ML20148F524
| ML20148F524 | |
| Person / Time | |
|---|---|
| Site: | Comanche Peak  |
| Issue date: | 03/31/1988 |
| From: | Scott Moore OAK RIDGE NATIONAL LABORATORY |
| To: | |
| Shared Package | |
| ML17303B208 | List: |
| References | |
| RTR-NUREG-0797, RTR-NUREG-797 NUDOCS 8803280194 | |
| Download: ML20148F524 (43) | |
Text
{{#Wiki_filter:.o, Og f! 4,.lq..o.
~
-t *
,- {
ASME CODE STRESS INTENSIFICATION FACTORS FOR COMANCHE PEAK STEAM ELECTRIC STATION, UNITS 1 AND 2 l S. E. Moore h
~
- 1. INTRODUCTIC-N The official Code-of-Record for the p' ping stress analysis effort being
/
performed by the Stone and Webster Engineering Corporation (S'a'EC) for Comanche Peak, Units 1 and 2, is the ASME Code,I 1974 edition including the Summer 1974 Addenda.* That edition of the ASME Code contains a complete set of rules for the design stress analysis of ASME Class 2 and 3 nuclear power plant piping in Articles SC-3600 and ND-3600, respectively. Further, those rules include a set of design criteria statements (equations) written in terms of nominal stresses, stress intensification fag. tors (SIFs) and stress limits. The intent of the criteria statements is that the c. .tservative combination of internal pressure C i stresses, cross-section moment stresses, and thermal stresses represented by the left-hand side of the equations will not be greater than the allowable *
.t stress limits on the right-hand side. ) j Since 1974, the design criteria equatior.o. including the ASME Code re-t quired SIFs, of Articles NC-3600 and ND-3600 have been the same for Class 2 \
and Class 3 piping. Howe er, ever the years' a number of changes in the design rules and SIFs have been adopted by the ASME Code Committee and published in later editions and addenda to the Code. Some of the revisions have been more restrictive and some have been less restrictive than the Comanche Peak 1974 Code-of-Record. Application of a specific change must, therefore, be justi- 7 fied and acceptable to the enforcement authorities having jurisdiction at the a
*The terms ASME Code. ASME-III, or Code as used herein refer to Section III of the ASME Boiler and Pressure Vessel Code, Ref. 1. Specific editions of i L
the Code and the relevant Addenda are identified by date, enclosed in brackets. e.g., [1974-S74]. Sections, Articles. Subarticles. Paragraphs, and Subpara-graphs are identified according to the Code numbering system, e.g., Article NC-3000, t
=- - -
?* '
\' t
. g* ',
a ;y
. /
nuclear plant site. The purpose of this study is to investigate the accepta-A bility of using SIFs from Subparagraphs NC/ND-3673.2(b) of the 1983 ASME Code j edition (1983-V84) in the Stone and Vebster, Comanche Peak, piping stress analysis effort _or to provide suitable alternatives for.the following items
\
in conformance with the Cc,de-of-Record provision of Subarticle NA-Il40.
- 1. Branch Connections,
- 2. Girth Buttseld (mismatch formulation),
- 3. Circumferential Fillet Welded or Socket Welded Joints,
, Subarticle NA-1140 (1974-S74] reads as follows:
"NA-1140 Effective Dates of Code Editions, Addenda, and Cases. -
(a) Code Editions become mandatory on July 1 of the publication year printed on the cover. Addenda may be used on or after the date of issue and become mandatory six months after the date of
- issue.
I ff)CodeEditions,' Addenda,andCaseswhichhavenotbecomemanda-tory on the contrac.t date for a component may be used by mutual consent of the Owner or his agent and Manufacturer or Installer c t , or after the dates permitted by (a) through (d) above. It is per-mitted to use specific provisions within an Edition or Addenda pro-vided that all related requirements are met. (g) Caution is advised when using Addenda or Cases that are less
~
t restrictive than former requirements without having assurance that they are acceptable to the enforcement authorities having jurisdic-tion at the nuclear plant site. (h) ..." . Figures 1 to 3 show the specific SIFs under consideration with the appro-1 lA priate notes and sketches as they were given in 1974 Figures 4 to 6 show corresponding SIFs, notes, and sketches from (1983-V84). The current SIFs, l notes, and sketches (1986-SS6] are identical to tnose in Figs. 4-6, where l ' we have identified the pa',11caton dates of the latest rule change in the margin, i l l I
\
.f
_3 + s
$ / 4..
. .. u~
Note (U ), for example, jas added in the Su:nmer 198?j AS.)e tda, i'.e. , S83. Each of these4hcuses jlus any related material, such as fabrication t requirements, that may havr' arc impact on' the acceptability of th'e , current, SILf for use with the 1974 design basis is discussed in the foliowtng' sections. 04 recommenda-tions are summarized in Sect 6.
- 2. DESIGN BASIS F,0R CLASS 2 AND 3 PIPING
~
"n 2.1 Design Stress Analvsis Criteria , -
a Prior to 1974 cuclear Class 2 and Class ; piping was designed according to the rules of Section B31,1 of the USA Standard Code for Pressure it.in g . The only loads other than internal pressure that weie evaluiatyd were the bending noments obtained fror a b'. ping system flexibilit3.anaVais for dead weight and thermal expansion. The acceptbnco criteria veze based on ma7.inum stress theory and the fatigue correlations tid i factors (S' ifs) developOj by Markl. 8 Marki and George, and Pod naugh and Georse in the late 1940s and 1950s. Nuclear power pjping was first treated differently from industrial power piping when Section B31.7 of the USA Standard Code for Pressure Piping was 1 ' ( t published in 1969. That code gave design rules fo; taree,different nuclear l piping safety classes, with the rules for Class 1 bt ing the most stringent.
, The design rules for Class 1 piping were patterned after the design rules for i
Class 1 nuclear vessels with one important and fundamental difference. A 'simpli-
, - - ~
n fied method of design st cess analysis consisting ei :very simple design criteria equations, written in terms of nominal stresses and leading coefficient multi-I pliers called stress ~ indices, was introduced. These stress indices are different from, but similar in nature to, the SIFs developed by Mark 1 for incustrial piping. l The design roles in B31.7 f6 Class 2 and Class 3 piping were pattirned af ter
, Jf , .
c' { , l
,f. <, , :l ,, , , . , - -. . . , , ,n.. - - - . , ~ . - - - - -
k the B31.1 industrial piping code. In fact, for the design stress analysis requirements, the B31.7 code simply refers to B31.1. In (1971-571] the ASME Code extended its coverage to include nuclear piping by simply incorporating the B31.7 rules for Class 1 piping and the B31.1-1967 rules for Class 2 and 3 piping. In [1974-S74), however, a complete new set of
> design criteria equations was introduced for the required design stress analy-sis of Class 2 and Class 3 piping. This new stress analysis procedure was a hybred between the B31.7 Class 1 analysis and the B31.1 Class 2 analysis of the 1971 Code. Subsequent changes in the design analysis rules for Class 2 and 3 piping have retained this dual nature. Although the Code makes some distinctions between Class 2 and Class 3 piping, the design basis stress analysis require-ments for the two Classes have been identical since [1971-571]. We.shall, therefore, refer only to the Class 2 rules hereinafter.
One particularly important relation between the Class 1 and Class 2 design stress analysis criteria is the following: i = 1/2 C K2 2 > 1.0 , (1) where i is the Class 2 SIFs, and C2 and K 2are the Class 1 "primary plus secon-l
- j. dary" and "peak" stress indices for piping system moments, respectively.
l Although-this re..ationship was used by the writers of B31.7 (see Foreword to B31.7, Ref. 9) and has been used extensively in later ASME Code rule stress analysis developments, it was not formally incorporated into the Code until l [1977-577], h'C-3673.2(b) . The lower limit of i = 1.0 was added in [1983-S84]. l It is important to realize, however, that Eq. (1) has been a valid relationship l (~ since before 1969 when the Class 1 simplified analysis method was introduced.
i o Rules for the design and construction of Class 2 nuclear piping are given in Article NC-3600, in Subarticles NC-3640 and NC-3650. SIFs are used specif-ically in the design criteria equatior.s that must be satisfied at every point in the piping system by a suitable stress analysis. In 1974 those equations were given in Paragraph NC-3652 as follows (using the Code nomenclature and numbering): NC-3652.1 Sustained Loads PD 0.751 M A
= + (0)
S gg 4t Z 1 1.0 S h n NC-3652.2 Occasional Loads P D S OL
=
[n + 0.751(M, 7 B
+ M h) 51.2 S (9)
NC-3652.3 Thermal Expansion iM gA 1S g. (10) and
^
S TE
=
n
+ 0.75i +1 f i'(Sh+S)'
A (II) where (1) is a unique stress intensification factor, SIF for each of the dif- 3 ferent types of piping products and welds that make up a piping system. SIFs i for commonly used products and welds are specilled in Subparagraph NC-3673.2. l l The SIFs under consideration in this report are given herein in Figs. I and 4 (See the Code for nomenclature definitions.) The intent of Code Eqs. (8)-{11) is that the conservative combination of pressure stresses (P terms) and cross-section moment stresses (M , M , and M 3 B C
terms) represented by the left-hand side of the equations will not be greater than the stress limits given on the right-hand side in terms of the allowable stress Shf r steady state and occasional loads and the allowable stress range S for cyclic loads. 3 Equations (8) and (9) are for protection against plastic limit-load collapse or rupture from primary loads. Equations (10) and (11) are for protection against large deformations and fatigue failure from cyclic loads. In [1974-W76), the Code added the concept of Service Level limits to the occasional load cateogry, Eq. (9), in recognition of the fact that more liberal allowable stresses could be tolerated without unacceptable damage - depending on the load source and post incident inspection and repair. In essence, the right-hand limit of Eq. (9) was multipled by a factor that depended on the severity classification-of the postulated incident loading. The four Service Levels and corresponding stress limits are shown below. Service Level Stress Limit Design /A (nromal) 1.0 S h B (upset) 1.2 S
- h C (euergency) 1.8 S h
D (faulted) 2. 0 S. n It is our understanding that the [1974-W76] Service Level stress limits have been specifically approved for use in the Comanche Peak piping stress analysis. The Class 2 piping design bases remained unchanged until (1980-W81] when the Code Committee made a major change in the stress evaluation criterion for primary loads based on studies by Rodabaugh and Moore. 1,12 Equations (9) and (9), written in terms of the fatigue data based SIFs, were replaced by new equartons written in terms of the Class 1 "B" stress indices developed from e
t plastic-collapse limit-load considerations. The new equations are:* NC-3652 Consideration of Design Conditions PD M S =B +B 2 11.5S h (Sa) gt 12t n NC-3653.1 Occasional Loads P D M 3
+M SOL " bl +B 2 $ 1.8 S h' (9^)
Z n , but not greater than 1.5 S y, where S y is the Code specified minimum material yield strength. The By and B 2 stress indices are taken from Subarticle NB-3680 for Class 1 piping design. The Service Level limits to be used on the right-hand side of Eqs. Ga) and (9a) were also changed at the same time to the following: Service Level Stress Lindt Design 1.5 S h B 1.8 Sh 5 1.5 S, C 2.25 Sh 5 1.8 S y D 3.0 Sh 5 2.0 S 7 Except for the addition of an equation similar to Eq. (10) for checking single nonrepeated anchor movements, the remaining design basis criteria have not been changed since (1974-S74].
*We have numbered the (1980-W81) equations as (8a) and (9a) to distinguish them from Eqs. (8) and (9) introduced earlier.
- ,. ~
December 31, 1981, corresponding to the publication of (1980-W81] and its mandatory application date of July 1, 1982, as required by Subparagraph NA-1140(a), are important to this investigation because changes in the SIFs made after those dates may not be appropriate for use with Eqs. (8) and (9) [1974-S74] for the primary load caregory stress evaluations. Other parts of the design basis criteria, however, are not effected provided that all other related requirements are also met. 2.2 Functional Capability Criteria In a somewhat related move, the Nuclear Regulatory Commission (NRC) . imposed supplemental requirements for assuring the "functional capability" of essential safety related nuclear piping.I ' ' The criteria to be satis-fied are given in Ref. 13 (1978), subject to constraints imposed by the NRC 14 staff that were based in part on Ref. 11 (July 1978). For Class 2 and 3 essential piping, those criteria made use of Eq. (9) and SIFs from NC-3672 (1977-S78]. According to Ref. 14. functional capability for Class 2 or 3 essential piping with D,/t < 50, except branch' connections, is assured without further proof if the following equation is sat:sfied:* P D M 0.5 + 0.751 -f- < 1. 5y, S (9f) n
*We have numbered this equation as (9f) for functional capability to l
distinguish it from the other Eqs. (9) and (9a). It is the same as Eq. (9) with a different right-hand limit. l l l I i
./ .
-9 where M equals the resultant moment due to weight, earthquake (considering 1
only one-half the range and excluding anchor displacements), and other sus-tained mechanical loads. Specified values for 0.751 are given in Ref. 14 Equation (9f) is the same as Eq. (9) (1974-W76] with Service Level B limit of 1.2 Sh replaced by the more liberal limit 1.5 S .y For Class 2 or 3 branch connections, Ref. 14 requires the Class 1 cri-teria to be satisfied. Because those criteria do not make use of SIFs, func-tional capability for branch connections is not addressed in the present study.
- 3. BRANCH CONNECTION SIFs SIFs for branch connecticas with the approp'riate notes and sketches from (1974-574] for both primary load and thermal expansion fatigue evaluations are shown in Figs. 1 and 2. Since that time the following chances have been made as noted in the margin of Figs. 4 and 5.
(1) (1977-579] when a new equation was added for calculating the SIF for checking run-end moment loads. At the same time defining equations for calcu-lating the branch-end section modulus Z3 and the run-end section modulus Z r were moved from Paragraph NC-3652.4 to Fig. NC-3673.2(b)-1 [ Fig. 4 herein}. 1 (2) (1980-S80] when note 6(d) was revised by the Code Committee to l exempt branch pipe sizes less than 4-in. NPS from the inside corner radius l ! requirement,to remove a difficult and expensive fabrication step that was judged to be almost unenforceable. The revision had no impact on the calcu-lated SIF values because the variable yr is not included in the SIF equations.
~
(3) (1980-S80] when note 6(h) was added to exempt branch connections from the outside corner radius requirements of note 6(e) provided that the SIFs are multiplied by an additional factor of 2.0, with a minimum SIF = 2.1. l l \ l
J (4) [1980-S82) when General Note 2 was added to Fig. NC-3673.2(b)-2 (Fig. 5 herein) to redefine the midwall radius ry of the branch. .. (5) (1980-S83] when a minor editoral correction of.no technical con-sequence.was made to note 6(c). Revisions (1) and (4) are less restrictive and according to NA-1140(g) must be justified for use with the [1974-S74] design basis equations. Revision (3) is more restrictive. It is our unders tanding, however, that - note 6(h) is being used in the SWEC piping stressinalysis s for Comanche Peak. Revision (1) is discussed below in Sect. 3.1. Revisions (4) and (3) are discussed in Sect. 3.3. 3.1 Run-End SIFs From 19'/4 to 1979 the following equation [1074-S74] was specified for calculating SIFs for both the branch-end and the run-end, R 2/3 c' 1/2 T{ r;
> 1.0 , (2) i=1.5(T r a I
r r p subject to the restrictions identified in Fig. 1 under note 6. According to Subparagraph NC-3652.4, the branch-end and the run-end were to be evaluated ( separately using Eqs. (8)-(11) with the appropriate moments and section modulus l Z b
#2'r Equation (2) was developed by Rodabaugh in 1970 for branch connections with moment loadings on the branch. His data base consisted of measured maxi-mum stress data for 23 models, fatigue test data from 8 models, and Bijlaard's
- theory for correlation guidance. However, in the absence of sufficient data for branch connections with moment loadings through the run, Eq. (2) was also 1
specified for calculating run-end SIFs on the considered judgment of the Code Committee that it was adequately conservative. (It is conservative with respect to the single data point that existed in 1970.)
In (1977-S79] a distinct SIF for checking the run ends was introduced I7 based on Class 1 stress index studies by Rodabaugh and Moore which in-cluded additional stress analysis data not available for the earlier study. The new run-end SIF equation was
/R 2/3 /r')
E * (3) i = 0.4 1 l\R,j -> 1.5 . I T
\r This is the same equation that appears in [1983-W84] and in the present Code
[1986-586]. Although Eq. (3) does not correspond exactly to the Class 1 stress indices C2 and K g (see Eq. (1)] proposed in Ref.17 'and adopted in (1980-S81], it is conservative with respect to the same data base. Equation (3) actually cor-responds with an earlier unpublished draft proposal for C2 and K2 . Table 1 shows comparisons between the SIFs from Eqs. (2) and (3), the corresponding Class 1 stress index product C22 K and the Ref. 17 data base for coment load-ings en the run. These data consist of fintte element results obtained by 18 for a series of unreinforced nozzles (U Models), Bryson, Johnson, and 3 ass uniform thickness reinforced nozzles (S1 Models), and compact tapered thick-ness reinforced nozzles (P30 Models); and experimental data reported by Rodabaugh 15 (Weldolet), Corum et al.I9 (ORNL-1), and Gwaltney et al.20,21 (ORNL-3, ORNL-4). . Table 1 shows that Eq. (2) gives a poor fit to the run moment data, being excessively conservative for many of the models and unconservative at the lower limit of i = 1.0 (21 = 2.0). Equation (3) provides a much better fit to the da'ta, being conservative with respect to all the data J except for the unreinforced U models and ORNL-4. None of these particular models met the reinforcement requirements of note o(a), and the ORNL models also failed to meet the outside corner radius requirement of note 6(e). If
the data from models that failed to meet Code requirerents are neglected, The then Eq. (3) is conservative with respect to all the relevant data. Class 1 stress indices, however, give the best fit t'o the data and are actually less conservative than Eq. (3), the current run-end SIF. Because Eq. (3) was introduced into the Code prior to [1980-W81] and because it is conservative with respect to the available stress analysis data base, it is clear that its use for checking run-end stresses with the [1974-574] piping stress analysis criteria is justified even though it is less restrictive than the previous SIF, Eq. (2). 3.2 Branch-End SIFs In addition to developing the run-end Class 1 stress indices, Rodabaugh and Moore also checked the Class 1 equivalent of Eq. (2) for branch-end mo-ment loadings against the additional finite element stress anlaysis data ob-tained by Bryson, Johnson, and Bass. O Table 2 shows the comparison between 21 > 2.0 from Eq. (2) with the finite element data from Table 10 of Ref. 17. l Equation (2) is conservative with respect to all the finite element data except for the unreinforced U modelsr' As noted earlier, those models did not i meet the reinforcement requirements of note 6(a). The small amount of uncon-servatism for those models, however, is not of direct consequence in validat-ing the SIFs for use in checking the branch-end moment stresses. 3.3 General Note 2. Fic. NC-3673.2(b)-2 In [1980-582] General Notes (1) and (2) were added to Fig. NC-3673.2(b)-2 l l ! as shown in Fig. 5. General Note (1) simply states that certain variables used in the branch connection SIFs are defined in the figure. General Note (2) re-defines the midwall radius of the branch r' as the midwall radius of the l i i
reinforced portion of the nozzle if the reinforcement length Lg is greater than 0.5(rgb T) , where r is the inside radius of the branch and T is the wall 1 b thickness of the reinforced nozzle. General Note (2) is technically in error and cannot be justified for use in the SWEC Comanche Peak piping stress analy-sis effort, either for primary load evaluations or for thermal expansion fatigue evaluations - independent of when the revision was introduced into the Code. Note (2) was originally perceived as solving a definition problem to distinguish betweenb T andT{forbranchconnectionsthatlooklikesketch(d), Fig. 5, but are considerably thicker walled than needed to satisfy the internal pressure reinforcement requirements of NC-3643 as required by note 6(a). As originally proposed, note (2) was only to be used in calculating the SIFs for the branch-end, Eq. (2), and the run-end, Eq. (3); but n_ot for calculating the branch-end section modulus. As published, however, it is clearly evident that General Note (2) may be used in calculating the branch-end section modulus as well, i.e., (4) Zb = (r') T{ . , The result is that the calculated moment stresses (0.751 M/Z) and (i M/Z) as used in the stress criteria Eqs. (8)-(11) can be reduced as a function of (r') simply by increasing the wall thickness of the branch. As pointed out in Ref. 17, however, the maximum stress in nozzle connec-tions from moment loadings on the branch pipe is not a function of branch pipe wal'. thickness., The maximum stress generally occurs in the shell side and is inversely related to the outside diameter of the nozzle reinforcement. The variable rpis included in SIF Eq. (2) to account for that influence. In
a recent study of branch connection SIFs conducted for the Pressure Vessel Research Committee, Rodabaugh 22 compared various correlating equations from the technical literature and the different codes. All the correlating equa-tions are indepen' dent of the branch wall thickness. 3.4 An Alternate to Using General Note (2) In the event that not using General Note (2) would cause an undue hard-ship in the SWEC effort, we can offer the following for consideration. Ref-erence 22 contains a substantial amount of fatigue test data on specialty product branch connection fittings (WFI products) that were not previously - available. Those data show that the current branch connection SIFs, i.e., Eq. (2) and (3), are overly conservative for fatigue evaluations when the multiplying factor of 2.0 from note 6(h) [1950-S80] is included. As a con-sequence, Rodabaugh recommends that note 6(h) be deleted from the Code. If the Ref. 22 recommendation is adopted by the ASME Code Committee, it would apply for thermal expansion fatigue evaluations using Eqs. (10), (11) [1974-S74] but not for primary load evaluations using Eqs. (8), (9). This is proper because note 6(h) [1980-S80] was originally introduced as a fatigue I reduction factor to account for the local stress concentration caused by an undressed weld at the branch-run intersection. Thus removing note 6(h) on the basis of new fatigue test data should only effect the Code fatigue evaluatin procedures. Moreover the present Code (1986-S86] requires the l use of Eqs. (8a), (9a), which do not involve SIFs, for primary load evaluations. A suitable alternative to using General Note (2), Fig. NC-3673.2(b)-2, is to perform the Comanche Peak piping stress analysis as in the past using note l l l
_15 6(h) but not using General. Note (2) for the primary load evaluations, Eqs. (8)- and (9). Then reanalyze only those cases that fail to meet the required stress limits using Eqs. (8a) and (9a) (1980-W81] and the moment combination procedure for Class 1 branch connections as required by Subparagraph NC-3653.3(b) [1980-V81], but not using either General Note (2) or note 6(h). The thermal expansion fatigue evaluations may be performed as in the past but without using either General Note (2) or note 6(h). 3.5 Related Requirements A thorough review of Articles NC-2000 Materials, NC-4000 Fabrication and Installation Requirements, NC-5000 Examination, NC-6000 Testing, NC-7000 Protection Against Overpressure, and NC-8000 Nameplates, Stamping, and Reports; and comparisons of the contents of {1983-W84] with those of (1974-S74] was con-ducted to determine whether any changes would impact the SIT calculations in the piping stress analysis. None were found. It is thus concluded that all related requirements for branch connections, other than those from Article NC-3000 specifically discussed above, are met as required by Subparagraph NA-1140(f) [1974-S74].
- 4. GIRDi BITTT WELD SIFs 4.1 Mismatch Prior to [1974-S74] the SIF for a Class 2 butt welding joint between sections of straight pipe, between straight pipe and reducers, and between straight pfpe and welding neck flanges was simply i = 1.0. In [1974-S74] an entirely new design analysis procedure for Class 2 piping was introduced. This new procedure included the use of SIPS from the earlier Codes, but they took
on new meaning in parallel with the Class I stress indices through the relation i = 1/2 C 22 K f Eq. (1). For girth butt welds, three different values for the SIF were given depending on whether the nominal wall' thickness t was greater n than or less than 3/16 in. , and whether the mismatch 6 was greater than or less than 0.1 tn, as shown in Fig. 1. Mismatch 6 was defined in note (1) and the sketch as shown. These SIFs corresponded exactly with the (1971-573] Class I stress indices as noted below. Type of weld C 2 K 2 1 2 1.0 a) flush 1.0 1.1 1.0 b) as-welded, 1.0 1.8 1.0 t, 2 3/16 in and /t n5 0.1 c) as-welded, 1.8 2.5 1.8 t < 3/16 in, or /t > 0.1 n n At the same time, [1971-S73], footnote (12) was added to the Class 1 piping stress index table, Table NB-3683.2-1, defining 5 as follows. Figure NB-4233.1 from the 1971 Code is included herein as Fig. 7. is defined as the maximum permissible mismatch as shown in Fig. NB-4233.1. A value of 6 less than 3/32 in. may be used provided the smaller mismatch is specified for fabrication." The following sentence was added in the 1977 Code. '
"For flush welds, defined in footnote (2), 6 may be taken as zero."
In [1980-S81] the definition of 5 was moved from footnote (12), Table NB-36S3.2-1, to Subparagraph NB-3683.1(a) and the last sentence was revised to read: i "For flush welds, defined in NB-3683.1(c) and for t > 0.237 in. 5 may be taken as zero." l i
This last sentence and the final form of the definition was taken from recom-mendations by Rodabaugh and Moore published in September 1978. Although the above definition for 6 was not added to the Class 2 SIF formulation until (1983-W84], as shown under note (1), Fig. 4, it is clear that its intended application, for flush welds and for welds with 6/tn 1 0.1 was developed as early as 1971 when the Class 1 stress indices were introduced into the ASME Code. The exact correlation between the (1971-S73] Class 1 stress indices and the (1974-S74] Class 2 SIFs for girth butt welds confirms the con-clusion. Thus, this particular Code revision is neither more nor less restric- -- tive than the Comanche Peak Code-of-Record. No other changes in the Class 2 girth-butt weld SIF were made until (1953-W84] when the mismatch restriction was dropped entirely for pipe with nominal wall t > 0.237 in, based on the recommendation of Ref. 23. The SIF for - n pipe with t < n 0.237 was expressed in equation form, as shown in Fig. 4, i.e.,
' ~~~
(5)
' ~
..# 1.0 $ i = 0.9 (1 + 3 6/tn) 5 1.9 . -
___ _.E This SIF corresponds exactly with the (1980-S81] Class I stress indices for as-welded girth butt welds joining items with nominal wall thickness nt < * ' l l NB-3683.4(b ), i.e. , - l 1 - , C., = 1. 7 + 3 ( 6/ t ) < 2.-1 , ,
"' (6) j ,
K, = 1.8 , .
~ . .
r as recommended by Rodabaugh and Moore in 1978. Reference 23 includes an extensive study of girth butt veld stress indices including the effects of weld reinforcement, mismatch, and abrupt thickness
change resulting from joining two pipes with different wall thickness. The recommended2 C and2 K stress indices are "... deemed to be adequate to take care of both Code permitted weld reinforcements and mismatch." It is thus clear that the [1983-W84] SIF formulation for girth butt velds, including mismatch 6, is also appropriate for use in the SWEC piping stress analysis effort for Comanche Peak. l l l l l
4.2 Radial b' eld Shrinkane Radial weld shrinkage at girth butt welds is no.t generally included in the design stress analysis of nuclear piping, even though it appears obvious that excessive drawdown would effect the stresses in the immediate region of the weld. That concern could be relieved to a certain degree by imposing fabrication tolerances on the amount of radial weld shrinkage permitted dur-ing construction. That approach was apparently taken for the Comanche Peak 24 Class 2 and 3 piping fabrication where the radial weld shrinkage a was limited to a$t I# * < 0.375 in., or n n A $ 0.1815 for en > 0.375 in. The question, "Should that amount of radial weld shrinkage be considered
~
explicitly in the piping stress analysis?", naturally arises. The answer to this question is developed below. Reference 23 includes an extensive study on the effects of radial weld shrinkage on the stresses at girth butt welds. Based on an analytical param-eter study, using a conservative mode) for the weld joint and thin-shell theory, the report shows rather high stresses at the joint, particularly-for for moment loading. Figure 16 from Ref. 23, included herein as Fig. 8 shows the maximum normalized axial stresses from a bending moment, corresponding to the effect on the C stress index, as a function of the veld shrinkage ratio 2 3 / t. The bounding equation from that figure, i.e., c/S = 1.0 + 2.9 a/t , (7) could be used to characterize the influence on C . Equation (7), along with 2 K and Eq. (1) could then be used to characterize the SIF. Using 4/t = 0.5 7 n and K 3
= 1.8 would then give i = 2.2 (for piping with t >n 0.237 in. so that
- e ,,- -,- .
-,y _# ,,- - , - ..-,,,,,v_ _ - . , _ , _ , , _ _ _ . _ _ .
mismatch could be ignored). If D/t for specific pipe sizes were included in the calculation, Fig. 8 indicates that the"SIF would be so~ewhat less than 2.0. Reference 23 goes on, however, to point out the conservative and uncertain aspects of the study, and finally concludes that veld shrinkage a should be ignored in the stress analysis of Class I piping, provided that 4/en < 0.25. That recommendation was incorporated into the Class 1 rules in [1980-581} by addition of the following in NB-3683.4(b):
"... Girth welds may also exhibit a reduction in diameter due to shrinkage of the veld material during cooling. The indices are not applicable if /t is greater than 0.25 where a is the radial shrinkage measured from the outside surface. " +
The question for Comanche Peak Class 2 and 3 piping thus beco=es "Should radial weld shrinkage be considered in the stress analysis if 0.25 < na/t < 0. F Consider the following, however. A restriction on radial weld shrinkage for the stress analysis of Class 2 and 3 piping has never been included in the Code; and was not added in (1980-S81] when consideration of mismatch was added for Class 2 piping and restrictions on weld shrinkage were added for Class 1 piping, in spite of the fact that both revisions were based on the same reference study (i.e., Ref. 23). This indicates that the Code Committee in-tended that radial weld shrinkage not be considercd in the' design ~' stress -
,;,.... . .' , i . ,, ,-
J,' ' : "' " ' analysis of Class 2 or 3 piping. 3, 1, ; H . FM 33- A at
'I ,z .;' s u ~ ,, , , . . -r 25 g Vs. .,
, .' 1 . ,--a --
In a reore recent study, Rodabaugh evaluated the fatigue design margin ; ~,
~~
for both Class 1 and Class 2 nuclear piping against moment loading fatigue test data on girth butt welded pipe (see Ref. 5). Figure 3 from Ref. 25, included here as Fig. 9, shows !! ark 1's f atigue-to-failure ' correlation line and the Class 2 design stress-range limit lines as a function of fatigue cycles for SA 106 Grade B carbon steel piping. The figure also shows a "design allowable stress-range" line based on a safety factor of 2.0 on stress-range.
Below 7,000 cycles the real design margin for Class 2 piping is considerably greater than 2.0, even for a sustained load sS = 0; the S3 = 0 line might represent the thermal expansion stresses in an empty pipe. Above 7,000 cycles the design margin is about equal to 2.0. The Ref. 25 study also considered th'e fatigue behavior of austenitic stainless steel and showed that the Class 2 design rules are more conserva-tive than for carbon steel. We, therefore, conclude that the design margin in the Class 2 piping stress analysis rules is more than adequate to accom-modate weld shrinkage up to a/tn = 0.5 for fewer than 7,000 cycles of fully reversed loading. For more than 7,000 cycles, however, it would be prudent to increase the SIF by the factor represented by Eq. (7). 4.3 Functional Capability As noted in Sect. 2.2 complete instructions for assuring the functional capability of essential safety related nuclear piping were provided by the NRC in 1978. ' The requirements on mismatch in Ref. 13 are the same as in (1974-S74) and to our knowledge have not been altered. However, the argu-ments presented in Sect. 4.1 are as valid for functional capability evaluation as for structural evaluations. The'r'efore, these same arguments could be pre-sented to NRC in a request for permission to use the (1983-W84] mismatch for-mulation in the functional capability evaluations. Consideration of radial weld shrinkage is not included in Ref. 13. As we have shown in Sect. 4.2 radial weld shrinkage a 1 en/2 need not be con-sidered in the structural evaluations for piping with less than 7,000 loading cycles. However, plastic-collapse, and thus functional capability, is not influenced by cyclic loading within the limits permitted by the Code rules. Thus, we conclude that veld shrinkage 4 5 t n/2 need not be considered in functional capability evaluations even for girth butt welds with more than 7,000 loading cycles.
s 4.4 Related Requirements s We have examined all relevant Articles of the Code and have compared the contents of (1933-V84] with those of [1974-S74] to determine whether any chances. other than those discussed above, would impact the SIF calculations or the piping access analysis procedures for girth butt welds. None were found. In particular the permissible mismatch given in Fig. 7 from (1974-574] has not changed, i.e., the requirements are the same in [1983-W84]. It is thus concluded that all related requirements are satisfied as required by NA-1140(f) [1974-S74].
- 5. CIRCDIFERENTIAL FILLET WELD SIFs 5.1 Structural Criteria Prior to (1974-574] the SIF for fillet welded joints, socket welded flanges, and single welded slip-on flanges were i = 1.3 based on the earlier work of Markl4 ,5 and Markl and George.6 In (1974-S74) when the new piping stress analysis criteria were introduced into the Code, the SIF was increased to i = 2.1 for fillet welded socket joints and the appropriate notes and sketches shown in Figs. 1 and 3 were added, the SIF for a "full fillet veld" was left at i = 1.3. Since that time the following changes have been made as noted in the margin of Figs. 4 and 6.
(1) [1980-S83] when the two fillet weld categories were combined under a single category identified as "circumferential fillet welded or socket welded joints" and Brazed joints" were identified under a separate category with an SIF of i = 2.1 from the previous Code. The SIF for circumferential fillet welded ... joints was expressed by the following equation
~
i i = 2.1/(C,/tn) 2 1.3 , (12) as noted in Fig. 4, where C, is the length of the fillet weld leg and nt is the nominal pipe wall thickness. (2) [1980-583} when Note (11), which had been added in [1980-S82] to permit a lower SIF of 1.3 provided that both weld legs were greater than 1.6 t n' was revised to define C, for fillet welds with uneqnl leg lengths as the length of the shorter leg. (3) [1980-583] when Fig. NC-3673.2(b)-3 was deleted and the reference SIF sketch was changed to Fig. NC-4427-1. Reference to Fig. ND-3673.2(b)-3 was changed to Fig. NC-3673.2(b)-3 in [1977] to identify the intended Subarticle. (4) [1980-580] when Fig. NC-4427-1 was revised to its present form as shown in Fig. 6. Changes relevant to the fillet weld SIF are the following: (a) the definition for the size of an unequal leg fillet weld was changed from
"... the size of the weld is the leg lengths of the largest right triangle which can be inscribed within the fillet weld cross section."
to ,
"The size of an unequal leg fillet veld is the shorter leg length of the largest right triangle which can be inscribed within the fillet weld cross section."
..."1.09 t 2 1/8 in."
(b) The minimum value for C, was changed from E = n minal phe van tMchess? to "1.09 tn' " '#8 n Except for applicaton to "Full fillet welds" revisions (1), (2), and (4) cay be less restrict,ive than [1970-574] and therefore must be justified for use with the (1974-S74) desiFn basis. Revision (3) may be considered as editorial and of no technical consequence. Revisions (2), (4a), and (4b). are technically the same because the restrictions on C, identified in note (11)
e and in Fig. NC-4427-1 are necessary definitions for evaluation of the SIF Eq. (12) in revision (1). The restrictions on C,, however, are actually more restrictive than [1974-S74] because the earlier definition for fillet welds with unequal legs was incomplete. Thus only Eq. (12) is less restrictive than [1974-574]. Equation (12) was presented to the ASME Code Committee, Working Group on Piping Desing (WGFD) in September 19-82, in response to a request from the Main Committee to relieve the unrealistic abrupt change in the SIF from 1.3 for a full fillet weld to 2.1 for a fillet welded joint. Equation (12) was justified on the basis of studies of Class 1 stress indices for fillet welded joints described in NURFG/CR-0371,23 1978, and Markl's earlier fatigue test 4-6 In Ref. 23 (p. 28) it was concluded that the Code values of C2 = 2.1 data. and X = 2.0 [1977 ed.] were "amply conservative." Using Eq. (1) then gives 2 the upper SIF value of i - 2.1. The lower SIF value of i = 1.3 is consistent with the earlier value for fillet welds with suf ficiently large leg lengths, i.e. , Cx = 1.6 tn, fr m the earlier note (11) [1980-582}, and Markl's data. Equation (12) simply provides a linear relation between the upper and lower bounds. Reference 23 (p. 56, 57) also examined Class 1 B stress indices for fillet welds. It concluded that "If the fitting material and the weld material are as strong as the pipe material, the (fillet welded, ed.] joint ... is deemed to be as - strong as the pipe." \ Thus we conclude that the [1983-W84] fillet weld SIFs including the related l footnotes and reference sketch restrictions on the variable C ,x are adequate and appropriate for.use with the [1974-574] structural design basis, t 1
5.2 Functional Capability As noted previously complete instructions for assuring the functional capability of essential safety related nuclear piping were provided by the NRC in 1978. The more recent Code revisions for fillet welds discussed in this Section do not alter those instructions. Nevertheless, if the subject were of sufficient interest, a case could be made to the NRC to permit use of the (1983-W84] SIFs for fillet welded joints in functional capability evaluations. It is of interest to note that currently efforts are under way in the ASME Code Committee (WGFD item No.192) to modify the Class 1 B and C stress indices to conform with the equation format of the Class 2 fillet weld SIF, and to incorpo-rate Code Case N-316, into the Code. 5.3 Related Requirements We have examined all relevant Articles of the Code and have compareo the contents of [1984-W84] with those of (1974-574] to determine whether any changes, other than those discussed above, would impact the SIF calculations or the piping stress anlaysis procedures for circumferential fillet welds. None were found. In particular, although revisions have been made to Par.agraph NC-4427 "Shape and Size of Fillet Welds", the material relevant to the piping stress analysis is fully incorporated into Fig. NC-4427-1 and note (11) of Fig. NC-3673.2-(b)-1. That material was discussed. We therefore conclude
- that all related requirements are satisfied as required by NA-1140(f) [1974-574].
l l l l t l
~
e
- 6. REC 0FDIENDATIONS This Section is a summary of the recommendations developed in this re-port relative to the use of the (1983-W84] SIFs in the Stone and Vehster piping stress analysis effort for Comanche Peak Steam Electric Station, Units 1 and 2, as requested in Refs. 2 and 24. Recommendatidhs for each of the three types of components that we examined are summarized below.
6.1 Branch Connections The [1983-W84] Code involves 5 revisions potentially significant to the calculation of branch connection SIFs for use with the I1974-S74] piping stress analysis basis. These are (see Figs. 4 and 5): (1) [1977-579] when Eq. (2) was added for checking run-end moment loads. (2) [1980-S60] when note 6(d), Fig. NC-3673.2(b)-1, Fig. 4 herein, was revised to exempt branch pipe sizes less than 4-in. NPS from the inside cor-ner radius requirements. (3) (1980-5S0] when note 6(d), Fig NC-3673.2(b)-1, was added to exempt branch connections from the outside corner radius requirement provided that the SIFs are multiplied by an additional factor of 2.0, with a minimum value of i = 2.1. (4) [1980-S82] when General Note 2 was added to Fig. NC-3673.2(b)-2, Fig. 5 herein, to redefine the midwall radius r' of the branch, subject to
'estrictions r on the nozzle-wall reinforcement length, and (5) [1980-582] when a minor editorial correlation of no technical con-sequence was made to note 6(c), Fig. NC-3673.2(b)-1.
All these revisions, except No. 4, have been shown to be either not less restrive or technically justified for use with the [1974-574] design
4 basis. As an alternate to using revision (4), which is technically in error, we recommend that the piping branch connections be analyzed according to [1974-574' vith revisions (1), (2), (3), and (5). Reanalyze only those branch connections that do not meet the (1974-S74] acceptance criteria using the later [1980-W81] criteria, i.e., the B stress index formulation for primary loads, but without using either revision (4) or the revision (3) penalty factor of 2.0. This procedure is justified by fatigue test data not previously available. 6.2 Girth Butt Velds The [1933-W84] Code involves 4 revisions potentially significant to the calculation of girth butt weld SIFs for use with the (1974-574] piping stress analysis basis. These are: (1) [1983-W84] when the restriction on mismatch 5 was dropped for piping with nominal wall thickness t > 0.237 in, so that the SIF for girth butt n welded joints for such pipe is i = 1.0. (2) [1983-W84] when Eq. (5) was added for. calculating the SIF for,. girth butt welded. Joints..in pipiog with.t is.0.237. n (3) [1983-W84) when note (1), Fig. NC-3673.2(b)-1 was revised to permit the use of mismatch 6 < 1/32 in for the SIF calculation if the smaller mis -
~'
match is specified for construction. l (4) [1980-S81] when a restriction on radial weld shrinkage a was added to NB-3683.4(b) that prohibited use of the Class 1 stress indices as given in the Code when r">jD.75 t n
- t I
l i i I l E . ,. . _ _ . . . _ . . _ , _ . _ . _ __, - . _ _ _ . . _ _ _ . _ _ _ - _ _ _ _ _ . _ , . _
.. s Revisions (1), (2), and (3) have been shown to be either not less restrictive or technically justified for use with the [1974-S74 } design basis. Revision (4), concerning radial weld shrinkage 4, was examined because of concern that the less restrictive weld shrinkage criterion of 4/t 5 n
0.5 specified for construction of the Comanche Peak Class 2 and 3 It piping might adversely ef fect the stress analysis acceptance criteria. was determined that the Code revision applies only to Class 1 piping. It was also determined that for Class 2 and 3 piping, radial weld shrinkage A/t n1 0.5 need not be considered either in the functional capability evaluations or in the structural evaluations provided that the number of loading cycles is less than 7,000. For more than 7,000 cycles, it is recommended that the structural evaluations but not the functional capability evaluations include consideration of radial weld shrinkage 0.25 1 a/t n 10.5 oy applica-tion of Eq. (7), 6.3 Circumferential Fillet Welds The [1983-W84) Code involves 4 revisions potentially significant to the calculationofcircumferentialfilleIweldSIFsforusewiththd[1974-S74] piping stress analysis basis. Those are: (1) [1980-S83] when Eq. (12) written in terrrs of the fillet weld leg length was added for calculating the SIF for circumferential fillet velded joints and for socket welded joints. (2) [1980-S83} when note (11), Fig. NC-3673.2(b)-1, was revised to define C as the shorter leg length for fillet welds with unequal legs. X (3) [1980-S83) when the reference SIF sketch was changed to Fig. NC-4427-1. (4) [1980-S83] when Fig. NC-4427-1 was revised to include a corrected definition for C,.
.s,
- s. ..
N All four of these revisions have been shown to be either not less J t restrictive or technically justified for use with the [1974-S74] Jesign ' basis. It is also noted that the revisions could be justified for use - in functional capability. evaluations. Approval by the NRC, however, s would be needed. '- s 6.4 Related Requirements -s We have also examined all relevant Articles of the Code and have i determined by comparison of [1983-W84[ with (1974-S74] that no revisions,
\
other than those discussed herein, would adversely impact the.use of the subject S1Fs in the Stone and Webster, Comanche Peak, piping stress analysis efforts. We therefore conclude N that all related requirements are satisfied , 3 asrequiredbyNA-1140(f)[1\74-574]. l, , s 6 O l s\
's .,
I
\
a i w 3 w! .1 . s W
. . \
yf
- q. '
REFERENCES
- 1. ASME Boiler and Pressure Vessel Code, Section III, Div. 1, Nuclear Power Plant Components, ASME, New York.
- 2. Letter, A. W. Chan, Stone and Webster Engineering Corp, to S. E. Moore, Engineering Consultant, 1983 SIFs for CSPES, Comanche Peak Steam Electric.
Station, Units 1 and 2, Texas Utilities Generating r.;., J.0, Nos. 15454.05 and 15616.05, CH1-CPO-389, dated August 23, 1956.
- 3. USAS B31.1.0-1967, "USA Standard Code for Pressure n ping, Power Piping,"
ASME, New York, 1967. 4 A. R. C. Mark 1, "Fatigue Testing of Welding Elbows and Comparable Double-
, Mitre Bends, " Trans. ASME, Vol. 68, No. 8, 1947.
- 5. A. R. C. Mark 1, "Fatigue Tests of Piping Components," Trans. ASME, Vol. 74, No. 3, 1951.
- 6. A. R. C. Markl. "Piping-Flexibility Analysis," Trans. ASME, February 1955.
' 7. A. R. C. Marki and H. H. Garge, "Fatigue Tests on Flanged Assemblies,"
! Trans. ASME, January 1950. ,
- 3. E. C. Rodabaugh and H. H. George, "Effect of Internal Pressure on Flexi-
) bility and Stress Intensification Factors of Carved Pipe or Velding Elbows,"
Trans. ASME, May 1957.
- 9. USAS B31.7-1969, "USA Standard Code for Pressure Piping, Nuclear Power Piping," ASME, New York, 1969.
- 10. Private communication with W. Evans, Stone and Webster Engineering Corp.,
Sept. 17, 1986.
- 11. E. C. Rodabaugh and S. E. Moore, "Evaluation of the Plastic Characteris-tics of Piping Pr'oducts in Relation to ASME Code Criteria," NUREG/CR-0261, s O?3L/Sub-2913/8, Oak Ridge National Laboratory, July 1978.
i 12. S. E. Moore and E. C. Rodabaugh, "Background for Changes in the 1981 Edition of the ASME Nuclear Power Plant Components Code for Controlling Pricary Loads in Piping Systems," J. Press. 7essel Tech. , ASME Trans. 104: 351-61 (November 1982).
- 13. "Functional Capability Criteria for Essential Hirk II Piping," NED0-21985, 7SNED174, General Electric C< . , September 1978.
~
l N' *
; y e, a n --m + - ~ . - - ,.,- 2.n p X .
y
' - 1
,. e ,
' (i ,, 1
,h31- 3
@ , - / r , ( ,q l f, ' j 14 USNRC Memorandum, J. P. Knight, Assistant Director for Components and' l Structures Engineering, Division of Engineering to R. L.sTedesco, .
/ ' ,- Assistant Director for Licensing, Division of Licensing,'"Evaluation of
/ 3 Topical Report 4 Piping Functional Capability Criteria," dated July 17, s
,/ >? ' 1980, c i
- 15. E. C. Rodabaugh, "Stress Indices for Small-Branch Connections with External '
/
, , Loadings," ORNL/TM-3014. oar Ridge National Laboratory, August 1970.
16. P. P. Bijlaard, "Stresses from Lot'al Loadings in Cylindrical Pressure .
Vessels," Trans. ASME, Auguat 1955. I
- 17. E. C. . Rodabaugt and S. E. Moore, "Strera Indices and Flexibtlity Factors
'I for Nozzles in Pressure Mes'dels and Piping,",NUREG/CR-0778, ORNL/Sub-2913/10, Oak Ridge National Laboratory, June 1979. ,,
l\
- 18. . J. W. Bryson, W. G. Johnson, and B. R. Bas.N "Stresses in Reinforced c.
J ] Nozzle / Cylinder Attachments Under External Loadings Analyzed by the
/?initeElementMethod-AParameterStudy,'NUREG/CR-0506QRNL/NUREG-5}.
Oak Ridge National Laboratory. August 1979 , 19. ,'J. M. Corum et al. , "Theoretical and Experimental-Stress Akalhsis of
- i ' OENL Thin Shell Cylinder-to-Cylinder Model No.1," ORNL-4553, Oak Ridge National Laboratory, October 1972. ,
- 20. R. C. Gwahney fet al,, "Theoretical and Experimintal Ctress Analysis cf ORNL Thin-Shell Cylinder-to-Cylinder Model 3," ORN -5020, Oak Ridge Nati:nal Laboratory,< r June 1975.
- 21. R. C. Gwaltney et al. f"Iheoretical and Experimental Strass d.alvsir of l ORNL Thin Shell Cylinder-to'-Cylinder Model 4," ORNL-5019, Oak'Rfdge
- National Laboratory, Jury 1975. q
- 22. E. C. Rodabaugh "Accuracy of Stfess Intensification Factors for Branch ^
Connections," Draft WRC qulletin, May W M
- ),
l I 23 .' E. C. Rodabaugh and S. E. Moore, "Stress Indices for Girth Welded Joints, Including Radial Weld Shrinkage, Mismatch, and Tapered-Wall Transitions,"
, NUREG/CR-0371, ORNL/Sub-3913/9, Oak Ridge Natonal Laboratory, September s '1978.
;f .
24 Pri.vate communication with A. J. Cokonis, Stone and Webster Engineering l~ ; .f Orp.:c.oncerning Brown and Root Pipe Fabrication and Equipment Instal-l t lation Instruction No. QI-QAP-11.1-26, Rev. 19, on Nov. 18, 1996.
' 25. / E. C. Rodabaugn. "Compari. son of ASME Code Fatigut Evaluation Methods I
! for Nuclear Class 1 Piping with Class 2 or 3 Piping," NUREG/CR-3243, 1 ORNL/Sub/82-22252/1, Oak Ridge Nation'*1 ' Laboratory, June 1983. t
)
l - <
, /
i t t
- 26. Minutes of Sept. 13, 1982. Working Group on Piping Design (SGD) (SCIII).
ASME Boiler and Pressure Vessel Committee.
- 27. Case N-316 Alternate Rules for Fillet Veld Dimensions for Socket Welded Fittings Sect. III, Div. 1, Class 1, 2, and 3. Cases of the ASME Boiler and Pressure Vessel Code, Dec. 11, 1981.
Y wb
s* Table 1. Comparison of run-end SIFs with stress analysis data from Ref. 17 Model 21 2i D b R,/T r ry/R, y/T r #$# p max a Eq. (2) Eq.(3) 22E UA 50.5 0.50 0.50 0.990 4.62 l'4.35 5.46 5.33 UB 40.5 0.50 0.50 0.988 4.54 12.36 4.72 5.05 UC 20.5 0.50 0.50 0.976 4.09 7.75 3.00* 4.26 UD 10.5 0.50 0.50 0.955 3.60 4.86 3.00* 3.60 UE 5.5 0.50 0.50 0.917 3.17 3.03 3.00* 3.06* UF 5.5 0.08 0.08 0.917 3.11 2.00* 3.00* 3.06* SIA 50.5 0.50 0.50 0.861 2.98 12.48 5.46 3.11 SIB 40.5 0.50 0.50 0.843 2.87 10.54 4.72 3.00 SIC 20.5 0.50 0.50 0.780 2.43 6.20 3.00 2.69 SID 10.5 0.50 0.50 0.705 2.11 3.59 3.00 2.65 SIE 5.5 0.50 0.50 0.623 2.07 2.06 3.00 2.65 l Sif 20.5 0.32 0.32 0.732 2.36 2.98 3.00 2.65 S1G 10.5 0.32 0.32 0.649 2.23 2.00* 3.00 2.65 S1H 5.5 0.32 0.32 0.563 2.08 2.00* 3.00 2.65 SII 20.5 0.16 0.16 0.646 2.49 2.00* 3.00 2.65 S1J 10.5 0.16 0.16 0.555 2.38 2.00* 3.00 2,65 S1K 5.5 0.16 0.16 0.468 2.33 2.00* 3.00 2.65 S1L 20.5 0.00 0.08 0.551 2.52 2.00* 3.00 2.65 SIM 10.5 0.08 0.08 0.459 2.61 2.00* 3.00 2.65 SIN 5.5 0.0S 0.08 0.391 2.63 2.00* 3.00 2.65 P30A 50.5 0.32 0.32 0.808 2.30 5.99 3.50 3.00 P30B 20.5 0.32 0.32 0.743 2.39 3.02 3.00 2.65 P30C 10.5 0.32 0.32 0.695 2.27 2.00* 3.00 2.65 P30D 5.5 0.32 0.32 0.659 2.05 2.00* 3.00 2.65 P30E 5.5 0.08 0.08 0 J56 2.68 2.00* 3.00- 2.65 Veldolet 12.25 0.35 - - 2.15 - 3.00 2.65 OKNL-1 49.5 0.50 0.50 0.990 5.00 14.16 5.39 5.30 ORNL-3 24.5 0.115 0.84 0.870 3.20 6.25 9.06 2.70 ORNL-4 24.5 0.125 0.32 0.950 4.00 2.72* 3.45* 3.57* is the maximum stress intensity normalized to M/Z. b Values marked with an asterisk (*) in this column are less than the corresponding value of IT,,,. l .Z l l I
O. .. o Table 2. Comparison of branch-end SIFs with stress analysis data from Ref. 17a b Model R,/T r'/R, Tf/T r #'/I p max 21 UA 50.5 0.50 0.50 0.990 16.60 14.35' UB 40.5 0.50 0.50 0.988 15.05 12.36* UC 20.5 0.50 0.50 0.976 10.85 7.75* UD 10.5 0.50 0.50 0.955 5.77 4.86* UE 5.5 0.50 0.50 0.917 3.50 3.03* UF 5.5 0.08 0.08 0.917 1.36 (2.00) SIA 50.5 0.50 0.50 0.861 11.07 12.48 S1B 40.5 0.50 0.50 0.843 9.84 10.54 SIC 20.5 0.50 0.50 0.780 5.64 6.20 SID 10.5 0.50 0.50 0.705 2.81 3.59 SIE 5.5 0.50 0.50 0.623 1.56 2.06 S1F 20.5 0.32 0.32 0.732 2.56 2.98 SIG 10.5 0.32 0.32 0.649 1.43 (2.00) S1H 5.5 0.32 0.32 0.563 1.39 (2.00) S1I 20.5 0.16 0.16 0.646 1.22 (2.00) S1J 10.5 0.16 0.16 0.555 1.26 (2.00) S1K 5.5 0.16 0.16 0.468 1.33 (2.00) SIL 20.5 0.08 0.08 0.551 1.22 (2.00) SIM 10.5 0.08 0.08 0.459 1.21 (2.00) SIN 5.5 0.08 0.08 0.391 1.21 (2.00) P30A 50.5 0.32 0.32 0.808 3.73 5.99 P30B 20.5 0.32 0.32 0.743 1.84 3.02 P30C 10.5 0.32 0.32 0.695 1.39 (2.00) P30D 5.5 0.32 0.32 0.659 1.35 (2.00) P30E 5.5 0.08 0.08 0.556 1.20 (2.00) a See Table 10 of Ref. 17. b 21 > 2.C from Eq. (2). An asterisk (*) indi-cates that the calculated SIF is less than the 3' datum. Forentheses () around the number indicate!* that the calculated SIF is less than the permitted minimum.
t . ARTICLU NC.3000 . )EstGN Fig. NC 367).2(bl.( e F lo s eu.lify Secess anienget. cat.on Dem ui.on F ac tee. k F c ioe. . s6eich f gfr* o eeach <oaneci.on 161 1 15(64,,, '/. r'm*/,h',,g , m) - F eg. N o.36 73.2(ul-2 suit wed til 1 1.0 to > 3/ t 6 a nd [, 4 0.1 l \ / .a ~s~ ^ lt i n w l Butt weld (11 1.0 for fivia we'd r f 4 f
- 1.8 for as weided it. A 3/16 orto > 0.1 F eliet welded joint, socket Fig. NO.3673.2(bl 3 wedded flange, or singie 1 2.1 sk.'tches (a),(bl. (c),
weeded sho on fienge (el
- so (Il
%. NOW3.2(bb3.
Fo se f.siea wed i 1.3 ske tch (d) NOTES: (1) The following nomenctatwee apol.es. r o mean radeus of pepe, inches Imatch ng pioe ior sees and coowsl. t, e nomenal weil th. cham of o.oe. inches (maicheng p oe for tees and ebows. see note (91). R
- bend eadews of e60ow or pipe or so. enches.
#
- one half angle between ediacent rmter ases, s a meter soacing at cenier I ne, enenes.
f, e te nf orced th ck ness, enches. 4
- mesmatch. .r.ches.
O, e outs.de diameter, enches. (6) The edtion aophes only of the ionoeg cond.tions are met: (a) The re nlorcernent een reoverements of NO.3643 are met. (b) The amis of the branch pipe is normal to the martace of run poo wl. (c) For branch connections in a poe, the arc distence measured between the centers of ediacent branches along the surf ace of the l run pipe is not less than three times the surn of the.c inside rad i in the longitudinal direct.on or es not less than two times the 7 sum of their radie *long the circumf erence of the run pipe. l (d) The ens de cornet red us r, (F sG.NO 3673.2(bl.2)is berwan 10% and 50% of r,. (e) The outer rad.us.c ,is not less than the targer ofer /2.1Te
- yl/2 (FIG.NO 3673.2(bl.2 sketch (cIl or T,/2.
(f) The outer redeus . . es not less tasa the larger of (tl 0.002# do (2) 2 (sen fi' teen t%e offset for the configurations shown en Figs. NO.3673.2(bl.2 sketches Ni and (bl. (g) Mm/Tr4 50 end r% :stm 40.5. t l t . FIG. NC 36tl.2(b).1 FLEX (BILITY AND STRESS INTENSIFICATION FACTORS I \ . ( Figure 1. 1 l l e: l l l l l l l t
o e ARTICLE NC 3000 1)ESIGN Fig. .NC 3673 21b12 e
-> Ib (-- T '
b BRANCH PIPE
-* (--
- 4--Ib do NE# -r 3 d h
r3 r
/ en 4 45' ,*
(
, s c' N
( ~
-en:90' 3 r
- OFFSET Yo y ? - OFFSET Tb ) % <2 i J b'2
$ hk A T, t 7 f, j_ LN L'i's Li" t 1 2 (a) (b)
BRANCH PIPE
-+ (- Ib -+ (I:b b BRANCH PIPE d0 g T:b T'b + 0.667 y do b t{ en 4 45' T
th ( q y fBRANCH r, f k s
&pn F
'~ (
, x\ -} 5 /2 kg_),- '
i! . Rm - Rm _- (c) (d)
, FIG. NC 3673.2(bb2 BRANCH DIMENSIONS NOTES:
r'm = mean esaius of branen pepe enches T, a nominal th.ekness og run o.pe inchen s r's = nominai in.cknen of branch p.oes. inches a, a ouis.oe o.ameier of branch p.pe. inchen Am = mean iad.us of run p os, enches Tc.8.r .r .'s.ro andra'eoeha* din'h'*fW'. e l35
Fig. NC 3673 2(b) 3 SECTION lli. DIVISION I - SUBSECTION NC SURFACE OF
-[ VERTICAL MEMBERS SURFACE OF 9 d4 / HORIZONTAL MEMBERS \ &_
\ l J 7 THEORETICAL THROAT #
(Al CONVEX EQUAL LEO FlLLET WELO 18) CONCAVE EQUAL LEO FILLET WELO NOTE: The "sire" of an equalleg finet weed es the length of me largest inscribed right nosceies triente. Theoretcal throat 0.7 x sae. SURFACE OF VERTICAL MEMBERS 4 y$o SURFACE OF db [ HORIZONTAL MEMBERS \ c- - THEORETICAL THROAT - (CI CONVEX UNEQUAL LEG FILLET WELD (D) CONCAVE UNEQUAL LEC FILLET WELO Note: For uneovat leg f dlet welds, the "sae** of the we4C es the leg lengths of the largest right trengte which can be specribed within the idlet weed cross section. Q 7, , -
-+ (--X
/ rX g
_s f Cn A AXXAx w i
] i Pl/16" APPROX.
BEFORE WELDING (E) SOCKET WELDING FLANGE l rna nommal o.oe wail thickness a m.n.
- 1.4 faor th.ckness of the hwD. whichever es smaller but not less ian 1/8" t nNOMINAL PIPE W/4LL THICKNESS X-E t/16" A PPROXIM ATE'.Y -
T ] J8EFORE WEL0 LNG 7'7
\
(F) SOCKET WELOING FITTING
~
s men = 1.09 ta but not seis tnan Ita* FIG. NC 3673-2(bi 3 FILLET AND SOCKET WELD DIMENSIONS (not permitted for connections over 2 inch nominal pipe size) 136 Figure 3.
&$l
- c. .
19tL3 Edition W84 NC4000 - DESIGN Fig. NC M73.24b).1 F ie s .b.s.ry Suess inteasif.caic
' Desce.ptu Factor a F actor , Sketch
,o, .. mne~
z..w.).Te
'5( F T @ (t) m .,m ac - -,.ent~o,ei.n t ,o.e-s ,rwne~,
z - . i A. ) T.
,m.un m r "(F(2) tnet not less than 1.5 Girth twet wow (Note (II) i 1.o t, t 0.2 3 7 sn.
1.9 enan. or wg4 Cirth bwit we6d (Note (1)] g o 9(1 + 34/t.) t c 0.23 7 in. but not less than 1.0 Circurnferential helet weideo c.
" f {
* 'd'd N"
g 2. I tic 11. ) F4 NC44271 ga, ice,, gg.il. (c.2)'
$$0 bwt not less than 1.3 and (c.31 stued point i 2.1 Fig. NC4511 1
- NOTES TO FIG. NC.3673.2(bh1:
W84 (1) The following nomenclature applies. t = mean radius of pipe,in. (matching pipe for tees and elbows) t, = nominal wall thickness of pipe, en. (matching pipe for tees and elbows, see Note (9)) R = tend radius of elbow or pipe bend,in. S4. # = ondalf angle between adjacent mater axes, deg. s = miter spacing at center line,in. t, = reinforced thickness,in. 8 = everage permissible mismatch at girth butt welds as shown in Fig. NC-4 2331. A value of 8 less than '/ in. may be used provided the smetter mismatch is specified for fabncation. For "flush" welds. as defined in Fig. NB.3683.1(cF1. 4 may be taken as rero. i = 1.0, and flush welds need not be ground. O, = outside diameter. in. (6) The equation applies onhr if the follownng conditions are met: (a) The reinforcement eres r.quirements of NC.3643 are met. . (b) The axis of the branch pipe is normal to the surf ace of run pipe wett.
$43 (c) For branch connections in a pipe, the arc distance measured between the centers of ediacent branches along the surf ace of the run pipe as not less than three times the sum of their inside rodii in the longrtudinal direction or not less than two times the sum of their inside radii along the circumference of the run pipe, go (d) The ineide corner radius r, (Fig. NC.3673.2(b).2) for nominel branch pipe site greater than 4 in, shall be between 10% and 50% T,. The radiue r, is not required for nominal branch pipe site smaller than 4 its.
(e) The outer redeus re is no less then the larger of T 12. (T. + vll2 (Fig. NC.3673 2(b).2 sketch (c)) or T,I2. (f) The outer radius r,is not less than the 'erger of (1) 0.002# cr. , (2) 2 (sin #)* times the offset for the configurations shown in Figs NC.3673.2(bh2 sketches (a) and (b). (g) A.lT, s 50 and r'.lA. s 0.5. N yo th) The outer radius te is not required provided en additional multipher of 2.0 is included in the equations for branch and and run end stress intensificat.on factors, in this case, the calculated value ofi for the brenets or run snail not be lese than 2.1.
$43 (11) C,is the fillet weld length. F'or unequel leg lengths, use the smaller tog length for C,.
, Figure 4.
hc ? I
e, a s
, 190 f.dition NC 3000 - DESIGN Fig. NC.367J.2(hi 2 o, T, .+ -o, T, etenen p.pe
-o. e. Yf +. +T c i i r d, ~ - 3 g a r3 l, 0 , < 45 oeg. (
f 'I" ' g .5(2
; ' N
[ g'\ h 0,,.Soe,,. T. ( ; .. 7 - offui -# c
,- of f sei q
To ' It
= , ,9 = , pp 2
s t 7' rr_ e ._t r'
. &\T ( Mad Rm l
To a{ R, T, a{ (si 2 tel 2 Branen pipe
+4 + Tf -o,.
- e. Tf = Te Stanch pepe o 3 x u #8 rf,, To
- I'b + 0.667 y 6,, < 4 5' T*g r p y
- ? W,
'l 8,enen /
2 w 6,, ' 3 N' , ,m _
\ T,yy -.
,y
[ 'I ,p Ty ,
^
, u f h
', ( kSs@\h " r- 'Id'h i' Wy& h f n .
*= G -
2 nm 2-2 Ici . (4)
- v. e owes 4e o.emeter os beenen poe. a CENERAL NOTES:
- r.
- mean red ws of beench ppe en. (Il I. 8. F.. fe. to. t, e# 3 y are genned in th.s 6gwte l'. e nomenal thechness of beencM peces. A (.'l if 1, ogwass or esceeds 0 5 % r, f. then t'. can t>e toten $32 A. = mean red.ws of twa poe, a as the rad.ws to the cmtet of I..
f, e nom.cel tm.ctness of twa p.pe en. FIG. NC 3673.2(b) 2 BRANCH OIMENSIONS Fig _ure 5. 169 7 h I
e
- e. - .
O 1983 Edition NC 4000 - FABRICATION AND INSTALLATION Hg. ACw27 3 4 , Theoret, cal inroat
, Theo,ei c # ihreat Surface of venecad memtwr
' _ s ,f ace of vert.ca4 memtwe
' - \
/
Conves fillet wetd ' S.se of wed s Cortave fdiet weed s y Surf ace of g {_ ' horizontal member l {
--. -- Site of weed NOTE The sete of en eoualleg fillet meid is the leg length of the largest inscribed right isosceles triangie.
Theoretical throet = 0.7 a size of wedd. (al Eeued Leg Feltet Wold [ p g Theoretical throat Surface of vertical member s ace of un.cW mer r
/
/'sg Conven f.iiei weid ( Concave fdlet weld N % '
s N Surface of horizontal member NOTE: The s.se of an unecas leg idiet need is tne shorter leg length of the largest right triangle wnich can be inscribed withen the f llet wed cross section. S$d (b) Unequal Log Fillet Weld
* --. - a min.
~*T4 e a men. _, a m,n, 7- a m.n.
- , .n. a m.6.
, dhN Y : /)hk ' '" '
1 l k\kkkM Y\YS\\YN MhM\\\M\M f
- r, or % in, whichever is erwier ht oro 4 in.ewharf> gr is smeder If -
- 1/16 in. aooros.
f Front and Back Weed - Fase and Back Weed (e.1) Slip-en Fiange (c.2) Socket Wetding Flange
--C# '- ' 'n nom nal p.pe I I west th cknese NOTE 3: a men.= 1.4in or th'ckness of the hub, dichever is smallere but not less than 1/8 in.
C# 'I'*I"'***'** C, men. = 1.09r, where t, = roninas pepe wall th.eknew 5T before vuesdeng It) Menemwm Wesdeng D monseons for Slepen 48uf
/
Seckee We4deng F1anges and Seeket We6deng Fsttangs (e 3) Sectet Weedeng Fittengs l ( FIG. NC44271 FILLET AND SOCKET WELO DETAILS AND DIMENSIONS Figure 6e r;e 6 23I
o y ~o Fig. N B.4 23 31 SECTION lit. DIVislON I - SUBSECTION NB w 11/2 ' - 450 M A XIMU M
' ; p WAX 2AM /
SLOPE f
/
-= .,
1/32 in. M AXIMUM UNIF ORM Mi!M ATCH BQRE DI AMETER 11/32 in. AROUNO JOINT PIPE ( , , _,
+ +-t OR 1/4 in. (LESSER)
/// S o I ALTERNATE F COUNTEft90RE
' 'o ($ NOT USED HiCKNEss. '-
IN. I j (c)
,j CONCENTRIC CENTERLlHES I
1 / l 3/32 6n.M AXIMUM AT ANY ONE POINT AROUND THE COMPONENT ( JOINT - J ! (b) OFFSET CENTERLINE 5 NOTE THE COMBINED INTERN AL AND EXTERNAL TR ANSITION OF THtCKNESS SHALL NOT EXCEED AN INCLUDED AN OF 30* AT ANY POINT WITHIN 1% i OF THE LAND. FIG. NB423$1 BUTTWELD ALIGNMENT TOLERANCES AND ACCEPTABLE SLOPES FOP. UNEQUAL 1.D. AND O.D.WHEN INSIDE SURF ACE IS I INACCESSIBLE FOR WELDING OR FAIRING Figure 7. 166 7.' ej
, 'I
!{. 79
. i e ,
/
6 -
/t = 20 5 -
40 - 80 4 - Q 00 Sb + 0 3 - l t
/
2 l I I l I~ i f i i o 2.o o o.5 i.o 1. 5 6/t FIGURE 8. OUTSIDE SURFACE AX1AL STRESSES, GIRTH BUTT WELD WITH RADIAL WELD SilRINKAGE, MOMENT LOADING.
.#.). f
s
.. ..D o
O CANL-owG 83-4597 ETO 3 l 1 3 i l l l g EO (9). CYCLES-TO-F ALLURE E0 (9),WITH F ACTOR-OF-
, 100 - 7 SAFETY OF 2.0 0N STRESS -
a d g 50 - Sg=0 -
/ Ss = Sn %
G - ~ { [EO. (6), CODE 2 STRESS R ANGE LIMITS . 10 - _ e FOR A106 GRADE 8 UP TO 6500F E - Se Sn 15 kss PER B31.1 5 , g I l I l t l t l t 10 102 103 toe tot 106 l CYCLES Fig. 9. Comperison of Eq. (9) with Code 2 allowable stresses for SA106 Grade B material. l l Fq, f -
-}}