This action might not be possible to undo. Are you sure you want to continue?

ORNL/Sub/2913-3

(NRC-5)

P.R. 115-7b

3

M45b

ODIATVO

E

EVALUATION OF THE BOLTING AND FLANGES OF ANSI B16.5 FLANGED JOINTS — ASME PART A DESIGN RULES

<V

E. C. Rodabaugh

S. E. Moore

Work funded by the Nuclear Regulatory Commission

Under Interagency Agreement 40-495-75

September 30, 1976

OAK RIDGE NATIONAL LABORATORY

**CENTRAL RESEARCH LIBRARY
**

CIRCULATION SECTION 4500N ROOM 175

**LIBRARY LOAN COPY
**

DO NOT TRANSFER TO ANOTHER PERSON.

if you wish someone else to see this report, send in name with report and the library will arrange a loan.

ORNL-118 (6-97)

Work Performed by

BATTELLE

Columbus Laboratories

**505 King Avenue
**

Columbus, Ohio 43201

for OAK RIDGE NATIONAL LABORATORY

**Oak Ridge, Tennessee 37830 operated by
**

UNION CARBIDE CORPORATION

for the

ENERGY RESEARCH AND DEVELOPMENT ADMINISTRATION Contract No. W-7405-26

t

**Printed in the United States of America
**

Available from

National Technical Information Service

U. S. Department of Commerce

Price:

5285 Port Royal Road Springfield, Virginia 22151 Printed Copy $7.50; Microfiche $2.25

This report was prepared as an account of work sponsored by the United States Government. Neither the United States nor the Energy Re search and Development Administration, nor the Nuclear Regulatory Commission, nor any of their employees, nor any of their contractors, subcontractors, or their employees, makes any warranty, express or im plied, or assumes any legal liability or responsibility for the accuracy, completeness of usefulness of any information, apparatus, product or process disclosed, or represents that its use would not infringe privately

owned rights.

**ORNL/Sub/2913-3
**

NRC-5 P.R. 115-7b

**EVALUATION OF THE BOLTING AND FLANGES OF ANSI B16.5 FLANGED JOINTS
**

ASME PART A DESIGN RULES

by , C. Rodabaugh

and

S.

E.

Moore

Manuscript Completed: August 20, 1976 Date Published: September, 1976

NOTICE This document contains information of a preliminary nature and was prepared primarily for internal use at the

Oak Ridge National Laboratory. It is subject to revision or correction and therefore does not represent a final report

OAK RIDGE NATIONAL LABORATORY LIBRARY

3

i4M5b

0Q1AT70

2

48 51 Basic Theory Flange Stresses 52 Flange Stresses at Preload Bolt Stress of 40. 54 . FLANGE STRESSES 40 41 Flange Stress Limits of NB-3647 Stresses of Rated Pressures. 3... Sizes. CODE REQUIREMENTS FOR FLANGED JOINTS SPECIFIC FLANGED JOINT CALCULATIONS 7 11 Classes.. BOLT AREAS AND BOLT STRESSES 14 Required Bolt Areas Per NB-3647.000 psi. . Zero Pipe Bending Moment.3 Fatigue Analysis of Bolts 37 Summary of Results on Bolt Areas and Bolt Stresses 5. 41 42 Allowable Pressures. Types H Pressure-Temperature Ratings Dimensions 12 12 4. .. Zero Pipe Bending Moment Allowable Pipe Bending Stresses at Rated Pressure .1(b) 14 Bolting Requirements of NB-3230 NB-3232. INTRODUCTION Nomenclature 1 3 2.1 Average Stress NB-3732.2 Maximum Stress 18 18 30 NB-3232.11 TABLE OF CONTENTS Page FOREWORD iv 1.

"DEVELOPMENT OF SIMPLE EQUATIONS FOR LIMITATION OF PIPE BENDING AND TORSIONAL MOMENTS APPENDIX C.iii TABLE OF CONTENTS (Continued) Page Stresses at Preload (S = 40.000) Plus Operating Conditions Effect of Pipe Bending Moment on Flange Stresses 63 72 Summary of Results on Flange Stresses 6. "RECOMMENDED CODE REVISIONS" . REFERENCES 91 APPENDIX A. 74 76 90 RECOMMENDATIONS ACKNOWLEDGEMENT 8. "TEST DATA ON FLANGED JOINTS" APPENDIX B. 7.

E. ORNL. S. is the program manager. C. Flanged Joints with Contact Outside the Bolt Circle — ASME Part B Design Rules. E. The study of ANSI B16. Rodabaugh and S. Moore. E. Other reports in this series are: 1. K. U. E. F. A Computer Program for the Analysis of Flanged Joints with Ring-Type Gaskets. and S. Reactor Division. Activities under the Design-Criteria program are coordinated with other safety-related piping and pressure-vessel research through the Design Division of the Pressure Vessel Research Committee (PVRC) of the Welding Research Council and through the ASME Boiler and Pressure Vessel Code Committee as part of a cooperative effort with industry to confirm and/or improve the adequacy of design criteria and analytical methods used to assure the safe design of nuclear power plants. Rodabaugh.5 flanged piping joints. ORNL/Sub/2913. Jr. M. 2. . Results from these studies will be used by appropriate ASME Code groups in drafting new and improved design rules. 1976).IV FOREWORD The work reported here was performed at the Oak Ridge National Laboratory and at Battelle-Columbus Laboratories under Union Carbide Corporation. 1 (May. 2913 in support of the ORNL Design Criteria for Piping and Nozzles Program. Lynn of the Metallurgy and Materials Branch is the cognizant RSR engineer. Moore.. and S. This program is sponsored by the Division of Reactor Safety Research. Office of Nuclear Regulatory Research. E. C. Nuclear Regulatory Commission. conducted using the ASME Part-A design rules described in this report. Nuclear Division. is a portion of a three- part study of flanged joints for the evaluation of design rules in current codes and standards. Subcontract No. 1976). E. O'Hara. FLANGE: ORNL-5055 (January. Report No. Moore.

in which maximum stress limits are imposed. P This was incorporated into the Code as Case 1677." Another problem involved the use of the equation con . by reference to NB-3647. Perhaps because of the unusual characteristics of a flanged joint. that P rather than P eq be used in calculating H . is somewhat difficult to intrepret when applied to a bolted flanged joint. * Code.1. was made in 1972 by one of the authors. This has been corrected by copying the text of NB-3647 into NC-3647 and into ND-3647. but leaving out the sentence "These minimum bolt areas may have to be increased to meet requirements of NB-3230.e. P It was noted that using P for checking the capacity of B16. several major problems in the Code rules were pointed out con cerning flanged joints. in this report. The criterion defining failure is Such leakage can often by stopped usually excessive leakage at the gasket. the major stresses occur due to tightening the bolts and. in effect. however.5 flanges are ordinarily used. At that time.1(a) for calculating the equivalent pressure.5 flanged joints for pipe It was recom bending-moment capacity would give very low allowable moments. Reference (1). One of these involved subarticles NC-3600 and ND-3600 which. are changed only slightly by pressure. . i. A review of the NB-3600 rules for Class 1 piping design. and "failure" occurs when the pressure boundary is penetrated by a crack. the problem is solved by imposing still higher stresses on the bolts and flanges. by additional tightening of the bolts. as a partial solution. eq tained in NB-3647. Section III. INTRODUCTION The bolted flanged joint represents perhaps the most complicated structure for which specific design rules are given in the Code (1)* For most structures (pressure boundaries) covered by the Code. as they ex isted in the 1971 edition of the Code. the basic philosophy of the Code. This is particularly true for piping where ANSI B16. refers specifically to the ASME Boiler and Pressure Vessel Code. in general. Nuclear Power Plant Components. mended.. For flanged joints. the stress is proportional to the pressure. required a fatigue analysis of the bolting of flanged joints in Class 2 and 3 piping.

calculations were made for a representative sampling of B16. but not necessarily accurate. These comparisons indicate that the calculation methods used in the text of the report are conservative. Available test data on flanged joints are presented and compared with results from FLANGE calculations in Appendix A. . The rather extensive and detailed evaluations of ANSI B16. and the computer program FLANGE which was developed to implement that analysis. Specific Flanged Joint Calculations".5 flanges as described in Chapter 3.5 flanged joints in the Code rules still remain. Both the Code procedures and the theoretical analysis of Reference (3) are based on simplified analytical methods. This report was prepared to investigate the use of B16.5 flanged joints developed in Appendix B. Class-2. Calculations were made using the Code procedures and also with the more basic theoretical analysis of Reference (3). Toward this end. and Class-3 piping systems.At the present time. a number of ambiguities and unnecesary re strictions on the use of B16.5 flanged joints in Code Applications and to make recommendations for appropriate modi fications of the Code rules. These simple equations from Appendix B are then included in Appendix C in the form of recommended revisions to the Code. One objective of this report is to develop relatively simple but conservative methods of evaluating the adequacy of piping products used in Code Class-1.5 flanged joints contained in the text of this report provide the basis for the development of re latively simple equations for ANSI B16.

Nomenclature A^ = A.^. Code definition j c. or W „/S . Code definition * For users of the computer program FLANGE: ax = (Wl-W2B)/Ab for P = P a = (Wl-W2C)Ab for P = P =P a = (Wl-W2A)/Ab for (Tb ~ Tf) = 50 F a4 = (Wl-W2D)/Ab for (Th . a . see Eq. A m Bolt cross sectional area (all bolts) Bolt cross sectional area of a single bolt = = Design bolt area = larger of W . mlb m2 a * a1 . a. Code definition Bolt circle diameter D o = Pipe outside diameter D d = = Mean (midwall) pipe diameter Bolt diameter E = Modulus of elasticity E1 = E = Modulus of elasticity at bolt preload temperature Modulus of elasticity at operating temperatures F F. a„. . b = = Axial force in attached pipe (see Table 1) Axial force in bolt G = Effective gasket diameter. Flange inside diameter b C = = Effective gasket seating width.Tf) = 100 F . (9) and Tables 5 and 6. = B = Bolt change factors./S.

M ..G )/2.Nomenclature (Continued) G o = = Gasket outside diameter Gasket inside diameter G. M M = = Total moment acting upon flange. k = = Tapered hub length.5 (equal to class designation) Pea = H t Pressure equivalent of pipe bending moment. Code definition m = Gasket factor. Code definition N P = = Design cyclic life as obtained by Code procedure Internal pressure P = Primary rating pressure per ANSI B16. gasket seating conditions. Code definition Total moment acting upon flange. welding neck flanges (C . P taken from the Code. M = Moment applied by couples to flange e. Elsewhere. where G = G. it is identified as This is to distinguish it from the moments applied by couples to flange (e.)/2 g H P = Attached pipe wall thickness. NB-3647 definition..g.g. bolt-load/gasket reaction couple). bolt load/ gasket reaction couple). taken herein as 2t + d + 1/8" The symbol M for pipe moment is used only in Table 1. = 16 Mpb ^G3 ppn q = 4S pb o /Do jg Pressure equivalent of pipe bending moment as used in FLANGE = PT = Leakage pressure . Code definition h h-. 1 G G = = Gasket mean diameter = (G Effective gasket diameter + G.. welding neck flanges 2b irGmP. G or G as indicated in text me I M = = Effective bolt length.

Code definition . = Initial or preload bolt stress S. bs = Shear bolt stress S . . = Allowable flange stress sa = Allowable bolt stress at atmospheric temperature. Code definition Bending stress in bolts S. operating conditions. including bending stress Sa T T = = = Stress amplitude (Used in conjunction with Code fatigue curves) Torque applied in tightening bolts Temperature of bolts T = Temperature of flange ring T t W = = = Temperature of flange hub and pipe Flange ring thickness and blind flange thickness Total bolt load W = Design bolt load.S H K 1 = Flange stresses.S . be = = Bolt stress after application of loads Critical bolt stress S. Code definition. „ S.Nomenclature (Continued) pb> Q = S = Stress categories. M . = Allowable bolt stress at design temperature.. PD = = = Maximum stress in attached pipe due to moment. Code definition S = Allowable stress intensity. Class 1 components S .. Code definitions Stress S . Code definition sb = S. pb S S Range of primary-plus-secondary stress Range of primary-plus-secondary-plus-peak stress S^ t = Stress intensity in a bolt.

Code definition y 0 = Gasket seating load.Nomenclature (Continued) m2 Design bolt load.V . Code definition Rotation in radians Q-^ = 02 = 0 m Rotation of flange ring of one flange in joint Rotation of flange ring of other flange in joint Rotation of flange ring due to moment 0 P Rotation of flange ring due to pressure ©„ Rotation of flange ring due to thermal gradient (ThP . gasket seating conditions.

However.As written.5. NB-3612. CODE REQUIREMENTS FOR FLANGED JOINTS In the Code. pressures. Accordingly. NB-3640 contains two paragraphs which are relevant to the subject of B16. moments and thermal gradients. A seemingly appropriate way to check for "consistency" .2. (2) (3) "This subsection". by Code definition.5 flanges meet the requirements of NB-3640 (pressure design) provided the design pressure/temperature is below the rated pressure/temperature combination given in B16. one of the objectives of the recommendations of this report is to provide rules under NB-3640 and NB-3650 that parallel those for all other products. it is not clear what is meant by "the design is consistent with the design philosophy of this subsection". NB-3640 (Pressure Design) requires that the product satisfy the stated criteria for the design pressure at the design temperature. the Code requires that the design of B16.5 flanged joints: states that: These are NB-3647 and NB-3649. The rules for flanged joints in the Code do not conform to this basic organization." There are three significant points with respect to the above quote. Under NB-3650.1 provides an additional restriction in stating that the pressure-temperature ratings of the standards shall not be exceeded.e. (1) ANSI B16.. i. NB-3649 "Other piping products manufactured in accordance with the standards listed in Table NB-3691-1 shall be considered suitable for use provided the design is consistent with the design philosophy of this subsection. is the entire NB-portion of the Code.5 is included in Table NB-3691-1. NB-3649 implies that B16. the product must satisfy the stated criteria for all loads.5 flanged joints be consistent with the NB subsection design philosophy.

NB-3647. .1 apparently does not include all of the significant sources of moment. A major portion of this report is directed towards the evaluation of B16.1. e.g. in particular. relief valve thrust. NB-3230 is shown herein as Table 2. if con sidered. "Pressure Design of Flanged Joints and Blanks". (2) The moment to be used in NB-3647. "Flanged Joints". NB-3647. At yet it gives explicit rules for moments applied to the flanged joint by the attached piping.5 flanged joints conform to the requirements of NB-3647. Moment loads. anchor movements or misalignment that is "pulled-out" during make up of the joint. for all other products. This type of load is not explicitly considered for other piping products and.1(b) gives methods to establish minimum bolt areas. are considered ex plicitly only under NB-3650 on Analysis of Piping Products..1 is different from requirements for other piping products.1 is a part of NB-3640 on Pressure Design. The Code requirements for what is needed to show that a B16. this time. (4) NB-3647 does not cover blind flanges.8 is to see if B16.5 flanged joints and their pressure-temperature ratings to see if they are consistent with the requirements of NB-3647.1. The detailed interpretation of the rules is covered in the text. but then states "These minimum bolt areas may have to be increased to meet requirements of NB-3230. would presumably be done under NB-3650. we only point out that: (1) NB-3647. (3) The explicit inclusion of axial force loads in NB-3647. NB-3647 is shown herein as Table 1. First.5 flanged joint meets NB-3650 appear to be quite clear.

lb C= diameter at location of effective gasket load reaction as defined in XI-3130. Other terms are defined in XI-3130. in. (c) Equation (6) in XI-3240 for longitudinal hub stress shall be revised to include the primary axial membrane stress as follows . in. in.3. S/f not greater than 1. Peq shall be determined by the equation 16A/ +_4F eq ttG3 ttG2 where: M= bending moment applied to the joint due to weight and thermal expansion of the piping. Flanged joints shall be analyzed and the stresses evaluated by using the methods given in (d) The allowable stress limits shall be Appendix XI. revised in accordance with (a) through (d) below. lb F= axial force applied to the joint due to weight and thermal expansion of the piping. shall be replaced by a flange design pressure. The designer may usean alternative P= design pressure as defined in NB-3112.5Sm and STnot greater than 1. procedure provided that it includes the modifications of this paragraph for the loads and stress limits for Class 2 piping. This paragraph presents an acceptable method forcalculating pressure stresses where: in flanged joints but should not be used for cal culating maximum service stresses which are required to meet NB-3230. (b) Equations (3) and (4) in XI-3220 shall be used to establish minimum bolt area required using allowable stress values equal to Sm asgiven inTable I1. The equivalent pressure.1 Flanged Joints.5Sm (a) The design pressure used for the calculation of loads in a flanged joint by the equations in Xl-3220 and Xl-3230. These minimum bolt areas may have to be increased to meet requirements of NB-3230.1. NB-3647 FROM ASME III NB-3647 Pressure Design of Flanged Joints and Blanks Sff .r where P is the design pressure as defined in NB-3112. PFD = P + P.5 Sm SRnot greater than 1.TABLE 1. psi.—— + — _ _ /M0 ^PB Lg^B 4g0 NB-3647.1 and Peq is an equivalent pressure to account for the moments and forces acting on the flanged joint due to weight and thermal expansion of the piping.

U.3 Fatigue Analysis of Bolts.0. rather than maximum stress. or other means which minimize residual torsion. Stress intensity. stretchers. However.10 TABLE 2. using the applicable design Appendix F may be used in evaluating these Faulted fatigue curve of Fig. 1-9. (d) Effect of Elastic Modulus. the suitability of bolts for cyclic operation shall be determined inaccordance with the procedures of (a) through (e)below. m.0.003 in.4). If the operational cycle being considered is of service stress except as restricted by NB-3223. The bolts shall be ac ceptable for the specified cyclic application of loads values of Table 1-1.000 The stress limits of NB-3232. (c) When gaskets are used for preservice testing resulting from direct tension plus bending and neglecting stress concentration shall not exceed 2. Conditions. and the re quirements of NB-3232 are satisfied when the appropriate m and r factors are used for the test gasket.2 apply. The allowable bolt design stresses shall be the values given in Table 1-1.2 Maximum Stress.4(e)(5) does not exceed 1. the fatigue strength reduction factor used in the fatigue evaluation of threaded members shall not be less than4. .7 Sm.3. using the larger of the bolt loads given by the equations of Appendix E as a design mechanical load. The maximum value abscissas. (a) Bolting Having Less Than 100. when applying the rules of NB-3232. pressure. and the minimum design seating stress. and differential thermal expansion may be higher than the values given in Table 1-1.4 and an appropriate fatigue strength reduction factor (NB-3232. the gasket factor.4(e). independently of all other Design and Operating Conditions.000 psi Tensile Strength.4(e) using the design fatigue curve of Fig.3. for bolting materials. (b) When sealing is effected by a seal weld instead of a gasket.4 provided: (1) The maximum value of the service stress (a) The number and cross sectional area of bolts required to resist the design pressure shall be determined in accordance with the procedures of (NB-3232. High strength alloy steel bolts and studs may be evaluated for cyclic operation by the methods of NB-3222. the bolts are tightened by methods other than heaters. (2) Threads shall be of a Vee-type -having a minimum thread root radius no smaller than 0. Actual service stresses in bolts.17). such as those produced by the combination of preload.3(b) for high strength alloy steel bolts. (3) Fillet radii at the end of the shank shall be such that the ratio of fillet radius to shank diameter is not less than 0. the design is satisfactory if the above re quirements are satisfied for w = r = 0. the value used shall not be less than 4.1 or NB-3216. shall not exceed two times the stress values of Table 1-1. Unless it can be shown by analysis or tests that a lower value is appropriate.1 Average Stress.4 is used. the rules contained in psi shall be evaluated for cyclic operation by the methods of NB-3222.2) by the ratio of the modulus of elasticity given on the design fatigue curve to the value of the modulus ofelasticity used in the analysis. Enter the applicable design fatigue curve at this value on the ordinate axis and find the corresponding number of cycles on the axis of NB-3232.3. 1-9. NB-3235 Faulted Conditions If the Design Specifications specify anv Faulted Conditions (NB-3113. Multiply 5aU (as determined in NB-32I6. The maximum value of service stress. as determined in NB-3222. averaged across the bolt cross section and neglecting stress concentrations. if the higher ofthe two fatigue design curves given in Fig. (e) Cumulative Damage. NB-3232 Normal Conditions (c) Fatigue Strength Reduction Factor (NB-3213. Bolts made of materials which have spec ified minimum tensile strengths of less than 100. at the periphery of the bolt cross section resulting from direct tension plus bending and neglecting stress concentrations shall not exceed 3 times the stress the only one which produces significant fluctuating stresses.0. this is the allowable number of cycles. The 2 Sm limit for direct tension is unchanged. y.4(d) and thus require no fatigue analysis.1 and NB-3232. NB-3230 FROM ASME III NB-3230 STRESS LIMITS FOR BOLTS NB-3231 Design Conditions (b) High Strength Alloy Steel Bolting. may be taken as zero. NB-3234 Emergency Conditions components on which they are installed meet all the conditions of NB-3222.3.2) at the periphery of the bolt cross section Appendix E. NB-3232. shall be limited to this value when and thermal stresses provided the cumulative usage factor. 1-9.3. NB-3233 Upset Conditions NB-3232.3(c)).060. only. Unless the The stress limits for Normal Conditions (NB-3232) apply.

inside diameter (pipe wall thickness for integral flanges). with some restrictions. size. . i. analysis of the bolting is already covered by NB-3647. calculations for a fairly extensive sampling have been made. pressure class. this aspect is discussed in this report in Chapter 5.2-1... "Flange Stresses". Sizes. is deemed sufficient to provide the equivalent of stress indices for such joints. 1500 and Sizes: 4". the flanged joint must The be analyzed according to the detailed requirements of NB-3200. SPECIFIC FLANGED JOINT CALCULATIONS There are a large number of variables included in B16.e. 300. the Code user could develop stress indices or conduct an experimental analysis. and the results are presented herein. 150. are listed below. Types Pressure Class: 2500 lb. The sampling should be sufficient to give the designer a good feel for important aspects of B16. i.5 flanged joints and.1 through its reference to NB-3230. 900. 8".11 we note that stress indices for flanged joints (or flanges) are not given in Table NB-3683. 8". While a complete coverage of all these variables was not deemed feasible. 3. and 12". 600. This leaves the flanges to be analyzed.5 flanges. Types: (a) (b) Welding neck flange joined to another welding neck Welding neck flange joined to straight hub flange (c) Welding neck flange joined to a blind flange * In principle. For 2500 lb (maximum size covered is 12") the sizes are 4".e. The parameters covered Classes. gasket dimensions and materials. all. 16" and 24" in all but the 2500 lb. type. Accordingly.

5 to ASME Code allowable stresses.5 are used in this report. The primary rating pressure.. Ratings at other temperatures Dimensions Most dimensions of B16. P (psi) is equal to the class designation.1. g^ i.5.5. taken as 8750 psi. e. .5 flanged joints are given in B16.5. X°is given in B16. The value of S = 8750 psi is approximately equal to the allowable stress of flange materials at the primary rating temperature given in B16.e.where Y . P = 150 psi for the 150 lb class DQ = Pipe outside diameter S = allowable stress. The calculated value of gQ also establishes the flange inside diameter bore. Y and C are given in B16. are given in B16.5g . and the hub thickness. P = 600 psi for the 600 class. however. These were selected as follows.12 Pressure-Temperature Ratings The ratings given in the 1968 edition of B16.5. P. a few are not.C i< LS the total hub length. The tapered length of the hub was set equal to Y . is: P = class designation (psi) (1) For example. where X=hub diameter at juncture with ring. B..C .5-1968. Inside Diameter (Wall Thickness): the equation: PD The wall thickness of the pipe attached to the welding neck flange was established by go = 2S2" + °'05 where (2) P = primary rating pressure. It is also the "representative allowable stress" used in proportioning ratings in B16. B=Dq -2go and g =(X -B)/2. The temperature at which P applies depends upon the flange material.g.

5) and the inside diameter is equal to the pipe outside diameter. and gasket as a statically redundant structure. Calculations were also made using the more basic theoretical analysis described in Reference (3). Effective bolt length is taken as 2t + d + 1/8". corresponding to the Code values for 1/16" thick asbestos gasket material. Calculations were made for bolt and flange stresses in * accordance with Code procedures (Code Appendix XI) . assumed for these materials. The stresses due to the design loads were then calculated. This forms a reasonably good model of a flange cast integrally into a valve or fitting body. and gaskets. the thickness g1 was assumed to extend through the hub and pipe. These Code procedures lead to minimum bolt areas and design loads which. because it considers the flanges. d = bolt diameter. Bolt Dimensions: the thread. The values used in the cal culations were m = 2. . in part. bolts. are proportioned to the design pressure. uses the modulus of elasticity of the materials of the flanges. Gasket Dimensions: The gasket outside diameter is equal to the raised face diameter (given in B16.75. * Appendix XI does not cover blind flanges. The procedure of NC-3325 and Figure NC-3325. y = 3700. This analysis. The bolt areas are based on the root diameter of Threads are Coarse-thread series in sizes 1" and smaller. 8-pitch-thread series in sizes 1-1/8" and larger. A modulus of 3 x 10 psi was The Code calculations require the use of m and y factors which depend upon the gasket material.1(e) was used for code calculations of stresses in blind flanges. bolts. Gasket thickness is 1/16". where t = flange ring thickness.13 For straight hub flanges.

14

4.

BOLT AREAS AND BOLT STRESSES

Required Bolt Areas Per NB-3647.1(b)

NB-3647.1(b) states that "Equations (3) and (4) in XI-3220

**shall be used to establish minimum bolt area required using allowable
**

stress values equal to S

m

as given in Table 1-1.3.

**Equations (3) and (4) of XI-3220 are:
**

For operating conditions:

**Wml = (-7T/'4) c2p + 2b7:GmP
**

For gasket seating:

(3)

(4)

W - (Am + V

where Wm2 = irbGy.

factors.

V2

m ml b

**The required minimum bolt area, A , is the larger of W ,/S, or W ^/S ,
**

m2 a

The values given by equations (3) and (4) will,

of course, depend upon the gasket diameters and the Code m and y gasket

Table 3 shows values of W

ml

, and W

m2

**„ for a commonly used
**

J

gasket.

The required minimum bolt areas shown in Table 3 are based on

S

m

= 35000 psi for SA-193 Code B7 bolt material at 100 F.

The pressures

r

are the rated pressures at 100 F.

It can be seen that the actual bolt

areas are larger than the required minimum.

Because the rated pressures

decrease more with increasing temperature than the decrease in the allowable bolt stress with increasing temperature, it follows that the bolt area is adequate for all temperatures.

Table 3 is for pressure loading only.

When a bending moment

**is imposed on the joint by the attached pipe, NB-3647.1(a) requires
**

that the design pressure be obtained by:

16 M

,

p

FD

= p +

„3

P-2.

°;

e^

*

The direct axial force term is not included because its use is not

consistent with the present Code analysis of other piping products.

vO vO vO o rH rH 00 vO 00 m vO 00

o

00 CM is m CO

CO CO 00 00 CM

rs

CM

CM m

CM

o o

CM

St O

CM m

CM rH o VO rH in

O vO on CO o rH CM

m m

o

O

o

CM

o

CO vO SiCO CO m r-H CO 00

is-

CM

O

is 00 vO ON

CM vO o St 00 00 vO 00

St 00 St CM o CM 00 rH

« rH

CM

o

00 00

rH 00 CO rH CO

CO

VO

CO is CM

o

CM

00 ON

vO CO

rH

rH

rH

00

CO

vO

O vO 00 ON vO rH 00 00

co in co

00

U0 s t

CO VO is O

N m

NNOvO

CM vO rH 00

o

st

o

m

m

vo

on

cm

is rs vo m

VO cm m m

r-t

r-i o

CO

is l-H CO CO CM VO

CM st m st is is rH co rs

m cm st vo on co co cm st

« -H

O

H

CM in

rH CM Is in

rHCO o

$

o ^-i CM

ON rs CO O

o o o

W N (<1 O

O O rH 00 CM st 00 CO

(^ ts CI O

O O rH 00

CM st 00 CO

(Ti is ro O

O O rH 00

CM st 00 CO

w

3

is en O

O rH 00

.•N st 00 CO

On is co o

O O rH 00

CM st 00 CO

m

n

H

O O CM CM st vO

O

CM st 00 CO

O

rH 00

CM ON O

^

O

rH vO

rs CM O

O

Ct\ C^ st o

co a\ vo st CM m

CMM O

CM CM

O

CM o

-ct O

jn on

O

on

O

o m co m

r-i

CM

OOOO

is On CM 00 CO rs ON 00 CO O T-\ ON ON CM i-H CM IS vO rH

OOOO

00 vO CM ON ON CM rH st CO rH

OOO

O

O

, e

in

Ls

r-{

o o o

S

HNO

^-i m o

IS

rH CO O O r-t CM

CM

CO CM rH co oo m rH

« -H

m

O

O

Pi

CO P.

ts

CM

CM

IS

vO

ON

O st

st

O

vO

O O

vO CO

O O O

vO

O

fJ_

-H

N

•H

CO

st 00 VO st r-{ CM

st 00 vO st

rH CM

st 00 vO st r-i CM

st 00 vO st ^ CM

st 00 vO st

rH CM

st 00 vO st

rH CM

St CO CM

W

rJ

<

CO

CD

O

in

o o

CO

o o st

o o

vO

o o

ON

o o

in

o o

•n

CM

16

The question arises:

what bending stress can be carried by the flanged

joint as limited by the bolting of B16.5 flanged joints?

Noting that

2

**M , = S , Z , and using the approximation that Z
**

becomes:

= (ir/4) D

g , Equatic

*

Substituting P

from Equation (6) for P in the first term

of

Equation (3) and solving for

S

gives:

o

_ sm K b

(tt/4) G2 P + 2biTGmP

TTD g /G

(7)

Pb =

**Table 4 shows values of S
**

values of S

pt>

**calculated by Equation (7). The calculated
**

m

are based on S

**= 26,700 psi for SA-193 Grade B7 bolt
**

pb

material at 700 F.

i

Two sets of values for S , are shown in Table 4:

(1)

**S ,, with the attached pipe wall thickness obtained by
**

the equation: g = PD /17500+ 0.05. This gives a

pipe with about the same pressure rating as the flanges.

(2)

S ,

pb

**with the attached pipe wall thickness equal to the
**

_

**larger of std. wt. wall thickness or PD /17500 + 0.05.
**

This is significant, particularly for the 150 class, because most piping is std. wt. or heavier wall

thickness.

**Code Case 1677 states that in calculating H
**

pressure (not P™) shall be used.

= 2biTGmP, the design

2043 .3457 .5986 .4614 14900 22000 17300 31000 15300 9100 19000 24 400 4 8 16 635 17300 20000 . or S . S . . IN PIPE ATTACHED TO pb pb FLANGED JOINTS.1071 15000 24 13650 10550 10250 10400 2500 4 8 12 3920 . P = RATED PRESSURE AT 700 F.2557 16900 22950 24 300 4 8 16 23500 26400 470 .1 RULES BOLT STRESS CRITERIA i Class Size P.(psi) for larger of g std.4214 2.8729 15100 9600 16 24 1500 1.0886 36400 13600 6500 11450 16100 14200 .1871 .2471 .3243 .8729 12700 14900 12900 16 • 24 900 4 8 1410 . ALLOWABLE BENDING STRESS.1271 . psi go S . NB-3647.1529 .6929 1.5986 940 .(psi) for pb S .4157 11700 20300 19300 13800 12700 14900 12900 20300 19300 16000 24 600 4 8 . wall go or 150 4 8 16 110 .17 TABLE 4. wt.2821 1.4936 .1979 .2843 2350 .4357 .2814 24100 24100 15100 9600 .7893 13900 13800 11500 13900 4 8 16 13800 11500 15000 13650 1.1239 .8614 10550 10250 10400 .

Slow.800 psi at 700 F. The elastomeric 0-ring has substantial pressure-sealing capability. To examine the effect of this average stress limit on B16.5 flanged joints. differential thermal expansion. and differential thermal expansion".18 Pipe made of SA-106 Grade B material has an allowable stress intensity S m = 16. S . The contents of NB-3230 is included herein as Table 2. except the "ring joint". The "ring joint" may have some pressure-sealing capability. and pipe loads are applied to the flanged joint in operation. averaged across the bolt cross-section and neglecting stress concentrations. the joint will not leak excessively. pressure. The preload average bolt stress. ** The types of gaskets covered by Equation (8) includes all of those shown in Table XI-3221.1. It has been shown by Roberts that the theoretical leakage pressure of a flanged joint with a non-pressure-sealing gasket is given by the axial forces equilibrium equation: * Leakage is defined as the gross type of leakage that occurs when the load on the gasket is reduced to zero.6 as 3 S m = 50. NB-3232.1 Average Stress The maximum value of "service stress". should be such that when pressure. Service stress is defined as stresses'feuch as those produced by the combina tion of preload.5 flanged joints. Bolting Requirements of NB-3230 NB-3647. must not exceed two times the tabulated allowable stress value.1 rules) maximums of S for the flanged joints ranges from 13 to 48 percent of that for straight pipe. It can be seen in Table 4 that the calculated t- (by NB-3647.400 psi. . diffusion-type leakage may and generally will occur at lower pressures. it is necessary to determine the magnitude of preload used in tightening the bolts in B16.1(b) states that the minimum bolt areas (as establi shed by Equations (3) and (4) of XI-3220) "may have to be increased to meet requirements of NB-3230". The limit on bending stress due to moment loads in straight pipe is given by Equation (12) of NB-3653.

g. which is a function of S.19 . S = maximum stress in attached pipe due to P So = bending moment attached pipe wall thickness F P = primary rating pressure (class designation). e. Accordingly. and the applied loads. flanges. and gasket.. The computer program The results FLANGE was used to determine the change in bolt stress.. = coefficients given in Table 5 and 6 S. it is necessary to establish a relationship between Sbl (whicn> in principle. can be expressed by the following Equation (9). for 300 class. „ is the bolt stress after application of the loads. not S.g /D . a_. > the initial bolt stress applied in preloading. = 4S . 3bl " S .2 =Sbl(l -e2/E±) +a/eq/F +a2P/P + a3^(Tb-Tf)/50+a4 (Thp-Tf)/10° (9) where a . bolts 1 P 2 = eq and gaskets at initial bolt up tempera ture (operating temperature) pressure equivalent of bending moment applied by attached pipe.2 PL = Vb2/(7TGo /4) where P = leakage pressure <8> S. bz = = bolt stress after application of loads gasket outside contact diameter G It is important to note that S. bl = initial bolt stress S b2 = bolt stress after application of loads E (E ) = modulus of elasticity of flanges. a . and the change in modulus of elasticity of the bolts. a. is a controlled value) and S . P = 300 psi = internal pressure T = temperature of bolts .

although items (c) and (d) would seldom occur at the same time. An accurate way For the present purpose. In order to arrive at an estimate of a suitable preload average bolt stress. serve as the equivalent of stress indices in Equation (9). the value of S. a . Accordingly. to include this force is not known. a .20 T = temperature of flange ring T. The values of a1 and a are given in Table 5. = temperature of flange hub and pipe. The axial forces equilibrium Equation (8) must be extended to include the axial force on the tension side of the applied moment. The conditions shown in Table 7 are more-or-less representative of conditions that are applied to B16. values of a. the axial load which must be added to the axial pressure load is ^jjjSqSp^jJ where D^ ±s the mean diameter of attached pipe. The combination of loads shown in Table 7 include a bending moment applied by the attached pipe to the flanged joint. we assume the set of conditions shown in Table 7. The resulting equation for the critical bolt stress S . we make the conservative assumption that the maximum tensile stress due to the pipe-applied moment (which exists only at one point on the pipe circumference) acts around the complete circumference of the pipe. corresponding to the theoretical leakage pressure is: Sbc =(^ +*VoSpb) /Ab <1Q) .„ can be obtained for any combination and magnitude of the loads. and a. is the bending stress in the pb attached pipe. and a^ are given in Table 6. With Equation (9) and the a-values shown in Tables 5 and 6. The values of a .5 flanged joints. g is the attached pipe wall thickness and S .

21

TABLE 5. FACTORS a± AND &2 FOR USE IN EQUATION (9)

Class

P

al

Size

WN

SH

a2

Blind

WN

SH

Blind

150

4

8

500

400

900 800

200

600

500

700

1,100

600

700 900 200 500

500

16 24 300 4

8

1,100 1,500

400

800 800

1,100

300

**1,700 1,700 2,500
**

600

**1,300 1,200 1,800
**

500

2,400 4,200

5,500

700

600

500

1,400

1,500

16 24 400

4

8

1,100 1,100

400

900

800

800

2,100

2,100 2,600 1,600 2,000

700

**1,000 1,000 1,500 1,500 1,900 1,100 1,400
**

500

800

1,800 2,800 4,200

800

800

1,100

600

800

900 600

800

4,200 3,200

2,800 4,000

900

16 24

1,100

500

800 900

600

4

8 16 24

400

500 600 900

300 400

600

1,300

500

700

900

300

**1,200 1,700 2,300
**

600 900

1,100 1,500

400

500

**1,600 2,800 3,700
**

700

900

4

8 16 24

300

400

700 600 400

400

700

1,100

900 500

600

300

1,700 1,600

600 700

800

1,000 1,000

400

400 400

700

**1,500 3,000 2,600
**

700

1500

4

8 16 24

600

600

700

400

400

500

300 400

500

900

1,100

300 300

300

1,300 1,600

400 500

500

2500

4

8 12

400

400 400

300 400 300

200

200

200 200 200

200

WN = Welding neck flange mated to welding neck flange SH = Straight hub flange mated to welding neck flange Blind = Blind flange mated to welding neck flange

22

TABLE 6.

FACTORS a

AND a, FOR USE IN EQUATION (9)

a3

Class

Size WN

SH

a4

Blind

WN

SH

Blind

150

4

8

16

24 300 4

8 16 24

3,400 4,100 1,700 1,800

3,500 4,300 1,900 2,100 4,600 3,900 2,300 2,000 4,600 4,300 2,600 1,900 4,600 4,400 2,800 2,900

3,800 3,400 3,400 3,500

4,700 5,400 4,300 4,300

3,400 4,200 2,100 2,300

6,400

10,900 8,700

8,700

5,900 8,000 6,900 6,800 4,900 6,900 6,300 5,600 4,900 6,200 6,000 5,900 3,400 4,600 5,000 5,300

2,700 3,500

6,700 11,100 8,900 9,200

3,300 5,600 5,400 5,400 3,100 4,500 4,600 4,700 2,600 3,900 4,100 3,900 2,700 3,600 4,000 4,000 2,000 2,900 3,800 3,600

1,600 2,000 2,500 2,500

900

**4,400 3,600 2,000 1,600
**

4,500 4,000 2,200 1,600 4,500 4,200 2,500 2,400

3,800 3,200 3,200 3,200

4,600 5,300 4,100 4,200

**4,700 4,100 2,700 2,200
**

4,700 4,500 2,900 2,200

4,900 4,800 3,300 3,300 4,400 4,100 4,100 4,300

5,300 6,100 5,300 5,400

**6,000 8,200 7,200 7,200
**

5,000 7,100 6,600 6,000 4,900 6,300 6,100 6,400 3,500 4,600 4,800 5,500 2,800 3,600 3,800 4,000 1,700 2,200 2,300

400

4 8

16

24

600

4 8 16 24

900

4

8

16 24 1500 4

8 16

24 2500

4 8 12

3,800 3,900

1,700 2,100 2,300

5,600 6,000

6,300

5,600 6,100 6,300

6,300 6,800 7,100

1,200 1,300

WN = Welding neck flange mated to welding neck flange

SH = Straight hub flange mated to welding neck flange Blind = Blind flange mated to welding neck flange.

23

TABLE 7.

**REPRESENTATIVE SET OF LOADINGS USED FOR ESTIMATING
**

A SUITABLE PRELOAD AVERAGE BOLT STRESS

(a)

(b)

**Internal pressure equal to the primary rating pressure; P = P
**

Decrease in modulus of elasticity from 30,000,000 to 23,000,000

psi, E /E

(c)

= 0.767

**Bolt temperature 50 F higher than flange ring temperature,
**

(T,

b

- T-) = 50 F

r

(d)

**Pipe and hub temperature 100 F higher than flange ring temperature,
**

(T,

hp

- TJ = 100 F

r

(e)

**Pipe moment load that produces a bending stress, S
**

in the attached pipe.

, of 8,750 psi

(Tb .. must be high is equal to S Equation (9) gives: Sb2 = Sbl(W " alPeq/P " a2 P/P " a3(E2/El)(Tb " V/5° "a4(Thp " Tf)/10° Equating S.^. S.767 H (13) Values of a±.24 Values of S. gives the values of (S. neck-to-welding neck joints (WN) in Equation (13) . a.000.Tf)/50 = 1.750 nsi.243. gives: (11) . . P eq /P = 2. bbi Sbc +alPeq /P+ a2 P/P +a3(E2/El)(Tb ~Tf)/5° +a4<ThD ~Tf>/100 f^/e^ (12) t For the conditions assumed in Table 7. P/P = 1.Tf)/100= 1. calculated by Equation (10) with P = P and S . 0 to S and solving for S. PC + 2.000 and (Th . enough so that S The initial bolt stress. = 8. shown in bl mm oc along with the values of S . a^ and a^ are given in Tables 5 and 6. + a„ + 0.767a_ + a. .243a. 1 L 3 0. Using these values for welding Table 8. pb are shown in Table 8.767. be .000 ' ' E2/E1 = 0. Accordingly: o bl _ S.J .

sbc sbl bb (S = minimum preload bolt stress calculated by Equation (13)./A. (3) bb ASME (sbiV Petrie (4) 150 4 8 9400 17500 26300 46600 36500 12900 16900 9200 7400 8700 56900 52000 16 24 300 4 13800 13700 45000 40300 52000 48100 38600 27200 37000 30800 34800 28600 35800 31300 31700 29700 10100 8 16 14400 11900 8200 5900 6700 6600 40300 36700 48100 45000 38500 24 400 4 14000 9000 14000 12400 13000 12600 15100 8 16 8300 6400 24 600 4 8 16 6300 9900 34100 48100 42500 36700 32900 42500 35900 9500 7800 8100 8500 8900 9800 7900 14800 16400 10300 13700 18000 15400 12600 15000 34400 38200 23800 30300 38600 35000 26800 32100 30300 32700 26500 30000 24 900 4 8 38500 35300 28500 40300 16 24 1500 4 8 11100 10300 8000 35300 28500 16 24 14200 15300 13100 8000 12300 10800 24100 36700 31800 27200 2500 4 8 12 15100 15900 31500 10000 (1) (2) (3) (4) = critical bolt stress. Equation (10). ml. d = bolt diameter in inches. b m' m2 by Equations (3) and (4). id2. bl'p . . INITIAL BOLT STRESSES Class ize be s (2) bbl Minimum «. = larger of W .-p . and W „ are calculated ° .25 TABLE 8. or W „/A. where W . ) = 45000/ -fa .

.3. to prevent leakage as judged by comparison with (S. (S.26 Table 8 also shows b for comparison. Obviously.5 flanged joints is seldom con trolled. the pipe fitter simply tightens the bolts to what he con siders to be an appropriate amount.1.n) bl p is of interest in that the initial bolt stress used in assembling B16.45. it is intended to show that an initial bolt stress of about 40. Petrie indicates that the "appropriate amount" normally applied by a pipe fitter is approximately given by the equation: (s ) . for Bolting Materials for Class 1 Components". Based on an initial bolt stress of 40. is met? m In so far as bolt-up conditions are concerned. The value of S m for almost all bolt materials listed are greater than 20.000 psi at 100 F.5 flanged joints so that the criterion of NB-3232. -) .5 flanged joints. (14) is adequate in Table 8. hence. column in Table 8. that the average bolt service stress shall be less than two times S .000 psi.000 psi is usually needed and applied. the minimum values of S. inches.) where d = bolt diameter. S . headed (S. these values The last used as initial bolt stresses are entirely inadequate. Rather. . (He may further tighten the bolts if a leak is found in a hydrostatic test or subsequent opera tion) . Table 8 is not intended to give specific recommended values of the initial bolt stresses that should be used in assembly of B16. we may then answer the question: What bolting materials can be used in B16.000 psi bl P /d~ For most joints. the answer is ob tained from Table 1-1. as obtained by Equations (3) and (4). "Design Stress Intensity Values.

then S. m * The a-values shown in Table 5 and 6 are for a sampling of B16.e. knowing that history.. However. the computer program FLANGE will give similar information for any flanged joint of geometrical shape covered by the program. = 40.. i.880 is less than 2 x 26. P. .700 psi at 700 F) the average bolt service stress could still meet the criterion of S.T ) = 40 F and (T . this may be such that the average bolt stress at operating temperature is less than 2 S at the operating temperature.T ) = P = 0 and E. T. with P eq . f However. The answer to the question for operating conditions is r more complex and one must know the time history of P eq .700.000 psi is acceptable. However. b2 <2 S . let us consider the case where the flowing fluid is rapidly decreasing in temperature. „ = 40. bl bz If the bolt material were SA-193 Grade B7 (S m = 26.120 F.5 flanged joints. . . If S^. From Equation (9) and Table 6: Sbl " Sb2 = 1'600x 4£ + 6j80°X =^° = "6.Tf) and (T or cooled down. One notes that P and P will produce a decrease in bolt stress./En = 1.880 = 46. the quantities (T. The flanged joint „ 2 1 consists of a pair of 24 . T.000.300 class flanges. and that (T = .880 psi. 46.Tf) can be either posi tive or negative depending upon whether the pipeline is being heated up As an example of how average bolt stresses could increase.000 + 6. b A hp and T .00.27 an average bolt service stress of 40. the answer can be obtained » from Equation (9). Exceptions include the annealed austenitic stainless steel materials such as SA-193 Grade B8.880 psi.

i. and that leakage occurs when S..a0p/p~ 1 pb o o z ' (16) Solving Equation (16) for S . For these loadings.(tt/4) G2 P/A.000 . . + 4 a. as can be seen by comparing Equation (7) with Equation (17). gives: s . the rated pressure for carbon steel flange material at 700 F.000 . = ______ pb o m d 40. „ = pb o o bZ S.000 is a representative initial bolt stress for B16. be as defined by Equation (10).P/P b2 1 2 i (15) Noting that P & eq = 4 S . pressure and pipe moment.28 Now that we have established the point that S .P eq /P .a.5 flanged joints./D P) 1 For comparison with S as obtained by Equation (7). fortuitous. This is deemed to be purely The two methods are entirely different.a„P/P 2 ___ o (1?) g (ttD /A.a. It can be observed in Table 9 that for the 150 class flanges.= 40. Equation (11) becomes: Svo = 40..000 .e. M Pb . we use the same pressure.000 psi and when the assumed loadings are the same as those in Equation (5). i.. we obtain the equation: (ttG2 P/4 + ttDm°o S . For the higher pressure . The results are shown in Table 9.a.)/AD= g pb o 40. (4S . it is of interest to use Equations (9) and (10) to determine the attached pipe bending moment capacity of the joints when the bolts are initially tightened to 40.g D )/P . the two methods of calculating allowable pipe bending stresses in attached pipe give about the same answer.g /D .e.

2843 2350 21800 13900 21800 13900 13800 11500 . (NB-3647.4936 .4214 .5986 .1 Pb FLANGE NB-3647.1071 15000 13650 2500 4 8 3920 .2814 . psi Pb FLANGE NB-3647. for pipe of wall thickness g S .2043 16000 12700 14900 12900 24100 15100 9600 .7893 33700 25700 27100 23800 13800 11500 15000 13650 33700 25700 27100 23800 16 24 .1271 . .1529 .1239 16 24 300 4 8 16 24 400 16 . S and Spb. IN PIPE ATTACHED(1)TO FLANGED BOLT STRESS CRITERIA JOINTS.8729 24 1500 .4357 .0886 Eq.6929 1.1 DATA ARE FROM TABLE 4) Class Size P. ALLOWABLE STRESSES. 7 150 110 .5986 20100 19100 27000 24900 27000 15300 20300 19300 20300 19300 13800 24 600 940 16 24200 31300 23000 22600 24200 27000 .4614 14100 24800 22100 9100 19000 17300 20000 11700 28700 22100 31100 17300 31000 635 .1 Eq.1871 7500 12600 16300 20400 6500 11450 16100 14200 22950 23500 26400 14900 22000 . wt.29 TABLE 9.2471 .2821 1.3457 23000 22600 12700 14900 12900 24100 15100 9600 .8729 24 900 16 18100 40800 27200 17400 18100 40800 27200 17400 1410 . for pipe of wall thickness equal to the larger of g pb o or std. 17 13600 36700 19400 25300 23900 38000 22900 . 7 13700 Eq. P = RATED PRESSURE AT 700 F.3243 .1979 .2557 470 .4157 . 17 36400 16900 Eq.8714 34700 27900 25800 10550 10250 34700 27900 25800 10550 10250 10400 12 10400 (1) S .

NB-3232. ±2 given by: Sbb = 21 (G1 + V where E d <18> = = modulus of elasticity of bolt material bolt diameter I 6"l = = effective bolt length rotation of flange ring of one flange 9~ = rotation of flange ring of the other flange. or other means that minimize residual torsion. . shall be limited to these values when the bolts are tightened by methods other than heaters.500 lb flanged joint. Equation (17). stretchers.5 flanges are usually tightened with a wrench that does produce residual torsion in the bolts. rather than maximum stress. if included. stress in the bolts.since B16. as illustrated by Figure 1. To examine the effect of this maximum stress intensity limit on B16. predicts significantly higher allowable pipe bending stresses than Equation (7).2 Maximum Stress The maximum value of service stress at the periphery of the bolt cross section resulting from direct tension plus bending and neglecting stress concentrations must not exceed three times the tabulated allowable stress intensity values. Equation (17) does not include thermal gradient effects. it is necessary to establish (a) (b) The magnitude of the bending stresses in the bolts The stress intensity in the bolts . these. Stress intensity.5 flanged joints. An engineering evaluation of the bending stresses in the bolts can be obtained by assuming that the nuts on the bolt remain parallel with the flange ring. However.3 times as much for the 4" . up to 3.2. S Ed The bending . which is used in program FLANGE. would decrease the values of allowable pipe bending stress and the results from Equation (7) might not be overconservative.30 classes.

31 Nut face remains parallel to ring surface FIGURE 1 . ASSUMPTION USED IN CALCULATING BOLT BENDING STRESS .

welding neck 40.86 x (15. 3 x 10 For this joint.8.32 Values of 6.900 class flanged to straight hub. An adequate estimate of the torsional stress in the bolts can be obtained as follows. are shown in Table 10. lb. I = 6. inch t Ffe = axial force in bolt. consider an 8" .-lb -0.350 psi. M = Sih = 40.5 . lb. in.000 x 13. and taking E as and using values of 0„ from Table 10. For well-lubricated bolts and nuts of the type used with B16.554 +1.625)/2 = 1.2 d = bolt diameter. calculated using the computer program FLANGE.905.750 in.5 flanged joints. we obtain Sbb =3X10 x1'375 (1.375".000 psi. 2 x 6. Sbl A. d = 1.078) x10"3 =8. the relationship between tightening torque and axial force in the bolt is: T=kdFb =kdSblAb' where T k (19) = = torque.5". . As an example of their use. = initial (preload) bolt stress = bolt area of a single bolt. For an initial bolt stress of joint. = S A.5 Table 11 shows bolt bending stresses calculated by equation (18) for all of the flanged joints covered by this report.

33

TABLE 10.

ROTATIONS

(1)

OF B16.5 FLANGES

Welding Neck

M

Straight Hub

e

M M

Blind

3

Class

Size

x 10

x 10"

x 10"

xlO2

1.704 1.430

xlO"

xlO"

xlO"

x 10"

150

4 8

16

2.104

1.898

2.434 6.407 13.66

0.910 1.342

1.365

0.903 1.265

1.383

2.868 2.622

8.439 32.25 90.15

3.038

3.123

24

4.454 6.038

1.298

1.746 2.054

0.726 1.020 1.336

22.54

2.381

2.484 3.202

1.000 1.385 1.982 2.288 1.137

8.910 8.568

2.028

5.318 0.780

1.621 0.710

141.8 6.7 24

19.81

300

4

8

16

2.012 4.098 5.872 1.457

6.919

13.73 22.73

2.056

1.988

3.173

0.949

1.047 1.177

0.659 0.866

2.780

6.171

44.50

82.82

24 400 4

8 16

1.609

0.665

7.633 2.114 2.380

**2.629 7.189 14.55 23.78 2.151 6.280
**

12.55

1.963

4.029

0.904

1.225

1.389 2.020

2.666

0.882

0.827 2.173 2.146 3.439

6.736 17.20 43.24 85.05 7.395

15.33

0.996

1.122

5.596

8.285 1.972

2.316 4.751

24

600 4

8 16

6.543 1.110

1.579 3.128

**1.480 0.610 0.785
**

1.028

0.365

1.215

0.597 0.739 0.862 0.947 0.513 0.610 0.741 0.751

1.096 1.680

1.491

2.277

36.00 52.69

6.625

14.18

24

**4.227 1.340 1.554
**

2.109 2.586

19.36

1.453 3.917

1.161 0.524

0.674 0.863 0.855

1.781

1.079

4.648

900

4

8 16

**0.304 0.403 0.205
**

0.773

2.602

3.344 3.360 3.540 1.925

9.073

11.19

1.078 1.284 1.424

29.45

28.39

24

1500

4 8 16

0.963 0.924

0.611

1.415

24 2500

4 8 12

1.384 1.674

0.582

3.164

5.298 0.018

**0.420 0.469 0.548
**

0.578

0.277 0.317

**0.781 0.749 0.936
**

1.146 0.531 0.509 0.497

0.013

0.011

0.016

0.499 0.024 0.026

0.026

**0.416 0.464 0.509
**

0.557

5.981 7.673

11.27

14.23 3.251 4.483

1.577 2.455

2.415

1.282 1.167

0.280 0.319 0.326

0.567 0.550

0.018 0.017

0.324

1.064

4.501

meM -

rotation due to moment

applied to flange ring of 40,000

W

r adians

**rotation due to pressure equal to P, radians
**

9™, =

rotation due to thermal gradient (T

- Tf) = 100 F, radians

34

TABLE 11.

**BOLT BENDING STRESSES, ' S
**

40 ,000 PSI

AT INITIAL BOLT STRESS OF

d,

Class

Size in.

£,

in.

Ed/£

Bolt

psi _6

x 10

(2)

WN-WN

Bending Stresses (ps (2) ,(2)

WN-SH

WN-Blind

150

4

8

0.625

0.750

2.625

3.125 4.000 5.125

7.143

7.200

16 24 300 4

8 16

1.000

1.25 0.75

7.500

7.317

**15,030 13,660 33,400 44,180 8,650 12,430 26,160 37,090
**

10,200 12,080 25,570 43,620 7,280 9,470 18,460 23,780 9,520 9,860 11,750 14,230 6,420 5,000 7,350 8,960 3,430 2,810 2,610

13,600 11,980 26,020 33,800

**17,760 16,270 50,110 53,440 11,090 14,800
**

32,770

0.875

1.25 1.50

3.375 4.250 5.875 7.125 3.750

6.667 6.176

6.383 6.316

24

**7,660 10,490 19,400 25,770 9,080 10,310 19,190 30,700
**

6,540 8,030 14,190 16,900 8,590 8,350 9,450 11,040 5,810 4,530 5,890 7,540

42,650

12,500 13,360 30,540 49,430

400

4

8 16

0.875

1.000

1.375

**7.000 6.154 6.346 6.667
**

6.562 6.000

4.875 6.500

7.875 4.000 5.625 7.625 10.000

24

600

1.75

4

8

16

0.875

1.125 1.5 1.875

1.125

**5.902 5.625 7.105
**

6.346

5.571

24

**10,110 11,690 23,250 24,960
**

14,000 15,540 15,230 16,860 9,630 6,770 10,190

900

4

8 16 24

4.750

6.500 8.750 13.625 5.625

9.000

1.375 1.625

2.5

5.505 6.667 5.417 5.310 5.350

1500

4

8 16 24

1.25

1.625

2.5

14.125

19.625

3.5

10,940 5,500 4,290 3,830

2500

4

8 12

1.5

7.625

12.125 17.375

5.902 4.948

2.000

2.75

4.748

3,280 2,660 2,490

(1)

Stresses

calculated

by equat:ion (18) using values of 6

and

6

from

**Table 10, and with M = Sfo Afo hG; P and (Th - Tf) equal to zero.
**

(2) WN - WN = welding neck mated to welding neck flange WN - SH = welding neck mated to straight hub flange WN - Blind = welding neck mated to blind flange.

35

**The constant k represents frictional resistance between threads plus
**

frictional resistance between nut and flange. Approximately one-half

**of the torque is resisted by thread friction; the other half by nut/
**

flange friction. Accordingly, the bolt is subjected to a torsional

load of T = T/2.

The shear stress in the bolt, Sb2, is then:

**sb2 •\ Msb! *4-x -4 -2 ksbi
**

ird

(2o)

The stress intensity, including the bending stress, is then given by:

**\ - [(sbi +sbb)2 +4x(2ksbi)2] *2
**

According to NB-3232.2, the value of S must be less than 3Sm>

equating S, to 3S we obtain:

(2i)

From

**^ = r(3sm /SU1)2- 16k2"l S I bl J
**

bl L

"1

(22)

**For SA-193 Grade B7, bolt material at 100 F, S^ = 35,000 for d± 2.5";
**

S = 31,600 psi for d >2.5".

m

With an average initial bolt stress

**S, =40,000 psi, Equation (22) gives:
**

bl

**(S, ,./St1) maximum = 1.500 for d^ 2.5"
**

bb bl

**(S../S.,) maximum = 1.231 for d> 2.5"
**

bb bl

Using the results of Table 11 (i.e., S ), it is apparent that for an initial

bolt stress of S

conditions.

= 40,000 psi, the ratios of sbb/sbl are weH below the

ratios permitted by NB-3232.2 and that NB-3232.2 is met for preload bolting

To check whether the NB-3232.2 criterion that the maximum bolt

stress must be less than 3S

is met for operating conditions requires

56. can be obtained by Equation These three values can then be used in Equation (21) to determine whether the maximum bolt stress is less than 3S . 34. and the shear stress S (20). The assumed operating condition to be checked is: P P P eq = 1. consider an 8" .000 x0. but.. The initial bolt stress is 40. T.768 x10_5 6T =(0. and T .000 psi.780 psi meets NB-3232.800 psi.1200 - 783 + 3060 - 2300 = 34.400 .780 psi* 0p =(3. 6. P..610 x10"3 x-J| =0.1. welding neck-to-straight hub.078) x10~3 at Sfel =40. . the average bolt stress S can be obtained by Equation (11).36 a knowledge of the time history of P .e.917 +0.000 psi =2. M The computer program FLANGE gives the change in moment so that the more precise results can be used.642 x10~3 * The procedure used here is approximately correct. b T^ hp T = 600 F (Sm for bolts = 28.780 is less than 2 S = 2 x 28.900 class flanged joint. As an example.=6.2.554 +1.632 x10"3 x|^|| =2.400 psi) » r / = 700 F = 650 F E2/E1 = 0.90 x^~ . as required by NB-3232. more precisely.403) x10~5 x^-°.4600 x 50_ 100 = 36000 . Once these four quantities are known. T.90 -400 x~^ -500 x±^ -3400 x0. Using the values from Table 10 for the assumed 8" .289 x10_3 at Sb2 =34.700 psi r T.674 +0.410 psi = 900 psi =2. . i. the bolt bending stress Sbb can be obtained by Equation 11 and Equation (18). .780 psi We note that the average stress of S = 34. is proportional to the flange moment rather than bolt load.90 Equation (11) and the a-values from Tables 5 and 6 give: Sb2 =40.900 class welding neck-to-straight hub flanged joint and the assumed operating conditions: 6M =(1.

4(d) to bolted flanged joints.3 Fatigue Analysis of Bolts NB-3232.000 = 16. . a "fatigue strength reduction factor" of 4. Because 53.999 x10~3 =9.400 psi Maximum stress intensity = (44. The flange bending stress from equation (18) with E = 3 x 10 and I = 6.4(d) are met.375" =Ed (6) bb 2£ 6.2 x 40. d = 1.300) + 4 x 14. the maximum stress assumed for this example to be 0.346 x106 x2. In the following discussion of fatigue analysis. This shear stress will decrease by the ratio E /E . the definition of "adjacent points" does not seem to have any meaning because a major temperature gradient to be considered is that between the flanges and bolts.5" is psi.0 will be used to account for the "notch" at the root of the bolt threads. First.090 psi is less than 3S r m at 600 F (85.800 psi. Accordingly.000 psi. The fatigue analysis is not required if the conditions of NB-3222.2 is satisfied. It is difficult for the authors to see the relevance of NB-3222.400 = 52. Second. this is probably the major significant cycle involved.520 psi 2 From Equation (20).37 Accordingly.780 psi Bending bolt stress = Shear bolt stress 9. intensity at the assumed operating conditions is: Average bolt stress = 34. the flange rotation at the operating conditions is -3 2. NB-3232.3 requires that changes in bolt stresses due to the postulated history of operating conditions must be determined and evalua ted.520 psi 1/2 = 14.200 psi in this example) the criterion of NB-3232. the bolt tightening /untightening cycle is completely ignored. the initial shear stress is 2 x 0.90.999 x 10 radians.

The complete fatigue analysis would require that the history of P P. n = 40. with S (23) = S /2 p Values of Sbb are given in Table 11 for an initial bolt load of Using these in Equation (23) gives the peak stress a and using the curve labeled "Max.000 psi.000. multiplied by 4.38 One of the postulated cycles may consist of bolting and un bolting the flanged joint a postulated number of times during the service life.000 psi. S . From Code Figure 1-9. or more generally. The fatigue analysis procedure can be illustrated by continuing the example of the 8" . T T and T be known. gives: Sp =[(40.000 psi (see Table 12). is obtained from equation (21) using k = 0. In operation. the operating conditions give a maximum stress in tensity in the bolts of 53. welding neck-to-blind flanged joint). S. then the peak stress range. Assuming cycles between the operating conditions and preload . nominal stress = 3.0 for the notch at the root of the bolt threads. It can be seen in Table 12 that the number of tightening/untightening cycles. obtained by tightening with a wrench. If the average initial bolt stress is 40. ranges shown in Table 12. bl Equation (21).OS ".0 40. Knowing that history.090 psi. the bolt stress variations can be obtained from the data presented herein. The peak stress intensity is 4 x 53090 = 212.024 x109] 1/2 x4. N .4. As developed previously.2.5 flanged joints. by use of the computer program FLANGE. for the flanged joints covered herein. due to tightening/ untightening. N . we m obtain the permissible cycles. is large even for the worst flange (225 cycles for the 24" . The peak stress intensity at the 40. cyclic bolt stresses would be superimposed on the initial bolt stresses.000 psi preload condition is 232. The values of N shown in Table 12 are far in excess of the number of tightening/untightening cycles normally encountered in B16.400 class. shown in Table 12.000 +Sbb)2 +1.360 psi.900 class flanged joint and the assumed set of operating conditions.

BOLT FATIGUE ANALYSIS. STRESS RANGE AND ALLOWABLE DESIGN CYCLES. nominal stress = 3. Class Size ksi^ 255 Cycles. (2) N. D Stress. N . (2) N D 150 4 8 16 455 428 443 356 210 503 458 250 320 472 305 247 534 24 300 4 8 360 233 293 322 230 356 301 547 511 246 294 334 485 354 282 239 270 16 411 356 530 514 414 322 318 355 246 249 310 380 308 253 485 475 322 225 24 400 4 8 293 234 238 269 310 238 245 514 489 358 250 16 24 292 358 600 16 228 555 522 226 231 564 542 465 238 243 284 290 251 256 255 261 514 496 376 236 267 285 236 237 420 374 522 252 261 233 232 24 900 16 437 534 362 468 518 496 465 564 587 551 538 522 503 573 452 455 243 252 226 236 241 224 219 24 1500 16 437 522 559 236 227 221 229 596 224 24 2500 4 8 234 216 214 530 611 621 229 215 213 573 551 616 626 238 241 223 219 217 514 503 577 596 606 12 213 626 213 626 (1) (2) Stress Range from Equation (23) Allowable Design Cycles from Code Figure 1-9. Curve labeled "Max. FOR BOLT TIGHTENING/UNTIGHTENING CYCLE WN-WN WN-SH WN-Blind Stress.39 TABLE 12. ksi^ 250 244 <2) 471 492 264 259 293 395 241 254 Cycles.0 S m .4. Stress. Cycles.

000 = 20. it is apparent that exceeding the Code fatigue limitations for any credible operating history is unlikely.000/2 we find the design cycles to be 55.D. Recalling that the example conditions correspond to going from shutdown to operating and back to shutdown conditions. Gasket I.40 conditions then gives a peak stress intensity range for those cycles of 232.D. (2) The limitation on bending stresses imposed by the use of P is such that the allowable pipe bending stress in attached pipe ranges from 13 to 48 percent (average of 28 percent) of that permitted by Equation (12) in NB-3653.700 (c) Bolt material: SA-193 Grade B7 From the results of these calculations. (1) The bolting area is sufficient to meet the requirement of NB-3647.". S . m = 2.000 psi. Entering Code Figure 1-9. = 20.000. This result was probably obvious to many readers who recognize that in a flanged joint with adequately high preload bolt stresses.75.D.4. with S a.. Summary of Results on Bolt Areas and Bolt Stresses Calculations were made on a representative sampling of B16.1(b) which states that "Equations (3) and (4) of XI-3220 shall be used to establish minimum bolt area. = raised face O.5 . values calPb culated according to the rules of NB-3647.1.000 . the variations in bolt stresses due to subsequently applied loadings are relatively small. = pipe O.D.5 flanged joints with: (a) Rated pressures of ANSI B16. the following conclusions can be drawn.212.1968 for carbon steel flange material (b) Gasket O.. .6. . The values cited are specifically for it SA106 Grade B pipe at 700 F. y = 3. Table 4.

000 psi is appropriate for B16.6.41 (3) The limitation on bending stresses by a seemingly more rational analysis is such that the allowable pipe bending stress in attached pipe ranges from 15 to 81 percent (average of 45 percent) of that permitted by Equation (12) in NB-3653. S . (6) With a preload average bolt stress of 40. . bending and torsional stress is met. FLANGE STRESSES Flange Stress Limits of NB-3647 Flange stresses are to be calculated in accordance with Article XI-3000 of the Code.5 flanged joints. (5) With a preload average bolt stress of 40. values calculated using program FLANGE. P . ii The values cited are specifically for SA 106 Grade B pipe at 700 F. (4) Calculations indicate that a preload average bolt stress of around 40. shall be replaced by the where P "flange design pressure". P.000 psi. the Code limit of 3S for combined average. Table 9. (a) The "design pressure". with the following modifications. This initial bolt stress is permitted by the Code. /ttG pb + 4 F/ttG) + P (b) The equation for longitudinal hub stress is: fM H t 2 + 4g0 The second term on the right is an addition to the equation for SR as given in Article XI-3000.000 psi. the fatigue limits of the Code will be satisfied for any ordinarily anticipated loading history. 3 6q eq = (16M . 5.

The bending moment in the attached pipe is assumed to be zero. H o whereas in Table 14. following flanges have maximum stresses exceeding 34.950 psi. (1) In Table 13. 24" . 300.500. The calculated stresses in Table 14 are different than those in Table 13 for three reasons.500 and 4" . ni The stress limits in Article XI-3000 essentially are: SR and ST < S Su < 1. then S = 23. 400 and 600 classes. 1.42 (c) The allowable stress limits are: S . . 16" . the controlling stress is usually (SH+SR)/2.Div. If the flanges are made of SA-105 material.5S (SR + SR)/2<S (ST + SR)/2<S Stresses of Rated Pressures. the value of S b m at 100 F for SA 193 Grade B7 bolt material.300 psi at 100 F and the maximum allowable flange stress is It can be noted in Table 13 that the 1. The locations and directions of the calculated stresses are shown in Figure 2. Zero Pipe Bending Moment Table 13 shows the values of maximum flange stresses calculated according to the rules of NB-3647 for a pressure equal to the rated pressure at 100 F and with S a = Su = 35.1.400.5 x 23.950 psi: Welding Neck: Blind: 24".900.2. S H R and S 1 not greater than 1. the maximum stress is usually S + PB/4g . 4" .000 psi.300 = 34. Table 14 shows the calculated values of maximum flange stresses in accordance with Appendix II of Section VIII .5 S .

Welding Neck Stress Straight Max. OP indicates operating conditions Maximum stress is in center of blind flange. PRESSURE IS RATED PRESSURE AT 100 F. MAXIMUM CALCULATED FLANGE STRESSES. Condition 150 4 8 16 20600 HL GS OP 24200 15500 20300 20900 19100 21500 HL GS GS GS GS 13900 11300 GS OP OP OP (275) 15000 18400 17700 17000 21500 HL R HL HL HL GS OP 21200 22500 14900 18400 26000 33500 16900 13500 24 300 4 8 HL HL HL OP OP GS GS HL HL GS GS GS OP OP OP (720) 16 24 24400 HS HS 23200 23700 HL HL GS GS 42600 19100 400 4 8 HL HL HS OP 22600 24500 25700 28700 22500 22600 26600 25300 29700 HL HL HL HL GS GS GS OP OP OP (960) 23000 26900 50000 23600 OP OP GS 16 24 600 4 8 GS GS 28000 37500 HS HL HL OP OP OP HL HL HL HL GS GS GS 22700 23100 32900 33900 OP OP OP OP (1440) 23300 27800 16 24 900 4 8 HS HS 47300 27100 27000 OP GS HL R OP GS HL R GS GS 31500 34100 37900 GS OP OP (2160) 29800 27600 25900 29000 16 24 1500 4 26000 30500 29000 HL HS OP OP R HL GS GS OP OP 31400 35300 OP HL OP HL OP OP (3600) 8 25700 22600 HL OP GS 24600 26600 28500 HL R HL 27000 30400 29800 16 24 2500 R HS GS GS OP OP 24600 30600 25900 OP 4 8 12 HL HL HL OP OP OP 29700 24400 22200 HL HL HL OP OP OP 36900 28800 OP OP (6000) 24000 25100 OP . .43 TABLE 13.—. Stresses are in psi. (2) (3) GS indicates gasket seating conditions. Controlv .(1) Co„trol(2> Condition Stress(1^ Ident. Stress Hub Blind<3> (2) Max. NB-3647 RULES. (1) Stress location/directions are indicated in Figure 2. S 35. Condition Control .(2) ' Ident. .000 PSI = St Class Size Max. Stress (Pressure) Stress .—.

IDENTIFICATION OF STRESSES CALCULATED BY CODE PROCEDURE .44 Stress Identification Nomenclature ASME VIII-1 This Report FIGURE 2.

Iden. Blind Max Controlling Moment Stress^ .DIV. S a = 25. For blind flanges. R HT R' <SH + ST)/2 GS = gasket seating. but the 4" . the operating moment controls for all . (Pressure) Stress . Stress. Stra ight Control Hub .45 TABLE 14: CONTROLLING CALCULATED FLANGE STRESSES. PRESSURE IS RATED PRESSURE AT 100 F. Controlling moment for welding neck and straight hub. 1 RULES. Stress.000 PSI Class Control Size Welding Neck Stress(l) Iden. OP = operating.150 class. (2) psi 150 4 8 16 12400 8600 psi psi HR HR 14300 9600 HR HR HR HR 11000 11300 21200 22500 13800 18400 26000 33500 GS GS GS GS (275) 14400 12200 10800 12000 17800 R HR 15600 13900 12800 14500 18400 19000 14700 15300 24 300 4 8 16 HR HR HR HR HR HR HR HR GS GS GS GS (720) 24 25200 12300 12400 18900 29300 15100 13800 19300 26800 19700 22200 21800 19100 400 4 8 16 24 HT HT HR HR HR HR 15200 18500 28000 GS OP GS GS (960) HR HR 19700 22500 37500 22700 23100 33900 33900 29700 600 4 8 16 24 HR HR HR HT 17600 17100 HR HR HR HR OP OP GS OP (1440) 21600 19800 22800 900 4 HR HR R GS GS OP GS (2160) 8 16 24 R 24500 24500 20300 34100 37900 R HR R HR 31400 35300 27000 30400 29800 36900 28800 25100 1500 4 8 21300 16500 18400 16200 22600 17800 HR HR 2530u 19500 HR HR R HR OP OP (3600) 16 24 R HT 21700 20900 GS GS 2500 (6000) 4 8 12 HT HR HR 24600 19800 17100 HR HR HR OP OP 15500 OP (1) (2) HR = (St 'H ' "R""* " + SR)/2. SECTION VIII .

950 psi for the gasket seating condition. The maximum allowable flange stress is 1. 600. If the flanges are made of SA-105 material.950 psi at gasket seating conditions and 1. are made of SA-105 material. stresses due to internal pressure are completely ignored.300 = 34. i. .000 psi. It can be seen in Table 14 that all of the flanges that were unacceptable in Table 13 are also unacceptable in Table 14. It can be seen that the 24" . respectively. Table 15 shows only the stress for the condition (gasket seating or operating) which is the higher.5 x 23.5 x 17. then the allowable stress at 100 F is 17.. these are the only three flanges which are not acceptable by Code criteria for the 700 F ratings. and 900 class welding neck flanges have stresses which exceed 34.000 psi.300.300 psi at 700 F. Number Number Acceptable Welding Neck Straight Hub Unacceptable 14 17 13 10 Blind 4 23 Table 15 shows the calculated values of maximum flange stresses for a pressure equal to the rated pressure at 700 F and with S a = 35. then S m If the flanges = m = 23. Table 14 shows that the following number of flanges are either acceptable or unacceptable.46 (2) The allowable bolt stress in Table 13 is 35. While not shown in Table 15. 1 procedure. whereas in Table 14 the allowable bolt stress is 25.300 x 25.950 psi at operating conditions. (3) In Table 14. S. Section VIII . Indeed.800 psi.500 psi. the higher allowable bolt stress leads to a higher flange stress. Where gasket seating conditions control.Div.e.300 psi at 100 F and S 17.000 psi. = 26. the values of S d m for SA-193 Grade B7 bolt material at 100 F and 700 F.

S = 35. Max. NB-3647 RULES.000 PSI. Condition Stress(1) Control() Iden. GS indicates gasket seating conditions OP indicates operating conditions Maximum stress is in center of blind flange.47 TABLE 15: MAXIMUM CALCULATED FLANGE STRESSES. 26. (3) ' Stress Condition 150 4 8 20600 HL HL GS GS GS 24200 15500 19300 HL HL GS GS GS 13900 8100 16800 11900 GS GS GS (110) 13600 17500 16200 16 24 R HL HL GS 19200 18200 20400 22200 22600 HL GS GS 300 4 8 16 14800 16200 HL HL HS GS GS GS HL GS 142Q0 13600 20800 21800 16200 13600 GS GS GS (470) HL HL HL GS GS 23300 24 400 4 8 16 24 600 40600 17300 17100 HS GS GS OP HL HL HS GS GS GS GS 21700 22300 HL HL HL HL GS GS GS GS GS GS (635) 25700 47900 17600 24700 27500 21200 21300 25200 24000 21400 24800 19600 17100 24200 22100 GS OP HS 4 8 16 24 HL HL HS GS GS GS GS HL HL GS GS GS GS GS GS GS OP (940) 16300 23500 43900 23400 19100 23200 26100 21400 18200 HL HL HS 900 4 8 16 24 HL HL GS 28100 28100 HL R GS GS GS 29700 29600 GS GS GS (1410) GS GS R HS 25900 24500 26500 22400 25200 27000 25300 21000 R HL 24900 22400 30300 21500 26900 GS GS GS 1500 4 8 HL HL GS HL GS GS (2350) GS HL R GS GS GS GS GS GS GS GS 16 24 2500 4 8 12 21400 19100 22400 R HS GS GS GS GS GS HL 24000 31400 23900 21000 HL HL HL HL HL GS GS (3920) 18500 17300 HL 19500 GS GS (1) (2) (3) Stress location/directions are indicated in Figure 2. PRESSURE IS RATED PRESSURE AT 700 F.800 PSI — Welding Ne ck Class Size Straight Hub Max. Condition Control. Stress Blind(3) Max. . Stress Stress(1) Control(2) Iden. S.

Calculated maximum allowable pressures are shown in Table 16.150 class welding neck flange. flange stress = 25.600 class welding neck flange has a lower pressure rating than the 24" . 14. b = 25. Note that the 24" .950 psi.000 psi. flange stress = 34. The reason for these illogical pressure ratings is due to the Code method of calculating loads under gasket seating conditions using the equation: n (Am + V Sa w = _ (24) The value of A m is the larger of W = I(tt/4) G2 + 2TTbGm P/Sb •[• (25) and Wm2 = ^y/Sa (26) . S 3. max. S a = S a = S.950 psi (b) 100 F.000 psi. The 24" . S 3.400 class straight hub flange has a lower pressure rating than the 24" .150 class straight hub flange. = S. Zero Pipe Bending Moment It is pertinent to calculate the allowable pressure for the flanges with conditions as used in Tables 13.500 psi (c) 700 F. and 15. D = 35..400 class welding neck flange has a zero maximum pressure rating.48 Allowable Pressures. It is apparent that the "pressure ratings" are erratic and illogical. max. controlling flange stress 17.000 psi.800 psi. = 35.e. under the following conditions. (a) 100 F. i. The 24" . S D = 26.

MAXIMUM ALLOWABLE PRESSURES (PSI) Welding Neck 100 F Class Size St raight lub I Blind 100 F VIII-1 700 F NB-3647 100* F NB-3647 Yob F VIII-1 700 F NB-3647 100 F 100 F 700 F NB-3647 NB-3647 VIII-1 NB-3647 150 4 8 16 760 680 590 560 830 760 590 530 420 420 1120 620 560 560 480 1280 980 1150 990 1470 1170 1160 850 580 420 220 860 630 340 320 1350 1010 720 640 610 540 1480 1170 480 450 470 890 1050 0+ 0+ 450 400 1100 870 780 470 1310 1090 930 0+ 1210 770 650 1730 450 24 300 4 8 16 24 400 4 8 430 1830 1360 970 750 2210 1820 210 910 680 1320 1550 1330 870 1050 630 1760 86°* 570 1330 490 380 1110 910 600 560 1640 1350 1610 1350 0+ 0+ 1980 1570 1470 16 24 1250 0+ 1740 110°* 67°* 390 1290 1110 1750 1720 1200 890 2220 890 660 1650 1610 1140 1100 * 1490 2360 2320 2230 2140 450 1110 1090 770 740 600 4 8 16 1630 2160 1670 1430 1820* 1130 0+ 1600 1340 790 2070 2270 2160 1810* 450 2780 3060 2910 1470* 81°* 1060 2170 1530 1490 1650 1590 24 900 4 8 16 24 1630* °* 1520* 17 30 2980 3410 3080 3280 1050* 82°* 112°* 1480 2210 2530 2290 2540 2220 620* 650 1000 1890 1650 1480 1780 1990 2400 3560 4660 4140 4230 2480 4340 4890 5650 5110 1840 3220 2430 3220 3800 4310 3550 1200 * 1500 4 8 16 24 2960 4340 5110 5810 4780 2040* 3230* 2000* 2250 1130 2330 2070 2120 2640 3820* 3190 3960 3630 4200 3790 3460 3070 3140 2500 4 8 12 6850 8100 8740 4640 5890 6780 5090 6010 7050 8590 9450 4200* 5300 6120 5230 6380 7020 5680 7270 8340 1720* 3640 4180 4210 5400 6200 6490 + * Stress at gasket seating conditions exceeds allowable flange stress. Maximum allowable pressure controlled by gasket seating conditions. No symbol. maximum allowable pressure controlled by operating conditions. .49 TABLE 16.

where Sf ..009232 in. for the 24" . A^ = 55. Equation (29) gives: 69. i.000 x 3.2.296 -3 35.000 750.' f_ gs . Sr = K M (27) The flange stresses are proportional to M . the flange moment for seating conditions is: M = Wh = gs G |"(ir /4)G2 +2TTbGm| •] P/Sb +kA -|-~ gs .49 35."3 and [(tt/4)G2 + 2TrbGm] = 750.3257 x 9. K* Jf(7r/4)G2 +2-rTbGml P/Sb +Afc j S h„ a G (28) Solving equation (28) for P gives: 2S all P = (29) S *h h K \ (tt/4)G + 2irbGm As an example.950 psi and S = S = 35.e. h = 3.ing stress flange stress. For this flange.50 If W 1 is larger than W ml mz (it almost always is for Table 16 calculations). then: KM gs = s all .296 in. where K is the proportionality constant.55.232 x 10 = 455 psi . XX maximum (or controlling for VIII-1) flange stress and S . = allowable 3.000 psi.600 class welding neck flange with $a^ = 34..3257".. If S = K M . K* = 0. .49 in.900 . With this specific data.

950 psi. for this particular flange.040 lb. Allowable Pipe Bending Stresses at Rated Pressure Allowable pipe bending stresses as limited by the bolting are shown in Table 4 and Table 9.1 appears to be equal to the rated pressure at 700 F. W is larger than W .2 and h =2. s'. the bending moment in the pb & . .52 in.52) x 2.000 psi. /ttG eq pb and M .51 For the 24" . NB~3647. Because 37.400 class welding neck flange. written so that another limit is imposed on pipe bending stresses by the limits on allowable flange stresses.000 x 47.1(a) states Because XI-3230 is on is to be used in XI-3220 and XI-3230. "Flange Moments". the flange stress by Equation (30)is 34. = 16 M . It is interesting to note that even if y = 0.770 is greater than S = 34. K = 0.8257/2 = 37. nbGy = 138. Sf = 0. -3 . With S =35. made of material such that S = 23.K-f-f* +Ab SaV2 a (30) ' For the 24" . However.871 psi hence. These are based on internal pressure However.300 psi or less.8257 in. the flange stresses will obviously depend upon the 3 magnitude of P attached pipe.400 class welding neck flange.040 + 35.01484 in.18 psi. Equation (29) gives P = 5.770 psi. is not acceptable for any pressure using any gasket. that P_ Note that NB-3647. A^ =47. the Code evaluation indicates that this particular flange. Accordingly.01484 (138. the Code evaluation is that the 24" 400 class welding neck flange is not acceptable for any pressure.

are the allowable pipe bending stresses as limited by the bolt stress.600 class welding neck flanges have zero allowable pipe bending stress. we have followed the Code procedure for calculating flange stresses. . actually. Also shown.300 psi at 700 F) and bolts made of SA-193 Grade B7 (S at 700 F). = 26. as limited by the flange stress.5 flanges which have a long and satisfactory use history are shown to not meet the Code criteria and. even with zero pipe bending moment.5 flanges.52 The allowable pipe bending stresses. The large sizes of B16. the allowable pressure is shown as equal to the rated pressure. in Table 16. for comparison.800 psi It may be seen in Table 17 that the 24" . The Code restrictions on allowable pipe bending stresses appear to be highly questionable for large size flanges.400 class) the flanges will not meet Code criteria for any pressure with any gasket. in one extreme case (the 24" . the results give negative allowable pipe bending stresses because the rated pressure is higher than the maximum allowable pressure (see Table 16). are shown in the first two columns of Table 17. In B16.64 psi higher than the rated pressure. have never theless been used for many years in piping systems where significant bending moment loads were imposed on the flanged joints. we have arrived at some seemingly contradictory results. taken from Table 9. while probably less capable of carrying high bending stresses in the attached pipe than the smaller sizes.400 and 24" .300 class allowable pipe bending stress is essentially zero. so doing. but before round off the allowable pressure was calculated as 0. It can be seen in Table 17 that the NB-3647 rules place severe restrictions on permissible bending stresses in the attached pipe. Table 17 data are for flanges made of SA-105 material (S m = 17. Basic Theory Flange Stresses In the preceding section on "Flange Stress Limits of NB-3647". The 24" .

psi. as limited by bolt stress criterion. P = RATED PRESSURE AT 700 F. ALLOWABLE BENDING STRESSES IN PIPE ATTACHED TO FLANGED JOINT. From Table 9. Allowable pipe bending stress. Class Flange Welding Size Stress() Straight Hub Bolt NB- Stress(2) Program FLANGE (P) Neck 3647 150 4 8 28700 25000 19700 19700 15800 36400 36700 19400 25300 23900 38000 (110) 21500 21400 17700 16900 22950 23500 26400 16 24 300 4 8 16 24 32600 16600 11500 19 29300 15000 18000 13000 25000 12600 13500 9100 18700 (470) 14900 22000 17300 31000 15300 22900 28700 22100 31100 400 4 8 16 24 28400 15000 8000 0 (635) 24900 27000 20300 19300 16000 12700 24200 31300 23000 22600 18100 600 4 8 7600 16100 7700 0 (940) 13600 10300 8700 13500 16 24 900 4 14900 12900 24100 15100 14200 40800.53 TABLE 17. as limited by flange stress criterion. PIPE WALL THICKNESS = g . psi. . 27200 (1410) 8 16 24 14700 10100 5000 14000 8700 9600 9600 13900 17400 21800 33700 1500 4 8 16 24 11800 12900 16200 10900 9500 11300 12400 6900 13800 11500 (2350) 25700 27100 23800 15000 13650 10550 2500 4 8 12 9400 12000 13000 9100 11700 12800 34700 27900 25800 (3920) 10250 10400 (1) (2) Allowable pipe bending stress.

in Appendix A.000 psi.54 The observation that B16. the effective gasket width. However.5 flanges do not necessarily meet the Code criteria has been brought to the attention of the ASME Boiler Code Committee many times.000 psi In the basic theory approach.Gm)/2. as the G flange rotates the effective gasket diameter will tend to shift closer to Gq and the effective line of bolt load application will tend to shift inward. an assumption whose basis is partially indicated in Table 8. the standard response has been that B16. b . While all of these have a measure of engineering significance. we start with the assumption that the bolts are initially tightened to give an average bolt stress of 40. the effective gasket diameter. m and y. Some test data relevant to these effects are included . There is potentially a major conservatism in this value of h .h . GM = (Gq + G±)/2 and C is the bolt circle diameter. a number of somewhat arbitrary aspects are involved in the design procedure. and provide a good method of designing non-standard flanges. In the Code procedure. It should be noted that these results do not involve the gasket factors m and y.. the total design procedure of the Code is not necessarily a good way to evaluate the adequacy of B16.5 flanged joints. e. and the flange stress limits. the gasket factors.5 flanges are accepted by the Code on the basis of successful operating history. Flange Stresses at Preload Bolt Stress of 40. we will present in the following some flange stress calculations obtained using the computer program FLANGE. it is pertinent to look in a more realistic way at calculated stresses in flanges and their significance.g. G. The maximum calculated stresses are shown in Table 18. For welding neck and straight hub flanges.000 A. For this purpose. the moment applied to the flange ring is simply M = 40. The locations and directions of the calculated stresses are shown in Figure 3. where h b g G = (C .

LL LL Code 150 4 8 39360 24390 44520 27380 23190 12550 30400 22490 23220 21050 19680 12640 16 36400 37350 27160 28000 RH LL 39640 44720 31910 33940 41210 42080 38590 36980 LL LL 30040 25850 18340 24 300 4 8 LL LL LL LL LL LL 18530 32080 16 40920 66200 32610 30020 CSO LS 35340 24 400 16 24 34070 26610 20790 33040 21030 LL LL LL CSO LS LL LL LL 18300 32310 39130 42480 79020 27960 26700 44640 51610 32330 33170 35590 40290 28350 600 16 24 900 16 24 1500 16 LL LL LL RH LL 20930 20490 LL RH LS 25080 37770 31020 45370 44400 35950 35050 41860 30720 39370 66290 40780 43280 42640 40910 44530 45470 41080 40870 32420 29000 RH RH RH LS LL RH RH LL 31390 32630 30370 38810 36830 36760 28660 39920 34980 34610 30090 27390 29780 28030 22340 RH LL RH 38850 33440 42960 RH LL RH LL 24 2500 4 8 12 RH 43700 35850 41960 37380 41880 33420 30320 30730 28780 25380 22130 RH RH RH LL 31690 29510 RH LL 20380 (1) (2) Stress location/directions are indicated in Figure 3... Stress Stress .55 TABLE 18. BASIC THEORY FLANGE STRESSES (PSI) AT PRELOAD BOLT STRESS OF 40. Maximum stress is in center of flange. surface is not indicated because stresses are equal and opposite sign (pure bending). LL LL Iden. . Blind (2) FLANGE Class Size Stress Iden. Straight Hub Max. For LL and LS. Stress.U.000 PSI Welding Neck Max.U.

56 Sequence of Letters longitudinal (in hub) circumferential (in hub) small end of hub large end of hub outside surface of hub inside surface of hub (T = tangential (in ring) \R = radial (in ring) = hub side of ring face side of ring FIGURE 3. IDENTIFICATION OF STRESSES FROM COMPUTER PROGRAM FLANGE .

000 x Afax hg -© (n) G3 * Equation (31) car almost be derived by assuming that a blind flange is that the metal outside a flat circular plate of outside diameter equal to G of G is ignored. The stress at the center of a flat plate.3047 P(f) of 0. 6(1 -v2) /Wh + 1-7905^2 ' ^G With v = Poisson's ratio taken as 0. is theoretically given by n-3(3 +v) V(G)2 So 0. modeled as described above plus the assumption of simple support along diameter G.25. and that the plate is loaded by an edge moment of M/ttG.2(b) with P = 0 and W = 40. this gives 2 This is identical to the equation given in NC-3325 except that constants .57 For blind flanges.3 and 1. The "Code" stresses were calculated using the equation given in NC-3325.78 are used.000 A^: (31) p JG\ 1-78 x 40. and with pressure loading as well as edge moment loading. Table 18 shows stresses identified as "Code" and "FLANGE".

600 class flange has been frequently installed in gas transmission pipe lines. specified in B16. The question arises: Can the bolts in a 24" . G m Bolt loads are represented by line loads along and C.400 class. As can be seen in Table 18.438".873". A more significant aspect of Table 19 consists of an evaluation of the stresses shown for g o = 0.000 psi are shown in Table 19 in the column headed g 0.5 and it is of interest to see how the calculated stresses vary as g and h are varied. Let us look at the stresses in this flange (rather than the 24" . G . this makes a significant difference for the higher pressure classes.600 lb welding neck flanged joint (flanges made . Part 2 extends from G m to the bolt circle. h = 2. which is not used very much) in more detail. A.The FLANGE analysis method (3) treats the blind flange more completely.438" and h = 2. and Part 3 extends from C to the flange outside diameter.691". is that one cannot state what the magnitude of stresses are in B16. We note that the second highest stressed flange in Table 18 is the 24" . using a pipe with a wall thickness of around 0. It is apparent from Table 19 that large variations in maximum stresses ( a ratio of about 3) occurs within ranges of g and h that could be used in conjunction with the stress analysis of a 24" . C. The symbols used to define the stress location Now. The complete set of stresses calculated by FLANGE for a bolt stress of 40.600 class welding neck. The major difference between the two methods is that the FLANGE method takes into account the flange material outside of the gasket diameter. The 24" .691". An important conclusion. neither g nor h are and directions are identified in Figure 3.600 class welding neck flange.5 welding neck flanges unless one specifically states the values of g o and h used in the calculations. the circular plate is divided into three parts. m Part 1 extends from the center to the gasket diameter. therefore.

290 4.980 -36.650 -9.290 53.450 8.150 25. h.090 4.760 -24.530 6.560 -26. Following symbols are analogous thereto.310 7.970 -40.440 -25.860 -26.000 87.680 0 MLL MCL 12.990 16.950 -1.060 LLO LLI CLO 26.560 32.280 Membrane Stresses 0 MLS MCS 0 0 0 0 33.720 64.0" (1) LSO h = 2.650 35.480 -34.370 -47.310 -34.691" h LSI CSO CSI 66.250 23. .309" = 2.200 -9.580 -23.470 24.910 25.360 22.100 21. MLS is the membrane stress in the longitudinal direction at the small-end of the hub.070 -800 24.590 14.000 PSI Stress g o = 0.160 34.59 TABLE 19.980 -26.930 88.330 (1) See Figure 3.530 17.438" = 2.460 -43.300 11.000 -103.720 -88.010 30.240 0 23.180 MT MR -9.790 0 56.100 0 25.873" h = 4.360 0 37.640 7.691" go = 0.940 -22.870 103.860 30.330 -9.260 16. ON STRESSES (PSI) IN A 24" .250 -8.691" °o g = 1.250 -24.760 14.480 35.370 36.980 20.330 47.930 30.873" g h = 0.990 -30.340 28. gOJ AND HUB LENGTH.300 4. EFFECT OF VARIATION OF PIPE WALL THICKNESS.940 -50.670 13.700 15.670 4.900 CLI TH TF RH RF 30.890 34.290 -66.700 4.240 4.030 -19.650 17.0" g h = 0.873" = 2.600 CLASS WELDING NECK FLANGE LOADED WITH BOLT STRESS OF 40.900 26.

688".344/3. at high bolt loads. The maximum calculated stress is then 103. If both of the flanges in the joint are capable of sustaining the desired initial bolt stress without gross yielding.625)/2 = 3.25. bolts of SA-193 Grade B7 material) with attached pipe of 0. the answer is yes. the flange only place in the flanged joint where. let us assume that at high bolt loads the effective gasket diameter is (Gq . eventually.875 = 31. In borderline cases. no difficulty * was encountered and. tends to rotate and one would expect a decrease in hG as the gasket load shifts outward and the bolt load shifts inward. it may take 5 or 6 rounds of individual bolt tightening. wherein the flange strength is increasing due to strain-hardening and stress redistribution. Indeed.25" .344" rather than (33 . . insofar as we are aware. the bolts can be tightened to the desired stress in 3 or 4 "rounds" of individual bolt tightening.000 psi? Despite the calculated stress of 103. rather than at C = 33".000 psi. The magnitude of the calculated stress of the small end of the hub is deemed to be an accurate evaluation of the elastic stress in an actual flanged joint. Tightening flanged joint bolts is ordinarily done by step-tightening in a more-or-less regular criss-cross pattern on individual bolts..8125" = 26.625".4375)/2 = 2. To illustrate the possible magnitude of these shifts. rather than at the gasket mean diameter of 25. the outer edges of the flange rings contact each other. Continued tightening will simply rotate the flange ring until. Figure A10 in Appendix A gives some evidence of the inward shift of the bolt load.4375". the calculated stress should be accurate. Also assume that the bolt load acts at (C-d) = 33 . The value of hQ is then (31.688 = 65.125 .gasket width/2) = 27. the stresses at this location are probably the However.000 x 2.500 psi rather than 103. If the flange strength is inadequate.125.60 of SA-105 material.1.26. these flanged joints are performing satisfactorily some 15 years after installation.000 psi. One of the authors has supervised the installation of such flanged joints. provided h is actually as assumed in the analysis. theoretically. it will be impossible to reach the desired bolt stress.438" wall thickness be tightened to an average bolt stress of 40.

the characteristics of the stresses and their distribution help to explain the load capacity of the flanged joint. the highest stress is a longitudinal bending stress at the small end of the hub. Table 19 shows that.5 times the allowable stress. i. The circumferential stresses in the hub.Div. The tangential stresses. Because of the large variation in stresses throughout the flange structure.e.860 psi. they decrease to zero at the outer edge of the flange ring. provided there is reserve strength in the ring as shown by lower radial and tangential stresses. this stress has decreased to 26. which are the most significant of the stresses with respect to gross yielding. 1 in the use of a "controlling stress".61 In addition to the bolt load and gasket reaction shifts. From a theoretical standpoint. also decrease substantially from the small end of the hub to the large end of the hub. one would not expect gross yielding to occur when the highest stress reaches the yield strength of the material. for a welding neck flange. The stresses in the hub are permitted to reach 1. the calculated stresses can vary significantly * Some recognition of this aspect is included in Section VIII . . Table 18 shows high calculated (elastic) stresses for an applied bolt stress of 40. Some local yielding can and probably does occur. As shown in Table 19. but the lower-stressed portions of the flanges have the capability of picking up the load . The radial stresses in the flange ring are shown in Table 19 at the inside diameter of the flange. The membrane longitudinal stress in the hub is zero.000 psi. In summary. . are relatively low.50. + 30.. a limit-load or elastic-plastic analysis (taking into account the bolt load and gasket reaction shifts) would be required to evaluate the load capacity of flanged joints.940. at the large end of the hub.340 psi. both membrane and bending.

Test No. to the detailed stress distribution in a flanged joint. just as any other piping product (elbows. requires that a flanged joint be analyzed under the rules of NB-3650. 3) in which measurable yielding of the flange occurred even though the maximum calculated stress was less than the yield strength of the flange material.000 35.300 20. in the absence of a limit load analysis.62 with g and h.5 flanged joints in which the maximum calculated stresses were significantly higher than the flange material yield strength. Operating history with the 24" .000 25. However. N Temp. Appendix A includes the results of a number of tests on B16. we will discuss an analysis of the flanges.600 class welding neck flanges shows that high calculated maximum stresses do not necessarily indicate that a flanged joint with such high calculated stresses is not capable of withstanding an initial bolt stress of 40.500 F 100 400 . psi 23. it is deemed appropriate to categorize the stresses as primary plus secondary. tees) purchased to an ANSI standard must be.500 8. Design Cycles. N . 10 10 Min. NB-3600. it is necessary to categorize the stresses shown in Table 18 so that a fatigue analysis of the bolt tightening/untightening cycle can be performed. Sm . it is not possible to quantify the load capacity of a flanged joint. The high load capacity can be attributed to bolt load/gasket reaction shifts and Unfortunately.600 Welding__Neck_ ^Straight Hub Max 35. The analysis of the bolting in flanged joints was discussed earlier. 2. then are as follows.000 Min 650 350 Max 25. a result for which the author has no explanation. For flanges made of SA-105 material.000 psi. the allowable numbers of bolt tightening/untightening cycles. Because the stresses are mostly local bending and because the bolt load itself is displacement controlled. there was one test (Table A9.000 Blind Max. Now In terms of an NB-3650 analysis of the flanges.000 Min. 8.000 2. as interpreted by the authors.

63 The above tabulation is based on stresses shown in Table 18 and a stress concentration factor (K-index) of 2. it appears that permissible bolt tightening/untightening cycles are far greater than would be anticipated for any normal B16.5 . see footnote 1 to Code Table NB-3222-1. the maximum stress at pressure occurs on the small end of the hub. but the maximum stress range of 5. the maxi mum stress range and/or maximum stress is not necessarily at the loca tion where stresses are maximum due to preload.5 flanged joint instal lation. As was found in the fatigue analysis of bolting. The change in stress is due to (a) The moment on the flange ring changes because of internal pressure and. therefore. However. the stress from the moment changes. Table 21 shows the variations of all stresses in the 24" .0 to account for such aspects as the bolt holes and the fillet between the hub and ring. outside surface. (b) The stress produced by pressure is added to the stress due to the change in moment by pressure. bl = 40.400 psi represents a hydrostatic test pressure of 1. we will consider only pressure loading and the range of stresses resulting from pressure loading. where the pressure of 5.000) Plus Operating Conditions At this point.1500 class welding neck flange. The temperature to use for the tightening/untightening cycle.300 class welding neck flange. Stresses at Preload (S.550 psi occurs on the inside surface. Table 22 shows the variations in all stresses in the 16" . The combination of these two effects may either increase or decrease that stress which was a maximum under preload only. if operating temperatures get up to 700 F. is taken as 400 F. Table 20 shows how the maximum stresses change due to a pressure equal to the primary rating pressure. in this case.

64 TABLE 20. PLUS PRESSURE Weld ing Stress Due Meek to: Stress Straight Hub Stress Due to: Class Size Sbl 39360 24390 P Sb2 + P 39600 24930 35730 (1) Sbl 44520 27380 39640 44720 31910 33940 41250 42080 38590 36980 44600 51610 32330 33170 42640 40910 S + P P b2 Stress(1) Iden. S. 150 4 8 16 130 250 LLO LLO RH LLO 320 520 560 770 44820 27880 39470 36400 37350 27160 28000 -275 360 120 24 300 • 4 8 16 24 37160 27460 28170 44460 32430 34290 41110 LLO 430 650 690 850 LLO LLO LLO 160 4380 2570 130 LLO CSO 40920 66200 32610 44810 67610 32950 30270 46810 80830 31520 27050 38060 68400 LSO 41740 39150 37530 44540 51330 33020 33820 41180 41070 LLO 400 4 8 16 24 LLO LLO CSO LSO 540 790 850 1040 LLO LLO 30020 42480 79020 160 4820 LLO LLO 3080 3280 220 600 4 8 16 24 27960 26700 39370 66290 LLO LLO RH LSO 650 870 -670 1180 LLO LLO RH LLO -955 2940 900 4 8 40780 43280 33810 36830 36760 28660 -1000 -1080 -1210 3200 -1520 40060 42330 37620 39900 35780 29760 38550 RH RH RH 44530 45470 41080 760 -860 -910 1220 45320 44270 39600 LLO RH RH LLO 16 24 1500 4 8 16 LSO 40870 38850 33440 41490 37840 34870 41550 42220 37680 RH LLO -1300 1180 -1320 RH LLO 540 -1700 39920 34980 34610 30090 RH RH 42960 43700 35850 RH LLO 24 2500 4 8 12 -1880 -1980 -2060 -2090 33370 33400 28910 1440 1240 -1900 1380 RH LLO RH RH 31090 29510 29790 31500 RH LLO 27390 26150 (1) See Figure 3 . LLO LLO LLO LLO Iden. FLANGE STRESSES WITH PRELOAD.000 PSI. EQUAL TO PRIMARY RATING FRESSUR = 40.

65 TABLE 21. 10_6in.156 —*0 8. M. analogous thereto. S = 40. P = 600 PSI. psi (1) LSO LSI 66290 65450 2940 4710 5090 5620 530 1690 2750 3100 2270 68390 2100 -66290 53670 13900 26980 CSO CSI t65450 52990 13720 26640 -26640 20030 4050 21690 -60740 58080 19340 27170 5550 4410 5440 190 2030 LLO LLI -26980 20290 4100 21970 -24950 22780 7150 CLO CLI 2490 3050 1990 TH 23960 -39610 31110 -23770 3825 TF RH RF -40560 32580 -40050 32170 -23570 0 440 -1060 -200 950 1470 100 -23870 0 MLS 3825 — MCS 33790 0 12200 33360 0 12040 5360 1110 38710 1110 — MLL MCL — 2925 14965 — MT -9300 -9180 4300 1355 -630 -7825 3670 — MR 4360 (1) See Figure 3. .600 CLASS WELDING NECK FLANGE.053 600 Stress Range .8.053 0 0 600 8. FLANGE STRESSES IN A 24" .000 PSI. psi —>.-lb P. MLS is the membrane stress in the longitudinal Following symbols are direction at the small end of the hub.

290 5.710 1.442 0 1.220 -1.830 5.250 11.730 -5. .720 -37.830 17. analogous thereto.710 6. P = 5.500 19.620 39.500 19.670 5. Sh1 = 40.170 6. M x 10 in.270 910 - - MT MR -8.920 -28.880 12.270 4.540 -23.020 0 14.440 -14.150 -- (1) See Figure 3.870 -28.240 -31.66 TABLE 22.920 11.330 8.720 4.060 20.690 11.870 16. MLS is the membrane stress in the longitudinal Following symbols are direction at the small end of the hub.920 20. psi (1) LSO 5.000 PSI.290 -6.600 11.440 12. -lb -.490 0 0 1.120 -4.860 20.300 -5.980 18.840 41.560 6.740 31.950 0 31.930 .520 6.500 35.490 20.190 4.250 -28.770 35.140 11.130 CLO CLI 29.450 0 -- MLL -- MCL 5.360 12.950 5.070 -6.010 13.310 12.1500 CLASS WELDING NECK FLANGE.980 25.940 39.270 29.820 19.590 28.600 7.400 PSI.830 -29.440 20.440 -8.050 -36.290 -- MLS MCS 15.850 4.-1.400 Range LSI CSO CSI LLO LLI 14.400 11.490 Stress P.270 28.240 5.130 -33.100 2. FLANGE STRESSES IN A 16" .490 13.620 17.140 TH TF RH RF 9.560 10.920 0 16.920 -14.

1.T^ = -50 F. This gives a decrease in flange stresses . The computer program FLANGE evaluates the effects of change in modulus of elasticity of the flanges.600 psi at the location of maximum stress due to bolt load.55 x 10 in.6 times the 100 F rating.-lb.32 x 10 5 in. the temperature difference between flange-ring and bolts (which gives a change in bolt load) and temperature difference between the flange-ring and the hub (which gives both a change in bolt load and flange stress). T.-lb (preload to S = 40. and gaskets.330 psi (16" - 400 class WN). the allowable operating pressure cycles (Code Figure 1-9. stresses due to temperature effects must also be evaluated.000 psi. the stress range for the hydrostatic test does not exceed 3.67 times the 100 F rating pressure. . we add the stress due to the thermal gradient to obtain the stress at the assumed conditions. 21. the largest stress range is 4. stress amplitude = 10. the maximum stress range occurs at CSI.330 = 15. The main conclusion that can be drawn from Tables 20.). Because the stress from Equation (32) is limited to S m for the pipe material. S m for SA-106 Grade B pipe at r r 100 F is 20. In doing a complete NB-3650 analysis. and 22 is that the stress range due to changes in pressure is small. In Table 20. The stress range is proportional to the pressure range hence. attached pipe as calculated by the equation: A few spot checks suggest that the stress range due to pressure will not exceed that in the PD s-^ loading will be large. it is apparent that the number of design cycles of pressure For example.000 psi) to 3. bolts.6 x 4. for a hydrostatic test of 3. . Table 23 gives an example of stress ranges where it is assumed that ^2^1 = °-80 (for a11 parts of the flanged joint) and that the average temperature of the hub (T^ ) is 50 F lower than hp the average temperature of the flange ring (T. t hp . In the example. f The combination of change in modulus and the thermal gradient changes the moment from 4. for example. the maximum stress location changes from RH at zero pressure to CSO at the test pressure..000 psi) is greater than 106.

44.220 -7.080 -- MT MR -5.390 0 32.68 TABLE 23.010 800 — (1) See Figure 3.520 13. M x 10 5in.140 18.160 0 32.000 psi.010 -32.030 -- TH TF RH RF MLL MCL 5. Sbl =40.600 7.530 -45.940 13.55 — 0 3.180 4.740 0 -480 480 -4.280 -24.260 34.560 -5.760 2.290 -22. P=0.080 -32.200 6.470 -3.280 -1.670 8. MLL is the membrane stress in the longitudinal direction of the hub.880 -31. psi CLO CLI.860 3.380 -1.32 Stress Thp " Tf F Stress^ LLO LLI -50 -50 Range. FLANGE STRESSES IN A 4" .780 1.190 -24.520 9.490 13.670 10. .420 5. THERMAL GRADIENT (Th -T)=-50 F.130 0 12.020 31.580 -8.530 18.490 -32.990 -1.810 150 3.430 5.470 9.410 -22.-lb —-4.900 CLASS STRAIGHT HUB FLANGE.32 0 0 3.010 9.920 42.960 -6. Following symbols are analogous thereto.660 8.420 4.

the stress range is quite small. For pressure loading. To proceed with the NB-3650 analysis of blind flanges . we must categorize the stresses shown in Table 25. The Code limit on ° m P. except that S pb =0.3P + 1.2(b) with P = primary rating pressure and W = 40. B16.5 S ranges from 34.950 at 700 F. It is apparent from Table 25 that the stresses due to pressure are less than the allowable stresses. The stresses in Table 25 are for a pressure equal to the primary rating pressure.78 x 40.950 at 100 F to 25. d . It is apparent from Table 24 that.000 x A^ x h W computer program FLANGE. G3 (32) The stresses identified as "FLANGE" were calculated using the The treatment of blind flanges was described under Equation (31). SA-105 flange material) is around 40.5 S . primary membrane (which is zero) plus Primary bending.category stresses is 1.000 cycles.5 welding neck and straight hub flanged joints should be acceptable under the analysis requirements of NB-3650.69 Table 24 gives an example of stress ranges for the assumed set of operating conditions shown in Table 7. only. even for this severe set of assumed operating conditions. the stresses would be 2.000 A^: 0. the FLANGE analysis gives significantly lower stresses than those obtained using the Code Equation (32). Calculated stresses in blind flanges are shown in Table 25.0. The stress due to pressure is a local bending stress in the center of the flange and it appears appropriate to classify it as P . The main conclusion from Tables 20 thru 24 is that. particularly for the higher pressure classes. It can be seen in Table 25 that. the pressure acts on Part 1 The calculation method recognizes that the bolt load at pressure is not necessarily the same as the initial bolt load. the design cycles for this loading (with a stress concentration factor of 2. m For SA-105 flange material.4 times the values shown. 1. for any credible set of operating conditions. . at the 100 F rating pressure. The stresses identified as "Code" were calculated using the equation given in NC-3325.

450 1.590 4.660 17.000 PSI.390 23.70 TABLE 24.740 3.910 -29. (T hp . Sbl = 40.550 — — TH TF RH RF 20.810 16.200 20.050 100 36.000 60 2.4.560 -650 -3.560 -2.810 -42.767.580 -10. (Tfe .930 4.890 17.420 6. FLANGE STRESSES IN A 24" .150 -8.100 1.040 760 MLS MCS 40.500 1.650 0 .200 60.382 0 0 0 0 0 2.860 730 -300 -10.310 42. psi T.500 - _ (1) See Figure 3.290 14.390 14.820 -1.- MT MR -3.350 5.450 0 280 -10.200 -66.490 -26.910 18.382 300 0 300 100 Stress . E2/E1 = 0.T_.080 -10.570 10. F f (1) 100 Range.730 1.990 1.100 21. hp -*• —*0 0 2.670 -3.T ) = 50 F.520 -280 49.690 38.260 13.230 0 -600 0 19.780 1.220 16.520 -41.570 -16.350 0 3.299 P.850 15.390 330 10.350 16.680 5.340 0 22.570 4.670 -36.570 -10.610 . psi Stress LSO LSI CSO CSI 66.v = 100 F MxlO" 6in. P 300 psi.300 CLASS WELDING NECK FLANGE.750 5.030 20.030 -5.490 0 8.740 -14.670 33.420 5.350 11.190 1. .130 -29.- MLL MCL _ _ 9.370 16.470 1.620 -15.590 1.040 760 1.120 -16.--lb-*. MLS is the membrane stress in the longitudinal Following symbols are direction at the small end of the hub. analogous thereto.750 29.030 15.900 LLO LLI CLO CLI 1.290 0 11.870 2.570 4.370 -23.480 29.

Sbl = 40. BLIND FLANGE STRESSES(PSI) WITH PRELOAD. PLUS PRESSURE EQUAL TO PRIMARY RATING PRESSURE FLANGE CODE Stress Due to: Stress Due to: Class Size Sbl 19680 P Sb2 + P 20620 14930 32880 30270 Sbl 23190 12550 30400 22500 P Sb2 + P 24760 16050 37320 31390 150 4 8 16 24 1300 1570 3500 6830 8890 1770 3350 5570 12640 30040 25850 3040 6030 8010 300 4 8 16 24 18340 18530 1440 2850 4830 7210 1590 19450 20530 34700 36790 22260 20400 35300 23220 24990 24400 21050 35340 32080 33040 21030 18300 32310 39130 40910 42330 28560 34070 26610 20790 8260 1950 400 4 8 2850 5210 8080 3360 6020 9260 24150 41610 16 24 35590 40290 43350 22460 49550 30800 600 4 8 16 20930 20490 32420 29000 31390 1980 3090 5370 6780 2160 28350 25080 37770 31020 2450 3700 6270 7810 2700 22750 35520 33110 28780 44040 24 900 38830 48070 4 8 16 24 32990 34940 34030 33130 29970 24610 33310 31720 27150 24240 22500 45370 44400 35930 35050 41860 30720 41960 32630 30370 29780 28030 22340 30730 28780 25380 22130 3490 5900 5260 4250 6910 6200 3060 3370 48650 42860 41250 44920 34090 1500 4 8 16 24 2430 2760 3540 4270 4880 2560 2950 46230 42260 44440 4060 2010 37380 41880 2500 4 8 2380 33420 30320 36370 12 20380 2360 2150 32470 .000 PSI.71 TABLE 25.

5 flanges should be acceptable under the analysis requirements of NB-3650. to a good first approximation. However. We will confine the dis cussion to welding neck flanges because the limitations will be less severe for straight hub flanges. that the flange stresses are changed. i. The effect of this load on changing the bolt load is included in the computer program FLANGE. see a. we can say. B16. First. we reached the conclusion that several widely used flanges could not take any pipe bending moment. A fatigue analysis leads to the same conclusion reached for welding neck and straight hub flanges. Allowable Pipe Bending Stresses at Rated Pressures". in the authors' opinion. . the stress range due to pipe bending moment range is simply S . We will now discuss the effect of pipe bending moments. as a result of applying a pipe bending moment.e. for any credible set of operating conditions. only in the hub region. as limited by flange stresses.e. we obtain a stress of S .. The change in bolt load is theoretically significant but the change in moment acting on the flange ring is small. at the small end of the hub because the stress range at the large end of the hub is obviously smaller.72 The stress due to the bolt load and the sum of stresses due to bolt load and pressure are appropriately categorized. at the small end of the hub and S (g /g ) at the large end of the hub. We will now consider appropriate limits to pipe bending stresses. we note that at the junction of the flange to the pipe. i. Considering the pipe bending moment cycle separately. Because flange stresses are proportional to the moment acting on the flange ring. we would only pb consider the range of S .. . as primary plus secondary stresses. In the preceding section on "Flange Stress Limits of NB-3647. Effect of Pipe Bending Moment on Flange Stresses In the preceding we have discussed operating conditions of pressure and temperature effects. as judged by the NB-3647 criteria.. in Equation (9) and Table 5.

300 psi (at 700 F) for the 99 minor stress cycles.35 for the major cycle and 1. and is located at the small end of the hub. As a simple example.600 psi (at 400 F) for the one major stress cycle. column headed Bolt Stress.200 psi (See Table 17.020 + 24. is equal to 79.1.200 24.35 x 1. At this point.8/2 = 218.200 x 1. Table 18 shows that the stress due to the preload.2-1 for an "as welded" is then (79. and S = 17. the usage Stress Range of Cycles 1 79.00 for the minor cycles.73 in combining the bolt tightening/untightening cycle. Let us assume an operating history consisting of a bolt tightening.200 99 Assuming the flange material is SA-105 and that S m = 20. The value of S However. we will use the K-index of 1. = 40. and other operating conditions with the pipe bending moment. let us consider the 24" ..780 psi for the 99 minor cycles.3) is 2. 99/60. .014 hence.200 psi. The total usage fraction is 0. Turning to the Code Figure 1-9.8/2 = 21.310 psi for the one major cycle and 24. we are analyzing the flanges and not the girth butt weld between flange and pipe.400 lb welding neck flanged joint with S (range) limited by the bolts to 24. In general. S.200) x 2. 100 cycles of S = 24. Program FLANGE).8 given in Table NB-3683. for this example. the stress at the large end of the hub might be controlling. girth butt weld.020 psi.000 for the minor cycles. then the value of K (See NB-3228. fraction involves the cycles Number Accordingly. we must select a suitable stress concentration factor. and then bolt untightening. the postulated combination of bolt tightening/untightening with intervening pipe bending stress cycles can be repeated about 70 times in-so-far as flange stress limits are concerned. we find that the usage fraction is 1/80 for the major cycle.020 + 24.000 psi.

However. F Number Aceeptable Straight Welding Neck Hub 27 10 27 Blind 23 Number Nnt Acceptable Welding Straight Neck 3 14 3 Hub 0 17 0 Blind 4 23 0 NB-3647. This.D. y = 3. m = 2. data indicate 3.75.1 700 .000. Gasket I. (1) The number of acceptable flanges with zero pipe bending moment is tabulated below.700 (c) Bolt material: SA-193 Grade B7 From the results of these calculations. checks adequately with the Code procedure design life cycles of 60.D.5-1968 for carbon steel flange material.D. A total number of 27 flanges of each type were Code Reference Temp.200 psi. Summary of Results on Flange Stresses Calculations were made on a representative sampling of B16.1 VIII-Div. range of 24.74 There are no test data to check the fatigue life due to bolt tightening/untightening cycles.400.D. = raised face O. 1 100 100 24 13 24 4 27 NB-3647. there are some test data given The test in Appendix A for fatigue under cyclic pipe bending moments. the following conclusions can be drawn. checked. specifically SA-105 for checking Code flange stress criteria (b) Gasket O.5 flanged joints with: (a) Rated pressure of ANSI B16. = pipe O.000 cycles to failure for an S . divided by 20 to correspond to the Code factor of safety of 20 on cycles.

are erratic and in 8 flanged joints the allowable pressure is zero. they may either decrease or increase. .1 flange stress criteria. (4) Basic theory calculations of flange stresses with 40.5 flanged joints. (6) Application of the NB-3650 type of analysis indicates that. (3) Allowable bending stresses in pipe attached to flanged joints. (5) Basic theory calculations indicate that the stresses at points of maximum stress due to bolt load are changed only slightly by internal pressure. with 40. as evaluated by NB-3647.000 psi initial bolt load give high calculated stresses. for those flanges which are not acceptable at the rating pressure. An ex planation of why the flanges are capable of withstanding the bolt loads without gross yielding is presented.000 psi initial bolt stress.75 (2) The allowable pressures by Code criteria. the fatigue limits of the Code will be satisfied for any ordinarily anticipated loading history.5 flanged joints. it is apparent that the Code flange stress criteria are not accurate for evaluating the pressure capacity of B16. is zero for 3 flanged joints and very low for most of the flanged joints. In view of the satisfactory operating history of B16.

1 to read: NB-3647. followed by the reason for the recommendation. hand to follow the discussion. (2) Change NB-3647. It is necessary to have the Code at (1) Change NB-3651. or . In the following. analogous to the Code format for all other piping products.3000. (a) Flanged joints manufactured in accordance with the standards listed in Table NB-3132-1. including the use of the appropriate allowable stress given in Table 1-7. Reason: We wish to inform the Code user that. For analysis of flanged joints. changes contained in Appendix C are cited.1. while stress indices are not given for flanged joints. as limited by NB-3612.76 6.2 Piping Products for Which Stress Indices Are Not Available. RECOMMENDATIONS Specific recommended Code changes are included herein as Appendix C. special guidance is given in NB-3658. (b) Flanged joints not included in Table NB-3132-1 shall be designed in accordance with Article XI . shall be considered as meeting the requirements of NB-3640.1 Flanged Joints. Reason: This subparagraph is written so that it gives requirements for pressure design. For other piping products for which stress indices are not available. The Code user either: (a) Selects a standard flange with appropriate pressuretemperature ratings. see NB-3658.2 to read: NB-3651. see NB-3680.

go to NB-3200 Experimental analysis .5 (1968) and using a bolt material having an S m . bolting.1(b).000 psi may be analyzed in accordance with the following rules or in accordance with NB-3200.go to Appendix II The proposed NB-3658 is intended to gather in one place the rather specialized requirements for the analysis of flanged joints and to give relatively simple and explicit rules for ANSI B16.5 flanged joints. by the last sentence in NB-3647. =0.1. The proposed NB-3647. are handled under NB-3650. as the Code is presently The present written. Other flanged joints shall be analyzed in Reason: Analysis requirements for flanged joints.value at 100 F not less than 20.77 (b) Designs the flanges for the design pressure by the Code procedure. invokes operating condition evaluations which. (3) Add a New Paragraph NB-3658 to Read as Follows: NB-3658 Analysis of Flanged Joints Flanged joints using flanges. The analysis of the flanges (not the bolts or the weld between flanges and pipe) under NB-3650 seems to follow the path: NB-3651 . go to NB-3680 NB-3680 .1(b) because non standard flanges are rarely used in piping systems. . because no excep Article XI-3000 uses stresses for Class 2 or Class 3 components and the stress limits are different (in some cases.Experimental or theoretical analysis Theoretical analysis .go to (d) . The authors considered deletion of the proposed NB-3647. accordance with NB-3200. and gaskets as specified in ANSI B16.1.1(d).1(b) is not the same as the present NB-3647. are scattered about and difficult to interpret. for all other piping products. even for P tion is made to Article XI-3000. more conservative) than the present NB-3647. NB-3647.No stress indices.

5. the first sentence could be deleted and the authors would have no strong objection to doing so.5. NB-3233 simply says: "Use NB-3232 for upset conditions". From this standpoint. The quite extensive portion of this report on "Bolt Areas and Bolt Stresses" indicates that the rules of NB-3232 are readily met if the bolts are SA-193 Grade B7. In addition. Specific reference is made to NB-3232 (rather than NB-3230) because it covers normal conditions. NB-3234 and NB-3235 are not referred to because the proposed NB-3658. they are not mandatory. basis that (a) The sentence was left in on the the report evaluations do not cover all conceivable combina tions of loadings. and (b) the bolting evaluation is fairly simple to carry out using a computer program such as FLANGE. NB-3658.000 psi at 100 F is given in Appendix B. NB-3200 m The reason for restriction to bolt material with S 20. nor to flanges designed in accordance with Article XI3000.3 cover the emergency and faulted conditions. (15) The first sentence retains the present Code requirement now contained in NB-3647. giving rules for ANSI B16.5 will probably cover most needs. However. . Accordingly.1 Design.2 and NB-3658. The bolting shall meet the requirements of NB-3232. Normal and Upset Conditions (a) Bolting. (b) The data given in this report gives a basis for the proposed analysis rules for ANSI B16.78 The first sentence was motivated by: (a) What few flanged joints are used in Class 1 piping systems are likely to be ANSI B16. However. that basis does not necessarily apply to other standard flanges that are or may later be included in Table NB-3691-1. Mfg < 3125 (Sy/36)C ^ Reason: . the limitations given by Equations (15) and (16) shall be met.1. can be used.

Equation (15) (a) (b) is conservative with respect to available test on joint leakage.2). A^ = total cross-sectional area of bolts at root of thread or section of least diameter under stress. is a good approximation of the moment capability (as limited by joint leakage) as determined by the computer program FLANGE with an assumed bolt prestress of 40. in. relief valve steady-state thrust.bending or torsional moment (considered separately) as defined for Mf but including dynamic loadings. in. The value of (S /36) shall y not be taken as greater than unity.-lb. S = yield strength (ksi) of flange material at design temperature (Table 1-2. If cold springing is used. Mfd < 6250 (S /36) C A^ where . in. = bending or torsional moment (considered separately) applied to the joint due to weight. (c) except for the 150 class.79 The development of Equation (15) is described in Appendix B. and other sustained mechanical loads. As discussed in Appendix B. C = diameter of bolt circle. . sq.8. where it is identified as Equation (B5). the moment may be reduced to the extent permitted by NB-3672. (16) Mfd . (d) tends to discourage the use of the larger sizes of 150 class joints by giving a small permissible moment. thermal expansion of the piping. gives permissible moments essentially equal to or higher than permitted by the rules in the present NB-3647.000 psi.1. in.-lb. M. sustained anchor movements.

this may be the major source of moment loading on a flanged joint. (Ql) Why is relief valve thrust not included? The only application of flanged joints in Class 1 piping known to the authors involves relief valves. can be taken for "cold springing" (Q3) Why are torsional moments ignored? (Q4) Why are dynamic loads (e.8 (A3) As discussed in Appendix B.-lb. Until such time as axial loads are required to be con sidered for all piping products. but they can be considered separately from bending moments and the limit used for bending moments can also be used for torsional moments. NB-3647. if any.1. Further. (A4) Presumably the reason dynamic loads were ignored in the definition of M is that they are of short-time duration and flanged-joint leakage would be small during that time. In the present. The authors' answers to these questions. torsional moments should not be ignored. provided these dynamic loads did not break or yield the bolts or flanges.80 Reason: The present NB-3647. also (A4) See (A2) Credit should be given for "cold springing" to the extent provided in NB-3672.1 relates P eq to both M and F = axial load. Several questions arise concerning the definition. earthquake) ignored? (Q5) Why is moment loading due to use of the flanged joint to "pull out" fabrication mismatch ignored? In many commercial piping installations. M is defined as: M = bending moment applied to the joint due to weight and thermal expansion of the piping. This is the only piping product for which NB-3600 requires consideration of axial loads. (Q2) What credit. the .g.. in. and resulting recommenda tions are: (Al) The "steady state" relief valve thrust should be included. it is recommended that F be deleted from consideration for flanged joints.

However. Reason: The first sentence is based on one of the major aspects developed in the report. To provide a control for this. moments and thermal gradients and hence. there is nothing to guard against a very high dynamic moment which might cause the bolts or flanges to yield or crack.e.81 flanged joint would be "tight" after the dynamic loading stopped. there is no point in checking the flange stresses under NB-3650 unless we go to the extreme of checking for bolt tightening/ untightening cycles.1. are due to the bolt prestress load and not due to either pressure or moment loads.5 flanged joints meeting the requirements However. Equation (16) is included in the proposed NB-3658. other than at the pipe-to-flange weld area. as pointed out by the second sentence. using appropriate stress indices from Table NB-3683. (b) Flanges. to place a limit on static-plus-dynamic moments. any credible number of bolt tightening/untightening cycles gives only a small fatigue usage fraction. the stresses at the welds may vary significantly due to pressure. of NB-3612 are not required to be analyzed under NB-3650. i.2-1. This applies even for blind flanges with pressure loading. The basis of Equation (16) is discussed in Appendix B [identified as Equation (B6)] along with calculated pipe bending stresses resulting from its use. (A5) The quality controls for nuclear power plant piping system fabrication should be sufficient to prevent fabrication misalignment pull-out being a major source of moment loading on flanged joints. Flanges of ANSI B16. that flange stresses. should be analyzed just as any other weld in the piping system should be analyzed. While the equation is quite conservative. NB-3653. the pipe-to-flange welds shall meet the requirements of NB-3652. see Table 25. and NB-3654. . Accordingly... it nevertheless permits relatively high dynamic stresses in the attached pipe for all except the 150 class. As discussed in the text. However.

.2(b) shall be met. rd Mfd' C' Sy' the lim±tation on (S /36) and A^ are defined in NB-3658.0 times the design pressure.2(c) is the same as the present NB-3655. where Pfd and Mfd are pressures (psi) and moments (in. The limitation given by Equation (17) shall be met.-lb) occurring concurrently during the faulted condition.1. in. pfd = pressure (psi) during the emergency condition concurrent with M£. the pressure limit is deemed appropriate provided Equation (17) is also invoked.82 NB-3658.3 Faulted Conditions (a) The pressure shall not exceed 2. using a stress limit of 2. (b) The limitation given by Equation (17) of NB-3658. [identified as Equation (B7)] along with numerical examples of its application. As discussed in Appendix B. (c) Pipe-to-flange welds shall be evaluated by Equation (9) of NB-3652.2. NB-3658.1(a).5 times the design pressure..2 Emergency Conditions (a) (b) The pressure shall not exceed 1. The basis for Equation (17) is discussed in Appendix B. (17) where D^ = outside diameter of raised face. Mfd <[11250 Ab -(tt/16) dJ Pfd] C(S /36) . The proposed NB-3658. m Reason: The pressure limit of 1.25 S .5 times the design pressure is analogous to the limit of NB-3655.

it was deemed appropriate to supply specific rules for flanged joints under faulted conditions. through the pathway of NB-3200 rules.0 times the design pressure is analogous As discussed in Appendix B. the guidance given by Appendix F for flanged joints is nebulous. Analysis for Testing Conditions is not required. The NB-3658. However. NB-3657. conditions is not deemed to be needed. if the leakage is deemed excessive. one normally is looking for leakage at the gasket.1 that average bolt stress shall not exceed two times the allowable bolt stress. refers the Code user The requirements of NB-3226 are not relevant to flanged-joints. is deemed appropriate provided Equation (17) is also invoked.83 (c) Pipe-to-flange welds shall be evaluated by Equation (9) of NB-3652. "Testing Conditions". In testing flanged joints. ° m Reason: The pressure limit of 2. using a stress limit of 3. refers the Code user to Appendix F of the Code. Code user may use Appendix F. at best. Reason: At present.4 Testing Conditions.0 S . It may be noted that NB-3656. the pressure limit to the limit of F-1360(a). . There is a control on maximum preload through the requirements in NB-3232. the normal next step is to increase the preload on the bolts. and hence. Accordingly. Add: (c) For analysis of flanged joints. to NB-3226. Consideration of Faulted Conditions. analysis for testing (1) NC-3651. see NC-3658.

Reason: We wish to inform the Code user that special guidance for flanged It may be noted that we jump from NC-3652 joints is given in NC-3658. as limited by NC-3612.1.84 (2) NC-3671. Reason: Editorial to introduce new equations and their numbers in NC-3658. (a) Flanged joints manufactured in accordance with the standards listed in accordance with the standards listed in Table NC-3132-1.1. of the Code. (5) Change NC-3647. shall be considered as meeting the requirements of NC-3640. to NC-3658.5(a) Change equation numbers: (12) to (15) and (13) to (16). . Delete Period After "NC-3647" and Add: and NC-3658. This is to provide consistency in numbering with the NB portion (3) NC-3673. (4) NC-3673.1 to Read: NC-3647.5(b) Change "Formulas (12) and (13)" to "Equations (15) and (16)".1 Flanged Joints.

in Class 2 piping systems there may be other types of joints. bolting. in principle. Other flanged joints shall be analyzed If the NB-3200 analysis is used. in accordance with NB-3200.000 psi may be analyzed in accordance with the following rules or in accordance with NB-3200. A question arises as to what analysis should be made for flanged joints not conforming to the first sentence. Reason: To gather into one place the rather specialized requirements for the analysis of flanged joints and to give relatively simple and explicit rules for ANSI B16. the last two sentences were included which. "Analysis of Piping Systems".5 flanged joints.85 (b) Flanged joints not included in Table NC-3132-1 shall be designed in accordance with Article XI 3000. will give guidance for any type of flanged joint. (6) Add a New Paragraph NC-3658 to Read as Follows: NC-3658 Analysis of Flanged Joints Flanged joints using flanges. and gaskets as specified in ANSI B16.value at 100 F not less than 20.5 and SA-193 Grade B7 bolts are commonly used. particularly in sizes larger than 24". allowable m stresses from Table 1-7 and 1-8 shall be used in place of S . While most flanged joints are B16. For lack of a better available solution. . and eliminate those requirements which should go under NC-3650. "Pressure Design of Piping Products". present rules in NC-3650 are not relevant to any kind of flanged joint.5 (1968) and using a bolt material having an S . Reason: To give requirements for pressure design under NC-3640. The Code case 1744 provides some assistance but it is limited to maximum pressures of 100 psi and maximum temperatures of 200 F.

more likely. their equivalents called "design limits") will be used for both Class 2 and Class 3 components. NC-3600 does not define or use normal conditions. Normal and Upset Conditions (a) The limitations given by Equations (12) and (13) shall be met. If cold springing is used. The recommendations have been written on the assumption that the revised rules will be implemented in the near future. . S = yield strength (ksi) of flange material at design tempera ture (Table 1-2.-lb. total cross-sectional area of bolts at root of thread or \ section of least diameter under stress. = bending or torsional moment (considered separately) as defined for M but including dynamic loadings. work by Code committees is in advanced stages in which these terms (or.5.2). sustained anchor movements. in. in. sq.86 NC-3658.. relief valve steady-state thrust and other sustained mechanical loads. in. M where M 1 3125 (S /36) C A^ (12) fs bending or torsional moment (considered separately) applied to the joint due to weight. the moment may be reduced to the extent permitted by NC-3673. Reason: At present. The value of (S /36) shall not be taken as greater than unity. upset conditions. Mfd £ 6250 (S /36) C A^ where (13) M. C diameter of bolt circle. in.-lb. However. emergency conditions or faulted conditions.1 Design. thermal expansion of the piping.

using appropriate stress intensification factors from Figure NC-3673.1 is essentially identical to NB-3658. However.1 of "The bolting shall meet the requirements of NB-3232" is not imposed.1(b).1(a). Mfd <[11250 Ah -(tt/16) D^ Pfd] C(S /36) . Flanges of ANSI B16.5 flanged joints meeting the require ments of NC-3612. NC-3658.1 are not required to be analyzed under NC-3650. (b) Flanges.87 It may be noted that a requirement analogous to that in NB-3658.2(b)-l.. which uses stress intensification factors. in. this is not relevant to the problem.1 and the same reasons apply. NC-3658. The relevant aspect concerns the assumed preload stress of 40. (14) where D = outside diameter of raised face.2 Emergency Conditions (a) (b) The pressure shall not exceed 1.5 times the design pressure. However.000 psi and the assumption is deemed to be equally valid for Class 1 and Class 2. The second sentence is appropriate for Class 2 piping. This is deemed to be consistent with Class-2 design philosophy in which cyclic stresses. Otherwise.. Pfd = Pressure (Psi) during the emergency condition concurrent with M. rd Mfd' C' Sy' the limitation on (S /36) and A^ are defined in NC-3658. The limitation given by Equation (14) shall be met. the pipe-to-flange welds shall meet the requirements of NC-3652. .. It may be noted that the allowable bolt stresses in Class 2 are lower than in Class 1. Reason: See those cited under NB-3658. are not considered. other than those due to cyclic moments.

using a stress limit of 1. h Reason: The pressure limit in (a) and the stress limit in (c) conform to the general limitations for emergency conditions being established for Class 2 piping.4 S. .2(b).3 Faulted Conditions (a) (b) The pressure shall not exceed 2. . h Reason: The pressure limit in (a) and stress limit in (c) conform to the general limitations for faulted conditions being established for Class 2 piping.8 S. The reason for (b) is discussed under NB-3658. Analysis for Testing Conditions is not required.4 Testing Conditions. Testing conditions will be identified in the rules under develop ment but.88 (c) Pipe-to-flange welds shall be evaluated by Equation (9) of NC-3652.. The limitation given by Equation (14) shall be met. using a stress limit of 2. The reason for (b) is discussed under NB-3658. fd and Mf (c) are pressures (psi) and moments (in.-lb) occurring concurrently during the faulted condition.0 times the design pressure. . Reason: Analysis for testing conditions is not addressed in the present NC-3600. Pipe-to-flange welds shall be evaluated by Equation (9) of NC-3652. NC-3658.3(b). as presently drafted. NC-3658. analysis for testing conditions is not required. where V.

Do not change the phrase: "using Equations (9) and (10) (ND-3650). the present equation numbers (9) and (10) are to be changed to (15) and (16) in two places.89 ND-3600 It is recommended that ND-3600 be changed as shown for NC-3600. ND-3600 (Class 3 piping) The only exception is equation is identical to NC-3600 (Class 2 piping)." Reason: For the purpose of these recommendations. will make ND-3600 and NC-3600 identical in this respect. if the recommendations are followed. .5. In ND-3673. changing NC to ND in all places where NC occurs. numbers which.

Subcommittee on Piping. The authors wish to express their appreciation to Mr. McKeon. T.90 7. E. ACKNOWLEDGEMENT The author wishes to thank the members of Task Group 1 of the Pressure Vessel Research Committee. and the members of the ASME Boiler and Pressure Vessel Code Committee. M. F. . Branch. J. B. and Mr. E. F. Vinson for their many valuable comments and suggestions. for their review and valuable suggestions. Pumps and Valves. Mr. Jr. Subgroup on Piping (SGD) (SCIII). Mr. O'Hara.

May 2. 1955. 345 E. 9. ORNL-5035.. 1953. February 28. "Tests on 3" 150 lb ASA Welding Neck and Slip-on Flanges". Heating. Its Relative Merit and Application". 1961.. and Oliver. Tennessee. and Stuart. Inst. (4) Roberts. June 11. (1974 Ed. 8.300 lb ASA Welding Neck and Lap Joint Flanges". 1. ASME. Section III. 1950.C. Oak Ridge. Vol. (9) George.. of Applied Mechanics. "Performance of 6061-T6 Aluminum Flanged Pipe Assemblies under Hydrostatic Pressure". 2.. "Tests of Heat-Exchanger Flanges".900 lb .010a. (11) Gough.J.5-1968. Rossheim. Piping and Air Conditioning". (6) (7) Maney. Nuclear Power Plant Components.) (2) American National Standards Institute. REFERENCES (1) ASME Boiler and Pressure Vessel Code. .. E. New York 10017. "Gaskets and Bolted Joints". ASME J.G. H. 56-PET-19. "Steel Pipe Flanges and Flanged Fittings". 1938. Published by the American Society of Mechanical Engineers.030. 2.. Mechanical Engineers. September 8. Oak Ridge National Laboratory. (5) Petrie. Rodabaugh. 4. N. 45th Street. No. (13) Tube Turns Report No. ASME Paper No. 345 E. 1936. and Holt. Engineering. 2. G. and Moore. 45 Street. I. April. (3) Rodabaugh. "Behavior of Large Taper Hub Flanges". (10) Tube Turns Report No. Pressure Vessel Research. D.W. "Hydrostatic Test of a 4 Welding Neck Flanged Joint". Fasteners. USAS B16. (8) Murray. "Bolt Stress Measurement by Electrical Strain Gages". New York. New York. No. 1937. New York 10017. Published by the American Society of Mechanical Engineers. First Report of the Pipe Flanges Research Committee". (12) Tube Turns Report No. "Computer Program for the Analysis of Flanged Joints with Ring-Type Gaskets". "The Ring Joint. Vol. "Bending Tests on 4" .91 8. 1955. May. Trans. Gebhardt.024. O'Hara.

ASME Journal of Applied Mechanics. May 21. 2. A.. The M.Y.G. 1964..W. (20) API Std.033d. "Tests of Taper Face Welding Neck Flanges Under Gas and Hydrostatic Internal Pressure with and without External Bending Moment". 2nd Edition. Petroleum Refiner. Kellogg Co. (18) Blick. 2. "Bending Moments and Leakage at Flanged Joints". A. 605. "Large Diameter Carbon Steel Flanges". 10020. (19) Design of Piping Systems. 1958. 1958.R. (15) Tube Turns Report No.036. September. "Fatigue Tests on Flanged Assemblies".. H.H. "Analysis of Shells of Revolution Subjected to Symmetrical and Non-Symmetrical Loads". N. 1271 Avenue of the Americas. "Effect of Flange Material Yield Strength on the Pressure Capacity of a Flanged Joint". N.C. New York. 1950.92 8. April 25.Y. ASME Transactions. . New York. Published by John Wiley & Sons.. Division of Refining. REFERENCES (Continued) (14) Tube Turns Report No. 1950. (16) Markl.. and George. 1956. R. Published by the American Petroleum Institute. (17) Kalnins.

APPENDIX A TEST DATA ON FLANGED JOINTS TABLE OF CONTENTS Page INTRODUCTION A-l TEST DATA ON DISPLACEMENTS. A-2 Change in Bolt Length with Pressure Flange Rotation due to Bolt Load Flange Rotation due to Pressure A-4 A-10 A-14 Tests on a Large Flanged Joint with O-Ring Gasket Change in Bolt Stress with Pressure A-l7 A-23 Flange Rotation due to Pressure Flange Stresses due to Pressure Tests on ANSI B16. . AND STRESSES A-l Tests on Heat Exchanger Flanged Joint with a Metal Gasket. A-24 A-26 Bolt Stress Change with Pressure LEAKAGE PRESSURE TESTS A-30 A-33 TEST DATA WITH PIPE BENDING MOMENT A^i5 .5 Flanged Joints with Asbestos Gaskets. STRAINS. . .

.

The data are for internal pressure loading and. available test data are presented which give some indication of the effects of the variables mentioned above. STRAINS. AND STRESSES Because flanged joints are statically redundant structures. No test data TEST DATA ON DISPLACEMENTS. . i. pressure. cannot be expected to be very accurate. "effective gasket diameters" and "effective bolt circle diameters". is given in Reference (3). There are larger uncertainties involved in the selection of load application lines. * All calculations made with FLANGE used the hub-to-ring boundary condition represented by the option IBOND = 2 in FLANGE. gives better correlation with test data than does the hub-to-ring boundary condition used in the Code method (IBOND = 0 option in FLANGE). to a limited extent. as applied to actual flanged joints. This option. Three sets of data which can be used to check the validity of the theory as used in the computer program FLANGE are discussed in this section. it is based on a shell and plate discontinuity analysis and. their theoretical analysis involves the rotations of the flanges due to the loads (bolt load. In this Appendix.. The theoretical basis for this program Essentially.e. there are major uncertainties concerning the "effective modulus of elasticity" of the gasket. are available on thermal gradients in flanged joints. thermal gradients) and the axial strains in the bolts. particularly with re spect to flange stresses.INTRODUCTION The text of this report gives results of calculations using the computer program FLANGE . For any but flat metal gaskets. for internal pressure combined with a pipe bending moment. in general.

the interaction effects can be evaluated for low pressures. In the statically redundant portion of the tests. consists of selection of an appropriate effective modulus of elasticity for the gaskets under the test conditions. an effective gasket diameter must be selected. the change of bolt stress with The major uncertainty in these tests pressure is discussed. thus reducing uncertainties concerning the "effective gasket diameter" and modulus of elasticity of the gasket material. at higher pressures. In this section.A-2 (7) (1) Tests on Heat Exchanger Flanges with Metal Gasket These data are closest to "ideal" for the purpose of checking the theory because a narrow metal gasket was used. Further. Tests on Heat Exchanger Flanged Joint with a Metal Gasket Reference (7) gives test data for the flanged joint shown in Figure Al. /•ON (2) Tests on Large Flanges with O-Ring Gasket While a flanged joint without bolt preload is a statically determinate structure. The data include the best available measurements of stresses in a welding neck flange. (3) Tests on ANSI B16. this is difficult because the raised face is very wide compared to the radial distance between the center of the raised face and the bolt circle.5 Flanges With Asbestos Gaskets ^ The leakage aspect of these tests are discussed later under "Leakage Pressure Tests" and details of the gaskets are given thereunder. Experimental results consist of: . these tests involved a moderate preloadhence. flange rotations and flange stresses for the statically-determinant loads can be compared with theoretical predictions.

DETAILS OF FLANGED JOINT.L. . of Knuckle Radius on Head 3ii» 16 \ -4"16 4" 28-1 Diam Holes i i 2i"VM///M 4 l». Gasket Material Soft Iron FIGURE Al.L. f 25"ID Gasket > i oo 26I"CLof'4 274 284"BC 24"|D 304/bD Tonoue and Gosket Details of All Flanges.FLANGE C-l FLANGE C-2 7C. REFERENCE (7).of Knuckle (Radiuson Head C.

(b) Modulus of elasticity = 30 x 10 psi for flanges.D.950 in.D. Change in Bolt Length with Pressure The test data are shown in Figure A2.419 (root area per bolt) = 11.A-4 (1) Plot of bolt elongation versus internal pressure.D. = plate O. along with the calculated result obtained by the following equation. as might be expected using this method. Elongation Considerable scatter exists in the test data.000 psi. was measured with calipers. Moment = hQ x Afo x 40. XGI = 26. The somewhat arbitrary aspect of the input consists of appropriate modeling of the plate between the flanges so as to incorporate the elastic response of the plate.875". The gasket thickness is equal to the plate thickness. (2) (3) Flange rotation due to internal pressure. arbitrarily set at 26" as representing that part of the plate which deforms elastically under load.000 A^ = 469. gasket thickness (VO) of 1.750" . (d) Pressure = 1. bolts. This was done by using gasket dimensions of XG0 = 27.000 psi = 439. 6 = (W1-W2) P_ A = b 1000 A F b -6 5.000" and Gm = actual mean gasket diameter = 26. and gasket I. gasket O. and gasket (c) Bolt area = 28 (bolts) x 0.732 sq.-lb. Flange rotation due to bolt load. in. The printout from the computer program FLANGE for the flanged joint is shown in Table Al. Wl = initial bolt load = 40. The input data to FLANGE (printed out on the first page of Table Al) is based on: (a) Dimensions as shown in Figure Al.280 lb.500".W5 X iU * .

.3990 05 6.00000 07 QPHG* 1.O.G1 25.42200 1.000 ibond 2 MOMENT THERMAL EX".3S9? 05 6.50000 00 Hi 4.0.75000 IBOND 2 PRESSURE..0000D I On 3.25000 1.37500 00 BOLT AREA* 3.R THICK.0000 07 2.0 3.62500 COEFF.00000-06 X8 = 2.81250 I TYPE BOLT PRESSURE...00000 YG 0 7 E8 6. 7*000-01 VO 3.390 20-09 QPHG3.0000-06 1.04480-05 EF - 6.52740-05 EF - 6.5 000D 00 FLANGE JOINT SIDE TWO <UN°RIMED QUANTITIES) QF^- YM - 8.0 TG 0.8750D 01 6.50000 00 » 3. OF MEAN GASKET ELASTICITY DIAMETER MATE 4 4. FLANGE O.GO BASE.00000 07 TF 6.000D-03 3.T MALL.000D-06 l.6000D 01 YG2 » 3.00000 07 EG » 6. P 1000.0000D-06 XB = 2..0000 07 2.58100-06 OTHGYF2 - 3.50000 0.17320 01 0.2200D-01 TH 1. „ 33.OOOD-03 3.75000 1000.77500 01 FACE 0. .36030-09 BOLT ING 3.'5000 MOMENT BOLT CIRCLE.75000 ITYPE 0.0 YFP2 3.3000D-06 VO * YG 1.77500 01 XGI « 2..O. FIGURE Al (SHEET 1 OF 4).6928D 05 3.000 ICODE 0 COEFF.00000 2.00000 07 QTHG= 2.T HALL.C P 28.00000 07 1.C 28.50000 01 GO- 4.H 2. DELTA Mnn.42200 0.00000 07 XGI AB 2. OF THE=»AL EXP.0 0. INTEGER TO INTEGER PAIR FLANGE JOINT SIDE ONE (PRIMEO QUANTITIES* CFHG- YM . A .0 YF2 1.60000 01 PBE 1. CIRCLE.17320 01 3.Gl 2*.0 YB2 > YG2 3.H 0. OF MEAN GASKET ELASTICITY DIAMETER HUB LENGTH.00000-06 TFP 0.6880 01 FLANGE FLANGE FLANGE PIPE HUB AT I.00000 07 0 7 FLANGE JOINT BOLT LOAD CHANGE DUE TO APPLIED LOADS.0000D-06 . I CODE 0 MATE 3 i.0000D 07 Y92 - 1.00003-06 EG TB 0.0000D 07 1.56720-06 YF2 > 3.0000D 07 BOLT CIRCLE" EB 2.A 30.00000 XGO 07 2.50000 0.75000 FLANGE FLANGE PIPE HUB AT I«°"8 THICK.6880 01 bsize YB 8.42500 01 GO" 4. .2200D-01 TH 2. "F DELTA MOO.0303 0 07 XGO 2.5 0000 00 BOLT LENGTH* YB GASKET 6.TABLE Al: OUTPUT FROM COMPUTER PROGRAM FOR HEAT EXCHANGER FLANGED JOINT.GO BASE. 7.00000 HUB LENGTH.

03000-03 DELTAP.564°D 05 W1-W2A.50310 05 FLANGE ONE TEMP.W2» 1.12790 05 W1-W20.W2A- INITIAL BOLT LOAD.39950 05 M2B.1.00000 03 FLANGE TWO TEMP.33470-01 > I ON MlM2SP- 4. 5.1897D 05 W1-W2C.2462D 05 .6928D 05 BOLT TEMP.0...4.3.3.H2CCJMBINfcD.1.06460 00 W1-W2W2A/-1.1.5649D 05 OELTA THERMAL..W2D-= 4.39140 05 7.7.00000-03 PRESSURE2.2709D 05 M2CP.39140 05 M2P5.0.3347D-01 W20/W1.69283 05 MOMENT SHIFT.8.24620 05 M2D> 4.39950 05 M2A.3339D-01 W2C/W1.39950 05 M26.1.12580 05 M2C.00000 00 H2B/W1.12790 05 3.0.0 DELTA1.2.0.0 W1-W2P.0 RESIOUAL BOLT LOADS AFTER THERMAL-PRESSURE LOADS 4.TABLE Al: SHEET 2 OF 4. LOADINGS GASKET TE-P.00000 00 W2/W1INITIAL AND RESIOUAL MOMENTS AFTER THERMAL PRESSURE LOADS. W2 B- TOTAL PRESSURE.4.. 0. 3.0 AXIAL THERMAL.69280 05 1.0 1.5.

968tP.:.::::::::::::.12820 04 SCSI.6S09D-02 SC LO.93270-03 THETA.::^:.96 350-0 2 SLLO.:r-:. 1..58750 04 SC LO.41760 03 -—03 ZC -1. 9. -3. -1.6346D 02 SRH.64070 04 STH.23950 03 SRF.0 03 SLLO.-6.. 7>5540D_03 ^i.02 STF.tM« 04 SRH.3.1. 9.1.3.02 SRp. .7010r.3.4.„ SRH.1>21820.r.™.1.„ 33°2°"°3 SLSO.07420 03 ». 54450-03 QFHG.660.-1.-3.-1.02 SRH.72180 0! SrH.3.62B4D 03 ZG.26B1O-03 SLSI.TABLE Al: SHEET 3 OF 4. .. CALCULATIONS FOR PCESSURE LOADING .26. SLSO. 3. .09720 04 STF.29369-03 ZC.-6.17910 03 SRF.2.0109D 04 SCSO.::::::Ir.-2.07300-03 .-2.68090-02 SLLI.L28220-02 Yl.64990 04 SLLL -1.64! 8D04 SLLI.2092D 04 s~F.„«. -.56490-03 YO.UD-03 SCSO.64180 04 SCLO.-1.:-»•-» SLSO. 3070D-0 1 SCLI.2. |>9a83B04 s:so.64890 04 STH.5 19630-03 Y..28820 04 SCLI.64990 04 SC LO.-3.-7.1238D-02 OFHG.2.:.1.9370D04 SLSI.2. 5.6.. SLSr.21750-03 OFHG- ^ *** OVULATIONS FOR WBMTwe L0AD[NG l-58l0°03 V°" 6'^20-03 .:.0 2620-01 STH.' " 15Binnna v„ .93700 04 SCSO.2.56920 04 SCSI.-3.4. 2.T:. TAPERED HUB FLANGE CALCULATIONS FOR MOMENT LOADING SLSO.„„.97300 04 SLSL -2.-2.60570 03 SRF. -4.1...I.5.-2.12520 03 SLLI.96650 02 Z». 4.3. 0.2388D 04 SCL.63S5D-03 ZC.. uwiom scs[_ ^ ^ SLLO. 4. 3.59180 04 SCLI.25090-03 YO.3.47400 04 SLLO.

7.-1.3.37710 03 STH.6245D-03 YO.2. TAPERED HUB FLANGE CALCULATIONS FOR MOMENT LOADING SLSO.9083D 04 SCLO.2.3O-03 SLL !• -9.5337D 03 STH.-7.TABLE Al: SHEET 4 OF 4.26180-03 VI.13450-04 THETA.99630-03 S: LO.7.5797D 03 SCLI.235!O-03 ZC -3.4020D-03 THETA.98060 03 SCLI. -2.123 80 0* SLL 1= -2.5.3. 6.-1.2.96140 04 SCSI. 4.50950-03 OFHG. 1. 6.9.. SLSO= 3.O9.18460 04 SRF= -2.52680-02 Y1.4561D04 SLLO= 4.33610 03 STF. -2.56500-02 SCSI. .-1.0233D 04 7.66560 04 STF.-3. 2.072 2D 04 SCSO. 9.3575D 04 SRH.3.78240-03 ZC -2. 187 10-01 STH.43500-03 CALCULATIONS FOR PRESSURE LOAOING SLSO.0 Q3 SRp= _3.4247D 04 SCSI.86720 03 STF.8043D 03 SLLO.56720-03 Y0= 8.85230-03 OFHG. 127 10-0 1 SCL I. W2 FOR ITYPE-3.3.04920 04 STH.43770 04 SRF.0448D-08 YO.24650 04 SCLI= 1.16560 04 SC LO.63950-02 SCSO.0829D-03 THETA.1. 2.-1.40080-0 2 SRF= 4. M2 OR M2 P FCR ITYPE-1 OR 2.94540 04 SL SI.5339D 04 SRH.6395D-02 SLSI.1.-2.4.61490 04 SCSI.32380-08 CALCJLATIONS FOR COMBINED LOAOING.-1.2.90240 04 SLSI.1.6. 1.67650-08 THETA.00910-02 STF. 9.-1. -8.24620 05 ZC -1.2404D 04 SLSI.68„0 03 SRH= _.-1 .25400-03 V*.70810 03 SCSO. -1 .24040 04 SCSO.9141D03 SCLO.-4.1.93800 Q2 > ZC -2.61870-02 SLLO.67770-03 Y0.3.43870-03 .93300-09 Y1.78100-03 CALCULATIONS FOR TFMPERATURE LOADING » SLSO.02430-02 ZC -1. -1. 5.33540 04 SILO.210 90 03 SLLI.05160-02 SRH.02630-0. 8.4.20450 04 ZC -4. 7.7868D-02 QFHG..QFHG. 1.4.-5.18560 04 SLLI.83180-03 ZC -8.833.98150-08 ZC -5. 1..-2.

FIGURE A2. REFERENCE (7 ) TEST DATA. .0001 in. COMPARISON OF THEORETICAL AND MEASURED CHANGE IN BOLT LENGTH WITH PRESSURE. 8 16 24 32 40 Decrease in Bolt Length .A-9 700 o TEST DATA THEORY BASED ON W THEORY BASED ON W. .

900 lb as calculated by FLANGE. . The theoretical change in bolt length with pressure is shown in Figure A2. (W1-W2) = change in bolt load b due to 1. excluding Accordingly. the length of cylindrical shell between the flange and knuckle radius on the head should be greater than about 4 V dt = 4V25 x . E = 30 x 106 psi. the head probably restricted the radial expansion of the the radial effect of internal pressure. cylindrical shell.000 psi pressure = 314. indicates that the full radial effect of internal pressure would not be expected to fully develop in the test assembly. For this effect to fully develop. = decrease in bolt length.125".422 = 13". The theoretical change in bolt load. examination of Figure Al rate of decrease of bolt load. The theoretical rotation for the input moment of * The effective bolt length is taken as the length between nut surfaces plus one bolt diameter. A^ = bolt area = 11. is shown in Table Al as (WlW2B). It can be seen that the theory overpredicts the However. and £ = effective bolt length* = 6.A-10 where "?. this agrees reasonably well with the test data. Using this value gives the dashed line shown in Figure A2. This length in the test assembly was only 4". respectively. P = internal pressure.732 sq in. Flange Rotation due to Bolt Load Test data are shown in Figure A3 and A4 for flanges C-l and C-2.

COMPARISON OF THEORETICAL AND MEASURED FLANGE ROTATIONS DUE TO BOLT LOAD.A-ll 1000 o 800 TEST DATA THEORY o O o 600 T3 O O -J o 400 00 FLANGE c-i 200 16 24 32 40 Flange Rotation . FLANGE C-1 REFERENCE (7) TEST DATA . minutes FIGURE A3.

COMPARISON OF THEORETICAL AND MEASURED FLANGE ROTATIONS DUE TO BOLT LOAD.A-12 1000 o 800 TEST DATA THEORY o o o •o o o 600 o GO 400 FLANGE 200 c-2 16 24 32 40 Flange Rotation . minutes FIGURE A4. FLANGE C-2 REFERENCE (7) TEST DATA .

.573 x 10 for flange C-1.484 x 10 3W radians = 2. test data indicate At higher loads.9375" for M^G 3.9375 W radians 439. the that yielding of the flanges might have occurred. The calculated stresses are high enough to indicate the possibility of such yielding. the curvature of the test data is opposite to what would occur from this effect.900 7. * The negative signs in Table Al for "THETA" is the sign convention used in the basic theory indicating rotation in the direction of moment due to bolt load^ ** Shifting of the bolt load inward or gasket effective diameter outward would also produce a non-linear effect.5115 x 10 and 3. however.516 x 10 W minutes. respectively.900 in.5115 x 10 3 x 0. These theoretical lines are shown in Figures A3 and A4.A-13 439. and similarly for flange C-2 -3 W minutes = 2. The test data and theory agree reasonably well at low loads.4345 x 10 for flanges CI and C2. where hQ = 0. this joint: * With M = WhG.-lb is shown in Table Al as THETA = 3.

000 psi. The theoretical increase in flange rotation due to internal pressure is given by: AG = ep_ Ap + em AM AP P M P (Al) From Table Al. we obtain: pi Flange ep' /p 6 /M m AM/P 1. .000 lb bolt load.9375 +|^|^ x 650 = 42. outside SLSO =^3gq0 x850.3867-4.5115 x 10 3 (5.000 x .A-14 The tests were apparently run by increasing the bolt load in steps and. the maximum stresses Flange C-1. perhaps as high as 650 psi with 850. SLSO (Longitudinal.3995) x 105 C-2 100n 439900 1000 . large end of hub. small end of hub. Flange Rotation Due to Pressure Test data are shown in Figure A5 and A6 for flanges C-1 and C-2.7703 x 10 3 3.2416-4. at each step increasing the pressure in steps. are: Under these loads.000 to 50.900 x 850. Presumably they were carbon steel with a yield strength in the range of 30. outside surface) SLLO = 439. SLLO (longitudinal.000 The materials used for the flanges and attached shell are not described in Reference (7).3995) x 1Q5 C-1 1000 439900 1000 1.980 psi.4354 x 10 3 (7.qqq x650 = 62.000 x . respectively.7810 x 10 3 3.140 psi Flange C-2.2h8oL 1.9375 +13.

5 10 Flange Rotation. minutes FIGUREA5.A-15 700 o TEST DATA o o 600 THEORY 500 Q. REFERENCE (7) TEST DATA FLANGE C-1. COMPARISON OF THEORETICAL AND MEASURED FLANGE ROTATIONS DUE TO INTERNAL PRESSURE.5 5 7. a> i— 400 3 to 0> 300 CD C 200 FLANGE c-i 100 2. .

minutes FIGURE A6.5 10 Flange Rotation. COMPARISON OF THEORETICAL AND MEASURED FLANGE ROTATIONS DUE TO INTERNAL PRESSURE. REFERENCE (7) TEST DATA . £ 3 to to 400 a> a £ 300 200 - FLANGE c-2 2. FLANGE C-2.5 5 7.A-16 700 o TEST DATA THEORY 500 - to Q.

770 x 10~6 + 7.810 x 10_9 x 284. with again some test data evidence of yielding at high pressure (actually. The theory is in good agreement with the test data for flange C-1.558 x 10~6 AP radians = 0.006 x 10_6 AP rad = 0. The test data do not indicate yielding of flange C-2. Bolt stresses (2) Flange rotation due to internal pressure.781 x 10"6 + 7. high pressure combined with high bolt load) .A-17 Using these values in Equation (Al) gives: For Flange C-1 AO = (1. Experimental data consist of: (1) Plot of bolt stresses versus internal pressure. were determined by strain gages on the bolts.00879 AP minutes For Flange C-2 A6 = (1. The theory over predicts the flange rotation of flange C-2.2) AP = 4. (3) Longitudinal stress on outer surface of hub due to internal pressure.72)AP = 2.0138 AP minutes lans These lines are shown in Figures A5 and A6.982 x 10_9 x 98. . Tests on a Large Flanged Joint with O-Ring Gasket Reference (8) gives test data for the flanged joint shown in Figure A7.

DIMENSIONS OF REFERENCE (8) FLANGES.9375" £ 180 —l^g" Bolts 180.9375" 190. IDENTICAL PAIR USED IN JOINT.6875" 185" 137.9375" FIGURE A7.0625" <t> 177.A-18 180. .

920. Initial Moment.96875 3.4 183. The elastomeric 0-ring gasket is assumed to act only as a for flanges. (c) (d) Bolt area = 253. in.009.000 10.A-19 Printout from the computer program FLANGE for the flanged joint is shown in Table A2.6 180. bolts. one would intuitively expect that most of the load between flanges would be carried by a relatively narrow annular ring of the raised face near the outside of the raised face.616. Modulus of elasticity = 30 x 10 "gasket".000 in.392.6 181.000 13.61875 185. The input data to FLANGE (printed out on the first page of Table A2) is based on: (a) (b) Dimensions as shown in Figure A7. Effective Gasket Gasket in.1 7.000 .352 lb. Pressure = 100 psi. hence. O. the gasket is simply the 1/8" raised face.D.-lb 185 1.514. x S = 7. and seal.4 184.7 1. Moment = h x A. I. and considering the rotation of the flanges.312.000 16. To investigate this.6 182. Looking at Figure A7. in. Diameter.4 183.46875 186. the following effective gasket diameters were assumed.D.4 sq in.920.6 179.3 184 183 182 180.46875 2.4 182.-lb Wl = initial bolt load = A^ x S = 5. The somewhat arbitrary aspect of the input consists of appropriate modeling of the "effective gasket diameter".96875 2.000 19.

06250 HUB AT BASE.0000-03 3.93750 177. FLANGE O.30000 07 EG = 6.03000 07 YFP2 3.00000-06 .D. 25000-0 1 3.0 XGO 1.99240-05 3.000CD 03 (UNPRIMEO QUANTITIES) 3FHG- 5..93750 MOMENT 7. FIGURE A7.00300 XGO 07 1.00000-06 EG 0.2 5000-01 3. OOOOD 07 EF » XB 1.C P 100.GO 1.0625D 00 TH 5.000 ICODE 0 MATE 2 193.7794D 02 6. OF DELTA MOD.06250 00 TH 5.64260-10 QPHGy « 3.850D 02 BSIZE 1. OF MEAN GASKET THERMAL EXP.50000 187.75000 HUB LENGTH.A FLANGE I.93750 I TYPE IBOND 2 COEFF.G1 2.00300 07 fOLT ING 1.0000D-06 TFP 0.0 TB TG 0.00000-06 GO- 1.53*00 02 Y3 2 3.H BOLT PRESSURE.8640D 02 FACE 0.0 6.64260-10 QPHGYM = 3.63600 02 PBE 0.9230 10. CIRCLE.00000 07 CASK ET BOLT ARFA2.D.50000 VO 0 0 Y9 3.77940 02 6.0000 07 1.00000 07 XGI 1.00000 TF 07 Wl 5.0000D-06 GO" 1..0000-06 1.0000D YG E8 07 6.00000 PIPE WALL.83600 02 Y32 » 3. OOOOD 07 AB 2.B FLANGE THICK. OUTPUT FROM COMPUTER PROGRAM FOR LARGE FLANGED JOINT.TABLE A2.0 YB2 3.00000 07 FL4NGE JOINT BOLT LOAD CHANGE DUE TO APPLIED LOADS.9924D-05 3.86400 02 XGI » 1.53400 02 YB2 « 3.41000-05 YF2 • TWO QTHG5.87940 02 EB ' 6.41000-05 YF2 » QTHG= 5. IDENTICAL PAIR > I FLANGE JOINT SIDE ONE (PRIMED QUANTITIES) O 2FHG- 5.0 1.00000-06 VO » YG 1. OOOOD 07 FLANGE JOINT SIDE 1..OOOOD 07 BOLT CIRCLE" 1.OOOOD 00 BOLT LENGTH* 1.OOOOD 07 EF - XB 1.16250 01 Ye » 3.T 5.0 YF2 3. 39240 06 0. ELASTICITY DIAMETER 06 6.

W2CCD*!8!NE0.03860 07 M20- 7.39240 06 MOMENT SHIFT.W2» 1.0 1. LOADINGS CASKET TEMP.W2B- 2.03860 07 M2P- 1.9200D 06 M2- 1.9201D 06 M2B- 1.W2A- 5.0 RESIOUAL BOLT LOAOS AFTER THERMAL-PRESSURE LOADS AXIAL THERMAL..0..88350-01 W1-W2CW2C/W1- 3.00000-03 PRESSURE- 0.75900 06 4.0 1.6334D 06 TDTal PRESSURE.0386D 07 .1683D 07 M2A- 7.0 1.3924D 06 1. 1.03000-03 OELTAP- FLANGE ONE TEMP.- INITIAL BOLT LOAD- 0.7499D 06 W1-W2AW2A/rfl» 0.24510-01 W1-W20W20/W1- 3.74990 06 DELTA THERMAL.OOOOD 00 W1-W2BW2B/WI- 2.TABLE A2: SHEET 2 OF 3.16830 07 M2C- 1.64250 06 3.0.03860 07 M2CP- 1. HIK29P- 7.92010 06 1.24510-01 N> INITIAL ANO RESIDUAL MOMENTS AFTER THERMAL PRESSURE LOAOS.0 5. W2 D- 5.99990-01 WI-W2W2/W1- 3.75480 01 9..OOOOD 02 FLANGE TWO TEMP.392*0 06 DELTA- BOLT TEMP.64250 06 3.

-4.-3.57420-02 Yl.2220D 01 SRH.03860 07 SLSOSLLOSTH. 1.51360 03 SLLI.-2.1.-1.42700-03 Z5.4I79D 02 2.14580-08 CALCULATES FOR COMBINED LOAOING.65000 03 SLSI- 2.1217D-01 SRH.-3.«5340 03 SRH3.76960 02 STF.2.8517D 03 SRH- 1.-4.05320-07 OFHG.-9.80430-02 SCSI. TAP ERED HUB FLANGE CALCULATIONS FOR MOMENT LOAOING SLSOSLLOSTH.-1.78260-02 Y1= 4.37670 03 SLLI.51130-02 OFHG.-2.43380-07 THETA.M2 OR M2P FOR ITYPE-1 OR 2.-1.4650D-02 QFHG.-1.TABLE A2: SHEET 3 OF 3.9171D 03 SRF.-3.09870-08 Yl.03220-01 SLSI.5905D 03 1.0258D 03 SCSI8.37670 03 SCLO3.94520 03 1.-6.1556D 03 SCSO2.22310 03 ZC -1.96880 03 ZC -1.27020-03 YO. . 1.5. 1.73790-02 ZC -2.-4.20450-02 STF.9766D-03 CALCULATIONS FOR PRESSURE LOADING SLSO- 5.38900-02 SCLI.29750 03 SRF.48920 03 SCSI- 6.-1.6560D 02 Z3.80140 04 SCSI.56480-03 SCLO.-5.75290 04 SLSI.12370-03 THETA.4690D-03 YO.30820 03 SCLI.7.19400-01 SRF.3.7.72350 03 SCSO- 7.85770 03 SCLI.-3.41000-03 YO.-2.99240-03 YO.83100-03 QFHG. W2 FOR ITYPE-3.-8.5.45390-07 ZC -2.1.42110-03 ZC.20450-02 6.846 70 03 STF= -T.8788D-03 .1.2.-4.0008D 04 4.1.0539D 03 SLSI.40360 03 2.7548D-04 CALCULATIONS FOR TEMPERATURE LOAOING ^ to N> SLSOSLLOSTH.03220-01 1.91040-03 THETA.-9.25030 04 SLLI* -9.-1.4. -5.1.8688D 03 SCLI.71840-02 Yl.35240 03 S'F7.0589D 03 SCSO.8.6182D-03 THETA.1.0644^-02 ZC -1.2S07O-02 S: SO- 3.3.-1.91740-04 SLLI.15720 03 SRF.2679D 03 SCLO.61130 03 SLLOSTH. 9.-1.-9.7166D 03 SCLO2.

/„ r c #-. For G = 185". in these tests. from Equation (A3) into Equation (A2) and solving for P gives the critical pressure.800 lb for P = 100 psi. Change in Bolt Stress with Pressure There are two regimes involved in these tests.-. the pertinent re sults of the other calculation are included in the following discussion.e. For the elastomeric 0-ring.A-23 The "gasket" width is kept constant at 1. Pc is that pressure at which the metal-to-metal interface load is zero.447 sq in. = bolt stress. the results shown in Table A2 give W2 = 1. .. in terms of P. culated by the equation: The critical pressure can be cal pc =-£* p ^> = bolt area and A = pressure area = (tt/4) where S.4 x 2240 = 87. an effective diameter of 185" is shown in Table A2. (180) = 25.15 psi . q bb = q s . A.5 . giving c 2 Sb2 =1. ?c x 25>447 Pc .749.6. i. below and above the critical pressure.4". P . This is not significant The calculations for as the elastic response of the strip is negligible compared to that of the bolts for any reasonably large gasket width. at pressures above P a separation of the metal faces occurs.083 ton/in.417 m) 253.2 at 100 psi or. The results from FLANGE show that Sfe decreases as the pressure increases.4 x2240) =3.800/(253.5 tons/in.749. The initial value of Sfe in these tests was 9.6 417 -— 100 (A3) Substituting S.(9.

the increase in bolt stress is given by Equation (A2) Flange Rotation Due to Pressure Below P . Above P .13 110. V psi S.15 97. the theoretical increase in flange rotation due c to pressure is given by Equation (Al).9022 x 10"5 AP radians = 0.954 5.907 184 183 182 4. c' ton/in. the effective gasket diameter is somewhat smaller than the raised face outside diameter. as might intuitively be expected. The values of G .593 125.0654 AP minutes (A4) . Accordingly.0386 x 10 for AP = 100 psi. at P e e c b c are: Ge» in.M = 1. P .0.0386 . Table A2 gives 6 /P = 9.758 x 10_10(1. b at P .7548 x 10~6 Ap + 3. Table A2 also shows M = 1.354 4.51 3.7 It can be seen in Figure A8 that.2 185 87.792 x 107 for P = 100 psi.0386 x 107 . hence.07 180. an effective gasket diameter of about 183" agrees best with the test data.9766 x 10_3/7.92 x 106.75 147.7548 x 10_4/100 and 6 /M = 2.A-24 These relationships are shown graphically in Figure A8 for the various assumed values of G . AM = M .7920) x 10? AP_ 100 = 1. and S.638 6. Equation (Al) gives: AG = 9. P m For G = 185".0.

REFERENCE (8 ) TEST DATA. psi COMPARISON OF THEORETICAL AND MFASURED CHANGE IN BOLT STRESS WITH PRESSURE. Pressure.A-25 50 100 150 200 250 Internal FIGURE A8. .

183. 184.[C .1805 AP minutes. 182 and 180.7 inches.(177.722AP in. Figure A9 indicates that an effective gasket diameter of about 183" agrees best with the test data. Above P .A-26 Equation (A4) is shown in Figure A9.469" 2 With A = 25. The complete curves are shown in Figure A9. The value ofA6 / AP does not depend upon the choice of G but the value of P does.447 in. Flange Stresses Due to Pressure Test data are given (only) for longitudinal stresses on the outside surface of the hub. the rotation is given by Equation (Al) but AM is obtained by the equation: AM = AphpAP where II = moment arm to midsurface of attached shell (A5) .722)Ap. along with theoretical lines for G = 185.75 x 10 6 + 3. The theoretical value of stress is obtained by: . Equation (A5) gives AM = 113. radians = 5. As in Figure A8 for bolt stress change.9375 + 1.9375 .2492 x 10 5AP radians = 0.(B + go)] II = [187. These are stated to be stresses obtained above P and are for a pressure change of 100 psi. These lines terminate at P .0625)] /2 = 4.-lb Equation (Al) then gives A0 = (9. .758 x 10 10 x 113.

psi COMPARISON OF THEORETICAL AND MEASURED FLANGE ROTATION DUE TO INTERNAL PRESSURE.A-27 50 100 150 200 250 300 Internal FIGURE A9. . REFERENCE (8 ) TEST DATA. Pressure.

(A6) The value of the longitudinal stress at the juncture of the hub with the shell is given in Table A2 as SLSO = 5. Equation (A6) gives: a = 5.~n .92 x The calculated stresses.660 psi ' v Similarly. Accordingly.059 for a moment of 7. are shown in Figure A10. The theory assumes that this is along the bolt Calculated stresses assuming that the bolt load However.92 x 10 in. for this location and surface. circle.650 H . Additional data along the hub and attached shell were (17) calculated using the general axisymmetric shell program.650 psi for a pressure of 100 psi. 9059 7. application line lies along the inside of the bolts are also shown in .A-28 ° =-i.92 x 10 7- x 113. MOLSA It is apparent in Figure A10 that the calculated stresses are sub stantially higher than the measured stresses. C. as the flange ring rotates.550 psi 10 7.722 x 100 = 18.-lb.722 x 100 = 13. from Table A2 and Equation (A6) the longitudinal stress on the outside of the hub at the juncture with the flange ring (identified as SLLO in Table A2) is 8377 a = 1519 + -^^ j. along with the test data.x100 +^ *-f x100 The value of AM (above P ) is A L P = 113. the nuts on the bolts shift the load inward.x 113.722 P in. One of several possible reasons for this lies in the assumption as to the location of the bolt load application line.-lb. and SLSO = 9.

STRESSES ARE ALONG OUTSIDE SURFACE OF HUB . COMPARISON OF THEORETICAL AND MEASURED STRESSES DUE TO 100 PSI INTERNAL PRESSURE.inches FIGURE AlO.A-29 10 15 20 25 30 Distance from Flange Ring. REFERENCE (8 ) TEST DATA.

along with theoretical results obtained using the computer program FLANGE.2) What is the effective modulus of elasticity of the asbestos gasket material? The gasket widths in these tests are such that the value of G e is a lesser problem than in the previously discussed tests of the large flanges with 0-ring gaskets.0. These are calculated as previously discussed except using h^ = 4. gives test data on the variation of bolt stress (measured with strain gages) with internal pressure. Reference (3) suggests that the effective modulus of elasticity of asbestos gaskets may be around 1.000 psi for the tests under con sideration. there are two major uncertainties in application of the theory: (1) What is the effective gasket diameter. in this set of calculations e we have used the gasket mean diameter as G . time under load. The data for welding neck flanges are shown in Figure All.75 = 3.469 .469 . In this set of tests. Figure 10. and temperature. the agreement between measured and calculated Tests on ANSI B16. In some flanged joints. tested within a few hours after installation. the bolt stress variation with pressure would not be much different for . Reference (10) gives an equation for approximating the gasket modulus of such "newly installed" gaskets which is about 50. Accordingly.719 in.000. G ? (.5 Flanged Joints with Asbestos Gaskets.j = 4. Considering the basis of the theory.A-30 Figure AlO as C = C-d. stresses is as good as should be anticipated. The modulus of elasticity of an asbestos gasket varies with bolt stress. Bolt Stress Change with Pressure Reference (9). For new gaskets.000 to 100.000 psi.

Aluminum Test No. . Aluminum Test No. 9 -10 • • • • • • Eg =10 J L J L 6 _ 0 2 4 6 8 10 12 14 16 18 20 22 Internal FIGURE All. IP" 12" 300 Class. 6 10' 6 • • 10 h 10 u _ I • I' 0 Eg. 5 10 H 1 1 1 1 I 1 h o CD 12" 300 Class. I Eg= V. Aluminum Test No. 100 psi COMPARISONS OF THEORETICAL AND MEASURED CHANGE IN BOLT STRESS WITH PRESSURE.e "10 r^r"1 ' h • • • . REFERENCE (9 ) TEST DATA. 2 v> 1- H H a. Aluminum 10 No6 H 1 1 Test No. • Eg =10* I I 1 1 1 1- s? \ \ 10 8' 150 Class.A-31 r^rr n ( j r i 1 r t r V \ \ 10 • • 8 150 Class. O O O to M 12 300 Class. Aluminum • •••••• 10' c 10 - Test No. Pressure.

2 106 psi 19. p g However.3 8.29 5. AS.1 . or E = 10 If we assume that the effective modulus of elasticity of the asbestos gaskets is proportional to the initial gasket stress.0 8" Aluminum 12" Aluminum 12" Steel 4. for the particular flanged Figure All 6 S ioints under consideration. 10 and 30 x 10 as tabulated below. It is pertinent to note that further increase in the assumed value of E above 10 8 psi does not lead to a greatly increased prediction This point can be of the rate of bolt stress change with pressure. /AP for E b of: g Flanged Joint 105 psi 10. and the initial bolt stress is the same for all tests. This is because the ratio of bolt area to gasket area is 0.346 for the 12" flanges. then the 8" flanges should have a smaller modulus than the 12" flanges.2 6. 0.123 for the 8" flange. shows the calculated bolt stress variations assuming E = 10 psi.A-32 E = 105 psi than for E = 10 psi. A qualitative trend in this direction can be seen in Figure All.0 -0.9 7. there is a significant difference.5 30 x 106 psi 21. 6 illustrated by considering the values of AS / P for E = 10 .

are shown in Petrie introduced the idea of a "performance factor". Details of the joints that he tested are not given nor is the method of measuring the His results. but.A-33 LEAKAGE PRESSURE TESTS Because leakage is usually the "failure criterion" for flanged joints. However. Such is not the case. for flat gaskets. Back in 1936. they did find (1 find gaskets that the following relationship held for a variety of metal (tt/4) G2 Pt = W O L (A7) where G o = gasket outside diameter P J-i = leakage pressure W = total load Petrie published results of leakage pressure tests where the load was applied by tightening the bolts in a flanged joint. Table A3. bolt stress described. the British "Pipe Flanges Research Committee" published a few results of this type. . one might expect to find numerous test data in which the leakage pressure of a flanged joint was determined. shown in the last column of Table A3 and defined as: W (ttG //nPt JB. _£/ALJl (A8) wb 'bA . the tests were not really on flanged joints because the load was externally applied and controlled. not applied by tightening the bolts.

50 13.604 0.780 0.874 12.617 .D.D.300 4.500 35. and Gasket Material Sbl O.25 15.000 70.50 13.400 28.50 13. Bolts.908 0.500 48.643 0.000 21. LEAKAGE PRESSURE TESTS OF FLAT GASKETS IN FLANGED JOINTS.712 0.0625 40.' Size I.500 10.800 6.000 3. No.A-34 TABLE A3.900 5.00 Akro Metal 14.602 0.500 5. in.700 28..877 0.660 0.625 12-1 1/4 Akro Metal 9.000 5. Flange O.00 Copper Copper 46.50 13.375 tf psi PT .50 15.000 120.200 5.800 46. in.' 7. DATA FROM REFERENCE (5)..000 6.000 105.200 3.600 2.000 35.D.200 6.75 1.927 0. in.5 IT 16-1 7/8 It Asbestos ii 14.800 46.. L! psi P b 14.000 0.25 ti 13.000 66.800 0.5 16-5/8 Akro Metal 8.600 6.25 14.872 25.660 0.000 0.976 0.875 7.

of course. in contrast to Reference (11) tests where W /W. . P d consistently less than unity. Equation (A8) also assumes that a gasket cannot carry a tensile load. only a few papers have been published giving data on leakage pressures of flange joints. One of these (9) . hence no theoretical comparisons can be made. L bl a If S. The writer would guess that the fluid was water. Table A3 shows W /W. Reference (5) does not state what fluid was used nor how leakage was determined.„ decreases as P increases. are discussed in the following. Reference (11) states that leakage pressure tests were made using water as the pressurizing fluid and that "the first signs of hydraulic leakage were easily and accurately determined by visual ob servation". an assumption of doubtful general validity for asbestos gaskets. then W Ik. with leakage detected by visual observation. an axial force equilibrium equation. These are all tests on ANSI B16.A-35 where G P Li = gasket outside diameter = leakage pressure = initial bolt stress = total bolt area S. along with data from unpublished reports of Tube Turns (Division of Chemetron).5 flanged joints and thus are particularly relevant to this report. details of the flanges bolt load did decrease with pressure.. P D (with load controlled externally) was essentially equal to unity. are not known. During the years between 1937 and to-date. bl \ Equation (A8) is. This suggests that in Petrie's tests the However. The test procedure was generally as follows. . valid if S PT is equal to S. b2 p b will be less than unity.at P = 0.

the pressure was then reduced to zero. B16. The flange dimensions are specified in ANSI Each table gives: (a) The initial bolt stress. 1. S.A-36 (1) A flanged joint was made up with a new asbestos gasket. as defined by Equation (A8). bolt material. Tables A4 thru A9 give results from these tests. The bolts were tightened with a torque wrench to a desired stress level. (b) The leakage pressure. (2) Using water as the pressurizing fluid. It should be emphasized that this is not necessarily and generally is not the bolt stress at pressure. the pressure was raised in steps until leakage occurred as evidenced by drops of water emerging from the flanged joint within a short time after applying the pressure step. for pressure and for the sum of the two loads. The bolt stress was determined either directly with strain gages in the bolts or indirectly by calculations of torque versus bolt load. (3) Provided the gasket remained intact. Calculated (Computer program FLANGE. (c) (d) The performance factor. Each table shows the flange material.5. and gasket material and dimensions. The latter (combined) stress is valid only if the initial bolt load does not change with pressure. Stresses are shown for the initial bolt load. P . IBOND = 2) maximum flange stresses. W /W . . the bolts were tightened to a higher level and the leakage pressure was determined again. pipe material.

440 2.S.400 630 1.500 2.A-37 Table A4: LEAKAGE PRESSURE TESTS.S. see Equation (A8).2. U. Y.648 in.000 70.400 48.600 25. = 65. Reference (9). psi fi t—tl— (1) (2) Bolt Load Pressure Combined 10.100 2.360 1. G o = 15".000 psi Bolt Material A193 Grade B7 Pipe Material (.5" 1 o Total Bolt Area = 16 x .620 7.600 psi Gasket: 1/16" thick asbestos.95 0. .700 2.T. = 42.375" wall) A106 Grade B.T.920 16.000 (1) 700 1.= initial bolt stress (2) P J_i = leakage pressure P d (3) W /W.160 41.63 6.500 47.500 57.80 0.300 32.87 0. Flange Stress.000 24.728 = 11. 12" 300 class.000 psi Y.960 27.400 psi.200 35.900 39.S = 73.000 36.550 17.-.020 S. = performance factor.S.06 0.260 2.7 in.840 51. = 13.2 Ki bl psi pT L psi wp /K b (3) Calculated Max.900 1.500 2. U. WN to WN. 9 Flange Material A181 Grade 1. = 42. Test No.rrG2/4 = 176. G.72 0.900 2.

S.93 0.690 15.500" wall) Gasket: 6061-T6.330 19.2.020 11. Test No. Flange Stress.000 93.000 17.070 65.380 24.000 10.S.000 300 550 0.58 0.76 Bolt Load Combined 6.680 8.350 1.050 12.600 psi.850 2.S. psi Pressure Wtn/Wk P D 0.700 34.940 2.S. WN to WN.95 0.730 32.81 0. = 40.200 1.300 class.140 320 590 860 1.100 psi G o 1/16" thick asbestos. Y.728 = 11. U. Y.500" wall.100 1.600 1.220 .67 0.300 1.330 7. = 15".648 in.500 27. = 45.840 25.010 6.5" Total Bolt Area = 16 x. G 1 = 13.780 28.96 800 950 40.46 4.030 1. ttG2/4 = 176.000 13.A-3 8 Table A5: LEAKAGE PRESSURE TESTS.500 22.850 0. Pipe Material: (0.T.2.050 26. = 46.74 0.460 1.190 1. 5 Flange Material: 6061-T6. 12" .400 psi.700 10.080 4.560 11.000 3.000 48.800 psi Bolt Material: A193 Grade B7 0.68 0. o sbl psi PL psi Calculated Max.7 in.270 9. = 38. Reference (9).880 16.T. U.500 36.000 15.000 1.030 18.730 1.83 0.070 62.800 1.

16 1. 8" . U.30 1.000 psi (0.460 500 1.290 5.302 = 2.070 psi.322" wall) Gasket: 1/16" thick asbestos.070 * 3. G.S.790 6. WN to WN. P d Bolt Load Combined 9. 0. G o Y.375" 1 Total Bolt Area: = 8 x 0.000 31.380 22.66 in. 2 2 sbl psi PL psi Calculated Max. U. .150 class. ttG /4 = 88.890 3.T.625".09 600 920 1. .100 psi Bolt Material: A193 Grade B7 Pipe Material: 3003-F.630 1.750 20.S.500 17. 1 Flange Material: 6061-T6.350 * Pipe ruptured at 1.950 12.000 psi = 10.890 10. = 9.800 psi. Flange Pressure Stress W /W. = 43. Test No. = 47.000 300 1.S. Y.T.S.322 THICK ATTACHED PIPE (ALUMINUM) Reference (9). = 14.000 59.416 in.A-39 Table A6: LEAKAGE PRESSURE TESTS.060 1. = 5.

S.S.700 6.000 66. U.05 42.600 125.900 psi Bolt Material: A193 Grade B7 Pipe Material: A106 Grade B Gasket: 1/16" thick asbestos.T.040 28.600 143.060 59. Y.500 34.200 4.202 = 0.550 1.410 2.700 107.400 6. Bolt stresses have been converted to equivalent on 5/8" bolts.08 1.350 These and following tests were made using 3/4" rather than 5/8" bolts.270 340 390 1.760 49. sbl psi V psi Calculated Max.175 6.400 3.950 1.480 50 90 180 220 1.480 45.670 24.15 1.000 79.280 55.560 58.490 52. 150 CLASS.880 29.290 12. ttG2/4 = 19.16 1.500 54.700 5.370 38.700 89. 0. G.000 psi.55 1.030 52.110 300 320 350 390 3.600 134.140 32.80 1.700 98.160 2.200 12.71 1.500 80.11 1.21 29.680 35.170 675 1.64 1. Reference (12) Flange Material: A181 Grade 1.07 1.900 46.33 410 400 5.970 24.56" 1 Total bolt area = 4 x . .500 161.450 15. . = 35.216" THICK ATTACHED PIPE.760 56. psi Pressure Combined W /W P b Bolt Load 10.A-40 Table A7: LEAKAGE PRESSURE TESTS. G o o = 5".808 in.700 19.510 5. Flange Stress.600 116.690 8.790 31.830 420 450 460 480 500 520 42.630 15.950 6.960 38.740 33.500 5.44 1.640 8.39 1.71 42.360 29. = 3.000 4.303 87.44 1.35 19.63 in.990 4.400 32.150 5.350 5.500 152.210 49. 3" WELDING NECK TO SLIP-ON.000 22. = 77.070 36.25 1.500 1.200 4.

37 1.350 56.728 = 5.800 42.S.000 psi .1875". U.300 9.500 10.800 37.07 1.800 350 0. = 4.824 in.2.500 6.02 -1030 -1080 9.380 .420 55.920 -30 -440 -640 -810 -930 28.S. = 30. psi Pressure Combined 8.300 50.T. Y. G.27 4.A-41 Table A8: LEAKAGE PRESSURE TESTS.16 1.400 33.370 18. o sbl psi PL> psi P ° Calculated Bolt Load Max.22 1.63. G o = 6.900 25.400 18.32 1. CAPS WELDED TO FLANGES Reference (13) Flange Material: A105 Grade 2.480 27.000 9. Flange Stress. ttG2/4 = 30.600 psi Bolt Material: A193 Grade B7 Pipe Material: Gasket: A106 Grade B (caps) 1/16" thick asbestos.760 62.700 55.5" 1 Total Bolt Area = 8 x . .400 16.440 37.840 47.500 8.2.730 61. 4" 900 CLASS WELDING NECK TO WELDING NECK.500 11.030 46.07 in.460 1.

610 71.17 1.800 35.400* 3. 8" 150 CLASS.25 2. WELDING NECK TO WELDING NECK AND 8" FLANGES WITH REDUCED FLANGE THICKNESS Reference (14).000 78.400 27.625".416 in.800 88.930 96.500 1. and 7.17 1.500* 83.620 5110 4220 5030 4510 3990 1440 32.270 26. see Figure 3 of text.17 0. (2) Maximum stress is LLO for tests 1.30 0.400 103.000 96.75 0.250 1.50 0.800 35.96 33. Flange Thickness.930 3. .040 5580 38. G = 9.900 101. 2. .and 4.50 96.850 38.14 1.91 1. . 6..980 2.230 182.460 2.270 33.302 = 2. Bolt Material: A193 Grade B7 Pipe Material: Gasket: A106 Grade B.400 88.375" Total Bolt Area = 8 x .2 ttG2/4 = 88. 3. G = 10.000 1..24 1.500* 41.A-42 Table A9: LEAKAGE PRESSURE TESTS.220 3. ..600 (1) * Indicates that load/pressure was sufficient to produce measurable yielding of the flanges.17 1. Strength psi psi (1) W /W. RH for tests 5. bl PL> psi Calculated Max.322" thick 1/16" thick asbestos. 0.06 0. in.66 in 2 o Yield Test No.380 30. Flange Stress. p'"b Bolt Load psi Pressure Combined 55.000* 130.600 73.500 88.500 35.720 178.080 1..

5 times the 100 F rated pressure. in turn. 35. Material WN.9 + 7. P : h as shown by the following tabulation. may be due to either elastic effects or yielding of some part of the flanged joint. flanges.000 Y. or attached pipe. one would expect the effective modulus of the gasket to be small for . Factors higher than unity can be due to increase in bolt stress with pressure.S. Material 55. particularly if the effec tive modulus of elasticity of the gasket material is small.4 6.400 Y.S. Max.S. 6061-T6 Aluminum WN Slip-on A181 Gr.2 + (2) The "performance factor". i. Material 2900 2850 1070 + 6950 1000 1080 1080 2.500 Y.500 Y. 1 WN.9 3.6 2. 35.. bolts. P . V psi L. The performance factors less than unity may be due to decrease in bolt load which. generally decreases with increasing bolt load.6 + 413 413 16.A-43 Some significant aspects of the tests are (1) The flanges have a pressure capacity much greater than the standard hydrostatic test pressure of 1.e. Flanges psi 12 12 8 3 4 8 300 300 150 150 900 150 WN. A181 Grade 1 Steel WN. This factor is theoretically equal to unity if: (a) (b) The bolt load at pressure is equal to the initial bolt load. which is theoretically possible.S. The gasket cannot transfer tensile forces.7 2. Qualitatively.0 3240 413 2460 3250 + 2980 + 413 413 7. W /W. Material 88. 6061-T6 Aluminum WN.

. 3. flange material yield strength is 35. yielding at unexpectedly low loads. we ' speculate that some of the results were influenced by "sticking of. One exception to this is shown The calculated maximum stress is 26. (3) The values of maximum calculated flange stresses indicate that the flanges are capable of carrying loads substantially higher than the load to produce local yielding. in Table A9.303 = 3. 3. Test No.100 psi and hence.000 x 0. p b However.400 psi.A-44 low loads and large for high loads. leading to the trend of decreasing values of W /W. in Test No.the gasket" so that the gasket can transfer tensile forces. 3 of Table A9.322/8. it is often necessary to pry the joint apart. should not have yielded at the flanged-joint test pressure of 2. 4 was a It also indicated In several of the tests described in Tables A4 through A9. However.500 psi. yielding of the attached pipe may have occurred and thereby influenced the test results. Anyone who has taken apart a flanged joint using an asbestos gasket will recognize that. Test No.460 psi. retest of the flanged joint used in Test No. after removing the bolts. the yield pressure of the pipe would be expected to be about 2 x 40. with increasing initial bolt stress.

tightened with a torque wrench to a desired stress level.a-A^ TEST DATA WITH PIPE BENDING MOMENT We have not found any published data on the effect of pipe bending moments on flanged joints.237". (1) A flanged joint was made up with a new asbestos gasket. and (20) give theoretical approaches for estimating the effect of pipe bending moments. The bolts were The bolt stress was determined either directly with strain gages on the bolts or indirectly by calibration of torque vs bolt load. the addition of a pipe bending moment. 4" 300 3" 150 12" 150 6-3/16" 5" 15" 4" 3-9/16" 12. References (18). in. (19).D. in.5 welding neck flanged joints. 4" 300 class. bolts were SA-193 Grade B7 material. gaskets were 1/16" thick asbestos with diameters as shown below. and 0. 3-150. Data from unpublished reports of Tube Turns Division of Chemetron are presented in the following and compared with text These are "leakage pressure tests".25" The attached pipe was SA-106 Grade B material with wall thicknesses of 0. I.D. . respectively.250" for the 4-300. . and 12-150 flanged joints. 3" 150 class. but with The data are for tests on Equations (7) and (17). 0. and 12" 150 class B16. Thfi test procedure was generally as follows. but do not include test data. Gasket Flange O.216". . The flanges were made of SA-181 Grade 1 material.

The leakage pressure (4) Step (3) was repeated with higher bending forces. was then determined as in Step (2).150 class. 20".300 class and 625 psi for the 12" . Figures A12. and A14 also show the pipe bending moment versus leakage pressure relationships from text Equation (7). respectively. bl bl It is apparent that Equation (7). For example. is quite conservative with respect to the test data.1(b) limitation of pipe bending moments as controlled by bolt stress.650 psi pressure at these conditions.000 in. and 12-150 flanged joints. The results are shown in Figures A12 thru A17 as plots of pipe bending moment versus leakage pressure. using S m = S.. . moment is given in the References and is the force times the distance from the point of force application to the gasket. (3) The leakage pressure so obtained is the zero pipe bending moment leakage pressure plotted in Figure A12 through A17. . which is the Code NB-3647. and S. The pressure was then reduced to zero and a bending force was applied to one of the attached pipes. (6) An upper limit on applied pressures was imposed of 1. in Figure A12 Equation (7) implies that the joint is not capable of carrying any pressure with a pipe bending moment of 82.A-46 (2) Using water as a pressurizing fluid. = initial bolt stress. 3-150. The axial distance from the load application to the gasket of the flanged joint was 35". . the pressure was raised in steps until leakage occurred as evidenced by drops of water emerging from the flanged joint within a short time after applying the pressure step. and 60" for the The applied 4-300. A13.100 psi whereas the test data indicate that the flanged joint can hold more than 1. (5) The bolts were then further tightened and steps (2) thru (4) were repeated. the other attached pipe was anchored.650 psi for the 4" .-lb with an initial bolt stress of 25.

A-47 r Test Data Symbol D 24 X Sbl 8750 17800 25100 Did not leak O 22 t 20 a. O O 18 t ® 73 </> f ® © I6i- I G </> fi Q. THEORY FROM TEXT EQ. . (7) H-J40 60 80 100 120 Pipe Bending Moment. Eq. (7). a> o J* 14 - o 12 - Theory. 4"-300 FLANGED JOINT. 1000 in -lb FIGURE A12: LEAKAGE PRESSURES. TEST DATA FROM REFERENCE (10).

. 1000 in-lb FIGURE A13: LEAKAGE PRESSURES. 3"-150 FLANGED JOINT. (7). o 9 w I8h 16 Test Data o c cu l4 Symbols 12 a X sbl 9800 19985 •S 10 - o 29450 43120 A 8 CD- Theory.(7) 10 20 30 40 50 Pipe Bending Moment._ CO 20 Q. TEST DATA FROM REFERENCE (12). Eq. THEORY FROM TEXT EQ.A-48 32 30 A 28 A 26© r 24 22 .

A-49 Test Data o Sb. Eq. . TEST DATA FROM REFERENCE (15). (7). THEORY FROM TEXT EQ. FIGURE A14: 100.000 In-lb. 12"-150 FLANGED JOINT. LEAKAGE PRESSURES. =40000 psi t Did not leak 060 Theory.(7) Dl 2345 6789 Pipe Bending Moment.

TEST DATA FROM REFERENCE (10). 1000 in-lb FIGURE A15: LEAKAGE PRESSURES. THEORY FROM TEXT EQ. o O 2> 8 cr» o a 20 4 0 60 80 100 120 Pipe Bending Moment.A-50 Test Data Symbols a X Sw . psi 8750 17800 O 25100 t Did not leak w Q. (17). 4"-300 FLANGED JOINT. .

O O £ 3 (A CO CD Theory Eg=650l/S7. 1000 in-lb FIGURE A16: LEAKAGE PRESSURES. TEST DATA FROM REFERENCE (12). THEORY FROM TEXT EQ.A-51 32 30 & A Test Data ibc»ls D X Sb| . 3"-150 FLANGED JOINT. o» O a CD Theory Ea = 3xl07psi 20 30 4 0 5 0 60 Pipe Bending Moment. . ps« 9800 19985 0 A 29450 43120 co Q. (17).

12"-150 FLANGED JOINT. . 100.000 in-lb FIGURE A17: LEAKAGE PRESSURES. (17). TEST DATA FROM REFERENCE (15). THEORY FROM TEXT EQ.A-52 12 Test Data ° Sb|= 40000 psi t Did not leak Theory Theory Eg =650 -/S gi Ea 3x10 psi _L 234 56 78 9 10 Pipe Bending Moment.

and A17 show the test data along with two sets of theoretical pipe bending moment vs pressure relationships.. E approximately represented by the equation: E = 650 JS 1 g X gl (A9) where S .000 psi. The calculations using E = 3 x 10 psi were made because this is the assumed modulus used in calculations for the tables shown in the text. Reference (10) summarizes tests on various 1/16" and 1/8" thick asbestos gasket materials compressed between platens to determine the effective modulus of elasticity. A16. are based on a modulus of elasticity of the gasket of E = 3 x 10 psi. Both are based on Equation (17) of the text... over the range of S increases with gasket stress..g. bl The first set. used in the tests.000 psi and the joint has been in service for several years. = S. In our opinion.A-53 Figures A15.A. time-effects. shown as dashed lines in Figure A15. with the "40./A gl bit) f . E However... These lines are also quite conservative with respect to the test data. A16. However. S. Not all details of these tests are known (e.000" replaced by the initial bolt stress.5 flanged joint after the bolts have been tightened to 40. E = 3 x 10 psi is an appropriate representation of the characteristics of an asbestos gasket in a B16. and A17. at this point. when the tests were partly run with low bolt stresses and the tests were run immediately (within a few hours) after making up the joint. we are attempting to see how theory correlates with test data. S . size of test piece. if any) and they are non-linear. is at least from 200 to 6.

920 22. calculated by text Equations (7) and (17) are as folliDWS Sbl' psi Pressure.000 0 600 0 600 20. and A17 that use of these more realistic (but still very approximate) values for E with the test data.120 psi. the joint is "self-sealing" for pipe bending moments. Values of S .800 psi but such a trend is apparent for SK1 = 19. .300 class flanged joints with 0. = 9. .. The theory indicates that with S. psi Eq. (17) . but that at an initial bolt stress of 24.350 8.300 .320 15. and A17 are based on a. indicate that they imposed bending stress of S Their results up to 60. .000 psi was sufficient to pre vent leakage. an internal zero (25" head of water to detect leakage). . (7) 24.590 psi. In running cyclic pipe bending loading fatigue tests on 4" .237" wall attached pipe.800 bl slopes upward. We do not see such a trend in the test data for S.. Another bit of pertinent data is given in the paper by Markl and George .950 . pressure of 600 psi was applied. In most tests... gives significantly improved correlation One notes in Figure A16 that the solid line for S. . A16.720 20. It can be seen in Figures A15. the internal pressure was essentially In some tests. E = 18.A-54 The solid lines shown in Figures A15. A16. 34.985 psi and S = 43. and a„ values calculated using E from Equation (A9). they note that an initial bolt stress of 40.= 9.000 40. .000 psi was not sufficient for "specimens subjected to high deflections". 1 = 9.800 and the corresponding S = 818 psi. Calculated Spb (Psi) Eq.000 psi in the attached pipe.430 10.010 25..

200 psi and. using the Code evaluation method. From Figure 9 of Markl and George's paper.000/20 = 170. Equation (7). = 40.000 cycles.000 or a design life of 60.400.100 psi) is 3.000 psi.000 to 60.000 psi(P = 0) or around 20. the conclusion is reached that insofar as fatigue evaluation of the flange subjected to cyclic pipe bending moments is concerned. Effect of Pipe Bending Moment on Flange Stresses. Basic Theory Flange Stress.000 cycles. both equations are con servative with respect to the test data. * Chapter on FLANGE STRESSES. (range) pb was assumed to be 24. the usage fraction per cycle was 1/60. which indicated that. the value of 3. One additional aspect from the Markl and George paper that is worth comment is their fatigue tests results using welding neck flanges.A-55 Both Equations (7) and (17) agree with the test data in that they indicate that a bolt stress of 24. . An example is cited where S . with a bolt stress of 40.000 psi) deflections. In summary of the test data versus theory comparisons: (1) The Code theory. is around >n is arc pb 30. However.000 cycles. is conservative with respect to the test data. The equations indicate leakage when S . cognizing that the Code method incorporates a factor of safety of 20 on cycles.000 psi. the stresses in the flange due to tightening the bolts can be ignored. the joints did not leak at values of S pb up to 60.000 psi (P = 600 psi).000 compares adequately with the Code value of 60. the cycles to Re failure (stress amplitude of 12.400.000 psi is not sufficient to withstand high (S . In the text .

the agreement is about as good as should be expected. . Equation (17) using E = 3 x 10 psi. Equation (17) using E E g = 650^/S .A-56 (2) The basic theory. gives better agreement with the test data and. (3) The basic theory. is conservative with respect to the test data. considering the many uncertainties involved in the tests and approximations involved in the theory..instead of = 3 x 10 psi..

APPENDIX B DEVELOPMENT OF SIMPLE EQUATIONS FOR LIMITATION OF PIPE BENDING AND TORSIONAL MOMENTS TABLE OF CONTENTS INTRODUCTION Page B-l BOLTING STRENGTH REQUIREMENTS B-l FLANGE STRENGTH REQUIREMENTS EMERGENCY CONDITIONS B-10 B-12 FAULTED CONDITIONS TORSIONAL MOMENTS B-15 B-15 .

.

The aim of the Code rules is to provide a flanged joint which has sufficient strength so that the bolts can be prestressed sufficiently to prevent gross leakage under sustained loadings encountered in normal or upset conditions.1 method for limiting pipe bending moments. while theoretically sounder.B-l INTRODUCTION The text of this report presents two methods of evaluating the pressure and pipe bending moment capacity of flanged joints.. we have discussed criteria for minimum bolt area which involve the concept of a "leakage pressure". As discussed in Appendix A. We then consider the flange strength requirements and then examine limits for emergency and faulted conditions. i. so that gross leakage will not occur. the method given in NB-3647. In addition. can only be described as a very crude approxi mation. . the bolts will be prestressed to S.1 and the method incorporated in the computer program FLANGE. In this appendix. The computer program FLANGE. despite its complexity. never- the less uses rough approximations in the analysis for pipe bending moments. > S. Practically. the Code does not directly im pose any minimum on the bolt prestress. both methods appear to be highly conservative with respect to available test data. The equations developed herein are included as recommendations for Code revisions in Appendix C. a minimum bolt prestress becomes a necessity in order to pass the hydrostatic test and/ or not leak too much in service.. the Code rules should provide a flanged joint which will withstand dynamic loads en countered in normal or upset conditions and. The NB-3647. Equations (8) and (10).e.e. BOLTING STRENGTH REQUIREMENTS In the text. after the dynamic loads stop. we will develop simple equations for limitation of pipe bending moments starting with consideration of bolting strength re quirements for normal and upset conditions. i. Hopefully. the joint will return to a condition where gross leakage will cease. However.

Equation (B2) gives higher stresses than the discrete bolt area approach. . If the bolts straddle the neutral axis. regardless of the neutral axis location. by factors of 1. An equation analogous to (B2) could also be developed using the discrete bolt areas.08.04 for 4. = bolt stress due to pressure D_ = outside diameter of raised face P = internal pressure A. This is not necessarily the case for a flanged joint. 1. the Code rules should provide a flanged joint of sufficient strength so that the bolts do not break as a result of emergency or faulted conditions. then Equation (B2) gives the same answer as the discrete bolt area approach when a pair of bolts are on the neutral axis. C = bolt circle diameter. Pb = pb (B2) Equation (B2) is based on the assumption that the neutral axis coincides with a centerline of the imaginary pipe. and 1. = total bolt area The stress in the bolts due to a pipe bending moment can be developed by assuming that the bolt area is uniformly distributed along the bolt circle.8. If we assume the bolt diameter is small compared to the bolt circle dia meter.41. the value of o"bm would then depend on the orientation of the neutral axis. and 12 bolts. The stress in the bolts due to internal pressure is conser vatively given by the equation: °bp " <*'V Dr P/\ where (Bl) o. The imaginary cylindrical shell so formed has a thick ness of A^I-nC. = . respectively.B-2 Finally. The maximum stress in this imaginary cylindrical shell due to pipe bending moment is then: o bm Z (ir/4) C2 VOrC) ~C V4 = M. but Equation (B2) will not be unconservative by more than 11 percent. where A^ = total bolt area.

used in place of o.000 psi or higher.5 is essentially correct. for our purpose. we note that the loading. in relationship to Code allowable bolt stresses. S.000 psi.000 psi without violation of NB-3232. we will use S. of 20.000 = 16. The 100 F rating pressure does not exceed 2. in examples and Tables.9. is well below 7.6) the stress calculated by Equation (Bl) does not exceed 7.3 b However. we now seek a simple equation to establish the maximum bending moment that can be imposed on the flanged joint as limited by the strength of the bolting. for most sizes and classes. the upper bound to o" 100 F rating pressure is 2. S.0 S. from flange strength considerations.800 psi. Accordingly. we can conservatively assume that the bolt strength needed to hold the design pressure will not exceed the Code allowable bolt stress. greater than 25. In the remainder of this portion on bolting strength requirements.4 bp times the primary rating pressure. of the total bolt strength to pressure In NB-3232. only the 16" and 20" Class 900 slightly exceed 7. the value of o. hence. However.5 (Par. 6. It can be seen from Table Bl that the statement in B16.4 x 7. Having assigned 1. = 25. to indicate the allowable bolt strength as given in Code Table 1-1. . and solve . in our final recommendations.000 psi when the pressure is eaual to the primary rating pressure. using Equation (Bl).000 psi. what remains for moment loading? bolt stress is permitted to be 2.000 psi. The extra 1. .1. occurs at the 100 F rating pressure rather than at the primary rating pressure.000 psi. we will be using S.0 S . We then simply use Equation (B2).000 psi at 100 F is deemed desirable so that the bolts can be prestressed to 40. According to ANSI B16.000 psi and. will be replaced by 25. we cannot use a value of S. the highest bolt stress due to pressure.B-3 Having Equations (Bl) and (B2). b The lower limit on S. with S. b bm . can be allotted to bolt strength needed to hold the pipe bending moment. at the If the Code allowable bolt stress at 100 F is 20. It can also be seen from Table Bl that. Table Bl shows the calculated values of a.0 S.

520 5.000 4.340 4.440 2.250 5.480 3.630 8 10 12 14 6.660 2.360 4.650 5.960 6.210 4.200 2.930 6.150 5.920 1.070 5.740 6.320 — — — 3.200 7.200 6. psi A^ D = total bolt area.570 5.765 7.330 3.340 4.510 5.830 3.570 5.790 2.590 5.460 1.660 2.270 — 4.810 4. psi 1.250 5.600 6.250 6.640 3.590 5.650 5.460 4.640 2.060 — 3.700 5.890 6.640 7.170 6.380 4.500 Size 400 600 900 1/2 3/4 1 440 670 930 880 830 __ — 1-1/4 1-1/2 2 1.810 5.550 5. bp 150 = bolt stress at primary rat ing 300 pressure.710 1.780 2.460 4.870 6.880 4.920 2.280 4.770 1.420 4.190 4. = outside diameter of raised face.270 4.170 1.500 2.390 6.B-4 TABLE Bl: BOLT STRESSES BY THE EQUATION a = (ir/4) D^ P/A^ Nom.930 4.500 3.840 2.910 5.070 4.850 5.220 — — — — 1.850 5.620 4.090 5.930 1.730 5.010 4. sq-in.690 6.380 5.610 3.760 4.950 3.680 5.390 4.050 6.740 6.600 6.610 — — — 1.180 6.720 5. .110 6.150 5.090 5.270 — — 3.690 5.330 — P = primary rating pressure.290 4.660 5.630 4.570 4.760 6.820 1.710 5.140 6.220 6. inches.550 4.160 6.600 — 2-1/2 3 3-1/2 4 5 6 1.690 6.000 16 18 20 24 4.180 6.710 4.

e. pb It would be very useful to obtain a similar with simple equation for limiting M . Equation (B3) provides a very simple method for limiting M respect to bolt strength. but multiply the right hand side by one-half.1 but its complexity is not compatible The computer program FLANGE is deemed to be a more Let us see accurate method. A. i. V -c S V8 (B4) Table B2 compares allowable pipe bending stresses as obtained from Equation (B4) with allowable pipe bending stresses obtained by the NB-3647. we would like to have test data on flanged joints in which the bolt prestress is 40.B-5 for M . but it can not be made a part of Code rules. provides a method for limiting M with its accuracy.000 psi. It can be seen in Table B2 that Equation (B4) is a reasonably good approximation of the moment limits calculated using the computer program FLANGE. . what happens if we consider Equation (B3) as representing the essential variables (C. However.1 procedure and from the computer program FLANGE. so that gross leakage will not occur. There are no test data available in which flanged joints were loaded with pressure and/or pipe bending moment sufficient to break (or even yield) the bolts.. ) needed to represent joint leakage. pb to obtain: > ±CVb/4 Equation (B3) provides a limit on M it does not necessarily limit M <B3> based on strength of the bolts. S. . with respect to joint leakage. the data given in Appendix A show that moments well in excess of these permitted by Equation (B3) have been applied to flanged joints. NB-3647. To check Equation (B4) experimentally.

1 18.6 2.5 11.5 16.4 21.7 16 24 900 4 8 16 24 14.2 27.7 20.7 27.5 33.2 18.1(2) 13.1 19.1 15.4 (1) Typical pipe is defined as pipe with wall thickness equal to the larger of std.1 16 24 400 4 19.0 24.2 17.3 20.3 13.8 10.0 25.05.0 31.8 45.0 21.4 14.0 18.3 8.6 10.5 12.1 20.7 27.6 17.0 8 16 24 600 4 8 11.9 27.0 22.5 NB-3647. or F D /17.1 16 24 300 4 8 12.8 36.0 19.7 17.8 27.4 54.500 + 0.3 9.B-6 TABLE B2: COMPARISON OF BENDING STRESS IN TYPICAL (!) ATTACHED PIPE Spb (1 .1 27.5 15.1 24.1 19. (B4) FLANGE^ 13.000 Class Size psi) Eq.2 9.500 4 8 12 34.1 6.7 13.3 10.7 7.7 21.9 24.6 18.9 27.4 17.0 25.6 16.3 10.0 17.0 23.6 150 4 8 11. wt.8 12.3 20.8 5.9 1.3 19.500 4 8 42.7 27.9 12. ° Values from text Table 9.1 40.6 14.9 25.3 23.4 27.8 22.8 11.6 16. (2) .6 13.1 23.8 16 24 13.

of course. the 3" size . wt.000 b pb in. was 40. Table B3 shows the stresses in "typical" attached pipe where the maximum moment is calculated by Equation (B3) with S^ = 25.3 ksi. The m value of S m for A-106 Grade B pipe at 100 F is 23. As a yardstick for comparison. but it can be used. = 25.000 psi.g. Typical pipe is defined as pipe with wall thickness equal to the larger of std. e. wall or P D /17. the stress in straight pipe due to thermal expansion and anchor movements is permitted to be 3 S .000 psi.9 ksi. that Equation (B4) is highly conservative. A-106 Grade B pipe at 100 F. A-106 Grade B pipe at 100 F. the data suggest that the moment could be moment.. It is apparent from Table B3 that many small size flanges have an "allowable pipe bending stress" by Equation (B3) that is greater than 69. with S. the permissible moment will be severely restricted. wt. for many flanged joints. On the other hand.g. the 8" 150 class is limited to a moment of 0. Data given in Appendix A is not ideal in this sense.076 times that for 8" std. doubled before leakage at 625 psi occurs. wt. e. are simply one-half of those shown in Table B3. Table B4 shows the stresses in "typical" attached pipe where the maximum moment is calculated by Equation (B4) with S. the 8" .6/69. = 25.05. In the smaller sizes of the 150 class.. It can be seen in Figure A14 that there was no leakage at this Indeed. Figure A14 represents perhaps the best test for the purpose because the initial bolt stress Equation (B4). These It is apparent in Table B4 that only a few small size flanged joints will not be restricted in permissible moment as compared to "typical" straight pipe.500 + 0.9 times that permitted on 8" Std.000 psi.-lb.B-7 internal pressure is equal to the rated pressure and the pipe bending moment is increased until leakage occurs.000 psi. Figures A12 and A13 also indicate It is pertinent at this point to discuss the magnitudes of moments that can be applied to flanged joints in comparison to the moments that can be applied to straight pipe.9 ksi. gives M . values. 3 S r m =69.150 class where the allowable moment is 10. even though the pressure was 625 psi rather than the rated pressure of 275 psi at 100 F._ = 267. The larger sizes of the 150 class are severely restricted.

7 36.0 53.3 39.6 37.0 — 154 127 — — 86.3 54.6 39.8 49.000 PSI.2 90.1 41.0 12.9 46.05.5 55.1 42.6 37.9 110 23.1 55.0 — 83.5 14 16 18 20 14.2 48.6 16.0 — — — 156 168 147 117 — 192 178 — — 2-1/2 3 26.5 34.7 18.8 46.8 35.5 20.6 15.5 45.8 38.7 — — (1) Typical pipe is defined as pipe with wall thickness equal to the larger of std.8 38.6 34.5 62.7 34.1 37.4 55.500 2.3 32.500 1/2 3/4 1 184 123 203 232 133 — 203 232 133 — 564 327 276 157 519 292 239 192 — 1-1/4 1-1/2 2 74.0 83.5 72.8 — — 83.6 28.7 40.3 34.2 44.2 38.3 41.4 38.9 75.4 104. .3 85.1 17.6 — 83.6 55.3 51.9 53.9 — 63.0 37.4 21.2 25.2 36.7 /4.5 16.3 28.0 37.4 58.2 90.0 38.B-8 TABLE B3: BENDING STRESSES IN TYPICAL ( ) ATTACHED PIPE WITH MAXIMUM MOMENTS PERMITTED BY EQUATION (B3) USING S.5 58.1 47.9 10.1 60.6 — 3-1/2 4 5 6 8 10 12 29.500 + 0.6 88. b Nom Spb (ksi) V 150 300 for Class Indicated Size 400 600 900 1.0 24.4 58.4 104.5 23.9 — 29.4 42.2 90._ = 25.0 39.5 34.6 51.4 83. wt.2 46.0 — — 24 32.3 35.5 67.9 34. wall or F D /17.9 97.

6 17.0 26.6 18.4 13.2 14.6 — 2-1/2 3 — — 76.8 30.0 — — 24 10.2 12.2 31.0 18.7 27.2 16.1 88. b • •- '- • ' ~ S .9 24.9 25.7 19.4 5.3 (1) Typical pipe is defined as pipe with wall thickness equal to the larger of std.7 52.5 28. = 25.4 — 23.500 + 0.0 21.7 — 83.4 23.6 29.8 26.1 45.2 260 116 66.2 10.4 58.9 19.500 2.1 — _ 282 164 138 78.7 27.4 8.5 — 43.1 14 16 18 20 17.8 18.9 18.4 17.8 14.05 .0 41.7 — — 21.9 61.2 96.0 37.0 41.1 — 45.5 — — 1-1/4 1-1/2 2 41. wall or F D /17.6 19.8 29.1 7.0 31.7 52.0 19.7 23.B-9 TABLE B4: BENDING STRESS IN TYPICAL ( ' ATTACHED PIPE WITH MAXIMUM MOMENTS PERMITTED BY EQUATION (B4) USING S.4 37.0 14.0 23.8 63.5 42.0 22.3 19.0 41.2 19.3 7.500 1/2 3/4 1 91.6 19.6 16.3 27.7 — — 96.8 14.4 17.0 19.1 20.3 — 3-1/2 4 5 — 25.8 6 8 10 12 11.8 73.2 17.2 27.3 78.4 20.3 9.9 17.5 — — 146 119 37.8 19.3 25. Nom.4 — 18.2 21.5 41.2 36.7 — 8.1 44.8 48.8 8.7 54.8 33.7 29. wt.0 22.3 101 __ 101 116 66.1 45.2 17.5 6.8 17.3 12. Size Pb 150 (ksi) 400 for Claiss Indicated 900 300 600 1.000 PSI.7 11.3 18.

we note that the allowable If stress for SA-193 Grade B7 is 35.5 flanges are deemed to limit is 7. For short-time moments. . but the bolts will remain elastic because the total strength re quirement is. from Code Table 1-1. For moments which act over a short period of time. bp + a. we take S. As long as the pressure and pipe bending loads do not reduce the gasket stress to zero. If the bolts are prestressed to 40. Leakage may occur.000 psi. Now./2. In the above.780 psi. at most.B-10 is severely limited. the allowable value for SA-193 Grade B7 at temperatures up to 700 F in Code Table 1-7. have ample strength to withstand these loads.000 psi. These results correlate with field experience in that leakage problems occur relatively frequently with the 3" and 8" 150 class.000 psi . the higher moment limit of Equation (B3) is deemed appropriate.4 + 12.000 psi at 100 F.000 psi will not be exceeded.000 psi at 100 F and that other bolt materials have allowable stresses as high as 50.3.3 (Class 1). we are essentially limiting the loads so that a prestress of 40.000 psi. The worst case is the 16" 900 class at 100 F for which a.000 = 42.200 x 2. FLANGE STRENGTH REQUIREMENTS Table 18 of the text shows calculated flange stresses for a bolt preload stress of 40. 2 S. bm 7. the flanged joint would not be expected to exhibit gross leakage with pressure up to the rated pressure in combination with moments up to that permitted by Equation (B4). the flange stresses (except in the hub) will not change significantly. if the flanges were capable of supporting a bolt prestress of 70. = 25. The value of S.4 + 25.200 x 2. we have taken S.280 psi. the worst case B16.500 = 29. Equation (B4) is deemed to provide a suitable limit for sus tained moment loads. is not greater than one-third of the bolt material yield strength. By limiting the pressure to b the rated pressure and the sustained moment to S.

000 psi. The change in flange strength as temperature increases must also be considered.. Equations (B3) and (B4) should be restricted to bolt materials which have an al lowable bolt stress in Table 1-1.5 flanges are capable of supporting a bolt prestress of 40. B8M and B8T bolt materials. with a specified minimum yield strength of 36.000 psi.000 psi. siderably higher than 40.g. B8C. in Equations (B3) and (B4) be replaced with the specific value This does not prohibit the use of bolt material with higher Further.000 psi. that available test data and most field experience are with a material like SA-105. the values of S. in Table A9. along with experience with B16. the higher allowable bolt stresses given in Table 1-1.3 of 20.3 cannot be used because we are not sure of the strength of the flanges.000 psi.g. Class 1 bolt material. SA-193 Grade B8. Tables A4 through A9 show test data where bolt prestresses were much higher than 40. from Table 1-1. SA-182.B-ll or 100.000 psi prestress of 25. for SA-540. The preceding consideration also leads to the rejection of higher allowable bolt stresses given in Table 1-7. e.5 flange materials even for a bolt stress of 40. at this point. The pertinent flange material property is its yield strength.000 psi and that there are other flange materials with lower yield strength. This will rule out.3 could be used in Equations . it is noted that flange yielding occurred at a bolt stress of 78. allowable stresses. In addition. in order to assure that 40. Elastic theory cannot be used to justify bolt prestresses in excess of 40.3.1. in Table A7.000 psi. is sufficient to convince the authors that all B16. The (B3) and (B4) with correspondingly larger permissible values of M data in Table 18. e. However. in particular.000 psi. note that in Table 18 we have stresses far above the yield strength of many B16.. Type 304 with .000 psi. The flange strength considerations lead to the recommendation that S. We can speculate that most B16.000 psi.5 flanges are capable of carrying bolt prestresses con Indeed. Accordingly.500 psi. we recognize.5 flanged joints. the bolt prestress was as high as 161. Grade B21. can be applied without exceeding the requirement of NB-3232.

to be obtained from Code Table 1-2. M < 6.5 times the 100 F rating pressure.5 rated pressures decrease with increasing temperature approximately in proportion to the decrease in flange material yield strength with increasing temperature. it is recommended that the limit for sustained moment loading during normal M fs < 3. and upset conditions be given by the equation: Accordingly. the emer gency condition pressure cannot exceed the B16.000 psi.250 (S /36)CAb (B6) The value of (S /36) should not be taken as greater than unity.125 (S /36)CA^ ~~ y b (B5) Where S is the yield strength (ksi) of the flange material at the design For short-time moments: temperature. for normal and upset conditions we have used up about all of the assured * NB-3655.1 states that the pressure under emergency conditions shall not exceed the design pressure by more than 50 percent.1 is quite acceptable for B16. However.B-12 a specified minimum yield strength of 30.1 by more than 50 percent. . otherwise." The words "calculated. If the user does calculate P by Equation (2). in accordance with Equation (2) of NB-3641.1"should be deleted as they are at best. It is significant to note that the B16. useless. we might violate bolt prestress limitations and/or bolt strength requirements.5 flanged joints. y both allowable pressures and moments decrease with decreasing S .2. he will simply end up with the answer that P = P. margin of flange strength. the pressure limit of NB-3655. Accordingly. Because the design pressure cannot be higher than the B16. the question arises We note that as to what co-incident pipe bending moment can be permitted.1 states that "permissible pressure shall not exceed the design pressure (P) calculated in accordance with Equation (2) of NB-3641.5 test pressure of 1.5 rated pressure. EMERGENCY CONDITIONS NB-3655. Accordingly.

we can take advantage of the fact that for most B16. and (b) Mfd 1 [11. ksi.5 flanges- However. Bending stresses in typical attached pipe calculated by Equation (B7) are shown in Table B5. PP by Equation (Bl) is less than 7.000 x 1. (S /36) £1. gency conditions that: We can. if not all.5 = 63. . therefore.-lb. the permissible moment is higher for emergency conditions than for normal/upset conditions.5 times the design pressure.900 class) 7.250 A^ -(tt/16) D2. even though the emergency condition pressure is 1. there is no basis available for estimating the magnitude of the leak if.5 flanges.00. a. it could be contended that flange yielding would be acceptable and would be part of the "repair of damage".6 + 25.5 flanges can withstand this loading but there is neither test data nor theoretical calculations that can be used to defend such speculation for all B16.3). B16.5 times the rated pressure. However.420 psi. in.200 x 3. require for emer (a) The pressure shall not exceed 1. The authors can 16" . P ]C (S /36) where (B?) P.. = pressure under emergency conditions concurrent with M„ td fd M fd = bending or torsional moment (considered separately) as defined for M but including dynamic loadings. for example.B-13 Recognizing that emergency conditions require shutdown for cor rection of the conditions or repair of damage in the system (NB-3113. In most flanges. along with stresses calculated by Equation (B6) and results from the text Table 9. The bolt stress would then be (for the worst case. speculate that most of. D = outside diameter of raised face S = yield strength of flange material at design tempera ture. we allowed the pressure and the moment to be 50 percent higher than permitted for normal and upset conditions..200 psi (see Table Bl).

S = 36 ksi at 100 F at 100 F.1 14.2 39. y y Strength limits.000 FLANGE psi) Temp = 700 F Temp = 100 F = NB-3647.0 15.8 13.3 16.6 11.6 10.2 24.0 30.4 49.3 27.0 12.9 7.6 36.1 40.4 79.3 39.8 12.7 20.5 52.6 27.7 38.3 28.8 55.0 42.5 rated pressure for carbon steel flange material.2 9.1 24.9 55.8 31.2 32.0 41. (B7) (3) Class Size (1) (1) 150 4 8 16 24 13.6 30.2 51.9 54.3 56.1 Eq.1 27.6 46.8 22.7 110.8 91.0 65.7 27. (B6) (2) Eq.6 79.3 13.8 45.2 34.7 36.9 38.9 46.8 9.8 78.4 39.3 10.7 1. Spb (1'.8 34.1 16.1 16 17. from Table 9 of text.3 19.9 38.2 24.8 47.9 19.8 900 4 8 40.5 65.7 36.1 27.9 37.9 57.2 30.6 10.1 18.6 18.1 27.2 35.500 4 8 16 24 33. SA-105 FLANGE MATERIAL.8 27.9 46.8 32.5 10.4 33.5 11.0 27.5 33.4 40.2 26. normal and upset conditions.8 25.500 4 8 12 34.0 90. (B6) (2) Eq.2 23.2 26.7 24.6 21. emergency condition with co-incident pressure of 1.1 15. S = 25.6 27.6 6.8 14.7 21.9 23.3 61.1 24.7 7.0 22.4 55. .2 102.3 43. (B7) (3) Eq.1 300 4 8 16 24 20.6 16.9 34.5 17.7 24 13.0 62.7 25.0 11.6 16.5 13.1 51.3 (1) (2) (3) Leakage pressure limits. Strength limits.5 15.0 72.1 23.0 24.8 11.0 35.1 23.1 35.6 34.9 ksi at 700 F.9 25.7 72.1 40.4 42.5 85.0 43.4 21.6 12.0 37.6 48.3 24.0 17.8 23.1 19.5 41.1 12.0 33.2 400 4 8 16 24 600 4 8 16 24 19.3 37.3 20.9 12.8 53.7 21.1 41.1 20.4 14.2 27.7 36.8 30.8 2.7 27.6 25.5 times the B16.1 13.2 27.5 26.5 23.B-14 TABLE B5: COMPARISONS OF BENDING STRESSES IN TYPICAL ATTACHED PIPE AT 100 F AND 700 F.7 47.0 61.

Large rotations would be prevented by contact between the bolts and bolt holes but. the permissible moment for emergency conditions will not exceed 1. While this may be above the B16. TORSIONAL MOMENTS The discussion so far has concerned bending moments. how the permissible leakage rate could be correlated with the leakage characteristics of flanged joints where yielding of the flanges might occur. it is not apparent to the authors what that permissible leakage rate might be and.8 times that for normal/upset conditions. FAULTED CONDITIONS NB-3656.0 times the design pressure. we should not depend upon the bolt shear resistance when concerned about leakage.B-15 Usually. These torsional moments are resisted by frictional forces between the flange faces and the gasket.5 test pressure. by reference in Code Appendix F. because bolt diameters are 1/8" less than hole diameters. While under faulted conditions.5 rated pressure. Equation (B7) will tend to permit higher moments than Equation (B6). it is deemed acceptable for B16. for such cases. it is not apparent that anything more liberal than Equation (B7) can be recommended. all that is required is that the flanged joint restrict leakage enough to permit a safe shutdown. the design pressure is somewhat less than the B16. if it were known. With a coefficient of friction y between flange faces and gasket. However. the torque capacity M of a flanged joint is . this occurs for Pfd = 0. states that the pressure under faulted conditions shall not exceed 2. for both pressure and moment loadings. Flanged joints may also be subjected to torsional moments applied by the attached pipe. However.5 flanged joints both from a bolt strength and flange strength standpoint.

The (S /36) factor is not needed for torsional moments but. For high loads at the gasket interface.i ' \ ' ^Pc \ =3125 c\ <B10> Equation (B5) gives Mfg = 3125 C Afo (for S = 36 ksi) hence. equation (B9) gives M = 5250 C ^ in.000 psi. We note that G <C. The unit axial force (lb/inch) on the gasket is Sb2Ab/irG. we first observe that the assumptions that G/C =0. For pressure loading only.7. If we assume that G/C = 1/2 and use Sfe2 = 25. sb2/sbl is m°re typically about 0. a coefficient of friction of u = 1/2 appears reasonable.-lb. for simplicity.000 psi. hence. and G is the gasket diameter. G/C is more typically equal to about 0. equation (B5) will be recommended for both bending moments and torsional moments—considered separately.000 psi.5 flanged joints and most operating conditions.000 psi( the same S moment evaluation) then equation (B9) gives assumed for bending V . For short time loadings. we can apply the same limit to torsional moments as for bending moments. where <j> is the co-ordinate angle.dF o (B8) where dF is the torsional resistance per inch of gasket circumference.5 and Sbl = 25. if Sbl = 40. Accordingly: V -/(!)' o ttG dcf> = (B9) We want to compare equation (B5) with equation (B9). With these more typical ratios. and S (at operating conditions) may be less than the preload bolt stress of (assumed) 40. insofar as leakage is concerned. where we do not necessarily design to prevent temporary gross leakage. and dF = p^^/ttG) (G/2) d<f>. Sb2 = 30.75. These can be considered separately because the bending moment does not reduce the total load on the gasket.B-16 2-rr Mpt=J f.000 psi are quite conservative for most B16. it decreases the load on the tension side but increases the load on the compression side. This is almost as high as the .

large movements are prevented by the bolts. .000 psi. in. equation (B9) is deemed as a suitable limit for short-time torsional moments as well as short-time bending moments.-lb. = 6250 C A. no test data exists on the torsional moment capacity of flanged joints. Insofar as the authors are aware. the torsional moment capacity of the bolts alone is 13000 Ci/2 = 6250 C A^ in. Accordingly. it would be highly desirable to obtain some test data.B-17 limit given by the analogous equation (B6) of Mf. If we assign an allowable shear stress for the bolts of 13. While the limits developed in the preceding are deemed to be conservative.-lb. In addition.

.

APPENDIX C RECOMMENDED CODE REVISIONS .

C-1

NB-3600

It is recommended that the NB-3600 portion be revised as follows:

(1)

Change NB-3651.2 to read:

NB-3651.2

Piping Products for Which Stress Indices Are Not Available.

For analysis of flanged joints, see NB-3658.

For other piping products for

which stress indices are not available, see NB-3680.

(2)

Change NB-3647.1 to read:

NB-3647.1

Flanged Joints.

(a)

**Flanged joints manufactured in accordance with the standards listed in
**

Table NB-3132-1, as limited by NB-3612.1, shall be considered as meeting the requirements of NB-3640.

(b)

**Flanged joints not included in Table NB-3132-1 shall be designed in
**

accordance with Article XI - 3000, including the use of the appropriate

allowable stress given in Table 1-7.

(3)

Add a New Paragraph NB-3658 to Read as Follows:

NB-3658

Analysis of Flanged Joints

**Flanged joints using flanges, bolting, and gaskets as specified in
**

ANSI B16.5 (1968) and using a bolt material having an S - value at 100 F not

**less than 20,000 psi may be analyzed in accordance with the following rules or
**

in accordance with NB-3200.

ance with NB-3200.

Other flanged joints shall be analyzed in accord

NB-3658.1

Design, Normal and Upset Conditions

(a)

Bolting.

The bolting shall meet the requirements of NB-3232.

In addition,

the limitations given by Equations (15) and (16) shall be met.

M

fs < 3125 (S /36)C Ab ,

(15)

C-2

where

M.

rs

**= bending or torsional moment (considered separately)
**

applied to the joint due to weight, thermal expansion

**of the piping, sustained anchor movements, relief valve
**

steady-state thrust, and other sustained mechanical loads, in-lb. If cold springing is used, the moment may be

reduced to the extent permitted by NB-3672.8. S = yield strength (ksi) of flange material at design

temperature (Table 1-2.2).

The value of (S /36) shall

**not be taken as greater than unity.
**

C = diameter of bolt circle, in.

A^

**= total cross-sectional area of bolts at root of thread or
**

section of least diameter under stress, sq. in.

M

d ± 6250 (S /36) CAb

,

(16)

where

M

**= bending or torsional moment (considered separately) as
**

defined for M but including dynamic loadings, in-lb.

(b)

Flanges.

Flanges of ANSI B16.5 flanged joints meeting the requirements

of NB-3612.1 are not required to be analyzed under NB-3650.

However,

**the pipe-to-flange welds shall meet the requirements of NB-3652, NB-3653,
**

and NB-3654, using appropriate stress indices from Table NB-3683.2-1.

NB-3658.2

Emergency Conditions

(a) (b)

The pressure shall not exceed 1.5 times the design pressure. The limitation given by Equation (17) shall be met.

Mfd <[11250 ^ - (tt/16) D2 Pfd] C(S /36) ,

(17)

.0 S m . p fd = pressure (psi) during the emergency condition concurrent with M£. the limitation on (S /36) and A. m NB-3658. . (c) Pipe-to-flange welds shall be evaluated by Equation (9) of NB-3652.1(a).2(b) shall be met.C-3 where Df = outside diameter of raised face. (c) Pipe-to-flange welds shall be evaluated by Equation (9) of NB-3652. and Mf. S ..0 times the design pressure. Analysis for Testing Conditions is not required. NB-3658. C. using a stress limit of 3. (b) The limitation given by Equation (17) of NB-3658. are defined in NB-3658. fd Mf. are pressures (psi) and moments (in-lb) occurring concurrently during the faulted condition.3 Faulted Conditions (a) The pressure shall not exceed 2.4 Testing Conditions. using a stress limit of 2. where Pf. in.25 S .

(4) NC-3673.C-4 NC-3600 It is recommended that the NC-3600 portion be revised as follows. (3) NC-3673. (5) Change NC-3647. (a) Flanged joints manufactured in accordance with the standards listed in Table NC-3132-1.5(a) Change equation numbers: (12) to (15) and (13) to (16). bolting. (6) Add a New Paragraph NC-3658 to Read as Follows: NC-3658 Analysis of Flanged Joints Flanged joints using flanges. (1) NC-3651. Add: (c) For analysis of flanged joints. as limited by NC-3612.1 to Read: NC-3647. see NC-3658.1 Flanged Joints. and gaskets as specified in ANSI B16. (2) NC-3671.5(b) Change "Formulas (12) and (13)" to "Equations (15) and (16)".1. shall be considered as meeting the requirements of NC-3640.1.value at 100 F not . Delete Period After "NC-3647" and Add: and NC-3658. (b) Flanged joints not included in Table NC-3132-1 shall be designed in accordance with Article XI 3000.5 (1968) and using a bolt material having an S .

However. in-lb. allowable stresses from Table m 1-7 and 1-8 shall be used in place of S .2(b)-l. Normal and Upset Conditions (a) The limitations given by Equations (12) and (13) shall be met.2). the moment may be reduced to the extent permitted by NC-3673.5 flanged joints meeting the require ments of NC-3612. sustained anchor movements.C-5 less than 20. The value of (S /36) shall not be taken as greater than unity. the pipe-to-flange welds shall meet the requirements of NC-3652. NC-3658. relief valve steady-state thrust and other sustained mechanical loads. 6250 (S /36) C A^ (13) where M_. Flanges of ANSI B16. . = bending or torsional moment (considered separately) as fd defined for M- but including dynamic loadings. sq. in. (b) Flanges. in.000 psi may be analyzed in accordance with the following rules or in accordance with NB-3200.1 are not required to be analyzed under NC-3650. M <. in-lb. total cross-sectional area of bolts at root of thread or section of least diameter under stress. Other flanged joints shall be analyzed in accordance with NB-3200. If cold springing is used. If the NB-3200 analysis is used. y o . thermal expansion of the piping. M :. fs — < 3125 (S /36) C A. (12) where M bending or torsional moment (considered separately) fs applied to the joint due to weight. using appropriate stress intensification factors from Figure NC-3673.1 Design. C diameter of bolt circle.5. yield strength (ksi) of flange material at design tempera ture (Table 1-2.

Pf. using a stress limit of 1. Mfd <[11250 A^ . are defined in NC-3658.0 times the design pressure. In ND-3673.4 Testing Conditions.3 Faulted Conditions (a) The pressure shall not exceed 2. (c) Pipe-to-flange welds shall be evaluated by Equation (9) of NC-3652. NC-3658. . using a stress limit of 2. The limitation given by Equation (14) shall be met. (b) The limitation given by Equation (14) shall be met. = pressure (psi) during the emergency condition concurrent with M^-j. in. Do not change the phrase: "using Equations (9) and (10) (ND-3650).". . the present equation numbers (9) and (10) are to be changed to (15) and (16) in two places. where (14) Df = outside diameter of raised face. ND-3600 It is recommended that ND-3600 be changed as shown for NC-3600.C-6 NC-3658. (c) Pipe-to-flange welds shall be evaluated by Equation (9) of NC-3652..1(a).5.(tt/16) D2 Pfd] C(S /36) . Analysis for Testing Conditions is not required.4 S. changing NC to ND in all places where NC occurs. C. id Mf. S . where Pfd and M_.8 S . NC-3658.5 times the design pressure.2 Emergency Conditions (a) (b) The pressure shall not exceed 1. are pressures (psi) and moments (in-lb) occurring concurrently id during the faulted condition. the limitation on (S /36) and A.

- Dual Plate_wafer_21 11 12 (2)
- Test Aria Acqua
- Guida Piano Sanitario Metasalute
- Breve Manuale Epanet Ita
- practical galvanic series
- UNI en ISO 14122 - 4 Mezzi Di Accesso Al Macchinario
- calendario 2013
- INSTALLATION_O&M_GSM_RX800_ENG&ITA.pdf
- Installation_O
- guida piano sanitario metasalute.pdf
- Micro Central i
- Centrali Idroelettriche
- Centrali Idroelettriche
- Kaplan
- 04 South America Brazil
- 1990 - Io Speriamo Che Me La Cavo
- catalogo viteria
- A fiber reinforced plastic joint for filament wound pipes
- 10 TRAVATURE RETICOLARI
- Motore Asicrono - Spiegazione
- Manuale Sui Motori 2
- EN2manual
- EN2manual
- Painting Colours for Rady Mixed Paints and Enamels
- Butterfly Flu01 Omb 04 11

Sign up to vote on this title

UsefulNot usefulEVALUATION OF THE BOLTING AND FLANGES
OF ANSI B16.5 FLANGED JOINTS

EVALUATION OF THE BOLTING AND FLANGES

OF ANSI B16.5 FLANGED JOINTS

OF ANSI B16.5 FLANGED JOINTS

- A Stress Analysis of Pipe Flange Connections
- ANSI B16.5 1998
- BS EN 1515-4:2009 Flanges Bolting
- Gaskets
- Get WRC Bulletin 506 Half Bead Temper Bead
- Bolt and Nut Standard Size Catalog
- Bolted Flanged Joints New Methods Practices - Proceedings
- ASME B.16.5
- ASME B30.1 - 2004 (Jacks)
- ASME PTC17 Reciprocating Internal Combustion Engines
- ASME B16.5 - 2013
- Piping Joint Handbook (Flanges, Gaskets, Bolts)[1]
- A193 B7M BOLT PROPERTIES
- Bolt Design
- bolt & nuts (nice to know)
- ASME BTH-1-2011
- WRC-107
- asme b30.3 2009
- Asme Flange
- ASME SECTION IX QUESTIONS
- ASME B16.5-2013
- bolting
- BOLTS AND NUTS DETAILS
- ASME Section IX.updated
- tube
- Asme Bpvc Section II-d 2007
- Gasket Handbook
- API RP 1110 2012
- EVALUATION OF THE BOLTING AND FLANGES OF ANSI B16.5 FLANGED JOINTS