..(. M ~

\\

(,

.s

/,...,,,,,,

N Coavtn:1.n nl UT usir.g a 4.5 dur < e t.tnar. ave.

Upper sketch r. hows how sound cen be t: r.32itted i.cr.n ths C ? interft.ce when the uniu~ed faces are in intirt.te centac:.

I.c..cr natch chaus thi: with a ;,cp t.w-n the unluced faces, r.out! is rell*::cd.. bu; c es not retura to the UT t' tnaducer.

- ty-y ----

,, _~.x.

, ~. -

~..A,,,,.s.--._

%'~s%

s

" ~~

....:.:. s ~ %. N s

/s i

.~

f...-........'~~m.,,,:..'.._:..,,.-..-.....~.,.

p

'.. /-

d =. -- ?.,,

-~~~.-~

N

'.. ~

Y s N Speial!y developed UT technique. Note that sound inpin;;tc. nn C1.? at appror.iuately 90 dei;rees, so that sound can be reflected bak to.he trat Sdu:or when there is sete caparation betvaar, the enfused f acc. If the unf used f aces are 1a Inti. ate coatset, sound ia tranc=itted across the interf cce, as -in the upper sketch.,

N :h.r...'_t.i.c _o.f__ t he Co._n_v_ am...t..i e nni e. n.d

.S..e..c.i.a..l.U.l.t. n e e..r. '.c.. F.~. a_ ti n.c t. i.on.s Pe r 4 rt e r'

-b 6

D e

.g,

.. _ ~..

.'_d a.

6 e

. =.

e.

• =

e.

6

?

9

.' J..(.*~ ", :w.;~;.'.

Table 3 Sun : tu v c f W / w r. '. u w y,

1thwate-r ~.. s :

Pe.'.

Conventional Special UT Plate ie.tual UT 70 and Sic Thichnt

% CLP

!. 3 Endin >,rn nh r 5

.375 33 CLP not detec;v' Det ec t ad CLP first 1 luch e

.375 38 CLP not detected Detecad CLP first 1 inch 7

.37$

53 CLP not detected Detected CLP firs: 3/4 inch E

.3 73 60 OLP not detected-Datected C:..?

full length Tt: ' o 4

'.L.e._s.u. ~l p.-c i ': I. - : t.ri e__.

On r.L.i t e _T, t,r.. '_.ds *J.i..t,h "nt v CL?

Plate UT per Specic!

Spr:cial UT Prneeduru

/setual

5., m.i..

II.'.3__ _> v ? ~,

~.

c.. d >. s. '.. -

.... '.C

.c.m '.

3.'.n c.r i. i n '." *2.,..

CLP

/.- l 3/8 30% for first one inch CLP for first 3/4 inch 147.

L-1 3/8 30- 40.1 for fir.3 t one inch CLP f or first. 3/4 inch 24%

C-1 3/6 40- 60% for full length CU for full length 477.

A-2 1/4 (Not par.orned.

.N a CLP for first 3/8 inch 357.

standard available)

. ')

.j i d it il 91 19 M M fl n it e~

h o C~T *.3 O.'s'

• )19

'C-2 1/4 No CLP - OK.

32%

_.Tc_b.l.. 5

',-.%..%.n !..S A..,.4 0. ;..v t.e. W. A.u.w.: r.u....t :...S t..a.t r i e u. S t e n t.' l.u. -.P...i.N til t.h ".T. c. v i.r "r o_ f_. -

='*

-m <.e. z v

.e..r. t_o_. T. n.u.i..l. e_ _;..

m. _n, a. "..I. d_...._.P.c

.. J n... s....u.. h

..1

._/.

u

.1... f. _.,. i...

3...._J.m de

.:......_.s.1 w..... _... _...

' - ~

W.ld

>,-l. 3./8" Plct-

• L :k of Pcnotrn,t!.on: Appoxi.ntly 1 '

)....r y"J ',

Wald Se.pl a ti

, o.

.u....i u.

e

,.1

.t g

. G. wZ2~

.,.,.Wf.,t.y,'O_,t 'i T.S. (cai) 65,900

^3,003 05 E00 -

s

. g-(pni) 38,900 30,500 38,400

,,a;.'.# Y.9

. j.

,s 3.

,..>.. r E...

...,... :.- +

.s,.-+.. c.. u. : a~. ; J..,:- -,... c... - <

. e.1 E,on;. ()

48 40 7.,

r-.

R.A.

(,, )

0 39 67

~

g

<..g...

P f,

.+.

D P*D "D)W 'M" l3a w.k L L

..n e..

y I-Wald 3-1, 3/S" Plate Lesck of Panatratica: Appuirm.aly 24"

.._s

.b.

- ':1 J

t 7.' '

.. n.:.., _.

...e w,

,, ~.

t ggJd 3

,.31.

r e.a T r.:

-t

. i. y g

i

.........',g

- ~ ~ - -

~.

..,.,.....,.,uw.~.".....

n) rim-

.gg, s

,e.

300 4

_.e

-...t f,

...."r.- ^...., p 3... v.

., s

,.,.,a, T.S. (pr1) 53,510 S3,200 85.803

..e

...,...~.;

..g*

4..

7,,

e.s Y.S. (psi) 33,000 40,000

.33,400

...,,, - =

. s..

s.n.

.;.y.

.. s " riom.

in 37-7 "a-

. r...

s

.y

..m )

,.. ; n.. ', f,.. R. A. (:;)

36 36 67

..o a,.

. I.e c

x, y z

r i

. -e y.

...u v.,

p. l.' l-.

.a3,.-

.. m,,

.n 4

.w,.f : E. ".

, ~,

v

,r.

.'.;,.~ Wald C-1, 3/8" Plate r

a. :c~,,,. 4, i.p ' g - (..j.N,.s '..,7,.

. 1.ack of Penetration: Approxicm ely 47%

i

~.m w.w ~..,.e 1 :.. ", }i,D:,. 'g -M..., :.

~

s..

gi.$y-Veld Sanple 1i.. ;

-.o 3.,,-

, t

..n fy;. ;.. p c1-

1. 2 Pan Ma.':a1

.,,3

... - 4.-..s.; rv. - -

' Y ('E~' [ Z.'n.d.[hs-

'. c.' '.. ;i. ; it 44,900 45,700 E1,800 e ~f.< a. ;2. dj-$['. T. S. (p i)l$','" N N [.N 7

.l_

j

[

.33,503 34,700 38,400.

n.'. ;', -. g.

r g'.. ;.g

.3,.

Elor,.

(,. )

4.5 S.0 71

-..Y.('O.': ;y. q,,M,y... 'r

- P J..*., N. '..'M.q,jiL^fR.A. (%)

21 20 67 c.

s. ~. ;.

..^.

.. c.r.

n..

6.,

e.*[t 6

'.4*

' 8D

,t-A "dA t,

g6sj VTN

.', h.0.*.m.h 5

4 8

• Tabb I.

- s t W 1,f i, o f_r,4 2.' '

."'v. o. _n_3 0 5../. m.m t ic S t r_. t..r. ' w._n. _S _t.e..n_i.P. l.u.

.Ma. d.o..L_'i c.h. i'_e r v ! n..a~.

w a.t..a,.;._ _t. +...

'. t. o_.s... e.. !. v. _...... n...: 1 :.." o E. _. *. __.v..'!_e

. o..f_.. 4 i. _t d A....

..1.. r...a

..d o. _l _u m.' u,..a

-h-_t _, f ? _ _u.. _P_h.a. r._,.

.._ ".. - - ~...

.....o_2.;_m

.L a_.

..1 r

'.a.' ' A. 2, 1/4" Pla c 1,ack of Pene tra: Loa: Approxiaataly 357.

W1.! S.aple

. 1

1. 1 e

t.,

a,.,*

T

r. 3,,

5

. = =.

~.. -

9..

. #...r..<.:

"i -

. s T.S. (p;i) 56,10]

80,0.'1 82,000

..~._.y-.....,...c. c s:?.d....s.......;.y s '..

....s...

v, s, ( p.; :. )

40,900 41, C. '

36,000

?. '3.. 7' P' $ f?.r.'.g.,:,Y. g y..'...,..f 'j Elon3 (4) 9 31 68 r,

r,;-ga. #36; p...* - (** g *- ~. L,.i ' r m ?'? R. A. ( ".' )

22 32 11

.s

,..t

.,..s.

.--~.

-. r.

.s,

.>o g

-e 9..

28

.r.

,,g

.')

't t..'

.g ii"*]D *]D'3l s

o o Ju o Ju SJ.

=

~

Wld T. -2, 1/4" Pla c e a

.; lack of T.. net. ntion: f.pproxim:tely 21' c

,a.

.c * -

4..

• , ' +

.g.a

".. t.?

't -,. < ' J,t

'31-- : ' U

. ;. 35.,,

, S.f

. : : :.;p.i p"--

b,.

t n;

y,,

-n w g,,

,. <.,.,,....c.,

..,., *. e g,.,, 3'Y :'

~

a.

T.S. (psi) 74,'.00

?., '.., 3 0 0 82,000

' i l',

C Y.S. (pai) 39,900 40,300 36,000

s.,

s i

, _"3 ; ", " ';

Elong. (7. )

24 48 68 f,:

a.

'm>

4 R.A. (~)

30 37 71

,..,p w.. a., p.... s.,

9..... #.,-, f.,.., ' -

m y,.,,,.,

,1. -., 9

- i Weld C-2, 1/4" Pln.t u u.-

,t

/_..p.~.np I.ack of Penetraticc: Appre:dcately 327.

"i '

.i.

. t. G.,.

y

..,..h.,.,

v..,.

~ s. <,

. ::.. ~9. ; b.:,

a

-:-t, - m.

%1d Sa ole

.,. ;..v.

a.

,' w :-..,.

.,;.r.. u, :..z< #..n,

.. e. m.. t.

4: m ;

,4

..,.

. :m.

, c -.

,y,. n., ;.:= n -.7 p i, p2 Base :'ata1

.-,.. e.,.. :-.- :i c..,.....,,,. 3.,r.m._ a_,d.

- 4..

..,. clv. '.;c'.

o 62,0 0 82,100 82,000

.27

...,. p,s,.,,,+ p,., N q... @..t;.;.:g..

T.S. (psi)

Y.S. ( p.ii) 40,( ]

39,300 36,000 i

. q p s t...; *

..,.,. (., )

33 u,

- m, g.

R. A. (~)

25 29 71 1.3-

. Pit a.9 n r s t T_e c.

burg.

e.:n csra p:r' r d on pipe s. tio'.s with kna n CL?.

Th*:

. c te r.: u.

perfe.raod c a 'lcn..j to <a p; : Tuc ':ic. p i.pe to. T-2 uhich or; of c ry.1 1.aich 1.4 hr ra r" at. nad a b u.. p t o r. 6 F u '.1..a. Tha

-' id nd third touts ur e ph lort. 2 ou speciallymreld:A cipw

',3.

2I. 0 245 rasp. :tively, uit - intenticaally introduend CLP. Tt:a bace nc:2 rial d in the bc.rce tents van vuifiwl an ec20 ying with the %quircaentn l

N.," :

E.' 40 Typs 304 a' taiala.:a 3;wal.

Tab'.

! u a

'uia ti m o f the bn a

-- tal c'a

'.'.:n1 r.nayl.;a a und acc h:. ;c al p enpe ctIw.

three burst tect.a c a perfa:cual in an identic.! canaa; e. : Yam ;;nt... <a utre winnaaned by a Ee:h:e1 MDS ear,inver. Th pipac were 12 in bea in inal disneter by 3/S inch vall by 4a inchea long.

Fla: plusa vare

.. e t

.lded into ecch end snl the pipe (da a vertical paaltiotQ +:n tostatically pres aurizea until frac:ure occurt.w.

Thu.all thickne.us r:e.;sured b ef or.1, d uri n;;, a r.d a f t :.. thu :ss:.

' ' !1:inully, tha

.anical propertius : roca th,: pips veld. vere detunained on nu ndjacuat, 2: ion of e.ch pipe.

'rellc" aphic exaniaa:tona had basa pvrfarn+d on :'.ta pipe a di crocs-

f.na o f t ha pi p u uc '... prie. ;n :h burs: t ut s :c verify th+ pr2uenr CLP. The percer taps oC C:.2 in chis testi:,a of tha rejar: a:: : b a.:ed n the everage of twelve 67 ore ::ndin;;s tahe:. c.t one-luch it:re.: ats the fracture faca of :ha burst teated pipe.

hydroc:atic :Sc? prer nre for thin pipe cien and acha elo in

..ir.ee v':n the requiracar.: of SA-312 and SA-530 la eclc. '.lteu an

.110. :

2 x Pipe Un.'.1 Strean x Noainal Vall Thicknesa P

=

divic 1 br-Sp-2ified Cu:Lide 1:ina.: c 2 x 15,00J psi x 0.375/12.75 P

a S32 psi hydrosta:ic test prassura P

n first pipe Which was te ted, No. T-2 burst at an internal preanura of over m

- 00 p.1.

The cecund, No..,6, t.nd third, 'io. 245, pip.s bur:: at pr". urns of

.veen 3700-3S00 pet'and 3000-3100 pai, respactively. Theca values tay be apar:3d to the Code-required hydrostatic pracaura tcat of SS2 pai.

ure 3 chova photographs of the burst piren. All of the pipa sections iled throup,h the valds, na

I, as enpcuted, CL? vas visible th/caug?. oat leng:h cf~;he failed area. Thu amount of CLP in production pipe T-2 cied along the length of the weld at the failed area, but averc3ed 15 m en: of the vall thickness.

Essentially, this neant that the wald ickness waa 15 perennt less than the nominal. pipe us11 thichre.ss, yet l.a veld var :cpable of wi thananding an internal-hydroc:atic pressure several tinas the hydrotest prescure required by AS!!E Section III.

e 246 had 40 percen: CLP and Pipe 245 had' 55 percent CLP, which van

.entionally introduced.

t

'able 7

,.1_k_tal e&.. and.Ma.t.t.anien t Pronr :!es o f Yottav.s:- en Pion Un ? ' Ir *ha l'.urnt ':m- -

s.

'._a n. i c 11..Jc.. x,l.y s e. '...

.C..

.P.

.?

.O.

.S i

.S.i Cr t'a

-.u.

17 C

p !! cat 15337

.0% 1.59 0.028

.017

.33 8.33 1G.80

.32

.34

.075 i t T a c c T- )

m '! 2 c. t 53167

.375 1.E3 0.033 0.037 0.51 9.55 18.40

(

. : 'ie s c 24 5 )

.;heny INat 714835 0.053 1,62.

0.025 0.035 0.65 9.47 16.27 0.29 0.16 0.057 n: Test 245)

! % ' h 2 7. I.a n l. ? r ? ? c r '. $ -2.

l Tencile Sf rent'h Yinld i' t: enc *.!i Elon".:!.c a Jessop I!r r 25337 04,300 ps!

33,M O psi C'J.

era:p 'ha: 53167 ES,20a 40,000 fes d.12;;haay llent 714835 87,000 37,000 70 4

4ble 8 is a tabnintion of the burst' test d sta, includi.e; the calcularad

.arile (hoop) c:ress e. t the time of fracture and the nacha icr.1 x ;rtit, of the nid as deteni: ed f reu a sample :.: rcad f.a.

c.4 ch p:. -2

. ;t.

I.lso -

particular iutu ust 1.:

t'- d.

.'11;y

't or to tb p r e a r,e.,

<f the pipa. The d_ a00.a r of t c.3 liret pipe increas d 21 per.-it pri'r

> frac:ura, e' the rail thi:'...as;, as t;ea::ured t.wo iv: hen a.<ay f ru ' ;he 41d in redecad 16 percent.

..'ry close enrrel:..lan wss faut. bat.,ua w.hwiul prxu rtius a d.:t c.'.ned irra separr.ca t e n s.'. I - b.>. r s ni es

?ne:r -ntd fron calculs:'ona bened upon burst cus: data.

las,a refer to

nbit u fo eccple:t tcat data, and to Tabla 9 f or cross-sections cad

.i.ditionnt data.

se,:tir,u.i of pipe used for the burst teats unru exat:in3d by UT and RT

..u

';a.u a:...

citer :hs be: t tast.

Pric: to the bu r.t test,

• F f i;.i t pipe

=

te w m, So. T-2, wt c:,;.nined by ec:reantiocci t.' proevdtru9 1; accordance aith AS!$ III, N2-25 22 :nd tha 15 percent CLP war, not dere:tal.

UT using the special pro:eduru and n! det ected only si::-1: %n; of CL? conpared to L& ir.nes of wcld leagth. Af.2r the burst test, t ha unf r.'.c ture a weld

'ength u.i reexa::ined.

Conventional UT wa2 able to pick up a snall signal, hat only when the &>.i. was intra:se.. by a f ac:ar of five. UI using the special precedure c:a RT could detec: the sane six-iaches cf CL?; towever, it was at:h nere dis;f n t on the RT reexanination, indicatin; that the unfuced ed;aa had separatad un.ar the hydrcsca tf.c tent p r es s u :. e.

The ot'.cr two pip aa, Sea. 2/.5 and 243, containia; 40 p.arter.: ar.1 33

• are clno wrained. Conuntiomi UT d : m ad 3r cent CU resp'::it '. /,

th e prus 2nce of C' /; however, the r.;gnal did not e:::ewi 103 parce u: of the DAC cu:ve and both welds rat the accaptance J oe13 of A M Section ill Nt-2352.2.

Tr. CLP wrs detected -in bo:h welde, when exarined by '.7 using the special tac'u! qui and by RT.

Tha wd ds did not n.mt tha RT a c ce p e n a c.- levela of ::h-2563 and Ne 3320. Subsequuat to the burat tests, the unicactured ler.I;;ha of veld wer ' re e r. :ined by canvent%:.'. JT.

The signal f rom the CLP was son what g: aater; however, the ex:.c!n tion indicated that both vel ' were anill within the acceptancu levels ^.~

NS-2552.2.

Using th. sg.;cial UT p:ccedure and RT revealed that the gap had opened because t'm UT signals v.ere core distinct and the RT fudientions were vidar.

4.

4

-able T, T --

Ta bu in t._t_.n.. of Burn: ""ont Dar.a Tea: ':u ibe:

T2 246 245 Cuate.:-11:te Ls '. c I Pena tra tion 15%

40%

557; Md.ne %al Propertina scrona thu wild Ton:11c S :eny,th

$1,003 pst 57,200 p:1 43,700 psi Yinld S;rcag:h 46,400 psi 42,500 pst 36,500.. 1 Elu ;.tian 25%

3%

burat. Test Eu-'

Tec: Pressure 5300 psi 3700 psi 3030 psi llaap Stress in Mpe uall a.

tiu i r : iracem e C4,830 psi 55,200 psi 43,000 pai Pi p C' u:ar O c:'.3. 1 12.78 12.71 12.71 Finni 15.72 13.00 12.70

'arease 23%

5%

1%

. F' p k'n11. Thickness Original

.332 333

.3E; Finni

.320

.370

.301 Reduction 15%

5%

27.

"Streas in the pipe wall ir. based upon tio folleuin; fomula:

Stras a Bu rt. : ?nasura x !!o. trul Insieb 9tm e a -

(2 x Nc=inal Wall Thickaes )

u?

, '..*-l_**

.R k

% % b,*

~

j *. ~....

u i

.M

.-.e 4

i v.

..e 4

./..

8

{

4 6.*

t

..t+

t. ':..

. s,,.,

~.,

~

g

-of

>.,;~

-s z

.~

s u

....s/#

f3

..m 1

, i ?

-n

.n.

v.

T..

.3 n

,4

.s

~

-1 SX Croe. -c+:ction of burrt c'

't t, md pipa at ira:ture,

. ~...... _,

W6 - h.2 C ' "-

• p r -s

..s

.e i

s 3

a of frreture area of burat tocted Pipe T-2 C'?).ehich frtettr d at aa incurnal pressuta 5300 pai (Pha:o reduced 1/5)

'D M

?

9-lt e

YY

.d aS, a

.; p 1

Photo of f racturn tre c'

i b ur.. t tas ted pip i 2..b (407.

r

-. J-CLP) ttich fractured at an s a

' A..,

internal pt:snure el 37f t,,%..

psi.

(Photo reduc.ed 1/6).

N-

.w-.......

.a m-..

=; * ~.%

..rr-=%v y,,, A.lQ,,,;.,'~!.

' L.

. 1 ' <- ~~ -

s

- j.

s..

e

..v.,,.

m

,.2 Photo of fracture area of

~.. _...

-m, r.

.p,l,c e. : w.;* x.

.~..,..e.,,

burcc teated pipe 245 (55%

r

. 7. /w..

'_,a.

P.

.t-Q g..~

d CLP) which iractured a t an 4(

~

. s.,.a.:%. ;,;:.;.g,v.2. gg.; ; 'i ~.._.

/

c.

internal pressure o f 3000 psi.

(Pho;o reduced 1/6).

a Sections Cut F-?.:a l

x 3 /8 ' k'all ? ion ',Thich Has E.r.en n rs t Tea ed FIC"?_~. 3 ge g8-

.rT b..',: 9

._C.r..c.v. _S e..,t_i.ctu A..n.d.T.+19._t.D..a__t a__.O_f._'.'?.. _P_I a

_S.e. c. t i ons U.r. r. c'T.o....P e r_f o_ ru D.u..r. s.t l., ?

~

..._ t..". 1/ 8

_ n l. _l._. I:

I

._v e r n e..

'. s..

... ~. t_r_. '.,/ L v,,'.._r. e__.m. _.

.t. 2.

h_r c..n__e.

s t s.. _.., _P

.- r n : n...,. D a T..b... '.,. "..:.P., i.i e..I. ' A n i..-...a t_A.. : ': a ',.. : h E. :.d

• ..'..n,-... -

.t,._..-

.i 1

r 1 Productim.: Pipu T2 Lack of hinatr tion:

Ap; r.winar.21y 15%

Dirce Tact Pressnre:

5,300 pai

.e

• w.

g;np gg.as a t 5,300 pai:

04,800 poi

,,. 3,

1

.,,w a... - Tc;tsile T.>

DA a:

," y-.

Ter.".le Strenph:

81,000 ps:

.f,?

/.4

.j

.. [. m(n.

.7.

Via a s::.mgth:

46,400 p,1

- T

-3.,. C.

3y 25 v.

s.

.,...m.u...~,..

e,.,..

. e.

a.

.,.. C'\\

a +

s D (P r. ]g Ag G u Ju sJf g

n nJ~ s(

A, p

i I

t.;;:

I 4

/5

(

4 s e. ec ia).5 w.." ; '. ' Pi p a 0246 Lack of hnn. tio:'r

/ p " c::1:antely 40 '

4 Burst Tut Presst 3,U' psi Esop Stcc a at 3,700 pos:

59,290 pc1 Tencila Tac.

D.s ta :

Ten.ile S tr er,th :

57,200 pst Yield C :eup h:

42,500 psi Eloc pt',cn:

6%

. ~

..e

n +.-

y.,.n. ~

.~;,., s..,n.~.,,..,,.

~

e 3y.,.

e.

s.

r,.,.

~

m..

w

)

+~

.. m 5

g

.n m

1 L

c..

q.

'f.. a dd.s,.x id,,,. w. ".. d 'i / c'. P...,;-Spacially-Velded Pipe 0245 w

.. s.

Appro: ic:ately.55%

t- "%.W.* '"QAi,f Lack of Panet ation:

h ', >'..I[F J I'

.% ; f.

'; E.',h.'}'t '..$,

Burat T m r-ascure:

3,000 psi

'g Q..?'c,J/.7~ 9 '.c.:

.- I_* %,}-;.K'.".

Heo; St:es3 at 3,000 ps1: 48,000 pai c *:" c g,.. 3: '

• % K

.,u.> *A.~..:.f; ':'.?;: *. 's Iena11e Tnc Da.a:

j.

Tensilt Strea4th:

43,700 psi

,.,jf 4 g d +;--.}.,._ t<,, ;.g, g., t., ;.

a

. ~

- u

<.a 6,,

s Y101d BEY 0*MEh!

3bsbb) PSI

$b T h..N k

'.(2

,f E

~ ' !

Elonption:

37, 2,.?Ef.e.,I.

• l 1. M. * *..C ~.. JJ..,3 '%; 'jsh.$.,;g.bi-27 T*

.g...

7,

.%'.',~.,..~.T.4

.,f1.,>,- ".

s

.n y *

.+

r

,,. n., p '.-

'UI

',% e' y

d

. m $,.

)

cs,

q,n) lf 19-u..O ;...,,

w.

-]-

ca.

- - e. -. e.d. A_S.y#. S.4-332 Produetto:-Wr: nt P el b.:. a-

- - -b mi na t i c,

e...' Dot.h'.

• U ld July 26 and 27,1979, r.n evaluation was conc'ected of the

'.*m w,s *.owa uble valded SA 31'. p.pe that Pu1~. un ?:r.ar Pm ducc.s had set

.de et t

Mi r f a:f. ' :y in Pt-

.naun:, C2.if o rat.s.

App ro xt:.r.:. y 5'

11 a f 'rt 1

pipe which hac b e. cut ad uC. dad into pipe suba ceablics

5. e.: 2.n-ed.

'iha pipe let as rarq2d f.v: six inches to 18 f eet; tha t..>.e t a r s

. sed f.ca hight inc ;en to 12 irc:.es !:PS and the wall thic'c-na ms r. aged 0.25 to 0.373 1:ches.

• ib jority of the p.pa was 10 '.- hes t'S 4

..hudul. 40 (0.365 inch utli : hic'r.nasu).

ex?.aina:. ion cor. d.sted of poli.

Jas, etching with Marble's r.

ant and ing thu veld crosa sectica on c. 1 the pipe enda which vera acem.cible 14% annifica:1ca.

In tua cues, the weld away froc the pipe :nd wrs

1:

by grinu:; a groov-3/W in depth a:roaa t he v31-.

Ta side

. nl cf the gro-ave r,a. pollahed r.H c:ched to show *ha udd er we :tio.1 L'1t ra sor '

enaninctica usin; the 70 degree spacisi recedura v n; also used

o search for the wora: case of CL?. A total of 209 feet of unld.iere

.: nnin :

u aing this pr

.v. urn. Attenp;6 to correlate tha t~.

ex2aitation resul.a virh P21?.

.a Powe: Prot -ts UT results w.re unsuccesa?al esca ue they M.d not hu records of vc.are the unacengt able UT ind'.:c':ini ru ices.ua in :he s ctions of ptpa othe.: than it. the few instanc,.3 wher2 tha J.' operar.cr hH actually c:.W.ed the vald area ca tha pipe. Two such arean a re locttr1 r.nd were.e.unincJ b.' grinding and etching.

Jhe C'? r.: the s e : - o M ut i. :ut w.r. r pprc xt: auly 10 percen, o : lesn of the ac1?. thh!.a:ss, ya: :!.1 70 de3 me signal anplitude respc:t: e wa a l u c t.

higher 1:. anplituch :nnn ther obtained.ro:a the CL? cro.w-threa ti a

rection

.ha*- showed.M percean of the wal.1 thichnaso.

It did ppm tha:

rep bc~.,.2a th9 unf used ego s ce s greater th:.

it was at tha 26 purcent CI.? lo: a: ions,.

TF h would Octoun; for the diffet. n o in un;11:ude, sinca all thru creas did indicate C ? lengthu gren:.cr :han the t canauunr siz:'.

.\\ total of 71 cross-sutions of welds were exauined by etching, 25 of thera showins scue degree o; CLP.

In all of the pi; us emnined, the CL? varied fron -ne percen: to is percent. of :ha vall thichnw a.

Tab.'.e 10 stows :t.-

da.2 cLtn iad fron this evaluatio:.

4

.D e. on. e 4."... _m. _! :._a t...i.o,n_o.f._',.lo_r. cm a Ple.

pipo 10 A r chs 'in diace ;

n d tt fw in lems h, wit h a 3/ ; 1:nh '<,11

' 'ct neaa.a selec*;c J f ro the 520 feet of pfp.4 er nined t Peli om ~.

"":; m ca_e" r.e

w. a id ais.ti ;;d c. Imi n,,

f: m t' a t />3921., ::C

.,686 al "F" chert (224

r. pipa nection b-- a ben - subject d to RT, UT ruid etching macinations

' th the f ollwi.;; leaulta.

..e len ;itut.inni veld on this c ~ica was r d:encaphyd, t.

6 ennra'.1ed by l ut c h 'm h t he U d ei;rce S ta nt. JJ and 70 ce4c.* Spec..l. prc2edur s.

The

die;;r. hy detec ted one 1/2 inch long area of CLP in t..,: opprisite and of te pipe f rem tha t which showed CI P by e te M.og t he em].

51-inolated

're. (

• 3/32 inch li:: ater) vrc i nJ ao f orni ccettered alo.q', thn n1d.

45 a cce etcr '.m ' UT prc eb.

i l.id not da t..ci 4.ny inc i.ta tionu ;;cante" saa 20,arrent Lic.; (10 parcoat I:;d).

'h4 70 6.c;r~. r.pr ini UT procedura de tec ted indientior, of r.pproxiaa tely

' ') petcent U.S (K p reent FSH) fc,e 22 icchen of thn "4 im:he of weld.

1 thu

• O r! pr ?p n 't, the c'ad oppN '.te t h e rnJ1 1,:*.;hically C.etenc d CLP, a CD per ; cat Du, Ct.P iudiention a s f o u wj.

CH ading on' cr.ching confir: ed

..'..n c 1/1f CLP vas p to nen t.

On the end where t ha P.T dete :ed CLP, a 50

-arc.nt D.\\C indiention uns found.

'.3 nre ". vere s,lec t ed f or f ur t! r r iaentiga* :'c.,

by ultrunn!c nc ! e chin.;

c ' 'no t Set L.e ln t"n t o f t he 1nve r,t. ira tion um. t o dete. ine tL) offect

' r ecm) the t.

id i. : h:g er,,alaa th.a cot.ul. and ta note wheth a alevan t or non-relevaa t si;.;n;.lu u >uld ha ob tained.

% 519 11 cre. taint c nynopsf r of ':T, UT and etching rraulto. Area A and D e r +: tro Jnch ta n m cout..h hn Liccrcphia. : y detected p.iren of npr# a M y 1/M 1",h.

Ultrc mni. cnnin a tio'-

wer e p.?r.'o m' f ran

.,th clden of the vcid usina the 45 degr :e standard am! tha 70 de;ree

~pcial procedura.

On the pictorial viewa shorn on Lhle 11 th,i ulccc-onic cign.ala det. ted are shown above or below tha horizontal line

' r.a.,1-to re: resent th. weld to ir1-lcate tht'. detection '.au f ro:a o na cJde or th-o t h e.

or both sides of the veld.

T unnin,', uas perforaed i.t

.he requhnd sensitivi;y (1X) and at increased.manitivitea (27. and 4X).

L'he date sl.cun con firLv that en increase in <.eus t tivit/ cla renuit in

<rrornous depicting of indicatican while grf..df.n3 auj ctchin;;, confirma ely tl.e RT 'ndf ea tions. Ao nenaitivitlea increated froa.1X thraut;h

~ *: the amplitudes, lengths and numbers of indicationa all increaaa.

4 4

- I nt.n._. i.__on.."'." E.m..i_l e._. d.

.. _7 9...0 3

'. >;r D B2.ul:; tory can :h ion I & E ilui.le-tn 79-03 req, ire:' t iut ; v o 'm:.e : r #.:

_...2. :, o of the. r..::. _

<'ir..'. uaid s la _ St.-012 pip 2 i:ce

'.c ang tm.

be fa:

'. When en: ::h.31ve.:. ; U.e nc u aaved tha t con /en ;?.nc '. U v a.2 not u!. d ul ev.aa:.u tio n for CLP,

he :.nphe. sis of the nor c s hi " t o. ' to a

. arc.inat? ;n of the e.ticcts of t ' P unar, the pro;.ertie and t.'

f unc tican

?

the n!r-The restti-.> of. tha initial oe.chanice.1 propec:y and burat i:

-! r

-.:e enc re-. n !.r.;.

A.i ditior.: '. bu:::c t ects of pi.- w( 1 'tc:2as<n' i

y,rees o f intentio:'. '.ly in :ndu:ac! CL? vere perfo: c ed.

T'-

intic n-s a ti.ticy anl t >.t3hnars c '.motsai:1; attinleas c:cel r.rc t ha inab.,

a tributinl; fx o:n 1.2 the capnility of. the pipe to with -:en2 hi;h

. t e rn s..'. pre,;su.c priar to burs:in;;. Th e. burst tes: da tn,.ae:han k al

. n e ti-

'.a t i, te. J1ographic inv-u ttga tio 1 and th a f ra-cure ar.alf :ts

.3u'

. x Lanted is thi report pro.' ' N su'as tan tie.1 evidea:' :.ha t !SC

- M ?. p.

W.th wald C:.? Vill edaqu:.:ely perfor: tha raq..ind a 9. vie.a 13CCions.

' !:hae:.:h thin inventir.:: ion cc:.:lude s that convention tl UT and P.T vili ca t d c :c: tha t:.,jer;:y ct CLP e. tdi:leau, ET aad U; will

ve c a.:

.:tc;; oth::: abrict.;1:,n defo.:ta c.:h as 7$re sinJ.

e

~22-

_ TAM.E 10

.E.v _i l a. n.. t !_on.O f./.&_7S_A.. 3._1_2..D_:> ub. l_e...'l 310. 2 d._P_i_e e....? u_l l_e__r. t P c. e r

_.c..o..d.u..c.t.:.

Is'o. o f Total length

!!o. of No. of Cror.s-IU'S DI 1.

Pipe Wil of Pipe In le a g t h.3 C ro n t.

Sections P.ange of Incha.

luthv Ew:

.E_ncin e !.

S e c ti'r.e s

_te!. t.h._C '?-..

_C f.P_.."_i..;

JO

.365

$1 fce:

7 11 2

. 03 ?. to

.073 inch.4s 10

.250 49 fae:

9 6

3

.015 to

.03" it Aas 10

.365 256 fee.

40 34 17

.C32 to

.096 inchiu 1::

.375 34 feet 7

6 1

.015. inches 8

.322 1 foot 1

1 1

.032 inchea 12

.253 1 fe st 1

2 0

10

.365 7a feat 15 7

0 10

.365 10 feet 1

2 1

.032 inches 520 feet 81 71 25.

T:;'A. e 11 - n a r n i t s o f t',7 ur,.'

• F.tch: a_.Uxar.i.n...tia.m -

9

~ ~ -

" store.r. Cr. u" P1,,

Area A Area D 1/16 Pore 1/16 Pero

' ' "h 35t-l'a Ind.

1X 3hg C;;

j~0D ~' # '

~

i' No Ind.

1:a Ind.

W 3 0"; - - M 501 16 0% -

.n. a.

y.:.*-. 5 0's........n.-

p, 0%

lc 0'a.;

9 4

30#6

~~~

No lad.

b. u.

)

10.,, +

s, clot No Ind.

!;o Ind.

(

u 10t No Ind.

40t No Ind.

?X i

s 40's 50*6 70 r, No Ind.

Metal Noise to High to 70'3

701, 90%

Inspect Both Sides

.fX

-:1 q--y

.1 p _00? +x.

100t+

y 1

6 Pore Pore

cin?ing Scale 1" equals 1"

/d'PFiDICES 4

APPEN9IX 1 2,%

AES-79-00-8

~,,1.s i j a ; r*..,T*: m i-(

Er.GINE ERING CONSilLTfN'i SERVICE 3

' ',' ' I 3 ;-)

y,5 DisTEL Dagyr,, Los At fo;, CA 9:022 '(415) 909 0212

/

. } cud SIG?l'FICAtiCE OF CENTEP.Ll.!E LACM 0F PE:iE7P.ATION Ifl DOU.iLE ',.ELDED STAINLESS STEEL PIPE - FRACf'P.E AtlALYSIS J

.y 3:m ts.,4 -e j SU3 JECT TO CHAN$2 by Wstren P. I' cts:w;hton Geof frey R. Egan Russell C, Cipo'la Prepared for Bechtel National, Inc.

SO Beale Street San Francis:o, California 94119 Attn:

W. R. Smith, Sr.

i August 1979

'a vites in Mechanical atx! Metallurgical Enginaring.WSJing, Corrcsion, Fracture Mechanics, Stress Analysis

TAGLE OF CONTENTS

Tiin1 TITLc PACE SYNOP3IS NO.'! ENC'.ATUP.E 1.0 INTRC3JCTION 1-1

1.1 Background

1-1 1.2 Scope and Objectives 1-3 2.0 FAILURE CRITERIA 2-1 2.1 Lockground 2-1 3.0 MATERIAL PRO?ERTIES AND SPECIFICATIONS 3-1 a.0 PIP!N3 SYSTEM STRESS ANALYSIS 4-1 4.1 Source of Pipa'Stresse3 4-1 4.2 Hoop Stress Oue to Intarnal Prescure 4-2 5.0 LINEAR ELASTIC FRACTURE MECHANICS APPROACH 5-1 5.1 Intrc;.ution 5-1 5.2 Deternication of KI 5-3 5.3 Deter.mina tion of K;c.

5-5 5.4 Critical Flaw Size Calculation 5-C 6.0 CRACK CPENIN3 DISPLACEMENT (C00) ANALYSIS 6-1

.5.1

Background

6-1

'5.2 Crack Opening Displace.nent (C00) Analysis 6-2 7.0 LIMIT LCAD ANALYSI5 7-1 7.1 Introduction 7-1 7.2 Determination of Appropriate Section Propert cs 7-3 P,Mt L

7.3 Determination of Applied Loads 7-5 7.4 Numerical Results 7-5 B.O MECHANICAf. TEST CORRELATIONS 8-1 9.0 BURST TESTS 9-1 9.1 Introduction 9-1 9.2 Elastic-Plastic Shell Analysis 9-3 9.2.1.

Background and Co7.puter Results 9-3 9.3 Yield Point Pressure Calculation 9-9 c.4 Burst Presture Calculated by Membrane Theory for 9-10 Large " trains 9.5

' Standard Cylindrical Vessel Burst Strength Calcu-9-10 lations and Field Test Correlations i

9.6 Determ.ination of Burst Pressure in Field Investi-9-12 gations - Flawed Pipe

TABLE OF CO?tTEllTS (Continued)

' 2 ':T'ON T!Ti:

PA'iE CO:;CLU510NS REFEP,Eti;E3 APPEiCIX A:

Pipe Material Properties - Full Sections A-1 Summ ry APPENDIX 3:

Tensile Tests - Known lack-of-Penatration B-1 Sc:=ary APPEliDIX C:

Effect of Geom 9.tric floniinaarity

'C-1 APPEtIDIX D:

Effect of Shear Dafortration 0-1 i

SYi;0p5IS This report sumurizes the results of a fracture mechanics analysis

.::pleted to d:'armine tha significmcc of centarline lack-of panetration

LP) defects loccted in double welded stainless steel pipa. The analysis

0nsists of

(1)alinear estic fracture mechanics (LEFM) analysis,

.2) a crack opening displacerant (COD) calculation, and (3) a limit load or r.et section collapse muthod.. Verification of the analysis results was r.ade utilizing burst and tensile test data.

The LEFM approach 1.ndicates that the defects located to date would ie far balow the threshold to causa fracture, but because the stainless steel pipe ma:ccial is very ductile,. that type of fractura criterica is n;'

appropria te.

The COD analysis confirmed that the apprcariata failure criterion is plastic limit load rather th3a one based on fracture. The results of both the LEFM and C0D a thodologies assure leak-befora-braak failure codes.

Es;; mated design pressures were calculated for temocratures ranging fro:a 1000F to 609F for various pipe sizes.

The licit lo:J analysis was then perfor:s.ed to g9nerate a family of curves relating the non-dimensional pressure at failure, to critical defect length for various pipe sizes. The critical flaw depth was found to be a constant for all pipe si:as at a gi,en Operating temperature and limit stress. The tensile test date indicate that the limit stress appropriate to the critical flaw depth calculatien is a func-tion of the flaw depth. The failure will be controlled by ultimate strength

onsiderations for small flaws (<20% wall thickness), yield strength for large

' us (>50" wall thickness) and at intermediate conditions will 52 governed the flow stress.

Lurst pressure calculations f e unflawed pipe wire perforced using eral cnalytical techn ques.

The results were used to predict the field est data and were found to b? conservative by t.pproxic.ately 13%.

Thesc

ults c a now be applied to plant systems to assess the safety margin aiierent in the piping for a given design or ser/ ice stress level.

e h

liOMEitCLATURE A

Area a

Half crack length or dep'.h

+

c Reduced thickness

= = 21. _==;-

=

/12(1 - v)4 C00 Crack opaning displactmaat E

Yoc..;'s' modulus e

Applied strain e

Yield strain y

Fcyl Cylinder correctica term H

Radial stress resultant h

Radial displacement 1

Mocent of inerta J

J-integral resistance value R

K Stress intensity factor ic Critical fracture toughness K

M

~ Applied mcment M,

Edge c; ment resultant 9

t

?

Interral pressure

?

Orst pressore 3

?

Applied nynal load a

Bending stress decay distar.2 rj Inside radius

+ rg r

rm Mean icdius

=

2 ro Outside radius s

Distance to outer os t. fiber

= 2a Wail thick; ass W

Wall ratio outsi'de diameter /inside diameter

=

ic Critical crack opfning displacement

?

Non-dimensional crack cpening displacec:ent cu True strain maximum Toad o-Applied stress o

Critical stress c

ID Inner diameter stress 2

cj Limit stress Membrane stress P/A an

=

00

- Outer dia. meter stress

a c

omir.al ul tir: ate streng th u

,,o Shell stre;s resultantr.

g cy Yield strecs o

Gecretric nonlinae.rity paramator Edge rotation i

v Poisson's ratio

= :(5 A

Shell large parrr.ater c

O e

1-1

. ')

ljiTRODUCTION

.1 fg.kJround In September 1972, centec li..a lack-of-penetration (Ct.P) defects were iscovered in ASTM A312 Type 30? stainless steel piping supelied by Youngstown Jdir.g and Engineering Company (thi:. infor,ation, along wi a other input data applied by Dechtel Intt.rnational, tre hereaf ter documented by the appropriate latech Engineering Services controlled document number of (1)).' The pipes were f abricated us'ing double welded seu welds produced by an auto;aatic gas tung-sten arc process without filler ca urial. They were subsequently beat treated by solutica annealing The defects were d etected in the longitudinil

- sean weld nt.- the cantar line of th, pipe wall.

It has been found that a side range of weldin; parar.etars were used in past fabrication of the type of

Se in ques. ion,and the CLP defects are thou;'.c to have been caused in part S/ misaligned weld ar:s.

The pipe has been supplied through the various channels to several

~ achtel nuclear power projects and has been utilized in a variety of systems.

T".e range of pipe sizas where a potential defect may exist is large (Q) -

. ', D - 2 ).

Furthermore, pipe examined has indicated that the defect sice Q.eight through the thickness) may vary along the length of the pipe sections.

s typical defect is shown in Fig.1.1.

lion-destructive examination methods have been utilized in an attempt

) detect the defects. Todate,neitherradiographic(RT)examinationnor

.:nventional ultrasonic (UT) examination techniques are able to detect-the

d 4

1-2 D " ]Do#Ju ol l3U

• D' 3

a y,

. w _ v.

r'.

i t._,

~. -

c.-~.e*.,

w

v..s,.; :w ~,. t v

m.,

, +, r..c., /.,.; G., g;:.,.%..~,

- h,31 d l',.e tal

.o...

.. t - v $,

.. r

.....*...:,..,. ~~...., c; I

. ; ; r.r o;

e ;. a. -.

.?.

4

...u.

[ p.

..nm.,.....$ ;..

+. 4.;~;.;.e. x...,,~c: e + j w >.""... 9. 3. ";y, ?,.,. c.. <. '... s \\l

.u

..,y.. '.

J.

y

.u.....n..,'),% ;... _ _',.. n.ca,5.,, s

.l.aCh Or.

O,'" '.,.7,.'> , f D.'}'.-;.. - - y i.e, ( * 'E J pene ti'a t.i On i -(r,;

. *d. ',$'.'.'s.

..,.3,,.,....

n w-. :.:n..;,.z,.,.. ~. a,...,,,,.., -.

h.,......,. s..

v..._...a...,..>.<,,.,.: : *,. n,.. <;,

.i.,

i.s.-

..<..<.0....

. s *:..,.

..1 f e c +.

c~..

.. w.. :

.. ~.,

.n -:.,

~-

u -... m...,.v.. g,~.w,.n.,.., : n,

s..

i....t.. e. n., -

.m

.,.-..~.~..t.

~.

..r,...s,,..~,-

s<.

3,. - -..n.

--g g n

,,atal k

. - 1.

...n v.....,d~_

.\\.+.

r.

..s

..s. m..

1.

,.t q

't' n.

-l r.

r.

,a

. i.

c a

+

a..',

,', m,

,l

,q. :.....:

Figure 1.1 - Centerline tack of Penetration in Double Welded Stainless Steel Pipe - Typical Octail.

k

1-3

. fect 'ecliably ((,1_; - AESCD-16).

A spet.ially davalop :1 technic,Ja has been le to detect a lack of-penetratien but only if an actual gap exists between a L.ifmad faces. For rest of t;u pipe examined, this coidition has not

,sen met.

1.2 S m and OSPetivas Aptech Engineering Services was asked to evaluate the significance of the defects frc.n a fractura critical viewpoint.

This report assesses the significance of these d2 facts in str.:ight pipa under internal pressure only.

The te:hniques developed are applicable to pipe elbows and te2s and for other 10adigs, however, the affectad gacmitries and piping sys:cas have not yet been identified for ench prcject.

In the interim, tiv analysis perfor.ved Mrein will be aoproorlate to many situatice.s where the pipinj is being uti-lized.

The obje:tives of this invastigaticn are:

(1) To identify the relevant failure mode, of the pha, (2) To dev ;op the appropriate failure model to predict critical loads and flaw sizes,

(3) To correlate analytical predictions with expericantal results from the burst tests.

The_ report is organfzed in nine sections.

Secticn 2.] discusses the copropriate failure criteria Dr stcinless steel piping.

Sections 3.0 and 2.0 present the mechanical properties and expected pipe stresses.

Detailed i

14 t!!c ussions of the En31ysis nthods and the results analyzed by clastic alysis, crack opening displacement (COD), and limit load walysis are pre-

- - oted in See; ions 0.0, 5.0, and 7.0, respr.tively.

The corralctica of

..ptrimental test date, both tan-ile and burst tests, is presented in Sec-ions 8.0 cnd 9.0, 6

i

2-1

' 0.

FA" 'lT:E C'CTEa r -

.1 Bacb:r0und The failure b:6avfGr of pipes under monotonic leading can be classi-

"d into thr 2 regim s where a rpecific type of failura mode is appropriate.

lne disciplines requirt.! to assess these regimes are:

(1) Lineer Elastic Fracture Mechanics (LEFM) - Where the structure fails in a brittle cannar and,on a inacra scale, the load to failure c: curs under nominslly elastic ic; ding.

(2) Elastic-Plastic Fractura Mechanics (EPF:1) - WSere the strur:ure fails in a da::ile manner,end significant stable crack o.: tension by tearing r..ay precede ultinete failure.

(3) Fully Plcstic Instability or Load !.imi:

wn.i e the failure event is characterized by large deflections and plastic str ins associated with ultimate strength collapse.

A diagram showing the relationship between critical or failura stress

'.d flaw size for the. three failure modes is shown in Fig. 2.1.

The shape

'd position of the failure locus shm.an will d2pand on the fractuce tout;hnass

'Ic) and strength properties _(o )laf the material, as well as the structural I-xuatry-(w) and type of loading.

In general, linear elastic fracture mechanics is used nost a:.aropri-c aly to dc.cribe the behavior of low toughness /high strength ma terials where i

\\

5

's

\\-

l

[

.p I

1

'\\

I fc i~

l lI

\\

j-I j{3 I

I L:mit T

\\

l toad f

d':

\\

o = ie::1 rcr.e Stress = F/A

/

\\

sE l

cb = Uculing Stress = ID.

\\

/.7,

/ / 1 %x x I

i e#

N M

/ j' 7'j

_ Elastic-Plastic (E.1 '

j

/

/

N l

/:N y\\ sx\\ y' s l.!

j.

/

Fract x x N-c amu'src ar

/

u

.x

o.,. t tm o m

s i,

/ "" ";d //

\\

\\

x r, <

\\

'7

\\x i

L,

//

AN N3MN\\xNDsxAj 0

a /w l0 o

flan-Dicensicr. 1 F!r. Depth, n/w I

Figure 2.1 - Schematic Shouing the Relation: hip C2turen Failure 5'.ress cc,J Flaw Size for T.o

~

Limiti.;g Failure Modes.

f e

2-3 stic ror, size is smail rel:tive to the str'.,ctural geome ry and little

cility pr? cedes fractur. With tnis method, no ccount of increased ariel raistance b t-it:le failure is c.:n wh2n significant ple:ticity

. present.

Under LEFM ccndition3, the mest a:.cful paramete for cheracteri:-

the behavice of cr?.eks is the stnss intenity factor, K, which charec-

~i:es th' singult stresses nnr the crack tip.

In contrut, ples;ic

.;; ability, when it occurs without prior crack extension, is do:11nated by e flow orr?erties of the naterial.

In thesa circumstances, the failure

ndition 13 indapendent of fracture tougnness end crack tip charact2ristics,

-d limit load anclysi.: is used to define the failure conditions.

Finally, a the clar. tic-plutic regi:r.e, tha crack tip sirgularity, th naterial aughnesa, and r.et section str.m.;th we ralevant para.wters for failure

saument.

El.'stic-plastic frac.ura rcechanics principles u:ing a J-interc al apercch or C00 have been applid to prc'ict failure loads ur.hr elastic-

,lastic concitions.

Significant du: tile behavior has been chstrved in bc:n cyperimental data ((l)-AESCD-5)cndintheliterature(,2,).

In particular, it has been found in experimental program; on the integrity of pips centeining strus torrosion cracks in Type 334 stainless steel pipes (2_) that:

(1)

B ec tu a of the substantial crack tip blunting that precedes crack growth, the applied stress at failure is virtually inde-pendent of the sharpnass of the initial flaw introduced into the mterial."

(2)

"The presence of a weld and any sensitication of the r.at: rial i

surrounding M a flaw dcas not, significantly aff:ct the epplied stress at i

2-4 (3)

"The enct shape of the flew is of considerably less inpartance tha the are of th? flaw (or of the r;t fi'a crea whM tr:ltiple flaws are pruent) in d2ttraining the applied str;ss t.t failure."

.e for+gaing su;;ests that a brittle failure modo for this material would be

. appropric;.t2 and that E?FK 6nd limit loed methods are requireJ for tha

aalysis of Type 30f, stainicss steel pipe, fiowever, estiastes of fracture tourbnes,,

r..,, are mat.'a frem eier. tic-plastic paramaters, J and C03, and con-4

.itions for fracture using LEFM cra checied.

O i

1

3-1 3.0

!? :rt! AL '"

"'!Es t'O ' '7:!:IciT!'"'

Since tha lack-cf-;?netration def ? cts are complet:ly contained within the weld retal, the properties of that region will be appropriate to the un" lyses '.h:t follow. Af ter velding, solution ann:aling kan perfer ed on the subje:t pipe. This solution annealing process will lead to na:hanical pro-certies in the weld re:al very rearly identical to those of the bate netal (sea It: tion 8.0 far experimental justificatica of this prac.isa).

For this reason, base material proper:ics which are readil;. availeble in the litera-ture and tensile t7st dst 7ppc:;riate to the weld region are used inter-cnan;e251" throughout the analysis parforr.2d in this repert.

A tech Engineering Services was providad with several r.atarial data in:uts. A 52rics of three cat; rial test rep,-ts on particul;r si:2 pipe: for both plc:e cnd pipe test; is shown in AppenJix A.

This Appendix also providas tes: 4:ta from weld metal t2nsile taats with full penetration and the results of a determination of ~the statistical 90s, 955, and 995 occurrence level properties with 95% confidence for the base material. A discussion of these results is provided in Section 8.0, Appendix B su =arizes caterial properties obtained from tensile test specimens with a rcnge of defect si:cs.

Finally, the A5ME Coda minir.um values were obtained (,3,, 4), and these

roperties.are given in Te.ble 3.1.

The actual basa metal propertie: app'ar to be substantially better than the minimum requirerents. To take account of the variation in m2chani-i e

al properties, the analyses presented in this 1: port, primarily th-limit

e 3-2 LA E

.J u n CJ

^

N Q

to.

• -?

gg y[n 2 O

Cn t.

En 3

ta w

em e

y ;;

u, s e' M

. m E.2 c e r f -~

w u,

[ *^

LO ee ep L'

"C.

.,.-L' t%-

rs o

to C. H r:,

o

. aH

".) D m

er C1 t*J L&J CL.

t.

H v

N

~

r. O W d

t" tw

.. d :.

O N.

M 4-

~

O tri

+

g y,,,,, t -.

c1 N

O 03 D

C: L tu e.s p

.r.

P-M a-w r>

L.

(9 gW LJ D4 3

5

1. 3 tt

[2 W

O M

C:.!

C1.

,O C)

M.

m.

en c's LJ v s..

a O

e m5 o'

O O

O O

e LJ O

C 6.).

Vi<

t.

O C '"

e; ts 5 (9

m en m

{:'s.

O&

O O.*

O O

we en e

e-o e4

^

D MU

• ]?.

W I3 -

X X

X n

a

.J.M C

M.

P%

c: :

C.

e*

OJ t)

f%

C (f)

AW N.

N N

N Cd

'O i-u.

u ti.

p-C O

O e

O O

O LJ O

O O

b.u N

O 4

L.)

>=

r

> >i values, were Executed for a non-di 2r. sic '.21 stress parameter (P/c '.

i - this c anner, the appre:>riate limie pressur: can.>e 6temined f r:.n systera "i.5 5 and cparati.'

te r.:+ rat.. es, as well as th 2 actu.1 n hanh::1 pro-r

ies, Wi-ee a :ra;ricte, referance to the AS".E Co.2 c.inicum preparties svn ' ade.

G L

4-1

.]

F I ? "' 5 '5TP! 57:.555 AM.Yi'S 2.;

Sour e of Pioe Str95 es Sin:e the lack-of-penetration defects are located in and parallel to ce longitudinal weld seam, the stress component of interes: is the hoop L

3: ess. This c mponent, orie ad transverse to the defect plane, provides

• e driving for:e to cause crack ini:iation. Previous work performed in r, stems similar to nose involved here, such as nuclear steam lines (5), has

--dicated that tne greatest contributica to the hoop stress is that due to in er-al pressure.

The mej:rity of the remaining cossible loading concitions do no:

' :ose hoop stress in s;raign: pipe se; ents of mos; piping systems.

For

!x3mple, dead weignt and eerthquake loadings tend to result in bending and 3dal stresses, which will result in heco strains due to Poisson effects Out will no: ::n:ribu:e to the total hoop stress.

Therm 3i gradients along tne I!ngth of the pipe are likewise non-contributory. The presence of through-

. ickness thermal gradients will result in hoop stress comconents and may efer some con:rtbution. Similarly, small hoop stress changes may occur from

2am ham er lesding cr dynamic events associated with upset plant corditions 4

c-LOCA; however, these tend to be of very low magnituda compared to the 3 ess from pressure loading.

For the individual piping systems of interest, these additional s'.ress

.rditions may need to be examined, in particular, fittings such as elbows

ees may be subjected to several of these contributions and will require s*:

t-

4-2

-asrate evaluatien.

If it is later determined that some of these stress c:m;cnen 3 are ecolicable, they can be incorporated into the model by con-artir.; to equivalan: ; essures.

1. 2

}!_e:c St-*s: 32; Inter-4 onssure A sampling of pipe sirrs which contain lack-of-penetration defects is summeri:ed in Table 4.1.

The solution for the hoco stress in a thi:k walled cylir. der (21) was used to detarmine tne membrane stress in a cipe. The hoop stress at tne inside and outside pipe wall surfa:e is:

- 2+r?~

r P

(4*1)

ID 2

.o i.

~

2r,2 P

(4*2)

CD r'-r-

_o i.

where r., end rj are the out:ide and inside pipe radii, and p is the internal pressure.

The primary membrane stress was determined by averaging the ID and C3 surface stresses (see Eq. (7.10)). To esticata the design pressure of the pipe si:es listad in Table 4.1, tha pressure at which the primary cembrane stress is equal to the allowable stress at 4000F was computed and presented in Table 4.1.

Estimated design pressures at other tem:eraturas are shewn in Table 4.2.-

Actual design pressures and temperature must be quantified to determine critical flaw sizes, however, these design pressure estimates will serve to' scope the analysis cethods, s

t

4-3 TABLE 4.1 SUMM RY C.: 3??.. DIMEN5' OMS AND ESTIMATED ALLO'e.3LE DE5I'!N PRESSUP.E5 ALLO'4ABLE

' 4! M L ACTU t h;MINAL DE3IiN

':: ^':

OUTSIDE O!A' DITES.

WM.L ', JI:::'!E53 PRE 25'J?E*

(psi)

' S:n 20 10-3/A"'

O.250" S70

~:" Sen 20 10-3/4" 0.365" 1300

" 5:n ECs 10-3/4" 0.500" 1820

';' Sch 20 12-3/4"'

O.250" 730

' S:n 40s 12-3/A"'

O.375" 1100

.. 5:n BCs 12-3/4"'

O. 50')"

1500

" Sen 20 14"'

O.312" 330

. '.* Sch 10 24" 0.250" 330

'" Sen 20 24"'

O.375" 570

stimated from prir.ary stres: Ifmits with 5,at 400 4 o

4-4 TABLE 4.2 ESTIMTE] DE51~A PRE 55URI3 FOR '.'ARICU3 P:PE SIZE 3 E3TD%TED DE3IGN PRESSURE (psi)

I'!PF E:I-T = 1000 T = 702-T = 4C^0; T = 6000~

'0" Sch 20 1000:

950 870 E00

3" Sen 40 1500 1420 1300 1200 10" Scn 80s 2120 2000 1820 1630

" 5

5 20 840 800 730 670 12" Sch 40 1250 1220 1100 1020 12" Sch SCs 1740 1650 la00 1

30 14" Sen 20 950 910 830 760 24" Scn 10 440 420 330 350 24" Sch 20 650 630 570 530 i

s 4

4 4

5-1

t.."M :). r,.i t e. *

• r ? i-*O.r. ":.r u. N*-

10 ?^;".4.

~

I r i--d u c t d --

Ur. der the c:nditions of lineer elastic frae:ure mechanics, the most

eful parare:ar fer charactarizing the behavior of cracks is the stress

' tan 31:y f ac::c, K;, which describes the magnitude of singular stresses

-ad of a crt:k in a linear elas:ic body leaded in tension. The near-tip ess dis;ributi:.:s for the-three possible crack loading modas ore illustrated

.:ig. 5.'..

. or loading normal to the plane of the crack (l'. ode I), unstabla ic..re cc:urs when the applied: stress intansi:y fac:or K; equals or exceeds

. -t f r:cturn :Jugnness.

The f~ accura t:ughness, Kic, is a material p.operty r

s. en detor ' -d using :candard procacares ({,) and varies wit.1 tempers:ure and acing ra:a. The n; plied stress in:'ansity facter cari be written in the form:

Ccvii,

(5.1)

X.

=

A

.,*ere o is the acting stress, a is the characteristic flaw dimension, a*.d C

s a par e:er which c:: cunts fcr the flaw shape, structural geometry, and

• type of loading.

In general,. C is a variable and in many cases must M'

.aluated numeri: ally. Fracture in the structure will occur when:

K.

3 K (5.2) 4 Ic.

To determine the significance of defects in the pipe under censiders-n, it is r.ecessary to knew the fracture toughness and the a'cting stress

'..' e 1.

A critical flaw size to cause fracture is calculated from:

n

~

ntL

'f

• /

/

/,

f

's n

yo o

'f

, y',

nu

.sc i

f n

a hce l

i la

?

)

-.n r

'f x

o u

t i

t c

of r

r a

gii f

ot r

g c

i p

s.

n t

n e

s g

s L a

o i

f l

'"*/l

\\

kc E

p e

K a

/

i

(

r_(

e v

)

r U

a a

0 f'-

"Y" i7.Wt

.u o'.

n e

i s

t f

c L

n a) f e

f s o

a r,

n s

ep I'

ye t r u

m n

e it c

ss i

c n

v e

e

=

el R

s t a s

nn e

J ii r

i m

1 t

o so s

sn 5

e(

~

l r

e f

l t o r

A St ug i

f s

s 7

I

/

i I

E D

I

'I

I

>/

0 1

T.

,i o

W

5-2 fK 1

te a;

7-1 -g -

(5.3)

=

Cr

\\

i; r.ote t7e t0ughness (K;e) and stress level (c) are known. Also, f or a

' wn t:ughness, the variatien of critt :1 crack si:e :s a fun: tion of

.;14 ec s:ress can be de*. ermined.- Canversely, tne critical ap lied s.ress

, a ft.nc: ion of crack depth can be c mputed fr:fa:

I_

t e#

e-(5.4)

=

Cv?I?

• e recuirements to c:t.;iete this. type of analysis are:

(1)de:ernina: ion

<y :"or the gecmetry in quest:icn, (2) de:armination of X f:- the ca terial, ge 4 (3) laveis of stresses er flaw geceetry for the given situation.

These i:;rs i re di.;:ussed in the sections thst follow.

'.2 Deter-iaatde n o '._v.

For tr.e curren: problem,. the lack-of-penetration defects were

--ielled as a center cricked finits width plate. Tha model utilized is shown in Fig. 5.2.

Fer this geometry,. an accurate analytical rela:icaship for K.

.afsts (7).

The tppropriate equation is:

1 - 0.025 (i) 2 4

o/di d(aes 2w 1

(5.5)

X

=

+ 0.05 I

w w

o~ ire Uniform applied stress s =

2a Crack length

=

s 2w Plate width

=

r.:

L L

l l

c 1

--... L 1 I

i b = 2w s-l,,

e

_t Figura 5.2 - Ce.nter-Crack 3d Finite Width Plata Under Uniform Tension used in LEFM and C00 Analyses.

s

\\

,. ~... _.

5-5

< s corresponds to an infinitely long lack-of-penetration defect (e.g.,

.ei cran) of c:nstant size along th+ length of the pipe, centered in the

.;dle of the well tni:kness.

As an indemcent evaluation af K, a cerouter generated solution was g

so uud to analy:e the present flas. The SIGIF progran based on the luence function (IF) metnod was used to calcula:e X as a function of 3

1iedstr*s:.andcrackdepth-(3,9_,J0).

0 The result: of t. e analytical and com: uter solutions are shown in ig. 5.3.

Note that the estkated accuracy for the analytical solucien ([)

is 0.1t for dl values of a/w, From the plot show.

n Fig. 5.3. the aopropriata K value can be g

':vd giv+n :ne acoiied stnst and wall thickness for a given flaw siza.

.3 Cer$--i.astic,

^3 K.

c Type 304 s: sinless steel is a highly ductile material and no accurate walua: ions of K exist.

For the purposes of this study, K values are Ic Ic

5:imated by rela:ing J-integni and CCD measurements to Kgc, values.

Measurements from compact tension specimens (H) have shown JIc ues in the range:

in-1b J

7000 - 8000 (5.7)

=

Ic

.e.

in

,li:ing the relation between K and J asgivenby(H):

g g

JI.e 2

i l

(5.8)

K 3

(1-v")

5-6 7,a I

I I

s.0 -

l-I I

d% s.0-'

I I

-u3 I

u r3 L

/

>, 4. 0 --

-)

Analydcal Solution (7).-+/

5=

a

3. 0 -

3 m

C C33 m

y 2.0 -

.2

?

8 Influenca Function Metned 1.0 -

0' 0

0.2 0.4 0.6 0.3 1.0 Nomalized Flad Depth, a/w Figure 5.3 - Non-Dimensional Stress Intensi j Factor vs ': r ali.;ed Fles Deptn for a Finite Width Center-Cracked Plata.

5-7

3ults in a range of valuu for-K.

of:

2c 450 493 ksi/E K_

=

15 In anotner study (1)), J;"messurements were obtained on Type 315 p

.ainies, stavl. This raterial has similar properties to the Tyce 304 studied ae, ar.d th e result: of the study are included here as typical of the type

.! values which were obtained Measurements were taken from full taickness ingle edge rot:hed bend specimens. The value of J, at departure fr:; line-r, aity is taken as:

?

in 75 J R 1.3 joule /mm~

13,543

=

=

in' rjain, fr the correlation present:d above,

= E;a kstEi.

K;c rinally, K values wern obtained fr:- 000 data.

In particular, frcn (2),

g

.: crack initiation, the CCD, 6, was given by:

6 0.12A in..

=

,.; and 6 can be correlated as demonstrated in (Ji) utilizing the relation:

2 X I 6

E-cs (5.3)

=

y

'r a yield stress value of c

= 53.5 ksi at 70 F (2), the corresponding 0

y3

..value is:

' ' values are the value of J after some slow stable crack extension has

':urred, hence, the Ja resistan:2.

5-8 Ic

• 451 k iM,

.a a simibe ranner, f:m tne minimum C:de specified yield strec;th (c. = 30 y>

si) at emoient temperature, the corres:onding value of Kyg = 323 ksi G.

these tests Q), the bending moment increased after crack initiation to a

.1

<imum vslue at knich:

S 0.135 in and c

= 63.5 ksi.

=

e 73

rem the equation used above, therefore, Kye = 571 kstE.

o 0

~ tn of t're previcus estimates ar? for 75 7.

At 4000F, the corresponding values of fios stresses are (1):

cf3 40 ksi at enck initiation,

=

50.5 ksi at muimu:n load.

o

=

g 30 that, assuming C00 behavior at 4000F:

K 373 ksi5 at crack initiation, yc K.

511 ksiM at maximum load.

1 2.4 Crit :e1 F, law Si?.e Csiculation d

It is clear that '.he fracture toughness valtos of Type 304 obtained 1 Section 5.3 are unrealistically high and confirm th? hign cuctility of r stenitic stainless steels. Using applied stresses equal to the ultimate

=

5-9 er.;., of the r.suriel at room tem cature (c = 8? ksi), the icwes t tough-as value de.arnined v2vicu31y (Klc = 323 ksi/i.i), LEFM (Eq. (5.3), criti-1 flea si:2s which are ;mtar thm 97" of the wall thich.ess are predicted, inte at thesse hign tou;hnest, levels tne LEFM :re dod is not' appropriata, the h

. as ti:-slasti: C33 ret.5od wes next examined.

/

\\

/.

e t

C-1

.,3 C??.CM CD ""1 OIRNTT;T (r:Cl *.'llLYt's

.1 Beckcround In this Section, the centerline la:k-of-penetration is analyzed to a-temine the ;citi:al cond.itions for fracture using C00 methocs.

The basic rinci?les of EpF?4 have been develoced over several years (H,16,17 H),

2nd cne na:f onal standard exict. for C00 testing (,1 ).

The methods used here J

follew : nose outlined in (JI) and (13) and only a brief surmary of the prin-ciples is presentad here.

In :en:ect, the CCD aperoach is similar to the LEFM approach in that

'e criti:al c:ndi:icn is ruched when the applied K, or C03 reaches the i

resis:anca level of t:ugnnes;. necessa y :o cause fracture (K r6).

The ic c

COD me: nod is cenoletely cercatible with the LEFM approach (E) and can be used in place of the :(ge approach.

For elastic behavior:

Ba a (6.1) 6 loge.Sec where 6 is the developed COD; e/, and o are the yield strain and stress y

respectively; o is the aoplied stress; and a is the half creck length of a canter crack plate modal..It can.be seen from Eq. (6.1) that as o approaches y, the deveicoed C03 becomes infinite. This only occurs for the elastic-perfectly _ plW c matariel behavior used to develop Eq (6.1).

Foe work

':rcening raterials, the relationship between developed CCJ and apolied

train (for stresses above yield) has been datarmined using analytical,

6-2

.rerical, tnd experimental methods (21, 22, 21). The relationships htween

D - d a;?iied strain are:

o (e/ey)",,for 0 c e/ey < 0.5, (6.2)

=

e/ey - 0.25, for 0.5 < e/ey, (6.3) 4here $ is the non-dimensicnal CC3, i.e.,

e (6.4)

=

2:e ya and e is the a:olied strain.

5.2 Cr?f< 0:a '-c 04 31&:2-a" * ("'7).in alv d The centerline lack-of-cenetration defec-is modelled in two dicen-sions as a flat piata with a central through r. rack.

The model is shewn in

~ig. 5.2.

The critical CCD value is detarmined from. (1],) for a sinilar custaniti: stainless steel and: value of S = 0.04 in. is used. (For this e

atarial, extensive plastic. strain is required, beyond the initiation of duc-tile ruo:urs, to initiate an i'nstability. The value of 6 = 0.0a in is, e

nerefore, a c

nservative measure of the available toughness.

This value is also mucn less than the value frnm (2,) discussed in Section 5.3.

In (1), no instrJmentati0n was used to define tha initiation point, and the difference in these values results frca measurements made at two different amount: of stable crack extension).

Using this value of 5 and Eq. (6.3), the critical flaw sine can be c

calculatec, and these results are shown in Table 6.1.

It can be seen that for the plate codel used even at strains as high as four times yield strain,

6-3 TABLE 5.1 CRITICAL FL;,. 5~E2 CALCUTE? F,0M CrACX OPDI' 3 03PLACOF.IIT METH005 LPP!ICOST:C'l CRITICAL F' J',! S

• I cley a (i.n) 0.5 17.0 0.5 12.0 0.7 g,4 0.3 7.7 0.9 6.5 1.0 5.7 1.2 4.5 1.4 3.7 1.5 3,1 1.3 2.7 23 2.4 2.2 2,3 2.4

- 2.0 25 1.3 2.8 1.7 3.0 1.5 3.5 1.3 4.0 1.1 a

4 1

'f*

s f

nr 4

i

6-4 the centar line lack of penetration has to be greater than the pipe thickness

o cause failure (i.e., 2a >> w) end leak before break is assured for tnis rodel.

The means that with this model, throug5-wall flaws would have to be present to cause failure. The ques: ion of how long a throu;h-wall flaw can be befor9 frac:ure will occur can be addressid by using the same model but c;nsidering the crack to represent an axial through crack in the pipe.

Usin; Eq. (5.3), it can be calculated that it would take about 21 times yield strain to :ause f ailure for a through--all crack ecual in length to the wall tnick-ness.

Sinca yield s:rtin in this material is about 0.2",,1: would require total strains of the order of 5% to cause failure. This ar.alysis coes not inclu:4 a bulging correction factor which would redu:e tha cricical flaw si:2s shewn in Table 5.1, particular,y as the flaw 5eco +s long in relation to pi;* radius. This analysis does show, however, tr : the limiting failure critacion will be plastic limit load rather than fracture.

5

7-1

7. 0 LIV LOC V;AYF:3 7.1 Intr?de:tien As dis:ussed in prwious stctions, a limit load analysis ess par-fan ed to determine the limit pressure or plastic burst pressure of pices

ntaining the lack-of-penetra: inn defects. A schematic shdi.; the # law odel anc gecmetrj is shown irr Fig. 7.1.

The critical internaT pipe pressure to cause failure was calculated

.. sing an intaracticn relation ccmmon in the analysis of steel structur:-5.

This relationship between tha apolied cembrane load (p) and bending rc er.t (ii) to pr-:::t failura in a beem or plate with a rectangular cross-sec:icn is:

+h=I (7.1) where P and 14 are the applied loads, and P1 and 114 ara the limiting values of

.t.

p and !1.

Equation (7.1) is used to estimate the internal pressure (p) to

ause failure of the pipe. The magnitude of Fj and i4j are functions of crack length, 22, flaw eccentricity, e, and material properties. Since the pipa

eometries of concern is a tnin-wall cylindar, it will be shown tha

the bending stress across the pipe wall is negligibly small.

The conditions which were assumed in the anal ' sis are semari:ed

elow

(1) The material behavior can be represented by a rigid-perfectly plastic solid (i.e., fully-plastic, non-strain-hardening behavier).

7-2 SECTICN OF PIPE WAl.L

, f' i

1

-a

.u e,

J

~,.

y l

1 s

s

.-._ _ f.,

,,.4

__a A

i l

a 1

rn = P.adius to Cantar of Wall Centeali i j

of Pips k

Figure -7.1 - Showing Para cetars of Limit Lead Model in Typic.'l Pipe Wall See:1on.

'. I y

-f

73 (2) Deformation and strass stata is two-dimensional under plane strain conditions (3) Flaws are infinitely long, pisnar defects with a uniform track lengtn (21) and can:erline eccentricity (e)

(4) Matarial yield locus is repr?sented by a von Mises criterion (5) Stress statas which are in equilitrium with the apolied pressure

.ad wnica everywnere satisfy the yield locus are assumed in crder to c:mputs tne limit pressure of the pipe.

Under these conditiens, tne ccm:uted limit pressure will be a low +r bound estirate to the theoretical limit 7.2 04 :a -m l.aa " e* - ' ':: -- - -i 1.3.ce c t ' o-P r " '. -i d e s F,_._.M._t The limit lead Pg and.orent Mg are determined from the g?ometry of the sec:itn (see Fig. 7.1) and One material properties. Accountir ' for the reduction in aren due to the fled, the linit load in pounds per i of weld Pg = t(1 - 2a/t) (2/v3)

(7.2) whers cg s the limit stress,. and the term (2/v3) is a factor arising from the i

von Misas yield critarion for uniaxial tansion under plane strain conditions.

The limi: stress is normally the material yield strength when the material behavior is perfectly plastic. However, for materials whicn exnibit signi-fic:nt strain hardening, ej could range somewher? betw+en e or o ts, and y

u the a:propriate value to use should be determined by tests. As an engineer-ing approximation, the average could be used:

8 6 g

7-4 of = (c

+ cuts)/2 y

(7.3)

' terial test data are : resented in Sections 3.0 and 9.0 to quaatify act 31 i

c.:llure behavice.

The limi: cceen: is ca.mautad from tne relationship:

Mj t(1 - 2a/t) (i1+ig)(2/6)cf (7.4)

=

"er* tne :ar s J and 9 are the centroid location of the unci :Wd or 3

2 re.aining arets above and belew the plastic neutral axis.

For positive ec:en-

-icities (see Fig. 7.1), forca equilibrium. requires that the shift in neutral

41s())is

}

0 (e

0)

=

=

(7.5)

= a (e > 0) y Theindividualcentreids,Jg andjg are com;;uted fro.m:

{k

• 7/

a)

(e 0)

(7.6)

J'y 12

=

(

trd e-a

+?+3-a

}y (e >

=

~

0)

(7.7)

ya (7.7) 2 (e

=

+

> 0)

..... ~....

.. ~

7-5

• 3 Detar-iaatinn -l Ae:Tind !. nadi The 3pplied tension and bending loads tcross the pipe wall w*re

. :.emin-d f rom t.% pre.ssur? strasses given in Se: tion 4.0.

Ey linerrizing a. stress distri.>ution, t.9e corres;ending membrane und bending stress ccm-ents are censervatively datamind frc: :

ICD +# 0- )/2 (7.8) e J

m (g-egg)/2 (7.9) c

=

3

'a:stituting E;s. (a.1) and (4.2) in Eqs. (7.8) and (7.9) gives the follo4ing

r;ressions for e and c :

m 5

(3r,2 + r,2) m

-- - p (7.10) e

=

2(e

- r()

g P/2 57'11) c b

.'ne applied force and moment are detarmined from the stresses using the fol-lowing expressions:

P t c, (7.12)

=

4 (t'/S)c II'13)

M

=

b

..4

!!urerical P.esults The ncn-dininstonal prassure (iii = p/cj) was cocputed for various pipe i:es (see Table 4.1) and flaw eccentricities using a ecmputer program. The i

fect of flaw eccentricity was small and, therefore, can be neglected. A

m

~

7-G vi3t of p vern;. f' n len;;h is shown in Fig. 7 for all pipe sizes. By corr-r;lati.ng-tt:e fs'lur.? diag am in Fig. 7.2 with the estimated unign pressures 1

tven in Table 4.2, tb r siind critical flaw si.:2s under these conditions a

,re de:arained to be independent of the pipe si:e. The computed resulis are iven in Fi;. 7.3, assuming failure ccnditions:

(1) cj = uts, (2) cf = c, and y

l3) cg defined as the cverage streng:n given by Eq. (7.3). The values of

- and s '

are the minimu:n cede levels at the appr priate tenversture, as j

u

,iven in Tabla 3.1.

For cf afined by Eq. (7.3), the critical crack penetra-d 5

-ions range few EB", for room tencerature to 48". at 6000F, Therefore, rather iar;e flaws are rcquired to cause 'tilure by the model developed herein.

7 The results of tais analysis will be used in conjunction witn tne Q

material perfor ance determined fr:a actual tests 9 resented in Sections 3.0 y

ad 9.0.

e t

a 1

3 w

e

a 1 } ' '., '. n. '. : ! p '.N, ; 1 h ', I I

l!;,.!'..

W:l.

. w.L.,..,..,;... :i.l;[ '

.l.

.. o.

..,._.n.

.;...e.s

~.

j!> ;[

.. _ ~

F ja 4

f

[^

n; ll'>

- ['

~ > o. ' D_

.l

.)

e y.. @. :,} 4.

o h' J.4

.,t.y,. 3

.m~

_ _ n. - 1. j

,1 p..

4

. r. _... 0.

., ~. -

o

.a. - l : 1 'l'.. : l c, h u. j i., '

U n p.

m i

.l.:;wy..J y,, }.; C..j. i l : p q,,

p, 1.,..

j' L n..

g ' :.u r

p.

i

-.'.g

,l.,.'.,

sy-

.!_m,

. f.. n. d. e.

.J.i j

.l 3

c < p>

m

*

-u-. -h ~

.... _ t

4. m..

.l:

1.

r._.

.l.,-.;'

4 e

T'. '___.

1....

'l

_'tj**._._'..._u-

. _'- Q g"__3.i % ).'"*"j

1l..,:i.~*t* ;i:

t l :!s.

.a-

. ; ',s I

.t.

.._..i.i f

p'1.11.4!' l~.-_..:...h ipi '; p"HO 1 J.

l_, /,2. 0 C. iF0,.~ -A M i'.:

.p,

._..s

-. F.'.j o w. m- ~ p-

.j T

P

.l.V f..._.6,'nl, afDip fa 9 fp lh:

Q'l.d #

.1.....

j.L..i.

.1;.

. h'..,m a.;. l

. : h.

-~_... l.

'. p_ : w a N.

F. a 1

i.

j..-

~._.. ".

n.

1.

... w. n l

I.

,"pc N m. n.. g.. n y.a n..n

. t;i.

s,

., < 1 v'

7h

.' :Fu i

n r

yn c,,. n..,. <.,. :

.. p! 4 : p: h,,..a.

_.e n m " h;-

n y h : n,

. J:

_~.

.'r r

.-.d._,.LN.'. '

.,., y mh.

W v. a w....;..

4, :,-

a

m. -

1._: n. i. l p.. : l p. "

m_

Vs.;','

o%jh

q. :.n..:l

,. _T n..h,a t:'

tw r u

l n,

3

.d2s -

v o.

l!F.

.fl:i;b.; -... _

'. d ph.

_\\

.I i m :F a.

.r'"

.......i n

p _

, :H.pp ' -}

y _ q..d"p p.

[.L' H 7 lh F.if. li y ; -[

t

h. q.

v.

-.nu 1: A /u - n : h,..., u ;l n.\\, ;. l.p e n. :.. I M,. Lra

' l. M..+.

.h, m

.p>

m m:

14 4.

"r *'. ' > +u

't -

d' e

i."-- i"n : '

.n a'. c"l.b h." n I

U

'N t, i

. i. m.

e,'.,e. 3,..

". k..,

,i. p. D ".l.-

.- c.. F. : L.dI.i : p i

.I s..

y %.a. :li ','.'. C, s.r u..

'. u p'

' @ & K.e r hi" ip !lf qp a

1 g @ ]i :fb

r

. f

',., ', -..-...ab..,....

.....,!m*,

nf ; k, h.N

.., _ _... ' h.

_Ifj:!'}.._..-

' f I : i.: f f '.; l

l..'

8 lIf s'j

.. l..i '

i

f

.L'

.i:

h N,,,.E

. lj. L - - i:: "\\/

ii}' Ui

..j! - E, ";lh ~ pIpf..

-t! l i

6. L

_ -] n. -

~. !. N.

i l

s s,

g W i p.. -.,.._.,... s,e : (

-s J... a N %

pr lN, 0, s.

n,.

a

.L 1-u

...._...-(...

o oi

+

k-..s.......'.i'.

t

.,> l

.h.

_%. x.,r e-

.c'..._ - ' / A. INN; a l;,

.p. [%

,d i

l 'Nt__.

N

j. jb p

-ji U

i

.m i J.

s Q >a p..p. m.x N

jk ( L. m.

s M

.pmp r :i.

.--..__-- -.w -..-.

~

a..1 _x_

l{H P - "

[ y, '. w.4_.p ' N,

['1.' y 'l:

~

Q]. y :.lb - I :lF l

k.' ',

[. 8 9 I

s'

. l..bl g'

q. g.%.

g5,i g.

4

j...

.m

I, _._..

,~.

.m

. m,.,.c-

..i-

,q c..n. m u:hs.. sn,..e. n ;.

. p. ~-

p.

x

,w m.

s

_,._,.,. f,;;

%. m -.,s r**9'n.

N.-

p.. sr....o : m. g h,, v_.y b.1 m r

s

_..,. s :.-

m.

l: g' !.

.l:':';'

-l' T.

i: 't

.. ~...

s..

. j

- ! N Q ' j..-

',,,a.g j, N g(..

. _ ~.. ' ~

m. s.,. x.c,. tn%...%..h.J.. a.,

u.

m i.

nt%. n n. e.o :p.u s.

l.h,..L.: p. w=.1. q...i. e.t

a. I.

- L

.l p.

..:.g.r.-.y.:

n, - p. n n..

. m, [ n,1.

q

.m...w

1. ,. m..... :,., [... n.,n u.

.].

g --

-.........-....s_.

o

..4

.-.. g*..i b;,i;j..b,,g'. a _j-l.

.b.,..j.;ii.

...iji... ':..

1 !

i g. : l.. M

.: i..

j c. h';[;

e

- :t:f n: :.HL t 1:

.i,jllif!f!ll,i'.fliflf!!f!'l.'f!!f.II f ! * ! !I.,.!' l' * !' ' ;? :

l.j.:ff f

$!! *!f!I!

.,._...r

/

1

,x

~

'f!$5lf!

.-_.a.-

.t.A b.il.!'. l. I I N.*.Y M d/ # @--t, k, N, (I I h.. d.ek h!

. '7 l :

. i, ;

1.

_. I :o.32 I'

r a e

.iWi:"M.. - i ih' _:iint!W 4 tiMU.' !!hhMih "HiHh E I + 1_. F :N 7p.pg :

pgg:p: ply pmyam 1 1 ea y

+

_s,

yu m;plM.'d-n,!' ::, g ! ! :

. p o.. o,j-it..;j' al.,

IM:

,,,P:

j-

,l',

'E h :L;.. :

'.f h t

' j l' : :U;)-l1:

l>

n.

-in

_.1 : ':

i D..

I; unn'

/c' / 6e. 9.y... s

.c./.:.: l du /, 'j Lu n &. m. ht. 18'4 a.= %..i m/Lc s<... s. w-s 4

r!

.i:a 1:: 1

.__g:

- 1:.:h:

/ :n a/ ws3t:

u}'o

~

c n

K i ~

Ii h

h h

k:bilN!

!!h. :

-. ". _T. -

l.h 11 h ; -

I

.7 i:..j!e..

., j @ i. y. l p:. : l.u.yb b_b h4 k h.:

1 f o r. '..:.l".', U,, l i. ' ' :-. -.L.p' "q.

.d..t'.I;..l.

G I.s.'

_.I: L l :.:.!.:

'H.it.

l.l...

n..

m ' t u l ' l ' n h i t i ! nl '... ' h. l ; ; ;. b:. '

'. ' l. _' _U lL 5.

h1 I. : :6.1.

,,n : :l

.o:

u.,, h u. h !": i t : h e m - '@n u L i.l.

';!a : :ci j:'rS' i-

1.. P.. I. :'

e e

...'.*% Og j

J J

n s

v 4.-.._.,.,.

.....-.m..

y......

y l

, l.i i

iya

.,I.-..

. l...

.9....

l I

,..u..,q 4

i.

L l

b, L

. m.....,..g.

....o..-.....

..g.-

a

-4 4'......

., l e...

.~.

e

_..........h

.s.

6

..i........

. _.= =..

g.

.r 4

t.

e

.e 4

-. i..

1.-

.e 4.

a

.t..

...e.,=.

....s e

t

..,...... 1.

,,_.,9

~

,..-..e6

,..g 6-a

.t..

g

..e.

+,

i.

e t.

.u.

a...

e

.h e.

...W......

i.-+... e i

..e..

4.--..

4

.m 1

~.

4.

.r

. e.,

. i

.. ee t

..,.... ~..

e.

,3

..........e...,............

~. ' -

p'.

- -.....,.w..

...... p..

.i, e

... l..

y

.e

..a.

..**6..-

• • a**e

• • ..e-..6-4.

.-.e a..,

.M I

e

.v.

.e s

1 p.---

.J.

,-W..

...i....

.-.+.

.4

v.......

e

...a.

........_e...

.. -......i...

........... m....

,4.

e

.g.............l

.i 9.

.....w..

-4..

I..

..p t

..v.

. w.+.

g.

.............L.,

4.........

5.,s

_.., f. 4. ?]

........t....._.......

w Z.

g..,._........

.i-.

j,.

e,'...........e

...=e.

.-i....

f.ffag/ e, ) / sy4

.6

.i.....

. w.

.e.pe ~,..

. ~..........,.. -..,...

e...,

...e,a-*6 1.

,a.

.v..

e..

+.

e..-

, -..... _ ~........,,.

~-m_~..,._..

~.

. s.

s,,.....,....

...r.

.4.

s.. -. -......

.. ~

.. ~.

.. l.

.. ~..... <......,..._... _..

t.1 t.

-... -...-....l.

.s.

,......_._.,............n....

..i.....

. s'..

w

.4 g,7.-(",".*..,,._...,._...

4

. t,.

g,(..>...

. -..- q........ _....%...,..

..m

...L.

. g. Nv, 1..

-g,,

~.

m,........

~..!._...,..._..._....

w....

.. _. ~.........-..

..i

(.....

~

. ~.

_w

...e...

p.

. t,........,

g; y, e,.

N._............... _.. ~. _ _.

,s

. y.

...l.........

....i.

5

.y

.. 1.

f.,.....-

'N,...

......l.......

....--....,..........L..

.s

......,......L..~,.i......,

g 4.

..f

.. l

--.f,..,

.,,. = = -

..6-

~. w,. p f........p.....

. d.,;. p g..

....I t

. rd ~.

j..[.,

t

.a==..,,,

..g r.%..,,, ;

m

.,+

5

. g.. t..

....J._

...........,,,,,Lv N.

u "*

.t..

....I....,.

n....

.I..

..... ~ l

......l.....

..,s j-_.-..

.. !.....,.............,............. _......i

.t

...T...,.

.a c-

. ~ -..

.l..

,t...i-..

1.

.. 1-

.6 s.

..I.

i.

.4

%,......m.

g t _..

... _..... -.. -.t...... _. -..

.i......i.

.i.,..

...f....

...=...

. s

..4

....s=.......

= -.

.....q-.g,-

e..

.c

..+.....o..

... 6.

A..b...i m.-.-.

.-..b /.,. w 1..,..

.=g.v"....*e..a r.

l-j i g,.e..

........-.,.i -

q pg

,g

.i,

es.f _

+.

..Aqsh.....> e

.. e 8ka.

I.

.p..

E... s e,.a.

........ i.

.i....

-....l...........

. l.

...t....f

-.......l......,......

.. 6

.}

.. + -

s t

.......W,4.,p.g,........4..=...

-^..

.....r.

f.a. -...

... *l.........

..b.............I*.....

a

.. l

.......l.

.... ~......,...

.(

..l..._...,.,.....,

i 6

- s.-..

..t.....................

...6.....s..I s

c.

l....I.

........ t

... l.

..a 4I e-

... -....... ~..

~... g

.f...

4

....n..

....i s.

,.-s....-... -....

-- i

. p..>

.s t

s.

-............a.......

...e

.-.,,.. +......

{........I

.t.s.

...e

..,....-..... -............,...l.

I..

6 4

... ~...

,,...t...... -

..........-...,.L.

.....m...I...J.....

i....

s

~... _.,.....

w

. f**I l

.t..1,....

wc.....

(

. e,

, - ~.... s O..,'2

+

.a

.t/.,.

i 6...

1

--..A,-=......J,

+

s. u,

.s i

,.. /.

d.V.. o.,.,e.,.

...e..

.J

> -i

,h...,)an.

...... - ~.

........a......s..

s~..-....i

u. I.

u...

.a

.,.I,....... -

.e.,

F,. --.. -

... I

• • 'h.,4.,....,,,..........,-......

=

.e,

. --,... c.

l..

[. ~ a. 3. n% -

I

.t

........a..

..,L. n.

i

.i.

.1.

..l...l f

........r..

~........

.i... d,.

l.

..f..,...

8

...t f..

.l o

i

.a.

.l.

.......),....a.l e.

. I.e e-.

..,...,..,.............I....... l

..r

..9..

. one - -..$.

..p e

. ;...p...

.l.

-l........

.v.

.a,..

i 6

I

_ _ =.

~........._.,~,

8-1

.0 fM c N Ni_t_r.=_ TE F CG oE U TInNs.

A series of mechan 4 cal pr:per:ies tests were conducted ((l) - AESCD-i, 'I300 ',, and AES:0-15) in order t: cuantify the effect of lack-of-senetration en yield stren.7:h, tansile strength, percent elongatico, and

,3rcertage reduction in areas as'a function of flaw si:e. The results of

nese test: are given in Appendices A and B.

Appendix A provides a summary of full sec: ion (e.g., no lack-of-penetration.) tensile tests on Typt 3C' material. This inforretion is

~

provided for tensile tes:s on botn pipe and plate material for caree hests of Youngs:0-n steel-((l) - AEICD-5) as well as two all weld metal tensile tests ((l) - AES 3-11). The solution anc.ealed weld metal properties are virtually equivalent to those of the base metal.

A sta;istical data reduction was peric ~ed, and the result: for the comoined basa and weld metal tests are given in Table 8.1.

All oc:urrence level values are for 95% :enfidence.

The occurranca level values provide an estimate, ta;ed on the mean and standard deviation, of the observed data of the value for which statistically 90%, 951, or 99% of the expected values will lie.above.

Appendix B provides a index of other tensile tests perforrec. T ese include-tensile samples. fabricatec with varying degrees of lack-of-cenetration and t?st coupons tested in conjunc:fon with full scale pipe burst test:.

In addition, the-all weld metal full penetration test results are repeated for ccmparison. The results of these tests are presented graphically in Figs.

. 3.1 and 8.2.

As can be seen in Fig.-8.1, the yield strength for a given

4 4

8-2 TA3LE C.I SU.t%;f if CATA Pd300 TION ANALYM3 TE.':51'_:

Y! ELD PERCFNT ST?.E fr ~ :

ST?I W u gLo,iq aicy (ksi)

(ksi)

(ksi) un Value-85.3 42.3 57.5 M Oc:' r*er02 Lavel 30.1 31.9 43.9 95". 0::urrencs level 78.5 29.E 40.6

>]f. 0::Lerence Level 75.3 24.1 34.2 S

w h

}

l

.1

.,~ ~ w

... _.,..... ~

/

.Q v'.4 m

$)

G o

a C1 C

O L.

.C e

m"

.>m T

U C

M r=

• s Q

n

.m M

Ce ti O

• )

8 c

m s

'J L1

• O Q

w U

U t

.~

D

• J b3

>=

F*

O O

n M

M L.

O O

C L.

p

.M U

.U r

_^*

94

=

e H

M e

v M

Q W

U N

"3 O

fl 4

94 at 3

k

• )

6 Q

O h

N C3 M

1.s 4

at

.C E)

C

• " C%

O O

U C

t.

O O

L C.

• J I

O Q

'O N

=

()

M

'M a,.e en a

g

(

U 0

r,

.4 a

O M

Q C

U t.

H e-

• e O

'N

~~

e-o DL 3

.C1 g'.

O-4

.p4

^

.- ~.

4 g

C O

O O

N O

CO W

v m

(m) 'nt.meu.ts 0131unistm-

4 4

8 0

', mee _.

! c.;

4>*O

• wee e

i e

C3-

~?

e

' t$

".*J

=

e Q

O

~

M M

41 Q

H U

e=

name O $

O t2 U

C H

M

~

.U.

O b

.::ll

• J

.N e-g d

N

.r..)

N

• ~

J 6

Oc y

O

'J 6

-i c3 U

cn Q c

t a

ti g

c g

U O y C

N L O

i O

c.

"m I

e o

L1 Q

C N

O eve a

sa d U C

p M

g)

U W-U gj L

Q Q

Q.

I N.

~

O O

M L

3

.Cn

~~ D O

6 Me'4S M o

- a e

o '~

O O

w w

n N

uo pv5uat3 quaoaad' J

4 s

/

8-5 52 ple is virtually indepencent of the size of defect to the limit of very h "ge defects. The variation over a range of 0 - 47P. Gefect size shows a

' eld streng-h range of 33.6 - 46.4 ksi.

The tensile strength data disolay a greater scatter with a decreasing value for increasing defect size with marked decreese-for defects granter

' nan 35" of the wall thickness. The corresponding percant elongation values also demonstrate much experimental scatter but even more dramatically, show

.ae exce: ed decline with increasing defect size. A comparison of the acclied s:ress to failure to the plane stress limit load prediction for these data and plate soecimen data from (2) is shown in Fig. 0.3.

This c:ecarison indicates that for small flaws:,. the locus based on c would be appropriatt uts snereas for large flaws (>50% penetration), the reduced ductility would indi-c2:e that limit load based :n' yield str:ngth seems valid.

In the transition region, an average stress or' flow stress, as proposed by Eq. (7.3), could be used anc may be a reasonable approach when comoined with specified minimum tecperties for all fias si:es of co'ncern.

The effect of flaw size on: the ultimate streng:n of pipes was de:er-nined by ccnducting full s12+ pipe burst tests. These tests are discussed in Section 9.0.

From the behavior of small tensile samples and the measure-ent of burst pressures from full scale tests, failure pressure models can be calibrated.

h

e gem

.._ m

. q.7y a

~

l:.

,b hf

!a

. a,9

}

't

.l l

l 1

i l

I

},

1 i

I i

I

._.._J'.__.II f;,

}

j' f I

I l

3

. _j -

q l

I l

l i

);

.- 1:

1-1

..j.

l l~

I I

i j'

i l

i-e.

t'

'.i 1..

I'.

,I t

i. '

ch.rr sc n... c.w -w._w-1.

~ 1,. a.__

. la, t l::

.i.

g - - 4

.g.

- ~~~-

i i

.1-g,,gy. w,.,,a y,,m_

l f,

l

! f.f [

-[

~!'

'}

~i 1 3 *! [/W ' C d l

l i:-

.e

..j. !

j' l

'p t

j.' t.

i.

.l::.

_. y.,.

,; 27 f*

f l

l l

l. di

.l.

'h

.. 'f;

.3.

l.

MF" ' 'd) ;

I l

y l'

-y*.p r

l-_

J All** /.hM', O \\

r.

j 5:

n 1:

f' l

,.l 3_l ll.

l' I

... l : - q, t,

l, 4

r :

r

.3 i

i..-

.1 t

' r i..

.A 1

1:

l e

l l

I

'N..I

'N, +.

l.

(

T'.

3 q

.f

. f.

I i.:l

'l l

i 1.

i.

l

.j.

'l I:

',\\ !

-l'.

- !^

n.-; ~

i.

f' l

i l

g l'

j f.3,_x..

j.

e I~D l l 1: I -li f i Uf

!.?.

d i

l l

1 e

(,e-i l *:

N I'

'l.'r l'

... !,,, i I,

S.

i.:. 4 :q u r'_.O fv I..

l.

.}'

-* w r,

q.X - 1

.wze en nw i i

'A.

i i

s N-g> _

N

.,w

-_. 'y,

,p,:

.j.

4_...e,

.. _. r_, (4 l

.,_ s

,, s l

t' N,;

- l i

j

,'.j:

l l

.1 pt

~

t.

^ _ _....

N 4

6 l(

]

._ _.J....'

. N.,

. _. f.

o A,

,f '

l l..

I.

.i

.l'

'h'

.a' N.

I l

.l'

.j l'

. T. :

i. l F,

y-

]:

}.

f N..

1

}

.i:

I i

1:

'l'

?

I l'

'. f;: / g l)f,c" ' /. 4. 2. (.

,v l

e-

+-

- l e

g l d.

i

[ --

i. p'7. t um,r.:r, '.t ; ' y

{-

3 p.

, s

~

.I.-

. i ol-I.i:

g g-l. I e.

t

.,,,i I

f lg-=e 6

.i m:

e. o!'

,3, W

• • U

.-~~'W

l f:

l

f:

.l.

l

'l f

l f.

I l

1 j 'j'

- -F

.l'

!,,,.gemm:-,chr=-w +- myl-ii j:

..;. }

l

.[

.j l.

.l#

l

-i f.

l l

.l~

l

! :l h.';f f

f l

I f

l.

I.

+

l

.o i.

r_G.:. sci".p.

1 G;.1._.MA.c/.._.5.,d:- 9.,)_e. ! Us _..'f,;_ - p.. -

.'l:

l' J

g Q'

j / /* _

t a. f -

t. i.. ' a -, - 1.11 f. a.:1 i.

l 1

j

{

^

f g

1; 1

p ip ' 1 l: m [l l

'f*

I '*

l '

l I

i l i

i: @ W j.

l' h 1 l. i 1

l l

g.1

?.0 EU?.ST TEST 3

9.1 Intr-du

tion To assess the burst behavior o.f the flawed oipe ac:urataly, a series

,f pressurized tests were performed by foungstown W+1 ding and Engineering

ompany under the supervision of Bechtel personnel. The results of this test

ro.

ram have been provided to Aotaca Engineering Services ((l) - AES:D-12 AESCO-18. AESCO-19, AE.5CD-20). A summary of these rasults is provided in Fig.

3.1. The fir.31 result ebtained. during the first t..r:

tes: sh:wn in Fig. 9.1 has been de:er.ined to be unrealistically high. A review of the test docu-

ants incicatas an accurate bur : pressure of at least 5000 psi was attained but ceyonc :na: point, test shor : mings may have sreventad an accurate final failure pressure.

An analysis of the the.oretical burst pressure was performed to evalur:e the test data obtained and substantiate the-limic load analysis results. An elastic-plastic. Targe strain, large daflection computer program oas run to evaluate the theoretical pressure to failure. This was performed ice bo:n a membrane solution valid in the middle pipe siction (without end affects) and for a combined membrane and bending solution in effect at the ends.

Com:arision of these results was then made with analytical methods,

he Svensson equation, and literature test' data from similar pipe material and geccetry. ~ The appropriate methodology and results for each of these-analytical tacnniques is presented below. The initial focus was to pro-vide a theoretical burst pres:.ure for 100% weld penetration. From.this value, pressures corresponcing to given percentage lack-of-;..enetration could

l

~

.e..,.

.... -...~. <

u.

m.., 2 -, :. u. fn, 1.:; : n 5:

..w r

m.-

..3 m..,1lu,,.y~ M,c..: ~121 O. q.. :,.

g.s. %

,,.:..Ti.; p.

. n+s.

TUCE 2 m..,,.. > (,3..n.. w z.

.... :., o<

.c.. v

.,.,,o.N.lgloi..h/:

Lack of Penetration: Approximtely 15%

..):

',?.'iIl/

i

,.'y.. p<

Oc,.

' H - l 7. ? - '..'.,,,

Burs: Test Pressura:

5000+ psi u

- M A-} 5.,f.N ; M I.f; Q$.',:, %,u

.U.d N:.

. ~.

$\\.'".!

Tensile Test 0ata:

TM

'S$

$. ce

~1 Tensile Stren,:.th:

81,000 psi

$ [lM. q..d.W O 0.".x:):e., jV..Q.. :.

U

-Q.:.

i Yield Strength:

45,400 psi pG r.~,:

3.,.e, s.,

.;..'a ;.y., R o a. :,,,.,.,, ',, y '.,

?.*e,.

• :: 9 t n.y; n.

4,. a

,. a*

~

w.

i,. i

. 3+ s.

.s L.)..1

~.:.,..

m..

1 ypr.:

,u.

).

a.. :::

y %cL,,

9~,.r,:,T.L.,',

~.* ;.;;,v%lu.W. :. :

1. ., M

~~

. %. s', @s#

9 L.ck of Penetration: Approximately 40%

. !.. - @ ' ?

.

i

.p j -

.. ;m

. m +.. u.

;.;.., q M Burst Test Pres

ure:

3700 psi d....,.Ng...i.b..,.,'e ',Y. !*'.'

~.t e. E...O,..'.).f, d,. s.-.. e t.3

.>....m..

,.c ;.

inh, Tensile Test 0ata:

I. Q.... ; l.- E.#. Ly...,

f. _..,.M., f.. n. !J

'M'.y.h 'r4. 2.&,$.,

Tensi1e 5trencth:

57,200 psi

.-;,.. l g *.,

. d..i.;;

^

,p a

e.

'.c.@.. u...M?. s!%,..,. f(

Yicid Strength:

42,500 psi

.y.,.

.l;..+.i.V.u.'.?;,:sp'n.;..n,'.z,s%.-~G!N....

.s,...v. v a. r.:f U., Qc.,

<.r,-n,.~...*..,.-,...,

..sf,^.W

.en..,

.Off

. : 4.e Elongation:

8",

lv n

.3

.: a.

o.w.

~.;:-

. a.:g

.. n. J..,n,,........

.~n. L : e... d.: >-

d + ?.

,:. r :.

4.' m. :.n.. :t.. s.,. v.%..,,..

...,,.f, :

.. <.,1.,

s

.y

.. r:.

.3

.~t, -

.,a,'.,.

.g.9

^.. 'v' w

=,*

.~,

0

'9>

pi f

o L.

.co R

$ 2[,, {

J W,..~.; q..a,#.','.w.-... Q.C..~ Q.. h

.'.'.2. 9.*.

n

y.r 9 &...

r,J\\ ;q a ~.c 80 0 Lack of Penetration: Acaroxiiately 555 m

..v..u..?.o. p....p(s L...;<.,.,. m...en,t;m:Jd x,;

.w.s.

s F Burst Test Pressure:

3C03 psi

,s,

q

~. Ul"_.lj :q....

sms.7, '.1 9 '.m.f 'W.;,99'S Tensi1e Test 0ata:

.: ;;;v--;.a..,. A 5 g.-

g ; q ;'. *

<:sr r-~y b..v @...-I. e:.

7'hM.N.go'N.~i.}l.~j.

Tensile 5trengtb:

43,700 psi M..ld.Mh$.A.. h('.".,

.h..

m n.

.,,.a a na 3.,

5 2.. A:

Yteld Strength:

35.500 psi

.*.:. -c.? v.;.-/.

. g; y,d+fy+.p~.%r..,.',f;f +v.

9.,p.p..

Elongation:

3%

&.w v. u _,:4. m,.,z. a

. -.. Wr. ';, :W p.%

4,9 n.w..,. y :...

..... s.,

. a.-..s.

.,...:-.,u

. ~.

i <. ? ny-4.,.,f..n.,A.<;ipv'.1. %e,., s #:'ta,,:i.r: e.

y

~

,N.

w...-]

.) W, e*. g\\ 8

,g,

.Q. a \\,2 u,.} %

1;/, y%

y-Y-. - w a :,, ? ' sQ. f'N 1,W" a

,.: y$ "g.

vf 2f, j'.

s

..m.

4.

..o,a;i.

,g...,,,

W M

Figure 9.1 - Crm-Sections of Burst Test Pipe Sarapias and P. elated Material Properties.

9-3 evalua od. Once these limiting valuer, are determined, it will be possible ivaluate field situations to access the margins of safety inherent as a un:: ion.cf operating ce design stress levels.

.2 El utic-Ma;t4c Wil Analvti 3.2.1 2a:k;rounOnd Conouter Results The dicensiors of the shell analyzed and the nominil material para-eters utili:ec are displayed in Fig. 9.2.

To account fur the larg deflec-

ions cbserved in the tests, a computer analysis using the 8050R5 code developed e, Lockheed Palo Alto Rasearch Laborat. cry (25) was perfor ed.

7050R5 is an elastic-plastic program for axisymetric shells which can also andle geow:ric r.on-linearttfes. The von Mises yield criterion and flow uie er) u.iii:ad. A bilinear s:rass-strain curve was utilized in the CBSDR5 run. Thi deviate slightly fr:m literature data which have been developed

nrough tensile tests on Type 3% stainless steel (2). This deviation is

,nown in Fig. 9.3.

An idealized yield point of 45 ksi was used hero (should it be found that substantial variation from these values could be encountered, l'stnll correction to the theoretical burst pressure can be made). The.com-wter program is accurate up to strains of approximately 18", however, due to

he generally linear behavior beyond this point,. extrapolations can be made sith 900d accuracy.

A distinction between two regions in the shell can be made.

Away 9om the ends, the shell will operate with nearly pure n eebrane stresses.

~ield burst investigations performed by other investigators on Type,304

sinless pipe (25) have noted that even four-foot lengths of pfpe may not i

..., ~ -

,...y.---

Symetry t = 2w =.375 Plane (ficminal) i Y

C A

'.ise.nd to be clarnd A' i

End Plug 12.75" CD (Nominal) 1 I

~N

'(

y,,i n '

3

- 21"

.l flominal Material Pare.:eters:

c 45 k,si yield

=

1. ul timate 85 ksi

=

0.3 v =

5 E

29.3 x 10 p39

=

i Figure 9.2 - Theoretical Burst Pressure Model and Material Parameters.

,i

).

i 8-=

/

/

/

80

/

/

/

/

/

Power-Law Sebavfor Typicald 70..

of Litarature Data (1) y

/

/

/

/

60 --

/

/

,,\\

Linear Behavior Assumed For

/

Compu er Analysis

/

5 3 --

/

?,o.

9 s

.g

.?

33 -'

a 20..,

10 1 10 20 30 40 50 60 Engineering Strain, ceng IO Figure 9,3 - Stress vs Strain for Computer !!odel; 4

i 9-6

e suf ':ien to mitigste entirely restraint imposed by end conditions

actusa of the large expected deflections. Ho,,ever, in general, only small corrections are necess 3ry.

In contrast near the pip.; ends, there is e boun-JLry leyer effect of ber.cing stress due to constraint at thesa locations.

7ne decey distance for this banding stress is found from:

~~~

n )I 2r (9.1) q

=

2 12(1.-v) from :ne c.~isymetric problem where q

Bo'adary layer thickness (.in.) or distance from the clamped end

=

within which tha bending stresses and strains decay to nyligi-ble value Ncminal radius (fn.)

r =

Wall thickness (in.)

t

=

Poisson's ratin v.

=

Tne value is found to be q = 3.72 inches As the length of the tube between welds is 46 inches, tne two boundary layers at the ends do not interact, and, hence, the bending solution due to one clamped end can be obtained indepen-den; of the other end.

In summary, the two separate failure conditions ;crresponding to the appro?riate pipe region, car, be anelyzed. First, the membrane stress field in effect at the pipe center, which does not include a significant bending component, was computed.

In the first burst test, plastic failure strain of 23.3". was measured in this region. This provides sufficient data to obtain

9-7

neoretical burst pressure.

Second, the ccr.bined bending cnd cambrane ass field at the pipe ends c:n be expected to fait under e strain condi-

.:1 equivalent to tha-level found in the. tensile tests dic.:ussed in Section 1.

This strain corresponds to the elongations seen _of approximately 60%.

The results of this two-part analysis are shown in Fig. 9.4.

Both e effec:!ve strain (in percant) at the clamped end and the circumferential

rain in the cembrane region are shown. The results indicate that the esti-

ed tu st pressure thus obtained is equivaient for both regions and their

orcpria:a failure strain levels. The results incicate a failure bur. t essure of 5050 to 5225 psi.

Calculations were completed to assess the correction required by the "Nt of geometri: non-linearity and shear deforration.

It was found that

.e-effec: of gac etric non-lineeri:y may be neglected until the plastic

rain in the tube bec;m a large frac: ion of the ultimate strain (in excess

-' the 13% strain for wh::h the code is accurata). The analysis perfomed is

resented,in Appendix C.

The effe:t of shear deformation may be significant for a thick shell.

.aever, it was found that for the pipes analyzed here, the effect is small.

" e inclusion of shear cefor.ation will result in a lower stress, t: :3, by

'eglecting shear, a higner stress for a given prassure will result or an

ropriately lower burst pressure will be ecmputed. ' The exclusion of shear Mo: nation thus represents a conserva: ism in this analysis. The details cre snown in Appendix 0.

In su: mary,. the computer calculations are conservative, i.e., they wedict a burst pressure less than would be observed in actual tests because:

a, 70

-h

- + --

60% 5:rrin-Te, sile Test n

5 0 -

7 Fatlure Level

/

/

/

/

50 ~

y

~~

/

/

7; Ef fectiva Strain at Ciam pd En:

~l

~

9 40 -

I 2

/

am

/

i c) l s

t U'

30.-

/

uu

/

23.3% Obser-ed

/

Midspan Strain I

/

20 -

t

/

/

/

/

/

/

10--

/

/

Circumferential L

f Strain in Pipe Midsection

-. - -,n.

p....

-- = p u.....= -.. /..

==

.1000 2000

. '.100 4000 5000 6300 Internal Pressura, p (.i)

Figure.9.4 - Effective Strain vs Internal Pressure for Unflawed Pipe.

J 9-9 (1) The boundary conditions bere assumed to tar clamped, whereas, the 1

welded end caps (see Fig. 9.2) may allow some rotation. Hence, for a given prassura level, the stresses in the tube would be

/

slightly lower than those calculated for the clamped end tube.

(?) The shear defermation effact has been neglected.

If this war.

included, the stresses developed at a certain pressure would be lower than those calculated as shown above.

(3) The bilinear stress-st-ain curve used here yields higher strains at a' given pressuca 12 vel than a power-las (or actual) curve would Therefore, the bur:: pr233ure calculated by the bilinear stress-strain curve wl:1.1 be lower.

' : wever, on the other hand, tne o050?,5 code does not account for the change in thickness as the plastic flow occurs. This will result in a higher pre-

. d'cted than actual burst pressure. The magnitude of the corrections obtained on botn sides of the actual valua should have ecuivalent effects, such that the accuracy of the above analysi: is ' considered appropriate to the require-ments.

Supporting analytical elastic-plastic solutions were also perfont.ed, and these results are presented in the following two sections.

h.3.

Yield' Point Dressure C31ctlation

~ Two yield point pressures ware calculated for comparison with the computer generated data. The membrane yield (valid for the middle pipe region) was obtained from the initial conditions using:

9 13 6

P#o

1. y Y

t (9.')

g

'a yield pressure was c:mo.c.ed to be 3149 psi. The yield stress at the

smped and (which woeld occur earlier in the pressurizatio".) was found from re equr. ion o,2 (e total)(c total)

(9.3) c u

total v total y

e g

4

.here c and o componen stresses include tambrane and bending cem:onents.

g 4

The value was confirmed to be 15'r 7si. These yield pressures are shown in Fig. 9.4.

9.4 Eurst Pre tt 3 CalcuTr%d by War'ne T'aorv 'or ' arce Stra4s large deflec: ion and/ sis requires tne use of true stress versus true strain valt.es for the approprfata uterial.. The engineering stress versus strain curve is shown in Fig 9.3.

The conversion to true stress-strain-for input to B050?.5 can be four.d using elastic-plastic.methocs and the appro-priate deriations. For an effective engineering strain of 26%, the mem.brane burst pressure is datermined to ba 5030 psi. This is in good agreement with the values obtained by computer analysis. The following Section discusses

,wo other burst pressure derivations for cylinders.

E.5 Standard Cvlindrical Vns;21 Burst Strencth Calculations and rield Te;t Corre!at4 's

In additicn to the detailed elastic-plastic analysis which ::as descrised in the crevious sections, literatsre solutions to the burst pres--

ture-of cylindrical pressure: vessels were atained.

In particular, two

9-11

' :!.tions are presented fcr obtaining the required values ([6,, 2J_).

These

re:

' I #'

(9)

P

=

g u

aad the 3 ensson equath n c'F in W (9.5)

P

=

g u

eyl

..w r a Cu 0.2 0 a

=

(9.5)

F y) p_g g+. - g f-e u

S r.d F,urst pressure (psi)

P

=

3 S5 ksi (assumed value) a' ibminal ultimate strength

=

=

u Wall ratto outside diameter /inside diameter W

=

=

Correction tem for cylinder F

=

cyl c

= True strain at maximca load u

1ese equations yield theoretical pre sure values of 5150 psi and 5051 psi, espectively.

In the burst pressure testing by Royer, Rolfe, and Fesley. (2_5,), tests 5

+re perfomed on four-f:ot leng-hs of Type 30? stainless steel.

It was uncluded that the Svensson equation was conservative, that is, apredicted 1:w theoretical burst pressure for this material. compared to the observed

9-12

.dllure pressures. Ho;iever, that experimental test data did fall within the canda provided by the two Eqs (9.4) and (9.5).

t A summary of th9 burs pres;urs analysis cocoletad is given in Table

1.long wi :n limit load prediction from Section 7.0.

.5 Ceterminat'nn c' Burst Pres _sure in Field Invan cations,- Flawed Dice The analyses discussed in the preceding sections wara performed to Letermine the burs: pr:ssu.re of an unflawed cylindrica. pipe section. This Section discussions the analysis for flaw-d pipe of known lack-of-penetration.

't draws on the results of the previous sections and is modelled similar to

+.e analysis performed in the limit loed section.

Since the burst pressure

.alcula' ion eethods used above all gave consistent results, the basic analy-

• ical caustian was used to credict the coserved test d-.ta.

A com arison of i-

he results of this equation i,d the bur:t test data are presented in Table 7.2.

The ac:ual burs t ter c values are compared witn the predic:ed limit load s alues for the'10"' Senedule 40 pipe in Fig. 9.5.

It should be noted that the standard burst strength prediction gives values which are slightly conservative for tha flawed nipe tests.

It is c:ncluded'that an accurate burst pressure calculation can be performed on

'lewed pipu using standard equations since these w-re confir.ed by (1) exten-f ve elastic-plastic analysis and (2) burst test data.

Further, the theore-tical results will be conservative by approxicately 10% when compared to

tual test data.. These results can now be applie

to plant systems to assess

ne safety margin inherent for a given design or 3arvice stress level when

. ampared to.the burst pressure.

9-13 L

0.12 4..-...---.-4-....--.----

- iT.M.,...

i- -.

fl0TE: of = S:i ks i

• p > 5000 psi 0.101

~~

n p = 3700 ; si a p = 3000 psi

[ 0.03-n

~ *L i

i 0.05 -

s 0

T 9

0.04 -

1.:

4 0.02 2 t

o n

0 0.2 0.4 0.6

0. Ta 1.0 flor altrad Crack Length, 2a/t rigure 9.5 - Comparison of Calculated Burst Pressures With Te:t Data for 12" 0 Senedule 40 Pipes based on o ts fCf " Outd '

u t

s

9-14 TABLE 9.1 SL'M:4A?' 0F 3U45T PRI550?.E CA'.CL'LATIC15 F0~ CNF1. AWED TYPE 304 PIPis Th:nRe. t...,L e.i.e:S i AM4.v7T:AL M: 900

ME35URE foril I.

Elastic 31as-it C cuter Analvs'1 Mem. crane Regico (Midspan) 5225

?:.ading P'us Memorane Region 5050 (A Clamped End)

II.

• nal v-i c E' :3-': 011s*i: amilvsis Mecerane Regica 5050 Large Strains III. Ba:3: 'malv:ical Ecult#: s P3 c. ' i n',l 5150

=

5'conssen E:uation 5051 IV.

Lini Load (Se:: ion 7.0 and Fig. 9.4)

Basedonej 7

= c

= 45 ksi 3195 a +

Basedoncf=-hc,,

65 ksi 4615

=

Basad en c.,

85 ksi 6035

= cu

=

s

9-15 s-TABLE 9.2 BURIT PRESf/j)E CALCULATI0tt FOR FLA',ED P! PES LACK CF ULTIMATF SURST PP.ESIU).-

ET? "2 * *'!

STREN'?1 TF00RY fisi)

'".T" 0

85,000 5150 15 81,000 4910 5000*

40 57,103 3467 3700 55 43,700 2549 3000 r.al bur;: pressure was greater than this value, but exac: level is unknown.

b t

i i

C01CLUSIC,1S i

(1) Centerline lack-of-psnatration weld defects in Type 304 stainless steel j

have been analyzed for failure by fracture under internal pressure i

l oe :!i r.g.

(2) Linear elastic fracture mechanics (LEFit) is not an appropriate method l

l because of the due:ile nature of Type 300 stainless steel.

High tough-ness values were realized hver 300 ksi/Iii), and the consequence was a prediction of leak-before-break pipe failure.

(3). In a similar manner, eTastic-plastic crack opening displacenent methods i

predic ed criti:al flaw si:es greater than the pipe wall thickness. The 1

leng:n cf a througn-enianus naw whien would be required, cause frac-ture was found as e function of applied-strain.

i (a) A limit load criterion was found to be most appropriate to this material.

C 7

Curvas designating non-diransional pressura versut normalized crack depth were generated for a range of pipe sizes. These can be ured in a

conjunc: ion with field matarlai cnd stress data to evaluate the severity of a given lack-of-penetration defect.

9 (5) The critical. flaw death was found to be constant for all pipe sizes i

i based on allowable stresses. Plots of norcalized critical flaw depth indicate a dependence on choice of appropriate limit stress and operating temperatura.

I i

(5)

For a limit stress equal to the Code allowable ultimate strength, a l

crit? al flaw dep:h of 75% of the through-wall thickness is ob'ained for 4

l

o.

0 operating temperatures fro:n 100 F to 6000 '.

For a limit ; tress based on the Code allowable yield, the critical finw is 20 40f, of the wall thickness over the sam + termature range.

Finally, for a limit stress eccal to the flow strr5s (the average'of the'two ;bev4), critics.1 flaws range frca 43 to 531 of the wall thickness.

'7)' The' tensile test data indicate ther a failure cri cerion based on ulti-mate strength is appropriate for small flaw.; (<20", wall thickness). A criterien based cn yield strength should be used for 1.rge flaws (>50" thickness), and the critical flaw sizes interradiate to these flaw limits should be deta.ined using the avarage of the yield and ultimate strengths as the ficw stress.

(S) The theoreti:ai bur:t pressure for an unflawad section c' pipe was c 1-ulatec usin;; (c) the UUR5 computer code, (b) elastic-plastic large strain theory, and (c) basic analytical rela tions.

Gcod cgree:ren' betueen methc:is wu obtained.

Anal' tical method ware then used to predic: the burst sressure for (3) y flawed pipas. The analysis predicted pressures, which were unifomly conservative by 10"..

.9) 'These results can now be applied to plant systems to assess the safety margin.iaherent for a given design or service stress level, as a function -

of operating temperature and pipe flaw depth.

4

(

REFERENCIS 1.

Correspondenc1 from Scatel Youngstown, in:1uding letters and stress repor'::, as f allows:

AESCD-1 Copias of B section of YWf.E Co Pipe (SA312 Type 30a) (2 copies)

AESC ?

Hero:

E. Tekorovic, OA Denecmen:, RE:

Telex dated 4/23/79 (April 25, 1979); Attachment AESCO-3 Radiographs (e - g)

AE3CC 2 Let:2r-G. t'gan to U. R. Smi t ' ; RI:

Outline of work pr:; ram Lack if ;usion in St.tialess Stad Pipe Welds; and Enclosura (lettar:

G. Egan to R. Rogers ( 7ted March 5,1979); RE: QA Pro:ecures f:r Engine tring Analysis)

AEI 0-5 Letter:

C. Stanley to W. M:Naugh:en (Dated 5/21/79) At'.ach-cen.s: Three carcificd mat rial tast reports re resenting pipe w+1ced by Youngsto,.n 1,rlding and Enginetring:

12" Sch>dule E'i, heat #10311, 10322, 39967 (ca;2d ~/C2/79) 10" ;3= ed;1e F.C3, heat #10511, 10222, 39957 (da:ed 1/02/79) 14" Sch +cule 20, heat 141000W (dated 8/07/75)

AIS:D-5 Motes: Telep"One Crnveesation between W, Mc:.3.ughton and R.

Whi:e(05/05/79)

AESCD-7 Letter:

W. McPanghton to W. R. Smith, Dated June 5, 1979.

RE: Finding 1 to cate of dataministic fracture mechanics analysis perforced to evaluate pipe condition.

AESCD-8 Memo:

R. H. Richman to File, RE:

" Fracture Toughness" of Type 304 Stainlass Steel AESCD-9 Notes: Telephone Conversation between W. M:::aughton and R. White (05/05/79).

AESCO-10 Bechtel " Wall Flaw Calculations (sent 05/05/79)

AESCD-11 Bechtel Tensite Test Data Sheet AESCD-12 Bechtel Burst Test Calculatian Sheet AESCD-13 NRC I&E Bulletin 79-03,."Lengitudinal Wald Defects in ASME

- SA-312 Type 304 Stainless Steel pipe Spools Manuf'ctured by Youngstown Welding and Engineering Company" March 12, 1979, i

e

\\

l

%[

1. +<>/

%*4 4A t

AeEE A<e T,e~

TEST TARGET (MT-3) i.0 5EMIE 5 RS DE-na m

l-l

$

• UN

\\

u 1.8 s-I.25 1.4 1.6 9

g hU N

MICROCOPY RESOLUTION TEST CHART Y

ll/

4ls %

ff qff '5

/'4 4

• ?kfb//h/

g w sy$

Q si, c

y3 y

1 a.aw.m.m._J

9 4

AESCD-l? Me. :o :

R. H. Rich::an to rile (Dated June 13,1979); RE:

Critical 8:Pec: si m in Type 301 st:inless stel pipe with incomalete f s n et longitudinal wiids.'

AESCO-15 Laboratocy Certifinta ( Anamet L,boratories, Inc.);

Resulte of ail

-sid ae ai tan;ile :asts, Da:ed June 11,197?.

AE5:3-15 Sechtel Pd. im cacy Repor: on Invesciga tion of '. eld imoir-fections in A5ME SA-312 1;uble Ueided Austeniti: Stainle:S Steel Pipe for Comp;ianca witn i;RC 1&I Bulle:in 79-03

(;ated June 1979).

AESCC

'.7 Appencix 1 to Bechtel Praliminary R2 port (listed above)

AESCC-13 Burs Test 1, 2, and 3 Field Data (Dated June 6 and 27, 1979).

AESCO-19 Sa cie Field Lack cf P2netrations (received at meeting with Becacel)

AESC3-20 Analysis of burs: pressure for cylindrical tube with clampad

ends, AES ',;-E l SOSF.5 C:mputer Result:, Parts I and II.

2.

"MecraT cal Tra.ctar Prodi~ctions for Sensitized Stainless Steel Pin % Wita i

Cir:rfemn:ial Cra,m:,

Final Report, EPR: Re:aarch Project RP

.J -1, Papor: 9 131 (Jun e 1.', 1975).

~.

American i;ational Standards Institu a, AfEI Standard B31.3-1973.

herican Socie:y of "ecnanical Engireer:, Boiler and Prescure Vessel Code, Section :::.

P.Nau;hton W. P. a.V G. R. Egen, "Significan:2 of Linear Incications in luin Staan 3iping of San Onofre fiuclear Generating Station (Units II and III)," Apteca Engineering Services Report AES-79-C5-3 (May 1979).

" Standard Tast MePod For Plane-Strain Fracture loughness of Metallic Materials, 'STM, 093-74.

Tada, H., ?. C. Paris, and G. R. Irwin, The Strus Analysis of Cracks Handcox, E111ertown, Pennsyivania, Del R2searcn Corpera: ion (June 1973).

Besuner, P. M.,

3. C. Peters, and R. C. Cipcila, "BIGIF:

Fracture Mech:.nics Coce for Structures, " EPRI Report NP-833 (July 1972).

Besuner. P. M., "Th? Influence Function Method for Fracture Mechanics and Resicaal Fa:igua Life Analysis of Cracked Components Unce* Cenpl::x Stress Fields," Electric Power Research Institute Tecnnical Report PR217-1-TR2 (July 1975).

P. M. Besuner, " Residual Life Estimates for Structures with Partial Thickness Cracks, ASTM Eighth flational Synposiur, on Fracture Meenanics May 1974; American Society idr Testing and Materials STP 590 (March 1976).

Cc.bined Ninth and Tenth Narte-ly Reports, EPRI Research Project RP 601-2, Gneral Electric Corpsrn. ion (1976)

~

Rice, J. R., "Ma thmical /sralysis in the P: chanics of Fracture,"

Frect

,ep. m..u,_r e_

Vol 2 (Ed. H. Liecowitz),1;ew York, Acadmic Press (1955),

.>.1 1.

Tana'<a, K. and J. D. Harrison, "An R-Curve toproach to C00 and J for an Austenitic Stecl, "Research Rapor: 7/1976/E, The Wlding institute (July 1975).

Rol'e, S. T. and J. M. Barsor.,

• cacture ad htic :+ CoM " "ructres,

Englewood Clif f:,.';J, Pren:f ce-islT, Inc. TJT/), p. HJ.

.~. '<.' ells, A. A., "?:otched Bar Test:, Fracture Mechanics and Brit:le Strength of Wided Stru::ures," Houdrxont Lecture 1954, Brithh ':eji~ h urr 31, I;3. 1 (January 19H).

.3.

Sc cter, J. D. G. and C. E. Turce', " Fracture Analys'is in Areas of High f;;minal 5 ; rain,

Pr?ceedf s D yd Int 0"a :ional Cor#' 7rce c-Presture h stel Temla;y, San And.11a, '.1(Oc:sfe"O73).

/.

Egan, G. R., The ::alication o' Fract'm Toughness Data To The Assess-cent of "ressure '.%ssei In:2gr::y, " Proceedings Second Inter 13 tic.lal Confe,renca on Pressure '!essel Technology, San An:onio, TX (0:::b:r 1-l&j.

U

5.

Burder,. F. M. an: 9 G. Oaaes, " Practical Use of Linear Elestic and Gr+rai rielc.y Frac ura Mac.1+.1ici Wi;h Par.icular Re'erence to Pressure huels," Conferaa:e on the Practical Apolice:ico of Frac ure Mechanics to Fre, sura Vessei Design Insti:stian of Ne:hanical Engineers, London, UK (1971).

9.

Critish Standards-Instituta, "Mathods for Cracking Openiq 01: placement (C00) Tas :ing" (1972).

J.

Egan, G. R., "Cr ;atibility of Linear Elastic (r'Ic) and General Yield-ing (CO3) Fracture Me:nanics, "Encin-. erin; rac ture Mec' vi cs, 'lol.15 (1973).

Merkle, J., " Analytical 4plications of the J In:egral," ASTM STP 536, American Society for Testing and Materials,

.l.

Hayes, D.

J., "Samt Ap;Tf::tions of Elastic Plastic Analysis to Fracture Mechanics," PhD Thesis, Londen University (J:te' er 1970).

c l' ?..

Dawes, M. G., "Fras ure Initiation in 51mm (2 Inch) Thick Weldr.ents of Quer.ched and Tempered Ste-ei, " Welding Institute Repor: E/35/70(1970).

Wang, C. T.. helied EIa:Mcity,, flew York, McGraw-Hill Book Company, (1953)

Bussnell, D, 5?50R5 Computer Code, Lockheed Palo Alto Research Laboratory.

a.

3.cy er, C. P., S. T. Rolfe, and J. T. Easley, " Effec: of Strain Hardening 1 Bursting Eeh:vior o ~ -r7ssure Yessels," 3 >cond Inte lati 131 Conference -

i Prenure Vessel hchnolo;7, Par

• II - Materiels, Freication, and

s:-c:icn, Tna Amr::an Society of Mechanical Enginears, !;ew YcM (1973).

na rv+y, J. F., T,.n_e.c.r,_v.i.nd ' ::icn of '6dten Pen;sure vess*1s, 2nd Edi tion, Van ';ostrand,.sinnol

,omp.any,:,Ycr4 ( 1:o i).

m r

c k

4

~

P -

ir TOA 2

i E

5 0 958 Uia 2 3 670 1

3 0

3' 1

0 0

1 7

CA X *J S

6 7

7 7

6 7

7 7 6 8 4 527 EDF 6

554 PE0R 1

T O 5

I N

5 3 774 T

5 5 E

7 852 5

CA 63 00 2 02 62 31 46 6

R2 55 66 E5 65 54 55 65 7 0 1

7 6 304 E

6 7 5

443 O

P L

E l

4 T

D;

)

2 5

1

- 3 7

9 3

3 625 L :t 90 65 36 60 99 35 50 0 4 8 3 021 i

E!

s 6

E I

k 35 61 19 65 34 1 8 62 6 8 1

2 5 R

1 5' 4 Y

(

44 3 i. 44 34 45 44 34 3 3 4

T 322 S

SdC 1

1 l

C Ei T 8

)

4 5

1 1

7 L

8 3 016

0. G 79 08 78 78

4. / 0 5 73 i

S 6 2 0 167 Sr s

f i k

7/ 41 69

3 73 5B 24 2 5 1

L u

6 3 035 E

(

8B 8U 88 U8 85 8G 88 8 8 L

T 8

877 T

U S

T

,a Y

Q Si e

e e

e a

a e

e e =

=

= : : :

T e

c e

e e

e e

i "d 5

p p

p p

p p

p t

t t

t t

t t

t t

ll l i.R l

E i

I l

a a

a a

a eee l i i

i a a a hEs i

i i

I 1

l e e e" vv P

P P ;P l

l i

P P l'

P P

P P

'P P

. U' P

P P P t S LLL r

G eee i:

ccc P

R 0

nnn 1

2 7

2 1

7 TE 0

eee 1

2 6

Z 1

6 L

A3 C

rrr A

I:

3 1:

9 0

0 9

9 C

3 9

1 r-r 0

0 I

! M 4

u u

i 1

1 3

1 1

3 f

1 c-c L

c:c O0O A

il'

%I%

E 059 E

L 999 P

• J s

s D

0 I

0 0

1 P

E

~

0 1

l 2

i t

C D

S e

C c

S n

E e

A d

5

)

)

i D

3 9 f dn O

0

)

n 0

o a

E 0

C "8

5 Z

4 4

4 I

/

/

/ /

O S

3 3

4 1

3 C

1 S

?

0

(

E I

1 A

(

(

mor f

E P

2 0

n I

1 1

e P

ka T

D C

1 1

E S

5 5

5 1

1 T

E 0

A

f

I': NILE ILSTS - la"";f: LAU.-UI -i'EitLI RATIOi!

SUP.; MRY PErcr!!T TEST DEFECT UI.LL

!!;.L L TDiSILE

't! ELD ELC::GATIGri AESCD f; UMBER SIZE T!!!CKriESS TillrXfir ss STREf!GTil STRFliGTl!

PERCffff

~

(inches)

(inches)

(ksi)

(i.si) 11 A-1-4

.090

.257 35 56.1 40.9 9

11 A-2-4 090 258 35 80.2 41.0 3i 11 0-1-4

,05*

.262 21 74,4 39.9 24 11 B-2-4 0S*

.259 71 84.5 40.3 47.5 11 C-I-4

. 03*

.2G1 32 62.6 40.6 13 c,

11 C-2-4

.03*

261 32 U2.1 39.3 37.5 d.

11 A-2-8

.05*

.376 14 85.9 38.9 47.5 11 A-3-8

.C5*

.379 14 83.0 33.5 40.0 11 B-1-8

.09*

.377 24 53.5 38.0 12 11 0-2-8

.09*

.3/6 24 83.2 40.6 36.5 11 C-1-8

.18*

.377 47 44.9 33.5 4.5 11 C-2-8

.17*

.377 47 45.7 31.7 5.0 15 E8^*

.00

.358 0

81./

39.3 52 15 F4**

.00

.256 0

82.3 30.5 G3 5, 18 1

.062

.384 15 81.0 45.4 25 40 57.2 42.5 8

18 246

.150'

.375 18 245

.206

.375 55 43.7 38.5 3

• Calculated from percent of wall thickness I;0TE: Talen frem AESCD-5, AESCD-11, AESCO-15, and AESCD-18.

• Ml l uelti metal tensile specimens

C-1 ADCEflDIX C EFFE'? CF GEC:4ETRIC N0?ll.NEAU TY The effect of geometric nonlinaarity has not been inco : orated into e com; uter progrcm run to calculate burs: pressure or the larp deflection clastic-plastic calculations parfoced in Section 9.0.

The effect can be repres-ented by tne paramet3r p and correction required

.atermined by substitution into tha appropriate matrix exoression for stresses,

' oments, disolacecents, and rotations.

ine coerricient, p, is used in the

- :iffness matrix (Eq. A.1) given hara in dimensicaless form.

It quantifies the chan;3 in radia.1 s ress: resultant, H, edge coment resultant, M, radial 4

fisplacemen:, h, and ed;e ro.tation, $.

M

~-

'c/At bending i

3

/2(i + p}

1 itc

<l (A.1)

J I

=

'h edge 1

/2(1 + ;)

f edge here E=

Ycung's modulus Wall thickness t =

-c

= Reduced thickness E._

/12(1 - v )

4 W

A=

Large parameter =p

c c.2

'ce paraneter a whi'ch providas t'2 nc.11inear ef fect is given by:

2 tr<c:--m 2.

o (A.2)

=

.,re Internal pressure p

=

"3 ob:ain 7.n appr:ximate valire for p, a burst pressure of 6000 p3i was chosen.

~'iis will reeult in; o :: 0.05 For raferer.ce, o = 0.0~, refers to a linear solution.

Since the approxi-

)

3:e value of a = 0.C5 is much less than 1.0, the effect of geometric non-linearity ay be r.Egle::ed until the ';1astic strain in the tut cecomes a larga fraction cf tre ultimata strain.

t D-1 APDEilDIX 0 EF ET OF SHEAR DEF00 TAT!0:1 The effect of including shair deformation may be significant for a tick shall. To check wnethe-th9 prasent tube is tnin enough to neglect the

1. fect of shear defor:1 tion,. the stiffness coefficients are examined with

he shear ter i included in a canner similar to that presented in Apcendix C.

M(1 +

)

l i,",

n,_

e 1

i i

l($/A)4 l

( 1 + t.Ar)

( i +..:"r) r--

i- -j i

---

l-------------------

l (0,1) i

=

,-a,ed;e l

1 2(1 +,Aa. ')

tu

\\

~

r m

2 I

l t (R

e i

.;_ )

2 l

(1+_q.)

edge 1

A4 aere u = 3 so that the tem 7

=. 0.94 (1 + 4)

(0.2)

I' Tnerefore, the reduction in s:fffnass coefficient which relates the edge

men: and the edge radial displacement is P.2%.

This is daem+d conservative since, for a'given displacement, inclusion of the shear effect will result in

' lower 5:ress. Thus, the neglec: of.the ef fect of shear defomation gives

ise to.a. higher stress in the pipe and required burst pressure is appropri-s:ely icwer.

i 4

I s'

t APPEND!X 2 Descriptica of Nondestructive E>a 'aations A2acc. ate: with Centerline o::k of Peae:ra:ica 1: S4-112 Doubla Welded Fipe Af:e: FRC I & E Bulle:1ru 79-03 was issued, one of haQS's first ultrasonic involvenen:s waa ;o exactne a shcr: len3 h of.2-3/4 inches OD,.375 inch nocinal wall tht:kanss pipe._ This pige had been previously considered to be ul::a,anically unaccap:able by :he 45 deg:ce, 5 percen: 1D and OD notch

cchnique ased by C1 :a L:.ba, Inc., Youngstown Welding and Engineering Ccupa y's SDE con:ractar. For purposes of consiu;ency, MiQS used the saae 45 degree (Ccde)- tachnique. While ID and OD indit::icns wer=a ce:ec:ad by MiyS at the saca 1;:a: ions aloa; tha weld sesa reported by Youngs:ovn, none of the in:ica:icna vera large enough in aaplituce to exceed the correct

nlibrs:isa referenc; level (DAC curve).

It waa apparent that Ultra Lab had been employing a sensitivity in eccess of exauina:ian requirenenza.

This weld seas was cross-saction i in seven difieren: '.oca:lons, and all seven crosa-sectiona were f ree of any def ects on ths ID or 03.

All sev<a cross-sec; ions tid have conterlinu 1 ch-of pen-:rn: ion (CLP), ranging fr:a app:rtinate-y 20 perce:: of the pipe wall thickness :o 1 css than three ;e::e.t. This deza:s::ated that :he UT :echaiques were ce:e::ing fabri:sti; 7:cducel refier:ars 'ro the planished vald ei.;ss on the 00 and :D sc: daces, bu: not froa thw CLP.

Br. sed ca these resuits, i: vas decidad to use a UT ang'.e bean exauina: ion w.k e re ! : the ultraceni: be an waald s:-ike this planer, verticelly oriented, cia-u.-

lack-of pan >: ation in as close to a perp+ndicular dir2ation as v.a pssaible. for e. 12-3/4 inch OD,.375 inch wall te.ickness pipe, this angle is approf. a:ely 70 degrees.

For calibra: ion purposes,13/03 no:ches were considered inappropriate f or de:ec:ing this sid-wall CL2, since cornar-type reflectors cro only ideal calib:10:en refle: ors whea crying :o de:ect flaws tha: crigint.:e at or are op-n :o ei:her the 1D or OD surfaces.

The use of a side-drilled ccli'::a: ion hole was.130 coosidered and discarded since a 360 degree type calibra:1on reflector will reflect sound back to the search uait regardless of the in:e; rating ansle of the transducer (i.e. O degrees through 70-degrees). This condition. is not representa:ive of the ver:10:lly oriented, nii-wall Ci? flava, which will reflect sound energy-back :o tne transducer only wheu the intergra:ing sound besa is normal to the CL?.-

In consideration of the above, mid-vall, ver:ically orients:cd notches =

were used for calib::: ion, Two notches were nachined in a 12-3/1 inch OD,

.375 inch wall segment of pipe. Eo:n notches were 1/2 inch long. One had a'1/16 inch ver:1 cal dimension, the other.a 1/8 inch vertical cimension.

i

1

..is 70 dagree =1d-wail no::h technique did not de:ect the known CLP

. the renalti.:g cA:crti, fro 2 ti.is s2 plc.

F.2diography of this weld was 4-f ormed p: lor to see:.!;c.ag and it also did not cetect CLP.

". an effar to determ.

vny

.:h U1' and.1T led :o detc.cc the presence a

known CLP, special

.: wela ver t fabric.

wl.th varyi 2 degrees of 4

.tentionally produced u F, i.e. 35 percen: :.

2gh 60 percen:.

One test uld had 50 perce : CL? for its en: Ire 2-1/2 in:.a length. Padiography

~:ecte: the CLP fe: only the first inch.

The 70 degras =id-wall no: h-

.2:nni;us also de:ce:ed :he CLP for only the first inch. The 15 cegree,

.//e per:ea: ID, CD no tch p rocedur a s (i.e. code procedurea) did no: detect

.y of the tanuf actured CL? in any c.

tha :es: welda.

These observations led to the p:eniss that the two'unfused usid prep 7.ces eay be in sufficiaatly inti. ace contact dus to wel; shrinksge 4:: esses tha: ultrason.: detect.an i; preven:ed by the sound beaa pissing

-..: cash the CL? it.s:aad

. beia4 ref'.20:ad back :o the t rar.sduc a : and radiographic de:ee: ion, prevented by the lack of a gap bet.*en the faces.

To inveatigata t

.s precisz, an addi:ional weld was cad s wi:h a 3.015 inch sco between :

joint surf aces and wi:n 35 per:en: of tha wall 1 Lness uno.2ed.

So:h :ndia3:aphy ana ths :pecial 70 de;rae UT exa -

na:12r we:e 2 ;e to detect :he CL? for the full 1sngth of this wel_.

Janven: tonal Coce shear w ive CT, 45 degree, five par:ent ID and OD notches, reper:ed "some initcations." Howevsr, an addi:ional 6db of jt in was required with this CoJs e.c.caina:ian to ob:ain a re:lection

- psi :o it J percan: of DAC.

These cratinations confitned that the in:1-nata con::.:: of :he unf us *c area is the re. sca for the inab:.:y of ei:'.sr ul rnaaniu or ra;;;3:4.0 a to ce:ec: CL?.

However, the geometry f t!e CLP is such tha: if sonnd is reflected f rom a CLP :ondician, :he majo:ity of the sound uculd not re:un :o the coaven:1onal 15 degree cinga : ::anscuca: :echnique and te d'. playe. en the (CRT) en:hoda ray

coe aa a LI indica: ion.

A es:ber of dif f erent UT techniques were also evaluated. higher frequencies, con:oured shoes, ana difieran: sizef transducars were tried.

Contoured shoes proviced the.only significan: improve ent by eliminn:ing couplant cauaed spur:cus incl: scions.

2.n examina: ton was =ad: using a 70 digree transdecar engle and calibra:-

ag from the five per unt 13 and OD notches. This techniqua is es:ita:ed

o be at leas: fou: times nora sensi*ive than the 70 degree nid-wall 1/16 nch notch :nchniqua and :he ena ination results confirmed this increased

.ua ina:lon sensitiv1:y.

Plot:ing the sound beau ener2y proille when enamining curved piping wi;h this technique bears ou: :ha: the cen:er et

..a cean nevar stri?es the corner of the ID or CD notches. These notch arners_are picked-up by the leading and.: railing sound rays of the over-

.il scund bea: ( bean spread ef fec;).

Becaust the sound beau energy

',tcasity is much less at the leadins; and trailir.g ray loca: ions than at ne cen:er of the bean, et;h : ore sensitivity is required to bring these

.:4e.h responses up to standard signal amplitudq height on the CRI.

This

2chnique does not work on an eight-inch OD pipe with.375 inch wall

.hickness, because the bea: spread po::Lon of cha 70 der ce sound bea:

.ius not strike the ID notch (Calibration sensitivity cannot be i

es:ablished wi:hout detecting the ID notch).

_o_

=

+

To de=onstrate the excesstve sensi:fylty of this technique, a 12-3/4 inch

~D pipe w h h a kno.n ptrer. fully penetrated weld was e:.a:ine with

nia :echnique, in a 17-inen. w:ctiva of veld that RT show+4 to be f ree f CL?, this technique indicated the existence of f our arcas th3: u:e e. ed

he fie* percent ID, OD no :: calibration auplitude resp:nna. One af

heSe areas was cross-sec:ioned and c:ched.

The welf was fully pr..t: rated 3: :his location. I: is ful: thn: because this technique enaminas at such L bigb snasitivi:y, tetallurgical grain struc ara changea c.used by :he weldin; Are bein; da:ac:cd.. Therefore, this technique is unable to discri:1'.'.1:e betwer.a fuiiy weld penetrated and par:ially weld penetrated productian velda.

These da:2 vere confir:ad by an enacination of 520 fee: of Toungstown n5 h5

- 12 aus ne'~ ~ stal:1ssa s:cel pipe remaiains a: Fullaan Iower

.a P roc a:: s, Para:oun:, California.

An additional ul::asonic examination was perfor:9. d o n oungstown pipe welds at the Pulitan Power Products f acili:y in Par.$.coun:, Calif ornia.

The primary objec::ve was :s exacine the longit..dinal pipe welds and to locate the "wors nase" cf CL?.

A::e:p;a were also cada to carrelata our UT findin; < with the Fullaan Power Products resul:s. The latter results ver2 no: cc:cessful bacnuse Pullaan's records dic no: indicati where the unacce;:abla UT irdi:a ions were lect:cd along the welcs; therefore, correlati : could at: ha naie.

A 70 degree UT precedure was used to s::a the welds and note ne treas of lart.sst signal asp!.itudes. Ixcept for :wo ins:;.r. u,

he ' it :: u si na'. t p'.itude cc e f raa ;'ipe sectica "F" 224.

o Grinding and e: chin, were performed at these :wo locations.

The C:2 at chase two loca.:1ons was approxima :aly 10 percen: or less of the vall thi:Lt.ess, yet the 70 degr24 signal responsa was c: Icas: three ti:es higher in unplituda then that cbtainad from the CL? cress-section tha:

nessurad 20 perceae of :he wall mhickness.

I did not appear that the g p of :he C ? was larger than 1: was at the.26 percent CLP locacion.

This coul; c.coun: f or the diffe rence in a:plitude, since all threw areas did indicato CL? lengths gralter thaa the transducer sise.

A de: ailed.er.a:1 nation was conduct ed in San Fran:'.sco on a tvo f oot section of the worst case of CL? tound at Fullman.

This piece was ;rlected fron hea: 39314,.NC?. f.68o, Drawing Sheet 224, "F."

RT ena=ina:ica revealed CLP approximately 1/2 inch long near ET See: ion #3, and si:e. isal::ad pores less

han-3/32 inch in dicce:er. Ultrast.i; exa:ina:Len ustaa tha 15 d23:ee s:anc rd procedure di. not reveal incications greater than 10 percent FSH.

Exanina: ion vi h the 70 degree procedure showed 10 par:ent DAC signals for approxicately 22 inches of the 24 inches and ca 50 percent signel one inch f c: ?.T Station 90, where RT did not indica:e anything, yet e::hing showed approxicately 1/16 inch CL?. On the opposite end nes:

RT station- #36, a 50 percent DAC sigaal was obtained f ro: CL?, which was confirued by the ?.T.

i Juo areas vare sele::ed for further investiga: ion by 'trasonic and

. :chich exaninti;ns. T r.4 i n:2.:t of-th investi,ation 2a to deter.:ine

he effect of scanning. he x-a?.ds at higher Sau. thaa.orn:a cnd to ate was:2.2 :clevan; :: nan-r e le '/act sign ls would be obtained.

".bla 11 cantsins the synopsis of the RT, UT and etching results. The altrassni: scanning at IX gala did no: raveal any si;;nifican: alsnals.

higher gain se::iags, nati:na'aly 22 and 4X, significan; signal cnplite.a. vere de:e::ed; ha 7-r :h-, e:ching c;;aaina:isa did no:

reveal any discon:iaulties in the veld other than the n1 ready known 1/16 inch ? ore. Consequen:ly, exanination at higher than norcal gain oetting2 produces no:-relevan: ul:rnaunic signals.-

In su_-...ry, when :he two n3used weld prap edgus are in in:ina:e contac:,

ts la aually :he cituatica fo: chis type of welded pipe, rediography vill n;; de:ec: the C1_? snd cone of :he ultrasonic techniques evaluated to data are capable o.' rol' ably detecting tha CL?.

1