Trevor Jee - Flange Stress Analysis Guidance

E N G I N E E R I N G

C O N S U L T A N T S

FLANGE STRESS ANALYSIS GUIDANCE NOTES

APRIL 2005

TECHNIP 35 R01A.DOC R01A.DOC

DOCUMENT CONTROL

REVISION HISTORY Rev Rev

Date Date

Descr Descr ipt ion

A

26/4/05

First issue

By

Check

AGK

PACM

DISTRIBUTION Mr Duncan Warwick Technip Offshore UK Limited Enterprise Drive Westhill Aberdeen Aberdeenshire AB32 6TQ Tel 01224 271703 Fax 01224 271271 E-mail [email protected]

ISSUED BY Eur Ing Alan Knowles Trevor Jee Associates on behalf of Jee Ltd 26 Camden Road Tunbridge Wells Kent TN1 2PT Tel 01892 500 775 Fax 01892 544 735 E-mail [email protected] © Trevor Jee Associates Associates 2005. The moral rights of the author have been been asserted. The text of this work work may be quoted in any form (written, visual, electronic or audio), up to and inclusive of five hundred (500) words without express written permission of the publisher. Notice of copyright must must appear on either the title, copyright or or acknowledgements acknowledgements page of the work in which it is quoted, as follows: “Quotations from [name of work] are copyright © 2005 Trevor Jee Associates. Associates. Used by permission”. permission”.

Technip 35 R01a flange stress analysis guidance.DOC

2/41

DOCUMENT CONTROL

REVISION HISTORY Rev Rev

Date Date

Descr Descr ipt ion

A

26/4/05

First issue

By

Check

AGK

PACM

DISTRIBUTION Mr Duncan Warwick Technip Offshore UK Limited Enterprise Drive Westhill Aberdeen Aberdeenshire AB32 6TQ Tel 01224 271703 Fax 01224 271271 E-mail [email protected]

ISSUED BY Eur Ing Alan Knowles Trevor Jee Associates on behalf of Jee Ltd 26 Camden Road Tunbridge Wells Kent TN1 2PT Tel 01892 500 775 Fax 01892 544 735 E-mail [email protected] © Trevor Jee Associates Associates 2005. The moral rights of the author have been been asserted. The text of this work work may be quoted in any form (written, visual, electronic or audio), up to and inclusive of five hundred (500) words without express written permission of the publisher. Notice of copyright must must appear on either the title, copyright or or acknowledgements acknowledgements page of the work in which it is quoted, as follows: “Quotations from [name of work] are copyright © 2005 Trevor Jee Associates. Associates. Used by permission”. permission”.


2/41

CONTENTS

1

INTRODUCTION......................... INTRODUCTION ......................... ............................. ............................ ............................ ........5 1.1 1.2

2

CONCLUSIONS ......................... .............................. ............................. .............................. .....6 2.1

3

MAXIMUM SHEAR STRESS ................................................................................... .....................................................................................................14 ..................14 VON MISES EQUIVALENT STRESS........................................................................................14

DESIGN METHODOLOGIES.................................................................................................17 6.1 6.2 6.3 6.4

7

SCOPE.........................................................................................................................................8 LOADING ON FLANGES................................................................................................... FLANGES.............................................................................................................8 ..........8 FLANGE FAILURE FAILURE MODES MODES ........................................................................................ ........................................................................................................9 ................9 DESIGN APPROACHES .................................................................................................. .............................................................................................................9 ...........9 SUMMARY OF OF CODES CODES AND METHODS METHODS .................................................................................10 22 ASME B16.5 CLASS SPECIFICATION SPECIFICATION ..................................................................................11 RECOMMENDED PROCEDURE PROCEDURE ..............................................................................................12

STRESSES USED IN ANALYSIS..........................................................................................14 5.1 5.2

6

FUTURE WORK ................................................................................................... ..........................................................................................................................7 .......................7

OVERVIEW OF DESIGN PROCESS.......................................................................................8 4.1 4.2 4.3 4.4 4.5 4.6 4.7

5

ALLOWABLE STRESSES ................................................................................................ ...........................................................................................................6 ...........6

RECOMMENDATIONS .............................. ............................. .............................. ................... 7 3.1

4

BACKGROUND ............................................................................................. ...........................................................................................................................5 ..............................5 OBJECTIVES...............................................................................................................................5

ORDER OF CONSIDERATION AND OF DESIGN....................................................................17 NON-STANDARD FLANGES ............................................................................................... ....................................................................................................17 .....17 STANDARD WELD-NECK FLANGES.......................................................................................18 SWIVEL RING FLANGES..........................................................................................................18

DISCUSSION OF DESIGN METHODS .......................... .............................. ......................... 20 7.1 7.2 7.3 7.4 7.5 7.6 7.7 7.8

OVERVIEW................................................................................................................................20 SIMPLIFIED STRESS ANALYSIS ANALYSIS .............................................................................................21 THERMAL GRADIENT ANALYSIS............................................................................................23 LOW TEMPERATURE FLANGES.............................................................................................24 AGGRESSIVE ENVIRONMENTS ENVIRONMENTS INCLUDING INCLUDING SOUR SERVICE ............................................24 FINITE ELEMENT ANALYSIS ANALYSIS ............................................................................................. ...................................................................................................25 ......25 COFLEXIP PROGRAMME PROGRAMME ............................................................................................. ........................................................................................................25 ...........25 TECHNIP MATHCAD SHEETS SHEETS .................................................................................................26


3/41

8

FLANGE GRADES, BOLTS AND GASKETS ............... ....................... ............... ............... ............... ............... ............... ............... ...........27 ...27 8.1 8.2 8.3 8.4

9

FLANGE MATERIAL MATERIAL GRADES .................................................................................................27 METRIC AND IMPERIAL STUD BOLT BOLT SIZES ..........................................................................28 BOLT TENSIONS ............................................................................................. ......................................................................................................................28 .........................28 GASKETS ....................................................................................... ..................................................................................................................................30 ...........................................30

RELEVANT READING...........................................................................................................36

10 REFERENCES........................ REFERENCES ........................ ............................. ............................ ............................ ..........39


4/41

1 INTRODUCTION

1.1

BACKGROUND Technip Offshore UK Limited (Technip) have been developing calculation methods to verify flange and bolt stresses on externally loaded flanges. This has included writing three MathCAD sheets 1, 2 & 3 for weld neck and swivel ring flanges, which cover both the PD 5500 4 and ASME VIII 5 codes. These sheets take as the starting point that a flange has been selected, so the type, material, and pressure/temperature rating are known. They then check whether that flange can withstand applied bending and tension. A software package developed by Coflexip 6 is also used by Technip to design fixed weld neck and swivel ring flanges to ASME VIII 5 code, using RX, R and BX gaskets. It again allows bending and tension to be applied at the joint but it ignores temperature derating. It calculates the bolt torque required during make up and stresses under operating and test conditions are examined. Technip have asked Trevor Jee Associates: 1. To write the technical guidance on how Technip designs flanges. 2. To research allowable stresses and bolt loads and to include advice on this in the technical guidance. Trevor Jee Associates is an independent company which carries out pipeline engineering consultancy and associated training for the oil and gas industry.

1.2

OBJECTIVES The overall objective of this study is to produce a guidance document covering the calculation methods used to verify flange and bolt stresses for externally loaded weld neck and swivel ring flanges. The main tasks are: ■ Determine allowable stresses ■ Write guidance document ■ Manage project


5/41

2 CONCLUSIONS

2.1

ALLOWABLE STRESSES The stresses which are allowed depend upon the method being used. With the standard flanges of ASME B16.5 22 and API TR 6AF 24, there is no need to examine the yield or permitted stress. The approach is to determine a classified flange for a particular temperature and pressure rating. The flange manufacturer selects suitable materials. With ASME VIII 5, the yield and membrane stresses are listed for US steels. These are derated depending on operating temperature. With PD 5500 4, the permissible stresses are listed for BS and European steels. These are derated depending on operating temperature. With FEA, providing all external forces and moments have been considered for installation, hydrotest and operational conditions, then it is permissible to assess the equivalent von Mises stress in the bolts and flange with the corresponding material yield value suitably de-rated for temperature.


6/41

3 RECOMMENDATIONS

3.1

FUTURE WORK 3.1.1 COFLEXIP METHOD It is recommended that a check be made on the existing Coflexip software. This requires a review of the coding pages of manual. From a reading of the Coflexip manual, it would appear that the minimum yield stress rather than the de-rated permissible stress intensity may have been used. It is recommended that a retrospective review be undertaken to clarify what permissible stresses were used as input to their flange design. If the minimum yield stress was used directly, then a list of currently installed flange locations should be drawn up.

3.1.2 MATHCAD SHEETS It would be convenient for designers to have assistance in selecting dimensions and stresses. For this, it is recommended that look up tables be written in Excel to select standard flange sizes, and to select suitable material stresses. These could be embedded into MathCAD. The existing MathCAD sheets are incomplete and unverified. It is recommended the MathCAD sheets for weldneck and swivel ring flanges be developed and verified.

3.1.3 THERMAL GRADIENT The thermal gradient can introduce additional bending stress in the flange faces. It is recommended that the critical temperature differential causing excessive additional stresses due to a thermal gradient be identified.


7/41

4 OVERVIEW OF DESIGN PROCESS

4.1

SCOPE This report limits itself to a discussion of subsea flanges for pipelines and spoolpieces. These are weld-neck and swivel flanges in carbon steel. For subsea use, it is common to use an iron or soft steel ring gasket located in a groove within the bolt circle diameter. Occasionally, corrosion resistant alloys are used for flanges and in this instance, matching suitable materials are used for these rings. The consequences of these materials’ properties are discussed.

4.2

LOADING ON FLANGES 4.2.1 FLA NGE FORCES AND LOA D COMBINATIONS Subsea flanges are subject to: Internal pressure due to the contained liquid  External hydrostatic pressure   Axial loading due to residual lay tension and temperature effects in pipeline  Bending moments due to thermal effects or misalignment of spool piece  Thermal reduction in strength of flange and bolting material due to the temperature of the contained liquid Thermal stresses due to the temperature gradient between the inside and outside  walls of the flange  Shear across the joint, which can only be resisted by the gasket It is normal to consider the following conditions as a minimum:  Bolt tightening at installation  Hydrotest Operational conditions 

4.2.2 DESIGN AND ASSEMBLY CONSIDERATIONS During initial assembly, there is no hydrostatic pressure differential and generally zero or very low axial forces and bending moments. Bolts are tensioned to a stated value for gasket seating.


8/41

The joint is then pressure tested, during which some movement of either the pipeline or spoolpiece may take place, subjecting the flanges to axial and bending displacement. Finally, in operation, temperature effects and further expansion may impose additional axial and bending movement. During these subsequent stages, the bolts are not normally adjusted in a subsea situation – although onshore, this can occur. This means that the initial tension applied to the bolts needs to be sufficient to maintain sufficient force on the gasket to ensure a seal without any increased bolt tension resulting in yielding. Any design process needs to verify as a minimum the three conditions of assembly, hydrotest and operation. Allowance should be made when assessing stresses for the reduction in crosssectional area at end of life due to corrosion.

4.3

FLANGE FAILURE MODES 4.3.1 INSTALLA TION FAILURES Flange leakage can be caused during installation by many conditions, including:  Incorrect flange or groove facing  Incorrect gasket size or material  Dirty or damaged flange faces  Inaccurate or uneven bolt stress – reliance on the torque of the nuts rather than measuring bolt extension may result in under- or over-tensioned bolts These can be classified as lack of care in specification or installation.

4.3.2 OPERATIONAL FAILURES Flanged joints can fail by:  Overstressing of the flange material Yielding of the bolt material  Leakage through the gasket seal interface  High vibration levels in adjacent pipeline and spool – causing repetitive damage  to the gasket and groove or loosening of the bolts In many cases, a combination of material overstress eventually results in gasket leakage. The last failure is due to inadequate support of the joint. Flange failures do not result in rupture of the pipeline itself. However, for hazardous product lines, any leak (however minor) should be considered as a failure, and must therefore be taken seriously.

4.4

DESIGN APPROACHES 4.4.1 NON-STANDARD WELD-NECK FLA NGES For non-standard weld-neck f langes, it is necessary to: Derive using ASME VIII 5 or PD 5500 4 or  Analyse using finite element analysis. 


9/41

4.4.2 STANDARD WELD-NECK FLA NGES For standard weld-neck flanges, it is possi ble to:  Select from ASME B16.5 22, API 605 28 or API TR 6AF 24,  Derive using ASME VIII 5 or PD 5500 4 or Analyse using finite element analysis.  4.4.3 NON-STANDARD SWIVEL RING FLA NGES There are no standard sizes for swivel ring flanges larger than 346 mm (13 5/8in) of API 17D 27. It is necessary to:  For small well head flanges, select directly from API 17D 27 or Derive using ASME VIII 5 or PD 5500 4 or  Analyse using finite element analysis. 

4.5

SUMMARY OF CODES AND METHODS Each of the codes and methods has pros and cons:

4.5.1 ASME B16.5 22  Is limited to standard Classified US weld neck flanges only (<24in ND)  Does not allow for applied moments or tension  Cannot be used for high pressure flanges such as API 6A 23 Provides a list of ASTM steel grades for flanges and bolts  4.5.2 API 605 28  Equivalent to above  Is limited to standard US weld neck flanges only (26in to 60in ND) 4.5.3 API 17D 27  Equivalent to above for weld-neck and tapered swivel ring flanges for subsea wellheads  Is limited to 5 000 and 10 000 class rating up to 346 mm (13 5/8in) ND 4.5.4 API TR 6AF 24  Is limited to standard API flanges only  Concentrates on small diameter at higher working pressures Only method to explicitly allow for thermal gradient stresses  Uses a graphical method based on results of FEA. This is difficult to adapt into  a convenient spreadsheet for design over a range of temperatures and pressures Reported that six of the standard flanges did not work! They leaked or yielded.  4.5.5 ASME VIII 5 Provides general approach to design of non-standard flanges  Uses yield and membrane stresses, derated depending upon operating  temperature Provides tables for US steels  Latest version is in metric but with imperial bolts  4.5.6 PD 5500 4 Provides general approach to design of non-standard flanges  Uses design stresses*, derated depending upon operating temperature  (*equivalent to ASME VIII 5 yield and membrane stresses)  Provides tables for UK and European steels  Does not give formulae for derivation of F, V and f (though same as ASME VIII 5 so the latter may be used) Requires reduction in stress for larger diameters (>1 m) See Clause 3.8.3.4.2. 


10/41

4.5.7 COFLEXIP PROGRAMME Calculates non-standar d weldneck and swivel ring flanges   Based on ASME VIII 5 method  Seems to use yield strength rather than the correct allowable temperaturederated values The swivel ring deviates slightly from ASME VIII 5 loose ring analysis  4.5.8 EN 1092 31, EN 1591 32 & 33 & EN 1759 34  Not yet fully released  Provides standard dimensions of Classified metric DIN type and imperialconverted ASME flanges Method of analysing non-standard flanges and gaskets  4.5.9 FINITE ELEMENT ANA LYSIS  Can calculate non-standard weldneck and swivel ring flanges  Can allow for all applied moments, thermal and tensile effects Compares von Mises equivalent stress with temperature de-rated yield  Needs a substantial investment of time and effort to set up and run the model  It is the only method available if the temperature gradient effect needs to be  included Can be used to design all flange shapes including the tapered swivel rings of  API 17D 27

4.6

ASME B16.5 22 CLASS SPECIFICATION 4.6.1 PROCEDURE AND LIMITA TIONS Most flanges are specified to a class by the pipeline engineer rather than fully designed by him. The manufacturer can then select suitable material and supply a flange to that class; a flange that has been certified as adequate for the particular service by the regulatory body. ASME B16.5 22 covers standard flanges in carbon steel for use with pipelines. It includes blank (blind) flanges, threaded, lapped and slip-on designs as well as the weld-neck flanges that are being considered in this document. It does not cover swivel ring flanges. The latest version of the document is in SI units with imperial unit conversions in Annex F. The dimensions of the flanges are provided for a number of classes: 150, 300, 400, 600, 900, 1500, 2500. It is common to see these written as 150# etc, and they are sometimes mistakenly referred to as 150 lb etc. Whereas there is somewhat of a correlation between working pressure and Class (in pounds) at around 441°C (825°F) for the more common materials, this is not a consistent for all materials and flange Classes (especially Class 150 flanges). The standard flanges are sized by consideration of the material, temperature and pressure, in order to determine the Class. No consideration is made of moment or tension from the pipeline, nor is there any allowance for thermal gradient through the flange.


11/41

To accommodate these other stresses, it is common to go up one class for subsea spools, where large moments or thermal loads may occur. That is to say, if Class 900 has been calculated, then Class 1500 would be specified.

4.7

RECOMMENDED PROCEDURE Figure 1 shows a simplified flow diagram of this recommended approach. Initially, all internal pressures, temperatures and pipeline bore should be determined. Select a standard flange if possible as the basis for the design. However, it should be noted that ASME B16.5 22 only specifies standard weld neck flanges up to 610 mm (24in) nominal diameter. Larger flanges must be to API 605 28. For standard flanges not subject to external moments or tension stresses, it is possible to look up their requirements in ASME B16.5 22 tables for a suitable class. There are no standards for the dimensions of swivel ring flanges other than the limited selection of small wellhead designs of API 17D 27. However the swivel ring flange must mate with the weld-neck flange and bolt sizes, so some standardisation is necessary. Use ASME VIII 5 or PD 5500 4 for design of non-standard weld necks or swivel ring flanges. The preference on which of these codes is used depends upon the materials being supplied and where the flanges will be installed. The former code should be used for American steel and US waters; the latter code is better for European steel. If the calculated stresses are too high compared with the de-rated maximum shear stresses (Secton 5.1) then it will be necessary to revisit the initial size of flange and reanalyse. If there is a severe thermal gradient (which could cause additional bending stress in the flange), then it may be possible to minimise this effect using insulation. However, if this is not possible then it may be necessary to check the stresses using Finite Element Analysis (FEA). Only use FEA as a last resort. It is the only method that can capture all effects but does involve a substantial design effort. Check the equivalent stress (Section 5.2) against yield. If this is too high then the flange dimensions need to be adjusted. A fundamental consideration with all the above design procedures is the selection of the correct bolts and gaskets. Guidance is given in Section 8. It is recommended that the design should follow the three stages – find a similar standard size of flange; develop using ASME VIII 5 or PD 5500 4; then check using FEA if necessary.


12/41

flange dimensions known bore, determine temperature and pressure of product

look-up table in Excel

swivel ring

match bolt diameter and positions of the mating standard flange

flange type?

weld-neck

2in to 24in

use standard sizes from API 605 flange stresses

yes

MathCAD US/UK steel?

UK

yes graphs or FEA

external momt or force?

design to PD 5500

design to ASME VIII

select from API TR6 AF2 or undertake FEA

26in to 60in

use sizes from ASME B16.5 or API RP6

flange stresses

US

nominal diameter

no

insulate?

severe thermal radient?

no select rating from ASME B16.5 or API 605

no

yes end

Figure 1 Flow chart for flange design Technip 35 R01a flange stress analysis guidance.DOC

13/41

5 STRESSES USED IN ANALYSIS

5.1

MAXIMUM SHEAR STRESS Both ASME VIII 5 and PD 5500 4 make use of the maximum shear stress. This is compared with a basic allowable stress equal to the lesser of 2/3 the yield or 1/3 the ultimate. The latter limit prevents rupture. It is an essential limit when following the method

5 STRESSES USED IN ANALYSIS

5.1

MAXIMUM SHEAR STRESS Both ASME VIII 5 and PD 5500 4 make use of the maximum shear stress. This is compared with a basic allowable stress equal to the lesser of 2/3 the yield or 1/3 the ultimate. The latter limit prevents rupture. It is an essential limit when following the method set out in these two codes. When calculations were undertaken by hand, this method was easier to use than the equivalent stress method. Maximum shear stress theory is slightly less accurate but is always conservative compared to von Mises for ductile materials. However, von Mises method should not be used for brittle materials that might be subject to rupture. A full explanation of the ASME VIII 5 stress definitions and combinations is described in Appendix 4-1 and Figure 4-130.1. See also Section 7.2 for a discussion of the stress intensity design approach.

5.2

VON MISES EQUIVALENT STRESS Finite Element Analysis (FEA) commonly determines the equivalent stress from the component stresses using von Mises equivalent stress theory. It then compares this to an allowable stress equal to 2/3 the minimum specified yield. It is common to modify this allowable stress by a design case factor based on the probability of occurrence of a particular load com bination. This may be greater or lesser than 1.0. For example, API RP 2A-WSD 7 permits an increase in allowable stress (from 60% to 80% of yield) for stresses due to combinations of environmental loads. Von Mises method has been shown to more accurately predict the onset of yield for ductile materials than the older method.


14/41

The three principal stresses should be calculated at all critical points of the flange. At locations with axisymmetric geometry such as found in plain pipe, the principal stresses will usually be in the axial, hoop and radial directions. (For nonaxisymmetric geometry, the directions may be different.) API 8 have defined that the principal stress components should be classified as one of the following:

Primary

Any normal or shear stress that is necess ary to have static equilibrium of the imposed forces and moments. A primary stress is not self-limiting. Thus, if a primary stress subsequently exceeds the yield strength, either failure or gross structural yielding will occur.

Membrane

Bending

Secondary

p is

the average value across the thickness of a solid section excluding the effects of discontinuities and stress concentrations. For example, the general primary membrane stress in a pipe loaded in pure tension is the tension divided by the cross-sectional area. σ p may include global bending as in the case of a simple pipe loaded by a bending moment. b is

the portion of primary stress proportional to the distance from the centroid of a cross section, excluding the effects of discontinuities and stress concentrations.

q is

any normal or shear stress that develops as a result of material restraint. This type of stress is self limiting, which means that local yielding can relieve the conditions that cause the stress, and a single application of load will not cause failure.

Table 1 API definition of stresses Combining the stresses at each cross-section is commonly done using von Mises yield criterion as defined by the following equation:

σe =

(σ1 − σ 2 )2 + (σ 2 − σ3 )2 + (σ 3 − σ1)2 2

Where: σe =

von Mises equivalent stress σ1, σ2, σ3 = principal stresses

5.2.1 API 6AF2 26 This adopts the same approach as ASME VIII 5 Division 2 Appendix 4. In Section 4.3, it recommends design stresses be limited to: ST = 0.83 · S Y and Sm = (2/3) · S Y


15/41

Where: Sm = design stress intensity at rated working pressure   ST = maximum allowable general primary membrane stress intensity at hydrostatic test pressure SY = material minimum specified yield stress  Alternatively, it permits combining triaxial stresses based on hydrostatic test pressures and limiting them to: SE = SY Where:  SE = maximum allowable equivalent stress at t he most highly stressed distance into the wall, computed by the distortion energy theory method  SY = material minimum specified yield stress The distortion energy theory method is more commonly known as the von Mises equivalent stress method.


16/41

6 DESIGN METHODOLOGIES

6.1

ORDER OF CONSIDERATION AND OF DESIGN We will discuss non-standard flanges first because the design of these is a general case. Standard flanges can then be understood as a special case. However, it is recognised that common design practice is to start with a standard flange and then check that it is suitable for any applied moments and forces.

6.2

NON-STANDARD FLANGES 6.2.1 RECOMMENDED STEPS FOR DESIGN Following the simplified flow diagram of Figure 1, the following steps should be undertaken for the design of non-standard weld-neck flanges: Determine the diameter, temperature, pressure, moments and axial tension  Determine flange dimensions by selecting a standard flange of the correct bore  Use ASME VIII 5 or PD 5500 4 to pick suitable material and to calculate the  flange stresses. Use could be made of a MathCAD sheet or the Coflexip programme 6  Increase flange dimensions if stresses exceed the maximum shear stress (Section 5.1) and repeat analysis Insulate to counter any severe thermal gradient or carry out a full FEA using the  flange sizes as a basis. 6.2.2 ASME VIII 5 AND PD 5500 4 These codes consider in detail the pressure, moment and loads on the flange. They provide a method of sizing weld-neck and the other flange types considered in ASME B16.5 22. But in addition, it provides a method of sizing swivel ring flanges by adapting the lap joint flange design. However, the method makes some assumptions:  For make-up there is no moment or tension in the system, only bolt tension.  There is no allowance for shear or thermal gradient stresses during operation.

6.2.3 COFLEXIP PROGRAMME 6 The Coflexip programme 6 generally follows the method described in ASME VIII 5 and PD 5500 4 to design non-standard flanges. Technip 35 R01a flange stress analysis guidance.DOC

17/41

However, the appendix at the end of the guidance notes seems to indicate the use of material yield stress rather than the temperature de-rated permissible stress. It is important that the correct material property be used – namely the basic allowable shear stress. Refer to Section 5.1.

6.3

STANDARD WELD-NECK FLANGES 6.3.1 DESIGN USING STANDARD FLA NGES Standard dimensions have been determined for weld-neck type flanges in US and European pipe diameter. With the standard flanges of ASME B16.5 22 and API TR 6AF 24, there is no need to examine the yield or permitted stress. The approach is to determine a classified flange for a particular temperature and pressure rating. The flange manufacturer selects suitable materials. However, the tables assume that there are no applied moments or axial forces acting on the flange. Nevertheless, it is possible to select a standard flange from the tables and then check to ensure it will withstand the applied forces and moments in operation. This would then assist with size selection for a non-standard flange design.

6.3.2 RECOMMENDED STEPS FOR DESIGN Following the simplified flow diagram of Figure 1, the following steps should be undertaken for the design of standard weld-neck flanges:  Determine the diameter, temperature and pressure  Determine the Classified flange dimensions by selecting a standard flange of the correct bore using ASME B16.5 22 , API 605 28 , API TR 6AF 24 or equivalent European standard.  If designing to API TR 6AF 24, then the effect of thermal gradient has been included. Otherwise, insulate to counter any severe thermal gradient. 6.3.3 WELD-NECK FLA NGE DIMENSIONS AND PRESSURES The dimensions for standard weld-neck flanges up to 24in diameter are given in Tables 5 and 8 through 22 of ASME 16.5 22. The dimensions are given in mm apart from the stud bolts, which are in inches. API 6A 23 gives standard flange dimensions for higher operating pressures. A new combined code is being produced which specifies both the US and modified DIN standards. This is EN 1092 31 and EN 1759 34.

6.4

SWIVEL RING FLANGES 6.4.1 DESIGN USING SWIVEL RING FLA NGES There are no standards for the dimensions of swivel ring flanges other than the limited ratings and sizes smaller than 346 mm (13 5/8in) of API 17D 27. Instead, these are usually proprietary items but must be designed to match the bolt diameter and spacing and the gasket size for that of the matching weld neck flange. They are similar in principle to lap joint flanges and can be designed to ASME VIII 5 or PD 5500 4. The swivel inner hub profile and full thickness of the swivel outer can give the flange the same strength and external load capabilities as a weld neck flange.


18/41

The figure to the left of Figure 2 shows how the flange hub is commonly socketed into a recess in the ring. The codes, however, only provide design methods for the simple ring-type, lap joint flange shown to the right. The additional steel of the socketed ring helps stiffen it to resist ‘toroidal’ bending of the flange faces.

Figure 2 Socketed and simple ring type swivel ring flanges The shape of flanges of API 17D 27 is similar to that on the left, but the tapered ring/hub interface is at an angle of 25° rather than vertical. If this style of flange were to be required at a non-standard size or rating, then full finite element analysis (FEA) would be required.

6.4.2 RECOMMENDED STEPS FOR DESIGN Following the simplified flow diagram of Figure 1, the following steps should be undertaken for the design of swivel ring flanges:  Determine the diameter, temperature, pressure, moments and axial tension  Design the mating weld-neck flange dimensions Use ASME VIII 5 or PD 5500 4 to pick suitable material and to calculate the  flange stresses. Use could be made of a MathCAD sheet or the Coflexip programme 6 Increase flange dimensions if stresses exceed the maximum shear stress  (Section 5.1) and repeat analysis  Insulate to counter any severe thermal gradient or carry out a full FEA using the flange sizes as a basis.


19/41

7 DISCUSSION OF DESIGN METHODS

7.1

OVERVIEW 7.1.1 WELD NECK FLA NGES Pipeline flanges 610 mm (24in) nominal diameter or less for hydrocarbon transport are commonly sized according to the standard imperial dimensions set out in ASME 16.5 22. For each diameter, there are a number of classes to match the pressure and temperature rating of the pipeline. Originally, the flanges were for land-lines but the same dimensions can be used with external overpressure for subsea use. However, no account is made by this method for the stresses induced by a thermal gradient. DIN produced an equivalent set of standard flanges for metric pipe diameters. Both are now listed in the new EN 1092 31 and EN 1759 34. API has produced ranges of flanges for high pressure hydrocarbon and wellhead use. Additional work has recently been undertaken by API to make allowance for thermal gradient. This was undertaken using finite element analysis and published as sets of graphs.

7.1.2 LA RGER FLA NGE DIAMETERS Some subsea pipelines used for hydrocarbons are larger than the standard diameters of flange given above. API 605 28 gives dimensions of flanges from 660 mm to 1524 mm (26in to 60in). Alternatively, two design codes, ASME VIII 5 and PD 5500 4 may be used for nonstandard weld-neck flanges. Both use a similar approach to determine the stresses.

7.1.3 SWIVEL RING FLA NGES There are no standard sizes for swivel ring flanges other than the limited selection from API 17D 27. However, for subsea use, it is necessary to be able to marry up the alignment of the boltholes on either side of the flange connection without imparting torque into the pipelines. The approach of ASME VIII 5 and PD 5500 4 can be used for swivel ring flanges. Technip 35 R01a flange stress analysis guidance.DOC

20/41

7.2

SIMPLIFIED STRESS ANALYSIS Both ASME VIII 5 and PD 5500 4 use an analysis method developed in the 1930s. It was based on the Taylor Forge method, further develope d by Walters et al 9 . Additional information on the method is given by Singh 10 . It simplifies the problem to a 2 dimensional balance of bolt, pressure and gasket reaction forces using plain shells.

Bolt load Flange hu b Tapered hu b

Shell

Gasket reaction

Pressure load

Figure 3 Simplified axial forces in flange The flange tends to rotate due to the moment from the combination of axial loads of the bolts, gasket and pressure. Refer to Figure 3. To this must be added the rotation due to the internal pressure, which is more pronounced in flanges without a tapered hub section. Refer to Figure 4.

Figure 4 Radial deflection of shell and flange with internal pressure load The bolts are deemed to be a point load as if they were spread in a thin continuous membrane around the flange at the bolt centreline. Other axial forces are deemed to act at the centreline of the gasket or shell (see Figure 3 for definition). To this is added the internal pressure during operations. An initial value of bolt force must be selected to account for makeup. This can then be checked for overstressing (or slackening) in operational use. The method can incorporate tensile forces and applied external moments. But no allowance is made for thermal gradient stresses.

7.2.1 ASME VIII 5 MAXIMUM SHEAR STRESS ASME VIII 5 uses the concept of stress intensity, which is defined as twice the maximum shear stress; which in turn is equal to the algebraic difference between the maximum and minimum principal stresses: SI = 2 τmax = σ1 - σ3


21/41

The stresses permitted by ASME VIII 5 are given in Tables 1A and 1B of Section II, Part D Material Requirements Tables for ferrous and non-ferrous alloys. The permitted operational tensile membrane and yield stresses (S m and Sy) for each material are reduced with increasing operating temperature. Values for a full selection of carbon, low and high alloy steels and nickel alloys are given. The same section also gives guidance for quenched and tempered steel and clad pipe. Section 3-350 of ASME VIII 5 provides the maximum allowable flange and nozzle stresses (Sf and Sn) are compared with the maximum longitudinal, radial and tangential (SH, SR , and ST) stresses, defined in Section 3-340 (a). Note that for weld-neck and swivel ring flanges, the flange and nozzle are normally from the same material; but the thickness or final treatment may cause the allowable stresses to differ slightly. In other types of welded flange, they may be made from two different materials. The code recommends that design stresses be limited to: SH ≤ the lesser of 1.5 S f or 2.5 S n SR ≤ Sf ST ≤ Sf (SH + SR )/2 ≤ Sf (SH + ST)/2 ≤ Sf

7.2.2 HYDROTEST Note that in the US, it is common to pressure test at 1.25 times the MAOP at the flange (API RP 1110 11 , ASME B31.4 12 and B31.8 13 ) rather than the 1.5 times specified elsewhere in the world (including the UK under the old BS 8010 14 ) and for some in-company procedures. The ratio of ST/Sm = 0.83/(2/3) = 1.245. This means that where the test pressure is the higher value, the flange membrane stress may reach yield.

7.2.3 PD 5500 4 NOMINAL DESIGN STRESS This British Standard (formerly BS 5500) closely follows the approach of ASME VIII 5. However, in Tables K.1-5 through K.1-7 of Appendix K, it provides the minimum tensile, the minimum yield and the nominal design strength values (R m, R e and f N) for UK grades of steel at a range of temperatures. This also gives guidance on the maximum design lifetimes for steel running at very hot temperatures (>350°C or more). Such temperatures are normally outwith the design condition of subsea pipeline flanges. The method used by PD 5500 4 for deriving minimum yield/proof stresses for elevated temperatures is that described in BS 3920 15 , now superseded by BS EN 10314 16 . Section 2.3 of the code shows the derivation of nominal design strength. In the following, variables are:  f E = nominal design strength  R m = minimum tensile strength specified at room temperature  R e = minimum specified yield strength at room temperature R e(T) = minimum specified yield strength at elevated operating temperature 


22/41

For carbon, carbon manganese and low alloy steels, the following strengths apply: Up to 50°C, f E = R e/1.5 or R m/2.35, whichever is the lower value   With a known value of elevated temperature R e(T), for 150°C and above, f E = R e(T)/1.5 or R m/2.35, whichever is the lower value Without a known value of elevated temperature R e(T), for 150°C and above,  f E = R e(T)/1.6 or R m/2.35, whichever is the lower value Between 50°C and 150°C, based on linear interpolation of the above  For austenitic stainless steels, the following strengths apply:  Up to 50°C, f E = R e/1.5 or R m/2.5, whichever is the lower value  With a known value of elevated temperature R e(T), for 150°C and above, f E = R e(T)/1.5 or R m/2.5, whichever is the lower value  Without a known value of elevated temperature R e(T), for 150°C and above, f E = R e(T)/1.45 or R m/2.35, whichever is the lower value Between 50°C and 150°C, based on linear interpolation of the above 

7.3

THERMAL GRADIENT ANALYSIS API TR 6AF1 25 and 6AF2 26 consider thermal forces due to the bending induced in flanges when the product is hot and the exterior is cold. They assessed the full range of standard flanges for leakage due to lack of pressure on the gasket or failure of the flange or bolt. In addition to thermal gradient, temperature derating combined with internal pressure and moments were applied. Tables were derived for the full range from finite element studies.

Figure 5 Finite element mesh used by API TR 6AF

24

The FE mesh in Figure 5 shows a simple flange with no taper to the hub. The element size is reduced at the stress concentration point between the flange and the shell. It seems to be common to model the pipeline thickness in two or three elements. It is important to model a quadrant of flange in 3D because of the effect of the applied moment and to examine bending in the space between the boltholes. At the holes themselves, the elements are adjusted to form a ring around the bolts.


23/41

However, in this model, the gasket has not been examined in detail. Compare this with the FE mesh shown in Figure 7, which models the groove and gasket in detail. Figure 6 shows a typical thermal gradient (in °F) and the resulting stresses in the flange obtained from the analysis. If the outside of the flange is insulated, the thermal gradient is more even (the outside is almost as warm as the inside) and there is little additional stress. Thus, it may be possible to ignore the effects of thermal gradient stresses if the flange is well insulated. However, the bolts need to be de-rated to take account of the hotter operating temperature.

Figure 6 Thermal gradient (in °F) and flange stress plots

7.4

LOW TEMPERATURE FLANGES Most concern is given to de-rating of material stresses for high temperatures due to the product. However, strength can also reduce with extreme low temperatures. A guidance document 17 is given for flanges made from ASTM A105 material subjected to low temperatures down to -29°C. Such flanges may be at risk from brittle failure. The document is valid for standard weld neck flanges conforming to dimensions given in ASME B16.5 22 and also non-standard flanges with a nominal thickness range of 32 mm to 102 mm.

7.5

AGGRESSIVE ENVIRONMENTS INCLUDING SOUR SERVICE Where the aggressive nature of the product demands additional corrosion resistance API recommends design to NACE MR 01 75 18 and the use of type 6B flanges. It is possible to replace the carbon steel flanges with CRA material or to use clad steel. The latter should be returned up around on the inside of the groove for the gasket. ASME VIII 5 gives guidance on clad pipe in Section AM-220.


24/41

7.6

FINITE ELEMENT ANALYSIS A proprietary FEA service offered by Welding Units (www.welding-units.co.uk) for flange design does model the gasket and adjusts the mesh locally in the hub to a smaller size around the groove. Refer to Figure 7.

Figure 7 FE mesh used by Welding Units Typical cell sizes are shown in Figure 5 and Figure 7. It is necessary to analyse a quarter of the flange in 3D to allow for externally applied moments as shown in Figure 5. For extreme conditions, it is usual to assume the worst position of the bolt aligned on the quarter axis. All conditions need assessing including the stress for thermal gradient (see example in Figure 6), which causes toroidal bending in the flange. Checks should be made for bolt, gasket, flange membrane stress and body stress under all load combinations. It is possible to include plasticity in design for secondary stresses.

7.7

COFLEXIP PROGRAMME Some guidance on materials is provided in Appendix C. Information is provided for stainless steel flanges because these are commonly used with flexible pipelines and risers. However, no values are given for temperature de-rating. Also, the values given are for material yield strengths rather than the correct figures for de-rated shear stress intensity required by ASME VIII 5 and PD 5500 4.


25/41

The programme simplifies the design of swivel ring flanges. It calculates the ring stresses following Section 3-340 (b) rather than (a). That is: it assumes the longitudinal and radial stresses in the ring to be zero. This then means that the tangential stress has no reduction from the radial stress moment contribution. It also assumes a simple ring rather than the socketed ring shown in Figure 2. There is a minor discrepancy in the manual for one value used to determine the value for Y. The ASME VIII 5 code on page 248 Figure 3-340.1 uses a constant of 5.71690 whereas the manual quotes 5.71 7690. Since these are very close, it is not possible to determine if this is a typographic error, or if the programme itself uses the wrong value.

7.8

TECHNIP MATHCAD SHEETS 7.8.1 INPUT DESIGN STRESSES It is recommended that design stresses be input from the tables rather than scaled from the ultimate stresses. These need to be de-rated for operating temperature. 7.8.2 EXTERNALL Y APPLIED MOMENT Because the ASME VIII 5 and PD 5500 4 methods both simplify the bolts to a single point load, there is no need to find the moment of inertia of the bolt plane. 7.8.3 DERIVATION OF HUB CONSTANTS Two of these sheets use the formulae from ASME VIII 5 to derive values for F, V and f. In PD 5500 4, the formulae are not provided, necessitating the use of the graphs. However, the graphs in the two codes are identical and are derived from the same underlying theory. For the ‘ASME VIII RTJ Flange_Master.mcd’ a hidden section on page 5 is used to derive values of F, V and f. However, values for F and V then seem to have been misread from the graphs and re-defined as new inputs. Also, the value for the longitudinal stress due to the applied bending moment (S H1) has been omitted from page 6. The note on page 252 of ASME VIII 5 Division 2 Table 3-340.1 referring to integral flanges states that certain values shall be given to F, V and f if there is no taper to the nozzle (that is, if g 1 = go). However, this hidden section would derive incorrect values for F and V in this condition. In the ‘12in PD 5500 sheet.mcd’ 1 sheet, despite the values of h/ho and g1/go being correct, it would appear that the value for V on sheet 7 of 8 has been misread from the graph. It is a factor of 10 too high. Subsequent variables are then incorrect.

7.8.4 COMPARISON OF CAL CULATED WITH PERMISSIBLE VALUES On the final line of ‘ASME VIII RTJ Flange_Master.mcd’ , the sheet derives stress limits which are both more than and less than 1.0. It is recommended that this line be an automatic check rather than an input.


26/41

8 FLANGE GRADES, BOLTS AND GASKETS

8.1

FLANGE MATERIAL GRADES 8.1.1 API API 6A 23 classifies suitable materials in Tables 5.1 and 5.2 by pressure. For flanges, only three grades are given, reproduced here in Table 2. They do not correspond to named steels, merely grades. API m ater ial designation

0.2% Minimu m yield strength

Minimum tensile strength

45K

248 MPa (45 ksi)

483 MPa (70 ksi)

60K

414 MPa (60 ksi)

586 MPa (85 ksi)

75K

517 MPa (75 ksi)

655 MPa (95 ksi)

Table 2 API material properties for flanges The allowable steel grade for each API pressure rating is reproduced in Table 3.

Pressure rating MPa (kip)

13.8

20.7

34.5

69.0

103.4

138.0

(2)

(3)

(5)

(10)

(15)

(20)

Material

45K

45K

45K

60K, 75K

75K

75K

Table 3 API material applications for weld neck flanges 8.1.2 US STEELS Tables 1A and 1B of Section II, Part D of ASME VIII 5 gives specified minimum yield and tensile along with appropriate design strength intensity values, S m for elevated temperatures to 482°C (900°F) or more. The list includes carbon and low alloy steels, high alloy steels, quenched and tempered steels, nickel and nickel alloys. PD 5500 4 enquiry case 5500/91 of September 2003 permits the use of ASTM and API materials. The response tabulates de-rated temperature values in metric units.


27/41

8.1.3 UK AND EUROPEAN STEELS Table K.1-6 of PD 5500 4 gives a ppropriate design strengths for steel forgings following the method of BS 1503 19 . These include values for carbon manganese steels and alloy, martensitic and austenitic stainless steels at elevated temperatures to 480°C or more. However, BS 1503 19 has now been replaced by BS EN 10222 20 .

8.2

METRIC AND IMPERIAL STUD BOLT SIZES It is common to find imperial flanges but with metric sizes of bolts. The preliminary standard prEN 1759-1 34 Annex F gives guidance as to the equivalents up to M39.

Imperial diameter (inch)

Pitch (threads per inch)

Metric equivalent (mm)

Pitch (mm)

1

/2

13 UNC

M14

2.0

5

/8

11 UNC

M18

2.5

3

/4

10 UNC

M20

2.5

7

/8

9 UNC

M24

3.0

1

8 UNC

M27

3.0

11/8

8 UNC

M30

3.5

11/4

8 UNC

M33

3.5

13/8

8 UNC

M36

4.0

11/2

8 UNC

M39

4.0

5

1 /8

8 UNC

M42

4.5

3

1 /4

8 UNC

M45

4.5

17/8

8 UNC

M48

5.0

2

8 UNC

M52

5.0

21/4

8 UNC

M56

5.5

21/2

8 UNC

M64

6.0

3

2 /4

8 UNC

M72

6.0

3

8 UNC

M76

6.0

Table 4 Suggested metric bolt sizes These are similar to the Technip recommendations given in Appendix C of the Coflexip work procedure 6, and reproduced here in Table 4. The only difference between the two is for 5/8in bolts, which the former recommends as M16 and Coflexip recommend as M18. Sizes smaller than M14 (½in) should not be used. Occasionally larger sizes than listed – such as M90 (3½in) – may need to be used but these are hard to handle by divers.

8.3

BOLT TENSIONS Bolt selection is an important consideration in the design of flanges. Although not part of the present scope, some information is supplied here as guidance.


28/41

8.3.1 ACCEPTABLE LOADS Bolts are generally tensioned to approximately 50% of yield. See discussion of bolt tensions in API TR 6AF2 21 . During operation, this is allowed to increase to 83% or reduce to 37.5%, depending upon the combination of loads. With increased temperature and internal pressure alone, the load in bolts tends to decrease. Eventually, the gasket will cease to seal. However, for subsea flanges, it is not common to further adjust the bolt tensions after the make-up operation, so the initial tension must maintain the seal during the subsequent hydrotesting and operation load combinations. Bolt re-adjustment occasionally does take place with landline flanges, in particular with large diameters.

8.3.2 TORQUE It is normal to specify a make-up tension for the bolts. The contractor can then pretension to this level prior to closing down the nuts. Alternatively, a torque wrench can be used to achieve this amount of pretension. However, the amount of torque that is applied is heavily dependent upon the lubrication applied to the nuts, washers and stud bolts. A procedure of tightening bolts generally follows a four stage make-up:  The bolts are brought to 30% of their required torque following a pattern (12, 6, 3, 9; 1, 7, 4, 10; 2, 8, 5 and 11 o’clock for a twelve bolt flange)  The same pattern brings the bolts to 60% of the specified torque The full torque is applied following the same order  Finally, full torque is repeated.  It should be noted that most codes have been written assuming that bolts will be torqued up using a wrench, or in some cases, simply using a spanner (with no check on their torque). Nevertheless, PD 5500 4 recognises that where controlled tensioning of bolts is undertaken (the normal subsea practice), then the values in the table may be increased by 20%. During testing, the design stress may be multiplied by 150%.

8.3.3 BOLT MATERIALS Permitted stresses in bolting materials are given in ASME VIII 5 Table 3 Section II Part D. These provide values for low and high alloy steels and nickel alloys, respectively. The allowable bolt stress is reduced depending upon the temperature and diameter of the bolt. Note that the temperature is not necessarily the same as that of the flange or product. PD 5500 4 (in Table 3.8-1) provides recommended design stresses for UK grades of steel bolts over a range of temperatures. The root areas for common sizes of both metric and imperial bolts are given in Table 3.8-2. Commonly, uncoated ASTM Type B7M or L7M studs are used with 2HM nuts.

8.3.4 CATHODIC PROTECTION It is possible to coat the bolts with PTFE but care should be taken to ensure they make electrical contact for cathodic protection.


29/41

Bolt yield strengths greater than around 700 MPa may suffer from hydrogen embrittlement. For higher strength bolts, Type L7 or B7 with 2H nuts can be used.

8.4

GASKETS 8.4.1 MAKEUP AND OPERATIONAL STRESS Care needs to be taken when loading the bolts to ensure that there is sufficient tension to guarantee no leakage after makeup and yet prevent over crushing of the gasket during operation. Refer to Figure 8 for the following discussion.

Gasket Stress

Cr ush in L im it Make-u Stress

Seatin

4 3

Stress 2

1

5 Gasket Deflecti on

Figure 8 Simplified comparison of gasket deflection and stress When the bolts are first tightened up, the gasket follows a non-recoverable path shown above between points 1 and 2. During this stage, the gasket is deformed to match the flange faces to make a seal. The point at which the gasket provides the minimum effective seal is known as the gasket seating stress. The region marked 2-4 is the useful sealing range of the gasket. The value of the seating stress can be found in the associated codes for the particular gasket. However, the crushing limit must usually be obtained from gasket manufacturers because it is not specified in the codes. When the gasket is compressed beyond its crushing limit, some form of breakdown usually occurs, varying according to the type of gasket. If the gasket is tightened to some value between points 2 and 4 and then the gasket is unloaded (by internal pressure or bolt loosening) then reloaded, it will follow a path such as points 3-5-3, rejoining the original curve.


30/41

8.4.2 GASKET MATERIAL For subsea flanges, there are three classes of gasket material listed in ASME VIII 5, Table 3-320.1. Gaskets are selected to be softer and more flexible than the material of the flanges in order that they seat properly, ensuring a good seal. The suggested design values for m and y given in the code and reproduced below in Table 5 are described as ‘generally proving satisfactory in service’. A similar table is given in PD 5500 4 (Table 3.8-4).

Ring joint material

Gasket factor, m

Minimum design seating s tress, y

Iron or soft steel

5.50

124 MPa (18.0 ksi)

Monel or 4% to 6% chrome

6.00

150 MPa (21.8 psi)

Stainless steels

6.50

180 MPa (26.0 ksi)

Table 5 Recommended gasket factors, m and y Design code API 6A 23 provides guidance on the Rockwell hardness for different gasket materials. This is reproduced in Table 6.

Ring gasket material

Maximum hardness

Soft iron

HRB 56

Carbon and low alloy steel

HRB 68

Stainless steel

HRB 83

Corrosion resistant alloys

To manufacturer’s specification

Table 6 Recommended gasket hardness For land-based facilities (where it is possible to re-seat flanges during maintenance operations), it is common to use other gasket materials and shapes. For offshore use, the approach normally demands a metal-to-metal seal and an octagonal or oval ring joint fitted into a groove in the face of the flange. In the event of an external fire, metal gaskets are able to resist temperatures similar to that of the flange material. Leakage of hydrocarbon product due to gasket failure would result in the line contents feeding the fire. Use of non-flammable metal gaskets helps prevent this.

8.4.3 API GASKET SHAPES Four cross-sections of gasket are permitted with appropriately shaped grooves. These are the BX, RX and R (octagona l and oval) from API 6A 23, which are reproduced in the figures below. API 17D 27 has similar shaped grooves and gasket types. Even when non-standard flanges are to be designed, it is recommended that the groove dimensions and tolerances should follow these standards.


31/41

OD ODT

G

C

23° 1

R

R D

R

H

/16" 45° max

R

E

23°

N Break sharp corner

/32 R

23° A Type BX Gasket

Type BX Groove

Figure 9 API Type BX pressure-energised gasket and groove API 6A 23 provides standard dimensions for Type BX gaskets, which are given up to a nominal size of 726 mm corresponding to a ring diameter of 852.75 mm. In the above figure, tolerances are as listed in Table 9:

Dimension

Descrip tion

Tolerances (mm)

A*

Width of ring

+0.20, -0

C

Width of flat

+0.15, -0

D

Hole size

E

Depth of groove

+0.5, -0

G

OD of groove

+0.10, -0

H*

Height of ring

+0.20, -0

N

Width of groove

+0.10, -0

OD

Outside diameter of ring

+0.0, -0.15

ODT

Outside diameter of flat

±0.05

R

Radius of ring

23°

Angle

±0.5

See note ±¼°

Table 7 Type BX tolerances * A plus tolerance of 0.20 mm for width A and height H is permitted, provided the variation in width or height of any ring does not exceed 0.10 mm throughout its entire circumference. Note: Radius R shall be 8% to 12% of the gasket height H. One pressure-passage hole is required per gasket on the centreline. The API Type SBX pressure-energised ring gaskets from API 17D Table 906.1 are of a similar profile as type BX but with slight differences in tolerances. Sizes up to 640 mm OD are specified. They also have a pair of pressure holes linking to the inside of the flange to prevent pressure lock when the connection is made underwater. Raised flange faces are assumed to touch with the Type BX gasket only.


32/41

23°

23°

H

D

P

23° F

R1

R2

E

C A OD Type RX Gasket

Type RX Groove

Figure 10 API Type RX pressure-energised gasket and groove API 6A 23 presents standard sizes for Type RX gaskets up to an outside diameter of ring of 600.87 mm. In the above figure, tolerances are as listed in Table 8:

Dimension

Descrip tion

Tolerances (mm)

A*

Width of ring

+0.20, -0

C

Width of flat

+0.15, -0

D

Height of chamfer

+0, -0.8

E

Depth of groove

+0.5, -0

F

Width of groove

±0.20

H*

Height of ring

+0.2, -0

OD

Outside diameter of ring

+0.5, -0

P

Average pitch diameter of groove

R 1

Radius in ring

±0.5

R 2

Radius in groove

max

23°

Angle

±½°

±0.13

Table 8 Type RX tolerances * A plus tolerance of 0.20 mm for width A and height H is permitted, provided the variation in width or height of any ring does not exceed 0.10 mm throughout its entire circumference. Note: The pressure passage hole illustrated in the RX ring cross-section is for rings RX-82 through RX-91 only. The centreline of the hole shall be located at the mid point of dimension C. The hole diameter shall be 1.5 mm for rings RX-82 through RX-85, rings RX-86 and RX-87; and 3.0 mm for rings RX-88 through RX-91. The API Type SRX pressure-energised ring gaskets from API 17D Table 906.2 are of a similar profile as type RX but with slight differences in tolerances. Sizes up to 140 mm OD are specified. They also have a pair of pressure holes linking to the inside of the flange to prevent pressure lock when the connection is made underwater.


33/41

P

23°

P

23°

P

23° F H

B

R1

R2

C A

E

A

Type R Octagonal

Type R Oval

Type R Groove

Figure 11 API Type R gaskets and groove API 6A 23 presents standard sizes for Type R gaskets up to a pitch diameter of groove and ring of 584.20 mm. In the above figures, tolerances are as listed in Table 9:

Dimension

Descrip tion

Tolerances (mm)

A

Width of ring

±0.20

B

Height of oval ring

±0.5

C

Width of flat on octagonal ring

±0.2

E

Depth of groove

+0.5, -0

F

Width of groove

±0.20

H

Height of octagonal ring

±0.5

P

Average pitch diameter of ring

±0.18

Average pitch diameter of groove

±0.13

R 1

Radius in ring

±0.5

R 2

Radius in groove

max

23°

Angle

±½°

Table 9 Type R tolerances For pipeline diameters requiring gaskets larger than given in the tables, the appropriate tolerances need to be adopted.

8.4.4 SELF-ENERGISING GASK ETS Gaskets Type BX and RX are deemed to be self-energising. Note that for these, the outside diameter of the gasket is specified. Compare this with Type R octagonal or oval gaskets where the mean diameter is specified. For self-energising gaskets, the pressure must be considered to act on the face of the flange out to the outside of the gasket groove.

8.4.5 PRE-TESTABLE GASKETS Although it is possible to adjust the bolts at the hydrostatic leak test stage, it is normal to tension them up only once during assembly, to a level that will need no further adjustment.


34/41

®

Figure 12 KaMOS gasket in groove Gaskets are available that can prove the leak tightness of the joint at assembly bly stage. stage. ® One such proprietary product is the patented KaMOS system shown in Figure 12. Details of this gasket are available at http://www.karmsund.no/eng_index.htm. These seal tests work by pressurising the gaps at the foot of the gasket grooves through holes in the octagonal gaskets, as shown in Figure 13. Any lack of seal between faces of the grooves and the gasket will then indicate adjustment is needed.

Seals

®

Figure 13 KaMOS testing of seal Other similar systems are available, such as from Grayloc, Vector and Oilstates. It should be noted, however, that not all clients accept this approach to testing.


35/41

9 RELEVANT READING This section lists the main sources of data regarding flanges and their design, along with a brief discussion of their content and relevance. ASME B16.5 22 – this publication provides a simplified method for determining the class of flange needed. It considers pressures and temperature derating for ten different flange material groups. No allowance is made for bending moments, tension or temperature. Dimensions for standard flanges up to 24in are tabulated. ASME VIII 5 – this is the main source of design for American flanges. The whole code covers the design of pressure vessels. Flange design is covered in Division 2. PD 5500 4 – this is the main UK source for the design of flanges. Again, like the ASME VIII 5 code, it covers the design of pressure vessels. The method for the design of flanges follows a similar underlying code to that of ASME VIII 5 Division 2. Section 3.8 is specific to bolted flanged connections and sixteen standard working forms are provided to help with flange designs of various types. Weld-neck flanges are Form 1 and swivel ring flanges can be determined by adapting Form 5. API 6A 23 this provides permissible stresses and standard sizes for flanges and gaskets based on ASME VIII 5 methods, specifically for wellhead equipment.

Pressures classes considered in API 6A 23 are 2, 3, 5, 10, 15 and 20 ksi with nominal sizes from 71/16in to 21¼in, though the larger diameters do not cover the maximum pressure classes. API TR 6AF 24 – this was developed from earlier work and provided an improved method of flange design as compared with traditional, more conservative methods, such as designing for pressure loads only. This covered the determination of rating charts for API 6B and API 6BX flanges but did not consider the effects of temperatures greater than 121°C.


36/41

The effect of temperature on load capacity affects: the yield strength (and thus the allowable stresses) of the flange and bolt  materials  thermal stresses caused by thermal gradients across the flange and between the bolts and the flange. This can cause bending of the flange faces. API TR 6AF1 25 – this study has examined the effect of elevated temperatures on inside bore of standard weld-neck flanges with the outside at 0°C and subject to wind-induced forced cooling. The elevated temperatures are 176.6°C and 343.3°C.

The analyses were undertaken using simple ANSYS axisymetric models and four different materials combinations. API TR 6AF2 26 – this study extends the work of API TR 6AF1 25. It considers the effect of applied moments and tension combined with the temperature profile through the flange. All seventy of the API 6A 23 flanges were modelled as a 3D finite element mesh.

For this work, the SESAM FE package was used. These flange and bolt material combinations are: Carbon and low alloy steel flanges with two types of bolt  Martensitic, ferritic and precipitation hardening stainless steel flanges  Austenitic and duplex stainless steel flanges  Suitable derating factors are derived. Graphs are presented for each flange diameter and material combination to enable the limiting loads to be determined. It should be noted that gasket leakage tended to be the governing case for the smaller flanges and for higher pressure operations. This may be due to conservative assumptions made in the analysis for gaskets. Stresses in flanges tended to have little reserve beyond the gasket leakage level. The bolt stresses did not govern for any case. They were typically within ⅔ of their yield strength due to make-up, pressure, tension and bending moment loads when they are made-up to around half their yield. With insulated flanges, the temperature profile is more constant across the thickness. This results in a higher capacity of the flanges to resist the other loads. API 17D 27 this provides sizes for weld neck and swivel ring flanges suitable for subsea wellhead equipment. The present study excludes these designs. Tapered swivel ring 346 mm (13 5/8in) flanges at Class 5000 and 10 000 are specified for wellheads. API 605

28

this provides sizes for standard flanges between 26in and 60in.

ISO 7005-1

29

– the ISO code covering both DIN and ASME flanges.

BS 1560-3.1 30 – This gives the dimensions of steel flanges for the petr oleum industry. Standard weld neck flanges are specified. It equates to ANSI B16.5 22.


37/41

BS EN 1092-1 31 – this is based on ISO 7005 29 but only specifies dimensions of DIN type flanges at temperatures and pressures from PN 2.5 to PN 100 and diameters DN 10 to DN 4000. It includes a range of different types of flanges including weld-neck, which it designates as Type 11. Note that these flanges have slightly changed from the older DIN pressure rating. This code also considers swivel ring flanges (combination Type 34 short weld neck with Type 03 loose plate flange). BS EN 1591-1 32 and BS EN 1591-2 33 these provide a detailed design method for flange and gasket selection. Blank and weld-neck flanges are included. Any size of flange can be designed using this code. prEN 1759-1 34 – the companion to EN 1092-1 31 covering ASME flanges. The prefix pr indicates it is not fully released.


38/41

10 REFERENCES

1 Technip MathCAD workbook 12in PD5500 Flange.mcd, dated 17/01/2005 – Design Gas Exp 12" ASME 16.5 1500# WN Flange (8 pages) 2 Technip MathCad workbook ASME VIII RTJ Flange_Master.mcd, undated – 14" ASME 16.5 600# RTJ Topside End - Installation Loads (6 pages) 3 Technip MathCad workbook ASME VIII RTJ Swivel Flange_Master.mcd, dated 17/01/2005 – Hub design for 40" ASME 16.5 900# RTJ Swivel Flange - Max Moment Capacity (9 pages) 4 PD 5500 : 2003 British Standards published document – Specification for unfired fusion-welded pressure vessels incorporating Amendment N° 1 and Corrigendum N° 1 (successor to BS 5500) – Section 3.8 Bolted flange connections 5 ASME VIII : 2004 by The American Society of Mechanical Engineers – Boiler and Pressure Vessel Code, Section VIII, Division 1: Design and Fabrication of Pressure Vessels and Division 2 and Division 2: Alternative rules 6 Detailed work procedure Ref 04 DIE T 202 revision 2 October 1997 – Fixed flange and swivel flange design – Software package: Flange Rel 1.2 7 API RP2A-WSD, Dec 2002 by American Petroleum Institute – Recommended practice for planning, designing and constructing fixed offshore platforms, working stress design 8 API Recommended practice 2RD by the American Petroleum Institute, June 1998 Design of risers for floating production systems (FPSs) and Tension-leg platforms (TLPs) 9 E O Walters, DB Wesstrom, DB Rossheim, F S G Williams – Trans ASME Vol 59, 1937 – Formulas for stresses in bolted flanged connections 10 K P Singh and A I Soler – Arcturus Publishers, New Jersey, 1984 – Mechanical design of heat exchangers and pressure vessel components 11 API RP 1110 by the American Petroleum Institute, March 1997 – Pressure testing of liquid petroleum pipelines 12 ASME B31.4 by The American Society of Mechanical Engineers, 2002 – Pipeline Transportation Systems for Liquid Hydrocarbons and Other Liquids 13 ASME B31.8 by The American Society of Mechanical Engineers, 2003/2004 – Gas transmission and distribution piping systems 14 BS 8010 Part 3 : 1993 Code of practice for pipelines. Pipelines subsea: design, construction and installation – replaced by PD 8010


39/41

15 BS 3920 : 1973 Derivation and verification of elevated temperature properties for steel products for pressure purposes – superseded by BS EN 10314 16 BS EN 10314 : 2002 Method for the derivation of minimum values of proof strength of steel at elevated temperatures 17 Un-numbered document by the Belgian Chemical Risks Directorate, Technical Inspectorate, Administration of Labour Safety, Federal Ministry of Employment and Labour, July 2002 – Recommendations for pipe flanges made in forged steel complying with ASTM A 105 18 NACE MR 01 75 Part 1 Rev 1 2001 (ISO 15156-1) – Petroleum and natural gas industries – Materials for use in H 2S-containing environments in oil and gas production – General principles for selection of cracking-resistant materials 19 BS 1503 : 1989 Specification for steel forgings for pressure purposes – now superseded by BS EN 10222 20 BS EN 10222 Steel forgings for pressure purposes Part 1 : 1998 general requirements for open die forgings; Part 2 : 2000 ferritic and martensitic steels with specified elevated temperature properties; Part -3 : 1999 Nickel steels with specified low temperature properties; Part 4 : 1999 Weldable fine-grain steels with high proof strength; and Part 5 : 2000 Martensitic, austenitic and austeniticferritic stainless steels 21 API TR 6AF2 Second edition : 1999 by the American Petroleum Institute – Technical Report on Capabilities of API Flanges under Combination of Loading – Phase II 22 ASME 16.5 : 2003 by The American Society of Mechanical Engineers – Pipe Flanges and Flanged Fittings NPS ½ through NPS 24 Metric/Inch Standard (Formerly published by the American National Standards Institute as ANSI 16.5.) 23 API 6A nineteenth edition, July 2004, by the American Petroleum Institute – Specification for wellhead and christmas tree equipment 24 API Bulletin 6AF First edition : 1989 by the American Petroleum Institute – Capabilities of API Flanges under Combination of Load 25 API TR 6AF1 Second edition : 1998 by the American Petroleum Institute – Technical Report on Temperature derating on API Flanges under Combinations of Loading 26 API TR 6AF2 Second edition : 1999 by the American Petroleum Institute – Technical Report on Capabilities of API Flanges under Combination of Loading – Phase II 27 API 17D (also known as ISO 13628-6) with two supplements to June 1996, by the American Petroleum Institute – Specification for subsea wellhead and christmas tree equipment 28 API 605 revision March 88, by the American Petroleum Institute – Large diameter carbon steel flanges (nominal pipe sizes 26in through 60in, Classes 75, 150, 300, 400, 600, and 900) 29 ISO 7005 Metallic flanges 30 BS 1560-3.1 : 1989 Circular flanges for pipes valves and fittings (Class designated) – Part 3: Steel, cast iron and copper alloy flanges – Section 3.2 Specification for steel flanges 31 BS EN 1092-1 : 2001 Flanges and their joints – circular flanges for pipes, valves, fittings and accessories, PN designated – Part 1 Steel flanges – incorporating Corrigendum N° 1 32 BS EN 1591-1 : 2001 Flanges and their joints – design rules for gasketed circular flange connections – Part 1 Calculation method


40/41

Trevor Jee - Flange Stress Analysis Guidance

Recommend Documents