...

.---------DISCLAIMER---------,

This book was prepared as an account ol work sponsored by an agency of the United States Government. Neither tl"le Umted States Government nor any •ncv thereof, nor any of the1r employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the CICC!Jracy,

completeness, Of usefulness of any information, apparatus, product, or PfoatSS disclosed, or represents that its use WOIJid not 1nfringe privately owned rights. Reference herein to any specific commercial prod1Jct, process, or service by trade name, trademark, manufacturer, or otherwise, does not necessarily constitute or imply its endorsement, recommendatk>o, or fiNOring by the United States Government or any agency thereof. The views and opinions of authors eJCpressed herein do not necessarily state Of reflect those of the United States GO\I'I!fnment or any .gency thereof.

EXPERIENCE WITH SIMPLIFIED

INELASTIC ANALYSIS OF PIPING

L. K. Seve::-ud Manager, Plant Analysis

Westinghouse Hanford Company Richland, Washington

Member , ASt.fE

~1arch 1980

For presentation at the Nuclear Engineering Division ASME Century 2 Conference to be

-held in San Francisco, CA., August 18, 1980.

· HANFORD ENGINEERING DEVELOPMENT LABORATORY Operated ·by Westinghouse Hanford Company, a subsidia!J of Westinghouse Electric Corporation, under the Department of

Energy Contract No. OE-AC14-76FF02170

COPYRIGHT LICENSE NOTICE a, acuotJIIIIU '' nus art •c 'c . '"'' '•Dl..U.t' '"' .. "' ''' '"' J(l" .. ltdtn IJrlt US tP '-wtalfttAt'l hfM le 'nltft I ,....tH IV'\.A't ,.,,ltp hH a.c:.ftV tft IIHI la lftJ (lffttiPi t

. ._ .... ""'lAP<'

DISTRIBUTIOK OF TillS DOCUMEij,T IS U lltMfTF.Il . -'"' ~

DISCLAIMER

This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency Thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.

DISCLAIMER

Portions of this document may be illegible in electronic image products. Images are produced from the best available original document.

, ...

'·

EXPERIENCE WITH SIMPLIFIED INELASTIC. ANALYSIS OF PIPING

DESIGNED FOR ELEVATED TEMPERATURE SERVICE

L. K. Severud Manager, Plant Analysis

Westinghouse Hanford Company Richland, Washington

Member, ASME

ABSTRACT

Screening rules and preliminary design of FFTF p1p1ng were developed in 1974 based on expected behavior and engineering judgment, approximate calculations, and a few detailed inelastic analyses of pipelines. This paper provides findings from six additional detailed inelastic analyses with-correlations to the simplified analysis screeDing rules. In addition, simplified analysis methods for treating weldment local stresses and strains as .w~ll. as fabrication induced flaws are described. Based on the FFTF experienc-e, recommendations for future· Code and technology work to reduce design _analysis costs are identified.

' . .,

NOMENCLATURE

a • Neuber equivalent grain half-length ~z • Stress index a 1.95/l2/3 c, • Weld shrinkage stress index Do • Outer diameter of pipe E • Young's modulus h • Notch depth KN • Fatigue strength reduction factor Kz • Local stress index Kt • Elastic stress concentration factor KT • Factor app 1 i ed to peak therma 1

strain component · Kc • Elastic strain concentration factor N · • Number of applied cycles . Nd • Number of design-allowed cycles P • Primary stress intensity · Q" • Secondary stress intensity R • Bend radius of elbow r • Mean pipe radius ~ • ·screening stress limit 3~ • Allowable design stress. intensity

range 1 imit Smc • Stress limit at cold· end of stress

range Srh • Stress limit at hot end of stress

. range T • Time of applied stress To • Allowable time for design 4T1. • Linear thermal gradient temperature

range through wall t • Pipe ·~all thickness a • Coefficient of thermal expansion. 8 • Carry over factor 1 • Pipe factor • tR

r2

v • Poisson's ratio e • Flow shape parameter 0 TE • Maximum thermal expansion secondary

bending stress 0 4T1 • Maximum radial gradient thermal

. stress . 0 ow ·• Dead weight stress Op a Pressure stress oeff • Effective stress ceff • Effective strain 'T • Total effective strain ce • Elastic strain 'P· • Plastic strain 'F · · ·• Peak thernial strain

INTRODUCTION · .. ·

The Fast Flux Test ·Facility (FFTF) piping design activities have progressed from the prelim­inary design in the· early 1970's through detailed ASME Section III code analyses including detailed· inelastic analyses for elevated temperature opera­tion. Design activities concluded in 1979 with final as-built reconciliation of stress reports for construction and installation modifications.· An overview of the flow of these activities is provided by Figure 1. As described in 1975 ( 1]*; signifi­cant project cost and schedule benef.i ts can be ob­tained if screening rules and simplified analyses can be used to confidently identify a· pipeline· con-· figuration that will pass detailed stress.~nalysis ~;ode. limits •. The screening rules for preliminary design [ 1] used on the FFTF project have served very well. Detailed ASME t;ode analyses using elas- · tic methods (2] and inela-stic methods (3- 5) have demonstrated that all Code design rules and limits are met. Accordingly, a correlation of the detailed inelastic analysis findings with the simplified analysis.screening rules will be presented.

;lfumb~rs in br.ackets indicate re·ferences.

1

"' .· ·.

i ..

• I

Before presenting the correlations, a short overview of the screening rules and background will be given. After the comparisons, simplifications tn the detailed inelastic analyses and supplementary simplified inelastic analyses will be discussed. Finally the paper concludes with recc:r.rnendations far fut~re code and technology work to reduce design ana l:ts is costs. · ·

SCREENING RULES AND BACKGROUND

The jreliminary design screening rules and limits [1 for primary stresses, shown in Figure 2, were easy to meet due to the low design pressures· of the FFTF piping. These low primary stress levels helped centro 1 ratche.ti ng and stress rupture damage at lew levels.

Screening limits for primary plus secondary stress ranges (See Figure 3} considered limits associated with creep ratchet, creep fatigue, and shakedown. These limits were applied to the very simple screening equation of:

(1)

where:

aTE • Maximum secondary bending stress in pipeline, usually at an elbow.

FIGURE 1. General Ingredients and Ffow of Piping System Design Analysis, A~E III, I.

. TEMFUAlUI!£ 1°Cl 432 533 593 704

100

··~ ~ 75 > !::'

'" z

~ 50

25

~-----~----~----~----~--~~--~0

FIGURE 2.

lOll 1100 l:axl UXJ lGl

TEMP£RAT\JRE I~ Preliminary Design Limits fnr Pressure .and Weight Stresses Using 50% of Code Case 1311-8 Primary Stress limits.

-:; .... ~ > !::: .,.. z

~ .,.. .... ... a: In

50 -· 3Sm P£R SEC VIII, OIV I

~-....,;:----~/ H.C. lJll·C

' ' ' ' ...... _ --

10 -~ 1.DW T£MP. HICII TEMP. CRITDUA CRITERIA .

a~~--~~--~~--~~~--~--~~~o 100 aoa a 1000 uoo 12t0 1300

TEMF£RA1VRE I~

FIGURE 3. Primary Plus Secondary Stress limits for 316 ss.

0 oT1 =Maximum thermal shock radial gradient stress considering all of the plant thermal transient and equal to-

E e~ (t~Tl) . m:vr

S =· An allowable stress intensity considering ·creep fatigue, creep ratcheting, and experi­ence factors.

A major feature of the screening rules and limits is the. shakedown and relaxation· of stress during the hold-time providing the transient stress ran~~ is always less than the "elastic action" . range. This is deOicted in Figure 4. "Ela~tic action" range for primary plus secondary stress range P + Q is defined by:

(2)'

where the standard ASME Code terminology of the ASME ·code Case 1592, Paragr<~ph T -l325 Test No. 4 [18] is used •

·If the .(P + Q)R exceeds 3 ~. then tile . relaxation of stress could be as shown in Figure 5 and not similar to monotonic relaxation and not · affected by the transients· as shown in Figure 4.

Based on approximate calculations, expected behavior, engineering judgment, and a few detailed inelastic analyses of pipelines (9],[3], 1t was deemed important to provide enough flexibility into the piping isometric designs to satisfy Equation (1)

· and k~p the ( P + Q)R 1 ess t~an 3 ~ •. Re~ults. from inelastic analyses of p1pelines 1s g1ven 1n Figures 6 and 7.

Additional background is presented in Reference [1].

: l.5SmH

FIGlRE 4.

s

srt!

rw· TIM£

--------------~--

--tJI4

Typical History for Pipeline Primary Plus Secondary Stress When (P+Q)R ~ l.SSmc + SrH·

5rH

·nMr

IP+QlR

FIGURE 5~ Typical History for Pipeline Primary Plus Secondary Stress When (P+Q)R !_l.SSmc · + SrH·

200 1000 2000 3000 (h) . I I ·I I I I

17 1- -

10 ~ r--1- _.("OUTS I DE -

.. MPa 69

-Vi ~

"' VI .0 '""' a:: ..... .,

Q..

0 0 ::t:

-10

-17

- t..,_----

I "'-INSIDE

-AB

0180 220

co

1000 1100

TIME lhl

I

2000

FROM A-B IS TEMPERATURE CHANGES:

1200°F _. 350° - 1200°F

FROM ~-:-D IS TEMPERATURE CHANGES:

1200° - · 700 - 1200°F

·-

-

.,

2860 3000

o·

-69

FIGURE 6. C~S ~i~eli~e Inelastic ~nalysis Results (Ref. 3). · ·

IS lCl

~8· 1D 09

)4

e ~ 0 0 v. ... ~ ... ·5 ... -)4

§ \

. ·10 ·69

·IS ·IOl

·lll ·I :IS 0 ~ 240

8 POSITION lclegl --... FIGURE 7. Hoop Stress Distribution in Criteria

Elbow Inside Surface (Ref. 3), ·

CORRELATIONS OF DETAILED INELASTIC ANALYSIS FIND­INGS WITH SIMPLIFIED ANALYSIS SCREENING RULES

·Table 1 presents the correlation data. The detailed inelast;c analyses include pipelines of s~mple. short runs as depicted in Figure 8 and ·

•• Q.

~ "' "' ! ... ... c ~

i I i i l

,_

. ·. ; 'i

' 1.

l ~ .

f

TABLE 1. CORRELATION OF DETAILED !~ELASTIC ANALYSIS FINDINGS WITH Sir~PLIFIED ANALYSIS SCREENING RULES

''!!" .. O.u:r•!!'_. h ....... ,t!'' ,.., .. ., , .. ,as's !tl 1au

••• ·- ... !2! ""!! !es~n t~ "\~ • •,r, i•hu•s1t••• ·t1 'P:r''~dnc?n~. 1=J :!U J.S.I ...!:J~ I "'"' I u ·-· ·-· • O.lt •l.lt .. .... I ·-· ·-· ·-· .... ,_, .... 0.12

• Uti

' ,_, 1-1 ,._, :J.I ,,_, .... O.JI

• .. 1.1 1.1 10-0 "-' .. _, .... 0.17 II .... • 1.1 ... 1~-· ~~..a "·' .... .. ,. ,. .... - ,_ . ... l.l .... ll.l .... O.JI

• an ,.,.u ... •• n1 u •--' '• t-. 1 ·~ ,,.,.u ... ••n •• JOI u - W..t ...... u '011 ••• !,O.Il": ~:; ::: : 1~: .~:= :: ~ R

• r.a .. , • '*''· us.G.ll •· Jll u • at.l ••• ' tzao'r

"II ..... ~ J ............. . • JD,I t.ll I 1G!QIIr

o.n .... .... ._,. 0.11 O.OCD

O.JJ o.zt ..... .... .... . .... 0.11 o.u .....

O.lt 0.1111

-c... 11•''' .. ,,....._,.,.., ,, a, .. ,~ ....,.....:::,.,. 1.0.,.. ce .... , .. ,, _.n ,,,, ... •• ca.••.­• cr- a. .... t.,. cr-.o-hll..,. ~ '" ,..,. l~looO ..,.. l~,..,. ''"''liM .CC¥"'• tt I'W 0111.• ll!r#ce •' ""' e:-.c- -.. .. len. All ere. ..... ,, .. ,, ..... ~ ~las• ... =---.- ,...,..... IJ' v.t ••eo.

c-..-tt• ,.,_..: 1 Ut • l.t ••; lltliiiiiJ • 1"JJIGGC. -~ • llllt, ~ • liiiiiiC.., '••· • n.••

IIIITJ

naNOOfs @IW'DITS

, •• Mlll'l P\IW CUI\(t

FIGURE 8. Pri~ary Crossover Piping Mesh (Ref. 4).

fairly complex, long runs as depicted in Figures 9 and 10.

The simplified analysis screening values for . the pipelines of Table 1 are cross-plotted on the scre_ening rule limit curves in Figures 11 and 12.

The:pipe11nes chosen·for inelastic analyses had the highest simplified analysis screening values. and the least design margins. The' design maroins were identified by detailed e.lastic ELTEMP [2] and simplified inelastic analyses such as the full · relaxation Bree [7] and the O'Donnell-Porowski (a] methods. Accordingly, since even the most severely loaded FFTF pipelines were demonstrated to meet all the code requirements by inelastic analyses. many ot~r.FFTF pipelines with similar but less severe loads and thennal transients are -also qualified.

-However, for these· pipelines, some of the very con­servative code elastic analysis limits (such as the Sq limits} were exceeded. See Figure 12. ·

Sit-IPLIFICATIONS IN DETAILED INELASTIC ANAlYSES

Oeta11ed inelastic analyses of piping do require some simplifying in modeling and analysis procedures to keep the analysis costs within rea­sonable limits·. These simplifications are technic­ally .justified by the satisfaction of ad hoc rules

. and conservative modeling. The simpl iff cations in modeling include:

• Use of constant .bending elbow elements (Type 17) of the MARC computer program [6]

•

•

Extrapolating elbow midsection stresses and strains to those for the elbow end weldments. by use of "carry-overu factors and indices to· account for nonuniformities introduced during the fabrication and welding of an elbow to a straight pipe section

Use of indices and fracture mechanics crack­growth models to assess local peak stresses and strains

In general, detailed inelastic analyses of a pipeline system provide primary and secondary stress effects. Substructuring techniques or use of in­dices are needed to account for peak stresses and strains. The simplification in inelastic analysis procedures include:

• Enveloping and lumping of thermal transients

• Extrapolating ratchet and .elastic followup strains to end of design life

The technical bases for some of the simplified methods identified above will be discussed in the

·following paragraphs •.

Constant Bending Elbow Elements

The technical justification for accepting use of the constant bending elbow elements depends on each pipeline analysis application. Considerations . include findings from prior elastic analyses of the pipeline such as the level of stress expected, the ratfo of in-plane to out-of-plane bending on each elbow, and the number of elbow element segments used to model· each elbow. Hibbitt [9] and P.an and Jetter [3] discuss the limitations of the constant bend element. Figure 13 shows a typi ca 1 mode·l of a 900 elbow using three JOO segments, 16 elements . around the circumference and 11 layers through the: wall. ·

A basic step in justifying the number of seg­ments, number of elements around the circumference,

. and the number of through-wall layers to model each elbow in a pipeline was the comparison of elastic. stress levels, moments, forces and displacements computed using the inelastic pipeline model to those computed using conventional elastic analysis pipe~ line models typical of analyses to NB-~600 of Sec­tion III of the ASME B&PV Code. Correlations to 'within 5% to 10% were judged acceptable. Moreover, for large diameter, thin wall elbows typical of

· breeder reactor plant piping, comparisons of elbow detailed shell finite element or finite difference .elastt~ analysis findings with constant bending elbow element findings were also uti 1 ized. As seen in Figure 14a, the degree of ovalization varies around·the elbow arc. Ovalization of 50% to 60% of the maximum at the elbow midsection exists at the junction of the straight tangent pipe •. However, the elbow net elastic flexibility has been shown by tests [10 - 12] to be adequately predicted by use of the simple formula of k = 1.65/~ given in the ASHE Code, NB-3600 (13]. Th1s formula neglects 'local flexibility distribution and varying ovality

''

• f

along the elbow arc. This ovality also penetrates one to two diameters into the tangent straight ·pipe. portion (Figure 14b & 14c). The code approach in­volves the flexibility factor as a constant factor applied to the elbow arc portion (Figure 14d). Accordingly, it is an "effective" flex.ibility·fac­tor far modeling the total elbc:M and effective tan­gent pipe flexibility for use in the total pipeline .system flexibility analysis.

,.

k: • li '

--' -. ~ .

•c.t•· .-.~ --· -... ,.,. ,_ ... _ ...

FIGURE 9. 8-inch SHL Inelastic Analysis Finite Element Model (Ref. 5).

As previously noted, test data [10 - 12] have been used to develop and have confirmed the simpli­fied flexibility m~thods and models for elbows with straight tangents. The highest elastic stress in­dice in the elbow midsection has also been shown to be adequately P.redicted by use of the Code formula [13] of C2 • 1.95/l 2/3. Therefor.e, by assuring that the constant bending elb~ elements used in the p1peli ne model provide numerical values of stress and deformation for elastic loading that correspond with sufficient accuracy to those computed using the standard Code formula and flexibility methods, the mode 1 is deemed adequate for use in the i ne 1 as- ·

. tic analyses, provided plasticity and creep effects !re li~ited as discussed below.

FIGURE 10. Ana lyti ca 1 Mode 1 of System 61 Primary ·Hot Leg Loop No. 1 (Ref. 5).

••

>4'1 ItA ~r-~~~~~~~_,~~_,~,-~~~,_~r-~

·Ull.;, S • • f 1ll5Sc • '-l5SHI ~ · )501 1(~ SlC Ill · -

/ 1510

?()I SlC VIII, DIY I • i.."!H

'

/ S LIMII·U

II EC. ~is :..~•-a

/ A.·•'MUX. CR£1· FATICU( LIMIT f();C ElASTIC ANAL

0 • llhb mn. D£TAII.£D tiA~Tit AHO SIMrliFIED INUASTIC ANALYSES r,AT PASSlD COD( LIMITS

• • llh£5 WITh DETAILED INUASTIC ANALYSES AH0 PASSED COD( LIMITS

Q~·L-_.---~~--~_.--,~~.---~-,~~---._-,~~---~-um~~

.TUIPlRATUR£ ("F) -·-···. FIGURE 11. Primary Plus Secondary Stress Range Limits for Preliminary Design of 316 SS Pipelines.

.;

.... ... = t;

1Vo\P£R4TURE ("c) m S'IJ r~

~r-~~---~~~~_,~~-7~,-~r--r~

· U7o'f l't'IIP., HICH TE/oiP. CllntRIA ~R!!IRIA

-JS., P£R SEC. VIII. DIY. I

• c.c. Ill!·(

I Sq UMIT P£R Eo;·u

Q1' CASlllll-1 .

• • •

+• LiNES 1111'11 Dt!AIUD INI:USTIC AIIALYSES AND PAS WI CODE LIMITS

OL--~--~_.~~~--~--~~--~~--~~0 lUI IDl 1101 • llOD UOD

1Vo\P£RATUR£ I'll

FIGURE 12. ·Preliminary Design Primary Plus Secondary Stress Limits for 316 SS Pipelines.

The thermal expansion and thermal radial grad­.tent maximum elastically-calculated stresses iml)osed on the elbows were, for FFTF, kept less than the 3 ~ level. This limited the plasticity and creep effects to local and small portions of the pipeline wall.

~ .,

• ;:, 'I /

:·I. .. '· ' j .

.•

; :

~

~

:

1 .i-..

• . . . Thus, many elbow element segments necessary.to cap­

ture stress redistribution associated witQ gross plasticity and creep throughout the elbow were not needed. .In addition, as the pipe 1 i ne response was expected to nshakedown" (Figures 4 and 6), good . . ... margins between the Code 1 imits and the calculated values of accumulated inelastic strain and creep­fatigue damage were expected to offset possible. limitations associated with the approximate elbow . model.

Sane details of the e 1 bow mode 1 s used in FFTF inelastic analyses are given·in Table 2. Generally, the FFTF analyses [3, 4, 5] found two or three elbow· element segments with 14 to 18 elements around the . circumferences and 10 or 11 1 ayers through the wa 1l were sufficient at reasonable cost. ·

, .. - ....

.I EJ1UI

I . . .

JILWHOI(, t D. & ... 1110' ICHI(, lll Jt II

FI~URE 13. Typical Constant Bending Elbow Elements.

0

SHORT RADIUS PIP£ BO~D. Rlr = 3

I ONE DIA. : ONE DIA. I ONE DIA.

z

0

• )

••

FIGI.RE 14a. :Typical Variations of Stress and Ovali­zation Along Pipe for In-Plane Bending.

'" "· ... ~ " ~ 0 s

A.

0

AHGI£ atD£CREASD

FIGURE 14b. Typical Distribution of Axial Surface Along Pipe for In-Plane Bending Load.

A

I ONE DIA. I ON£ DIA. I ON£ DIA. I

BEND

~I •• DISTANCE FROM CENT£R OF BEND .... -, ..

FIGURE 14c. Typical Variation of O.valization Along Pipe for In-Plane Bending Load.

BEND

VARI-'TION USED IN CODE Fl£XIBILITY ANAlYSIS

tR k. L65/~ ; ~. -2 . r

TYPICAL ACTUAL VARIATION INCLUDING OVALIZATION EFFECT:.:

•o ~~ •4 DISTANCE FROM CENTER OF BEND

FIGURE 14d. Flexibility Factors Along Pipe for In­Plane Bending Load.

Stresses and Strains at Elbow End Weldments

F:lbows are ·often attached to straight portions of piping by girth weldrnents located at the junc­tion of the elbow torus and the tangent straight pipe. Stresses and strains at the elbow midsection were found by Markl [14, 151 to govern the fatigue · life of pipe elbows tested below temperatures wh~rP. creep effects are significant. Accordingly, the

· ASME Code [13] in NB-3600 provides stress indices for butt welding elbows that are baseQ on midsec­tion stresses, but a Dolt~ .100 is required. For FFTF, the largest D0/t is 64. As the D01t ratio gets larger, the stresses at the elbow and weldments

J'

•

'z o L1 LOCAL l1tlD SCI

Zz o ~ Stt~II«NNl fACTOII

'z o QJOW IIIDICl

'•. 'z Zz 3 •• ' J. '·· ( •.

ONfnf

fOR OIA. o ltln& 21 in

'•,. '• . 'z FOil OIA. ~ lift

'fD"'• G -~ D Zl

•

FIGURE 15. '• '•

Relating Weld a and c to Elbow "t{idsec-tion a and c Cue to Moment Loading.

TABLE 2. ELBOW t400El DESCRIPTIONS ''MH• ,,..., Oft('f't!!l .. "•tt• (1~ ~ O.tC1'1!ttO'I ef lf'l'lld .... Zili. '· ......... tJ ... .... •· tf ~ofr-t.• G. •' {a,-n EO. of !t~l

L!!:.l L!!:.l ~ .!l!!:l -'!.. ,.,...,,o, I'....,.U ~~·· &IOI'IJI...-~

J t.nl 1-J .,_, l.tJ Ill II

J LUI 1.1 ·-· I.U Ill 16 u let'-'._ I I t1.._.....,. l

• LIJr l.t 1.0 ·-· ... tO .. u . ......... • G.IJJ ... ... 0.11 tO .. "" II u 1tllllelt .... , ............ I o.m 1.1 IIJI .. ,. . II u I t1r.c- .._I ...........

II LJJI t.l ,. .. 0.11 ... tO Ul u a .. ,,. I.J 41.0 .... Ill II • . .......... ,

,,,~ .... ' ... • a.•- ,,, ...... ,

·may becane 1 arger than at the ·elbow midsection. Considering the limitations of Markl's test data, it was decided to also calculate stresses and strains at the elbow end weldments, in addition to those at the elbow midsection for the FFTF.

Detailed inelastic analysis of pipelines using the constant bending elbow element (No. 17 of the MARC finite element computer program) [6, 9] do not account for the secondary stresses .due to fabrica­tion mismatch or radial shrinkage at the welds. In addition, peak stresses and strains due to weld sur­face irregularities, etc. are not ·directly included tn the pipeline inelastic model. Accordingly, a simplified method of evaluation was devised and con­ceptually described in Figure 15.

The simplified method consist~d of using.the . pipeline system model to predict inelastic response· for primary and secondary stresses excluding fabri­cation mismatch and weld local effects. Stresses and strains at the elbow end weld joints were then approximated using the values canputed for the elbow midsection combined with carryover and shrinkage factors •.

The carryover factors were determined from detailed ~hell finite element, finite difference analyses of the FFTF elbow designs, and from con-. si deration of the experimenta 1 . data [10, 11, 12] on ovality distribution such as shown in Figures l4b and 14c. For the FFTF applications, a carry­over factor of 1/2 was found conservative. How­ever, higher factors may be needed for larger and thinner-walled pipe elbows.

Radial -welds on"FFTf to join the seamless pipe and machined elbows ~re done by an automatic weld- . · 1ng machine. The weld reinforcel!lent and surface

irregularities were much milder than typical manual. welds. Fabrication alignme~t and mismatch toler­ances and the welds were all kept within Cooe limits. Accordingly, the Code stress indices tied to the fabrication limits were considered appropri­ate to account for local stress concentrations and · · fatigue strength reduction factors. Thus, a local indice for the girth welds based on the Code [13] was taken to be 1.8 in magnitude. The Code does not have a factor for radial weld shrinkage effects •

Discontinuity stresses of the .type depicted in Figure 16 and due to radial shrinkage in thinwall piping, were approximated by elastic analysis of a number of shell shrinkage distributions and R/t ratios. Based on these findings and code indices for girth welds, the special indices of Figure 17 were adopted. These indices were intended for use in predicting the maximum stresses and·strains at welds in the pipe axial direction because the K2 • 1.8 local factor was considered an axial fatigue strength reduction factor. An appropriate K2 value relative to the pipe/elbow hoop direction was judged to be-1o1 to 1.2.

• oll+6M I l - tZ

11 • -f (1 +Z. 9 t) MAX. AT ourut SURFACE

C1 • .I!. (U A) MAX." AT 0U1D1 SURFAC£ p l l

..... -.•. FIGURE 16. Discontinuity Stress Due to Radial

Shr~nkage in Welds.

To obtain approximate. ~~lues of elba.., .and weld maximum stresses and strains for comparing to the elbow midsection stresses, the method was as depicted in Figure 15. The 1/4 factor is based on the 1/2 carryover factor and the maximUm·axial stress of-1/2 of the maximum hoop stress. Table 3 shows combined indices for t_he various FFTF pipe sizes for both elbow midsections and elbow end we ldments. Note that as the diameter gets 1 arger, the weld indices are larger than for the bend mid­section. These indices were u·sed with the elastic flexibility analyses of the pipelines and are based on -shrinkage and mismatch data of Table 4.

To obtain· stresses and strains at elbow end weldments for use with the inelastic analysis code evaluation, a simplified method was used. The hoop and axial maximum stress and strain values computed for the elbow midsections, using_the inelastic pipe analysis, were first examined. The values· at the elbow end weldments, exclusive of weld shrinkage and configuration peak stress effects, were taken as 1/2 of the midsection values. 1he radial weld shrinkage produces shell bending under axial mem­brane load. One approach is t~ apply the stress indices of Figure 17 as multipliers with weld peak stress indices to the calculated effective stress arid strain, exclus"iv~ of weld effect!:. Thit is:

1 weld oeff • Cz- B t 2 K2) aeff at elbow midsectfon (3)

..

where: t 2 • indices of Figure 17

8 ··0.5 for in-plane bending

and 8 • 0.1 for out-of-plane bending.

The 8 factors relate the axial membrane · stress to the max im1111 surface stress

c1 • 1 o • o. 1 ( t ) . c

2 • L 0 + 2. 9 ( t)

c2 • L 4 + 2. 9 ( ~ )

t>3116 & r s 9.1

t~3116 OR ~ >0.1

A • RADIAL SHRINKAGE, ?:/o MAX.

T • WALl THICKNESS

6 • MAX. PERMISSIBLE MISMATCH, NB-3683, 2-1, .·SUMMER 73 ADDENDUM

HEDL IOIIZ-170.1

_FIGURE 17. Special Stress Indices for Radial Shrinkage in Welds.

Tl&l J. 0.11111 (liD WQ.DI'(It IIIGICU nPICA. Of fftr Pt'115

• 11&11 Outr ezr1 · c;"tztl 'z tz 't'· ·''"' Stn

nua .. u Ot_.t., ~ ·r¢ . Cz .;--· 't .

. -1.!!:.1,_ l.!!.J.~ l.!!.J.~ :::!: :=c: O.ID ,_, 1.11S· Jl.O z.a I; Ill l.zt !.ll J.IG 0.111

0,1!11 J.t !.115 IO.l J.!S 1.01 I. !It J.ot 0.91

.J 0.111 t.l 1.50 .. , J.ll Z.IO !.OJ J.M 0.95

• O.ZlJ 1.0 .. , u• o.zi 1.!1 1.9Z ,_ .. 0.111

• O.ZIII 7.1 I.IIS 151 •. II 1.11 !.IS •. u o •• I o.m u l.l!S Ut 1.11 1.111 L5! 0.911 O.llt ·

ll O.lJ\ t.l IJ.OO 001 1.91 Z.ll 4.07 7.JJ 1.QS

'31 0.171 t.S lUll 7ll 10.1 J.az 7.77 ll.ft 1.ll

<z- ASIC elaot tMtCII • l.IS/lZJJ. ~ •ITir ... tiH Aft. Sa. Ul• T .. lt III·Jtll.l·D

-lr ~~a~t or••..O '""'"' -'- tao to • ... "~~"" 111

..-crAM th'U .. u es .... ~- ,,..,ca tor...,., '~'~ • 1.1

~gr 1-t•. ptpe. tN t100I .tds«ttOl'l lndtCI Cz h 1•91"' tAl" tM elbolf tnd .. ld Oflll 1tre11 . •••ca 1t lz Cz t'z/~. for l ... in. :I'PI• ,,_.. '''''9't ;toe ttrta ,..,d tnc~tu of Czlz • l.S

TABLE 4. PIPE AND ELBOW WELD Sff!INKAGE AND MISMATCH VALUES

Parameter < 4-in. Oia ; 4-in. Piee

Radial Weld Shrinkage, ~ 1.i!l:.l ~ 1.i!l:.l 4/0~R.

Out-of-Roundness Run Pipe

. Elbow Bends

Mismatch (6) ThiCkness· Tolerance.

St

Conversion factors:

. 0.062 0.062 0.062 . 0.062

8.0 8.0 0.094 . 0.094

12.5 12.5

1 in. • 25.4 rrrn OR • Outer radius of pipe 00 • Outer diameter of pipe

K2 • 1.8, local weld peak stress indice.

1 weld ceff • (! B ~2 K 2 ~ teff at elbow midsection (4)

The above approach may b·e too conservative; therefore, another approach is to separate the mo­ment and pressure-loading indices and apply factors to hoop and axial stresses at both inner and outer surfaces of the pipe wall. See Figure 18. For the 28-inch FFTF pipeline of Table 1, Equation (3) gives 1/2 8 t2 K2 • 1.36, based on in~plane bending B • 0.5. The method of Figure 18 gave a comparable factor n • 1.45. For the 16-inch FFTF pipeline of Table 1, the 1/2 8 !"2 K2 • 1.05 and 1.47 for B = 0.5 and 0.7, respectively. Both the 16-inch and 28-inch · pipelines were demonstrated to pass Code·limits (see Table 1) rather easily because pressure and thermal expansion stresses were so low that plasticity and creep relaxation were limited to shakedown response.

CONSIDERATIONS FOR FABRICATION-INDUCED FLAWS

During fabrication and erection of FFTF piping, the pipe outer surface inadvertently received scratches, dings, drill-induced blemishes, and other surface flaws. Figure 19 shows a picture of the variety of surface defects taken on a workmanship sample used to measure and characterize typical f1 aw depths and shapes. Most of the f1 aws were <0.010 inches (0.025 rrm) but some were de"eper. Considering that such flaws can reduce the oper­ational fatigue capabilities, an acceptance criteria· was developed for identifying what flaws could be tolerated and what flaws should be removed prior to insul_ating and placing the pipe into operation.

The ability of the piping to withstand flaws depend on:

Operating stress magnitudes and cycles

• Operating temperatures

e ··Flaw shape, depth and orientation

e Material fatigue strengths and crack­growth rates

Two failure modes were evaluated:

• •

Low cycle fatigue, 843 plant events

High cycle fatigue, pump-induced vibra­tions of flow-induced vibrations up to to9 cycles ·

The steps in the analysis are depicted in Fig­ure 20. Criteria for acceptability and repair pro­cedures were developed using two types of analysis:.

•

•

Creep-fatigue life of notched surface (crack initiation) using ASME code methods

Crack-growth life using fracture mechan• ics methods

Maximum stress-range bounds for use in the creep-fatigue and crack-growth analyses are shown 1n Table 5. These values considered the ma~imum stress levels allowed by the applicable design rules including plant thermal transients, effects of

clamps on piping straight sections, and high-cycle vibratory stresses induced by the large sodium pump vibrations exciting piping to resonant response levels.

In order to carryout the analysis and obtain results applicable to all FFTF ASI~E Class 1 piping (sizes 1 inch through 28 inches) and temperatures up to l0500F (5660C), simplifying assumptions were needed. Far both creep-fatigue and crack­growth analyses it was assumed that the surface flaws are located at the highest stress locations on the pipe fitting and straight pipe sections.

Creep-Fatigue Analysis

The creep-fatigue analyses used the formula for evaluating theoretical elastic stress concentration factor Kt for shapes associated with Figure 20 and an idealized 0.010 inch (0.025 mm) deep drill­induced flaw with a 0.005 inch (0.012 mm) root radius. For the drill-induced flaw the Kt form­ula u;ed was [16]:

K •1+2 ~ t VT

For temperatures below BOOOF (4270C), where creep effects are not significant, the fatigue strength reduction factor KN was used in the fatigue analysis instead of Kt· The KN value was determined using [17]:

(5)

(6)

1 +

where:

R • root radius

a • Neuber equivalent grain half-length

Figure 21 shows that it is difficult to get KN values over 3 for drill-induced shallow blemishes of <.0.010 inches (0.025 mm) deep.

I ELIOW MIOSECTIOH • [L8(M MIOSECTIOH

'z. • 14 • Ul.l./11 OUlU SURfACl l • U • tt I .I. I II IIC£.l SURfAa

cac • u ,.~.Ill OUlU sU!IfAa ~i~"«N.r ll 1.1.111 INO S\JIIfACl

~ • Ll AHD ~ • LlwnD PW; STRUS INOIW

(•w) 1n~ [• '• , Z ., 1 ] Ill (• )tn~""".J(• ) ~ • A H A H ''I• fl'f ....,.. --., [ff _._ . t MIDSECTION

FIGlRE 18. Approximate Method for Evaluating Elbow Weld Stresses and Strains.

For ~emperatures above 8QQOF (4270C), the creep-fat1gue anaJ,..rses used the ASI~E Code Case 1331-5 or 1592-8 Ll8] to evaluate the peak strain range as follows. But the KE value was taken as KN instead of Kt based on the work of Maiya [19]:

'T • K,c, + KZ,cp + Kr'F•

where: (7)

'T fs the derived maximum strain for the loading condition.

fs the elastic strain fn the regions under consideration, exclusive of strain concentrations.

is the theoretical elastic strain concen­tration factor.

fs the inelastic strain in the region under consideration, exclusive of strain concentrations and peak thermal strains.

Peak thermal strain associated with the peak thermal stress intensity as defined in Section III.

Strain concentration factor applied to peak thermal strain component, 'F·

, FIGURE 19. Workmanship. Sample, 16-fnch 316 SS Pfpe.

A~E Code fatigue curves were used in the low cycle fatigue evalua~ion:.. rcJJ· VIlli ctLt, ·.Y I·" lrJIII! I'Villlla tions in the h1gh cycle range, the futigue strength~ of smooth bars reported by Chow [20] and Jaske drld O'Donnell [21] were util !zed.

! '··

.,

.. . .. ..!._

0r·

lta".IICICIII'UIAU IAUICAI'ION- llriCII..aD R..WS

FIGURE 20. Analysis Program for Developing Accept­ance Criteria for Fabrication-Induced Surf ace F 1 aws.

Crack-Growth Analysis

In the design of FFTF piping, the range· of pri­mary plus secondary stresses in pipe fittings sue~ as elbows and tees, were limited to a value of 3 Sm as given in Table 5. Due to the radial shrink-age of_the weldments joining the.fittings to the straight pipe sections, the welds also represent · location of increased stresses. Outside the fit­tings on straight pipe section surfaces, the maxi­mum applied stress range is about half that of the ·fitting (i.e., 1/2 • 3 5;,). · .: . . ._

The evaluation of the flaws were.divided into two stages; crack initiation and·crack propagation. Normally, a non-flawed smooth surface will require many cycles of stressing before a small crack will ·_ develop. However, a notched or flawed surface can initiate a crack very early in the part life and then tte question shifts to how fast will the crack grow.- Figure 22 shows the threshold flaw size cal­culated for the crack to grow under various applied stres~ ranges, For maximum allowable design stresses the threshold. sizes ate given ill Table 6.

Tl'e crack-growth fracture mechanics· ana lyses were·accanplished conservatively assuming that the surface flaw, which is normally not as sharp as a

. crack, to be a crack. The d~tenni l)ation of the applied stress intensity, 6K, was based on the methods of Section XI of the ASME Boiler and Pres­sure Vessel Code. Other formula based on the· work of Hsu and L1 u [24] and Shah and Kobayashi [25] . were also employed for further insight.

·The-crack-growth analyses indicate the growths are fairly sensitive to stress level (see Figure 23) but. very little growth is expected below ~QQOF (4270C). See Figure 24. · _ .

Tte crack-~rowth rate and threshO-ld stress intensity data (see Figure 25 and 26) used were· based on work by James (22, 23]. To account· for long-time high-temperature effects, an environmental rate acceleration factor (Figure 27) was obtained -by extrapohtion.

TABLE 5. PIPING STRESS.RANGE BOUNDS

PL + Ps + OR Max Value Max

Location ·Mer Temp

F !& lliii lt'i!hl Cycles

Elbow Midsec. 800 427 50 345 843

Elbow Midsec. 1050 566 35 240 843

Elbow Midsec. 800 & 427 & 4 28 109 1050 . 566

Elbow Ends 800 427 25 172 843

Elbow Ends 1050 566 18 124 843

Elbow Ends 800 & 427 & 2 14 109 1050 566

Straight Pipe 800 427 18 124 843

Straight Pipe 1050 566 14 97 843

Straight Pipe 800 & 427 & . 2 14 109 1050 566

·-· ..... ... «,.· t ..... \1.1

8 s:. L<

I • 0

. J.l -------LO

... ·-·

.· ·-· ·J.I

FIGLRE 21.

1\AW Olfhl ..... 1

II(De. Mt•l01.1

Typical Fatigue Strength Reduction Factors KN for Drill-Induced Flaw Shapes.

An example of the creep-fatigue and crack­growth analyses findings is given in Table 7. As shown in Table 7, the high-temperature elbow mid-

·sections are located where such flaws may cause non-satisfaction of the creep-fatigue criteria. However, the high~temperature elbow ends and straight pipe sections do have adequate creep­fatigue margins. When a drill-induced 0.010-inch (0.025-mm) deep blemish, which is not as sharp as a crack, was assumed a crack, its growth was predicted at -0.042 indies (0.10 mm) for the high temperature

•

i .

•

',~._~~~~--~~--c~~~~~~~~~~~o--~~_u

STIUS IAI4 6 • 1b11

FIGURE 22. Estimated Thresholds for Flaw Growth.

TABLE 6. T~ESt()LD FLAW SIZE FOR CRACK GROWTH .

Stress. 11a • 3 ~. Elbow, Tee, etc. ffL (mil)*

70 800

1000 1200

6 2.5 2.0 5.0

* 1 mil = 0.025 mm

Stress. 11a • 1/2 • 3 ~. · on Pipe s(rairht Section

. mi 1 *

25 10 8

18

elbow midsection~ The other locations had very gnall growth,< 0.001 inches. The wall thickness, in this case, was -3/8 inch (9.5 rrm).

The high-temperature crack-growth results are fairly uncertain due to the very rapid change. in growth rate as the applied stress and effective· . stress intensity change (see Figures 25 and 26) and the cyclic time change. This time-dependent effect fs often referred to as a "frequency" or "hold-time" effect (sec Figure 27). Thus, although the crack extension for a 0.010-inch (0.025 mm) crack-like flaw fn a high-temperature high-stressed elbow mid-· section is calculated to be 0.042 inch (0.10 mm), it could actually be much larger or much smaller. A 50% increase in stress .level would incre~se the

-predicted crack growth to 0.13 inch (3.3 mm). A 50% increase in depth, 0.010 inch (0.025 mm) to 0.015 inch), (0.037 mm} results in a predicted crack growth of 0.27 inch (6.9 mm}.

Above SQOOF (4270C), creep effects can greatly enhance the crack-growth rates and reduce the low-cycle fatigue life. If the maximum allow­able code design stress is developed during opera-. tfon. a very small flaw, as little as 2 to 4 mil deep, may grow during the design cyclic life to unacceptable levels. Thus, elbows, which do have local areas stressed to the Code limits, should have all surface flaws removed. In straight pi~e sect1ons, the operating stresses are.usuatiy less than half those in the elbows. Round bottom flaws

up to 10 mil deep may be tolerated with no signifi­cant adverse effects on the piping fatigue integ­

·rfty, provided all. operating vibratory induced· stresses are as low as expected.

For operation below SOOOF (4270C), where creep effects are insignificant, the crack growth fs slow and the low-cycle fatigue life for a given· cyclic stress level is greatly increased. Thus, · · low temperature (below SOOOF} piping flaws any­where on the piping, up to 10 mil in·depth, could be tolerated.

o.• cs=r· Jl6 u

! O.l

i .. ~ O.l

oJ

0.1

FIGURE 23. Crack-Growth Rate vs Stress for lOSOOF.

0.4

cu :!

. 0.1

1 in • 25.4 mm·

~ • 0.020 in., A~ • Z.S lui

H(Dl 7101-JOI. I 0

Figure 24. Crack Growth vs Temperature.

i I.

. . ...

z i

SCATTER S.ANO FOR ANNEAlED TYPE 316 SS

UPP£RIIOUN0 LINE FOR ANNEALED

TYPE 316 SS

e OATA AT LOWER EFFECTIVE STRESS INTENSITIES (Specimen No. 92,

I • 0.67, f • ~00 cpml

HEOL 7801·301.2

FIGURE 25. Upper Bound for Fatigue-Crack Propaga­tion Behavior of Annealed 316 SS in an Air· Environment at 10000F for the Full Range of Effective Stress Intensities.

From the findings presented and from other ana lyses and considerations, it was concluded that, in general, higli-temperature straiuht pipe will tolerate drill holes and surface t1aws· on the order of 0.015-inch (0.037 rrrn) deep. Low-temperature. (bela.~ SOOOF) large pipe and elbows will tolerate 0.025-inch (0.64-mm) deep drill holes with no ·need for blending.

Based on the creep-fatigue and crack growth fracture mechanics stress ana lyses,· 1 imits were developed that depend on whether the flaw is located on a piping fitting (such as an elbow.) or on a straight section of piping, .~nd on the intended opera·ti ng temperature. The acceptance criteria developed called for all of the following surface defects to be blended out:

1) . For 1C7ol temperature piping with operat·ing ·temperatures SOO.OF (4270), or below

~) Any surfaces defects over 0.010 inch (0.02S mm) in depth. ·

b) Any arc strikes or weld splatter.

.t.NNEALED TY'E J~ 55 . · TfSTfO AT IOOO"F

330 c f c .COO cpm, 0.05 < Rc O.SO

SODIUM ENVIRONMENT 0 SPECIMEN 23-', R • 0. QS e SPECIMEN 2~. R • O.SO

VACUUM t7.1·S.6xl0"6 tor) 0 SPECIMEN 264, I • 0.086 • SPECI~EN 264, R • 0.500 4 SPECIMEN 265, R • 0. 500

NITROGEN (I ohft ULTRA·HIGH PURITY A SPECIMEN 267, V SPECIMEN 267, Y SPECIMEN 271,

AIR ENVIRONMENT

NIUOGEN ENVIRONMENT

SODIUM ENVIRONMENT .

OR VACUUM

EfFECTIVE STRESS INTENSITY FACTOI, IC ll·RI 0•5, lb/(in. )3/% -· FIGURE 26. Fatigue-Crack Growth Behavior of · Annealed 304 SS fn Sodium Vaccuum Nitrogen, and Air Environment at lOQQOF.

'*~-o~-r----r---.----;----r----r--.---~ '\~ SLOPI • ·1.0

\ \

I.

\ IXIIIAOOU.IION

\/ ' ' ',

ANNIALID SfAINU!! Sflll . flUlD IN AN All (NVtiO~fNf

........... :s;:_TY•I )00 Af 10111°,,

. TYPI ll• · .., •CDf• .• c f 4

0 ','rr~s~...l.,rr"~-,o.~.,·l;--.:...,a~..;·z:--...l,rr"":''---...:..... ___ ,.~.,o'----,olL---I.'oi-..J•a"

1 ia • 0.4S.Ug

lin •25.4 mm ."lDI. 7101-lOI.I

FIGURE 27. Frequency Effects on Crack Growth for Stress Intensities Over 15000 lb/(in.)3/2

2) For hfgh temperature piping with operat­ing temperatures above BOOOF (4270C),

a)

b)

c)

Any surface defects of any percept-. able depth in elbows or fittings.

Any surf~ce defects with sharp bot­toms of any perceptable depth i·n pipe surfaces.

Any smooth bottom defect over 0.010 fnch (0.025 mm) in depth in pipe surfaces

d). Any arc strikes or weld splatter.

TABL£ 7. SUMMARY OF CREEP-FATIGUE AND CRACK-GROWTH ANALYSIS RESULTS

, ..... t4 111-. . ... , " .. ·- ·- - ..... (4 ... ·- -UU!!e .J:U i!J.!.L.· tal!& ~ ~ ,,,,.,;;, !!.!e !..!!.!!: i!!..L

,, .......... - II Ill .... lA u t. ar' 0.01111

(\eetlh_U_ - • .. , Q.)l u '·' o.n '·' • • r' 1ft L ...... ........ , .. -·- • ..,, o.n .. , -, . ..,, .. .. Zl Ill II.GIU LO u lo ..... 0.-,, ....... IIIII II Ill 0.11 L77 ..• t. w-• IJS Lll>ll

,, .. c. -·- I ... 0.77 LO -""' ... , .. .. II ., l.O ... -,..,. ........ - 10 .., LO 0." 1.1 -\trltpt,.. ....... I ... ' .... ... -

-.& • LJ 1.6•19'1' ._.. .. • 0..010 11111{8,

..,._ Ml• fll 1.011 t•. 10.1~ •t h ....... O.OlD ••· (Q.Oft _, t11ttel f1• -.ca. ,., • t•tttet n• •a e~ a.o11 ••· tM ~,..., ..,....u •• o.n •..a.

-- cr.a ..... --· .. tt .............

The depth of blend was also controlled so that the residual wall thickness was adequate to meet pri­mary and secondary stress limits. Defects greater than 0.025 inch (0.64 mm) in depth were given case­by-case evaluation and repair treatment.

The 10-mil limit in the acceptance criteria was a conservative limit chosen with the recogni­tion that considerably larger flaws could be toler­ated. But the field inspection technique was too crude to allow the limit to approach the maximum calculated capability any closer. Moreover, most of the flaw types experienced previously in con- . struction were less than 10 mil deep. Flaws greater than· 10. mil, but not greater than 25 mil deep, are blended out. Flaws deeper than 25 mil are given special evaluation and the appropriate action determined on a case-by-case basis·. ·

High-temperature piping elbows should be free of defects as they are generally the most vulnerable locations fa" effects of surface flaws on piping integrity. This is due to the uncertainty in the time-dependent crack-growth rates, and the elbows are locations of maximum stresses, and there is potential for vibration-induced high-cycle stresses. As it is difficult to accurately predict system vibrata"y response, measurements and insp_ection for pipe vibratory motion have been taken and will continue during FFTF plant startup testing. This will assure that the piping has sufficient margins against high-cycle fatigue.

CONCLUSIONS AND RECOMMENDATIONS

Simplified rules and preliminary design limits developed for FFTF piping in 1974, based on expected behavior, engineering judgment, approximate calcu­lations, and detailed inelastic analyses of three pipelines have served very well. All designs based on these simplified rules and limits have been con­firmed by detailed code and inelastic analyses.

Six additional FFTF pipelines have had detailed inelastic analyses and comparisons to simplified analysis finding have been performed. Accordingly, detailed system inelastic analyses of pipelines are practical for primary and secondary stress/strain

evaluations. However, simplified analysis methods were needed and developed for weldment radial shrinkag~ and local surface stresses and strains. The use of K indices, as in elastic analysis, seems to be the only practical way to treat weldments.

Simplified analyses and elastic analyses pro­vide significant insight for designing a pipeline and they help provide valuable data useful for com­paring with detailed inelastic analysis results. After adequate comparisons of detailed inelastic analysis response for lines limited to P + Q ~.3 ~. where the temperature hold-time relaxation continues monotonically unaffected by the thermal transient, .. the development of less conservative elastic analy­sis rules should be attempted by ASME Code bodies • Moreover, it is our experience that by keeping t.he pipeline primary plus secondary stresses in the range where shakedown in creep occurs, the (i.e, P + QR ~ 3 "S;,) creep-fatigue life will be governed by the stress-time history and very little usuage will be consumed by the cycle fraction related to the strain range. That is, the N/Nd fraction·is small and the T/TD is designed so that elastic fo1lowup is not significant and the P + Q stress ranges are less than the elastic shake­down range. Then the creep damage will correspond to monotonic relaxation during the service life and be acceptably low. Of course, elbow end weldment radial shrinkage, mismatch and configuration must be controlled or the weld will become design controlling.

.Scratches, d{ngs, chisel ~arks, etc. inadver­tently get imposed on piping and equipment while the plant is under construction. Accordingly, con­sidering that such flaws can reduce the operational fatigue capabilities, an acceptance criteria should be developed for identifying what flaws can be tol­erated and what flaws should be removed prior to insulating and placing the pipe into operation. For future designs, to provide a sacrificial layer of material that could be blended off, it is recom• mended that 0.025-inch {0.64-nm) allowance be applied in the design like a corrosion allowance o.r wall thickness tolerances. Moreover, more high tem­perature cycle fatigue and crack-growth data are desired in the ASME code to assess fabrication i ndi ced f1 aws and vibratory stresses.. In part i cu­lar, threshold oK and da/dN crack-growth rates up to 12QOOF are desired. Smooth bar high-cycle fatigue data to 109 cycles are also desired.

Significant ·advances in methods-and technology for elevated temperature pipei.ng design have occur­red in recent years but improvements are still expected and desired to reduce design costs and to. eJ1hance the reliabi_lity of the piping.

ACKNOWLEDGEMENT

This paper is ·based on· work performed at the Hanford Engineering Development Laboratory (HEDL), Richland, Washington operated by Westinghouse Han­ford Company, for the U.S. Department of Energy. The.author is grateful to the HEDL Plant Analysis staff who carried out many of the analyses provid­ing input to this paper. Special thanks go to · M.J. Anderson, W.L. Chen, S.N. Huang, H.R. Lindquist, G. D. Summers and E. o. weiner.

J

REFERENCES

1 Severud, L.K., aSimplified Methods and Application to Preliminary Design of Piping for Elevated Temperature Services,• Presented at the Second National Congress on Pressure Vessels and Piping, San Francisco, CA, June 23-27i 1975, Advances in Desian for Elevated Temoerature

· -. Environment, AS11E. New· Y ark,. NY. .

2 Sampson, R.C. and JagelS, R.E., "Stress Analysis for the Design of Liquid Metal Piping in the Fast Flux Test Facility," 78-PVP-21, Presented at the Joint ASHE/CSNE Pressure Vessels and Piping Conference, Montreal, Canada, June 25-30, 1978.

J Pan, Y.S. and Jetter, R.I., "Inelastic Analysis of Pipelines in FFTF CLS Module," Pressure Vessel and Piping Conference, Miami Beach, FL, June 24-28, 1974, Pressure Vessels and Pieing:· Analysis and Comouters, AS1·1E, New York, NY.

4 Chen, W.L. and Weiner, E.O., ·"Inelastic· Analysis of Pipeline in FFTF Heat Transport System," PVP-PB-028, Presented at the Joint ASl~E/CSME Pres­sure Vessels and Piping Conference, Montreal, Canada, June 25-30, 1978.

5 Huang, S.N., "Inelastic Analysis of Two Pipelines in the Fast Flux Test Facility," PVP-36, Presented at the Third U.S. National Congress on Pressure Vessels and Piping~ San Francisco, CA, June 25-29, 1979. ·

6 MARC-CDC, "Non-Linear Finite Element Analysis Program, (User's Information Manual)" Ref. F, Control Data Corporation, Minneapolis, r4N, May 1974.

7. Bree, J. "Elastic Plastic Behavior of Thin Tubes Subjected to Internal Pressure and Intermittent High-Heat Fluxes with Application to Fast Nuclear Re-actor Fuel Elements," Journal of Strain-Analysis, Vol. 2, 1967, pp. ·226-238.

8 O'Donnell, W.J. and Porowsk1, J., "Upper Bounds for Accumulated Strains Due to Creep Rat­cheting,a Welding R~search Council Bulletin No. 195 June, 1974. Also, Transact1ons ot the ASNE Wetsure Vessel, Pioino and l·laterials Conterence,. Miami FL, June 24-28,_ 1974.

9 H1bb1tt, H.D., •special Structural Elements for Piping Analysis,• Presented at Pressure Vessels and Piping Conference, Miami FL, June 24-28! 1974, Pressure Vessels and Piping: Analysis and Ccmputers~ A:)'~t:. New York., NY.

10 Pardue, P. and V1gness, I.; "Properties of Thin-Walled Curved Tubes of Short-Send Radius," Trans. ASME, Vol. 73, 1951, pp. 77-87.. ·

11 Gross, N. and Ford, H., •The Flexibility of Short-Radius Pipe Bends," Proceedin9s of the Institute of Mechanical Engineers, SerleS a. Vol. 1, 1952-1953, pp. 480-491.

12 Imamasa, J. and Uragami, K., aExperi­mental Study of Flexibility Factors and Stresses of Welding Elbows with End Effects," Proceedinos of the Se,ond International Conference on Pressure Vessel Technoloor, Part r~ San Antonio, IX, Oct. 1-4, 1973, pp. 4 7-426. . .

13 ASME Boiler and Pressure Vessel Code, Set-tion III: Nuclear ?ower Plant Components, Division 1: Metal Components, ASME, New York, NY, 1971.

14 r~arkl, A.R.C., "Fatigue Tests of Welding Elbows and Comparable Double-Mitre Bends," Trans. ASME, Vol. 69, 1947 p. 7. --

15 Markl,.A.R.C., "Fatigue Tests of Piping Components,• Trans. ASME, Vol. 74, 1952, p. 301.

16 · Liebowitz, H., Vanderveldt, H., and Sanford, R., "Stress Concentrations Due to Sharp Notches," Experimental Mechanics, Dec. 1967, pp. 513-517.

17- Juv1nall, R.C. Stress. Strain, and Strenath, HcGraw H-ill, New York, NY, 1967, pp. 2S2-259.

18 ASME Boiler and Pressure Vessel Code, "Cl~ss ·1 Case Interpretat1ons: Code Case 1331-5, Components in Elevated Temperature Service," and "Class III Case Interpretations: Code Case 1592," New York, NY, 1971.

19 r~aiya, P.S., ''Effects of Notches on Crack Initiation in Low-Cycle Fatigue," Materials Science and Engineering, ·Vol. 38, 1979, pp. 289-294.

20 Chow, J.G. Y. and Soo, P., "Development of a Procedure for Estimating the High Cycle Fatigue Strength of Some High Temperature Structural Alloys," Methods for Predictino Material Life in Fatigue, A$i"1E WA'1, Dec. 2-7, 1979, New York, NY.

21 Jaske, C.E. and O'Donnell, W.J., "Fatigue Design Criteria for Pressure Vessel Alloys, ASME Journal of Pressure Vessel Technoloay, Nov. 1977.

22 James, L.A., "Fatigue-Crack Propagation in Austenitic Stainless Steels,• HEDL-SA 1051 Atomic Energy Review, International Atomic Energy Agency, Vienna, Austria, Jan. 1976.

23 James, L.A., "Frequency-Effects in the Elevated Temperature Crack Growth Behavior of Austenitic Stainless Steel - A Oesign Approach," 78-PVP-97, Presented at the Joint AS~1E/CSl-1E Pres­sure Vessel and Piping Conference, Montreal, Canada, June 25-30, 1978.

I

24 Hsu, T.M. and Liu, A.F., "Stress Intensity Factors for Truncated Elliptical Cracks,• Handout at the· Seventh llational Symposium on FrJcture MQchanicl, .College Park, MD, August 27-29, 1973.

Z5 Shah, K.C. o1nd Kobaya,hi, A_S,, •on the Surface Flaw Problem," in The Surface Crack: Phy­sical Problems and Comoutational Solut1ons, ASME, New York, NY 1972, pp. 79-124.