Revised Guidance on Residual Stresses in BS7910

Serco Technical Consulting Services, Birchwood Park, Warrington, Cheshire WA3 6GA, UK.

Peter Gill
University of Manchester, George Begg Building, Manchester, M13, 9PL, UK.

Liwu Wei
TWI Ltd.

Steve Bate
Serco Technical Consulting Services, Birchwood Park, Warrington, Cheshire WA3 6GA, UK.

Proceedings of the ASME 2011 Pressure Vessels & Piping Division Conference PVP2011 July 17 - 21, 2011, Baltimore, Maryland, USA. Paper No. PVP2011-57071.

Abstract

A major revision of the British Standard BS7910 on 'Guide to Methods for Assessing the Acceptability of Flaws in Metallic Structures' is being planned for issue in 2012.

This paper provides an overview of the proposed revised guidance in relation to recommended weld residual stress profiles. As such, the paper is focussed on the proposed revised Annex Q of BS7910 which deals with residual stress distributions in as-welded joints.

1. Introduction

All aspects of the British Standards BS7910^[1] on 'Guide to Methods for Assessing the Acceptability of Flaws in Metallic Structures' are being reviewed with a view to updating the guidance for a revised version of the document planned to be issued in 2012.

The main requirements for providing revised guidance on residual stresses for BS7910 are centred on Annex Q that contains 'Residual Stress Distributions in As-Welded Joints'. For this Annex it has been required: (i) to incorporate the most relevant up-to-date upper bound profiles, (ii) to enable the user to readily evaluate stress intensity factors for the profiles without having to go outside the document (bearing in mind that the present BS7910 contains stress intensity factor values only for membrane and bending stress components, whereas the residual stress profiles are commonly represented by high order polynomial equations), and (iii) to make the Annex more user friendly for the assessment engineer to navigate.

2. Most up-to-date profiles

The through-thickness residual stress profiles provided in the current version of BS7910^[1], R6^[2], FITNET^[3] and API 579-1/ASME FFS-1^[4] procedures have been assessed in detail in terms of understanding their validation and the basis on which any recent changes may have been made. The profiles were considered for the following weld types for orientations transverse and longitudinal to the weld:

• Plate Butt and Pipe Seam Welds
• Pipe Butt Welds
• T-Butt Welds (Plate to Plate)
• T-Butt Welds (Tubular/Pipe)
• Repair Welds

The recommended profiles provided in BS7910, R6 and FITNET are all based on upper bound fits to experimental and analytical data and they originate from the same sources (e.g.^[5]). Not surprisingly, the profiles provided in these three procedures are very similar, but with some fairly minor differences seen in R6. These differences arise from the fact that the residual stress profiles in that procedure are reviewed on an annual basis with a view to recommending modifications, as required, on the basis of up-to-date experimental and/or analytical data.

It is the understanding of the authors of this paper that the profiles provided in the recently revised API 579-1/ASME FFS-1 procedure were developed from the results of a large analytical programme involving detailed finite element calculations. These profiles generally differ significantly from those in the above-mentioned three procedures.

On the basis of consistency with the current version of BS7910 and the need to have knowledge of and confidence in the validation data, it has been decided that the residual stress profiles for the revised version of BS7910 should be based on a combination of the R6 and current BS7910 profiles.

3. Stress intensity factors

The recommended residual stress profiles have been partitioned into Membrane (σ_m), Bending (σ_b) and Self-Balancing (σ_rs) components by the following equations for a given stress profile σ(z), where z varies from 0 to B through the wall thickness,

Figure 1 contains an example of such partitioned through-thickness transverse stresses for ferritic pipe butt welds made with a low heat input E₁ / B ≤ 50 J/mm². The stress, σ_R ^T, is normalised to yield stress, σ_y, and it is shown plotted against z/B where z is the distance from the bore through the wall thickness, B. The dark broken curve is the recommended residual stress profile represented by the equation:

The solid line is the membrane stress component evaluated from Eq. (1), the light broken line is the bending stress component evaluated from Eq. (2) and the dotted curve is the self-balancing component evaluated from Eq. (3).

Figure 1: Plot of decomposed components of transverse stresses in ferritic pipe butt welds made with a low heat input E₁ / B ≤ 50 J/mm²

Stress intensity factors for the Membrane (K_m) and Bending (K_b) Stress components can be evaluated from the existing solutions provided in Annex M of BS7910 which are planned to be also incorporated in the revised version. The question then remains as how to evaluate the stress intensity factor (K_sb) for the self-balancing stress component.

For surface flaws under a through-wall sinusoidal stress distribution, stress intensity factors for extended flaws bound those values for semi-elliptical flaws of all aspect ratios. This is illustrated in Figure 2 for a cylinder with an internal circumferential flaw, whereby stress intensity factor, K (normalised to the square root of wall-thickness B) is plotted against flaw depth, a (normalised to B) for a sinusoidal stress distribution (tensile near the wall surfaces, balanced by compression in the mid-wall region). These values of K are for a unity tensile stress level at the surfaces (and a unity compressive stress level at the mid-wall). Values are shown for flaw depth to flaw semi-length (a/c) values of 0.2 and 1.0 as well as for an extended flaw. The stress intensity factors were evaluated from the relevant solutions provided in the compendium contained in R6^[2] whereby the through-wall stress distribution over the flaw depth, but in the absence of the flaw, is represented by a polynomial equation. It may be noted that the K values shown for the surface flaw cases are for the deepest point of the flaw.

Figure 2: Normalised stress intensity factors for cylinder with internal circumferential flaw under sinusoidal through-wall stress distribution

Similar information to that provided in Figure 2 was evaluated for other cases as follows:

• Plate
• Cylinder with internal axial flaw
• Cylinder with external axial flaw
• Cylinder with external circumferential flaw

Figure 3 contains the highest normalised stress intensity curves for each of the five geometry cases considered. As can be inferred from above, all these curves represent the extended flaw cases. It can be seen that the curve for the plate provides an upper bound to the cylinder curves.

Figure 3: Highest normalised stress intensity factor curves for all cases under sinusoidal through-wall stress distribution - for extended surface flaw in all cases

Returning now to the self-balancing component of a residual stress field and it is evident that the stress intensity factor, K_sb, for this component will be significantly lower than for the membrane and bending components (i.e. K_m and K_b respectively) in the majority of situations. Therefore, it would be advantageous for a good and simple-to-use upper bound solution for K_sb to be developed which would not result in the overall stress intensity factor being overly-conservative.

For each residual stress profile corresponding to the various weld types as listed in section 2 above, the normalised stress intensity factor was evaluated for the self-balancing component following the same calculational route as that referred to above in relation to Figures 2 and 3. For conservatism, only the K solutions for extended flaws in plates have been considered. Examples of the resulting normalised stress intensity factors are shown in Figures 4 and 5 for transverse residual stresses in ferritic pipe butt welds made with a low heat input and longitudinal residual stresses in ferritic steel Tubular T-butt welds, respectively.

Figure 4: Normalised stress intensity factor for self-balancing stress component of transverse residual stresses in ferritic pipe butt welds made with a low heat input E ₁ / B ≤ 50 J/mm²

Figure 5: Normalised stress intensity factor for self-balancing stress component of longitudinal residual stresses in ferritic steel Tubular T-butt welds

For longitudinal stresses (ferritic steels) the following equation applies.

Figure Q.5.1 Typical longitudinal residual stress distribution for tubular T-Butt Welds

From the normalised stress intensity factor profiles such as those given in Figures 4 and 5, the peak value of K could be obtained and it is planned for such values to be incorporated into the revised Annex Q of BS7910. If the user requires a more precise solution, then it is proposed that Ksb can be evaluated using new solutions planned to be incorporated into BS7910. These solutions will be for a surface flaw or extended flaw in a plate under a stress field represented by a polynomial equation up to the 5th order. The solutions are taken from the R6 procedure. Only the plate solutions for a polynomial stress field are proposed to be included since they have been shown to be a reasonable upper bound to the solutions for cylindrical geometries (see Figure 3 for example). The user will of course be able to use the newly incorporated solution to evaluate the stress intensity factor for the full residual stress profile without having to consider the partitioned components. This approach may be somewhat overly conservative in some cases though.

In summary, the stress intensity factor (K_rs) for the residual stress profile in the revised Annex Q of BS7910 will be evaluated as:

K_rs = K_m + K_b + K_sb

where K_m and K_b are obtained from Annex M and K_sb can be conservatively evaluated from peak values provided, or more accurately determined by way of the solutions from R6 that are planned to be incorporated, using the self-balancing component of the residual stress profile.

 

4. User friendliness

In the revised Annex Q of BS7910, the recommended residual stress profiles will be illustrated graphically together with the σ_m, σ_b and σ_sb components. In addition, tabular information will be provided on the σ_m and σ_b values, the σ_sb equation and the peak value of K_sb that has been evaluated by the route described above.

For the tubular components, the residual stress profiles are presented in terms of z/B, where z=0 at the bore surface, and B is the thickness. For surface flaws emanating from the outer surface, the profiles can be re-characterised by replacing z/B by 1-z/B in the polynomial equations, where z would then be taken from the outer surface.

A further recommendation is that the profiles given are appropriate to ±1.5W, where W is the weld width. It will be recommended that flaws situated outside of this range need to be assessed on a case by case basis.

An example of the layout is given in Figure 6, representing the details for longitudinal stresses in ferritic steel Tubular T-butt welds.

Figure 6: Example of layout for longitudinal stresses in ferritic steel Tubular T-butt welds

5. Conclusions

A summary and justification has been presented of the proposed revised Annex Q of BS7910 which deals with residual stress distributions in as-welded joints. The information presented has been in terms of:

1. incorporating the most relevant up-to-date upper bound residual stress profiles,
2. enabling the user to readily evaluate stress intensity factors for the profiles without having to go outside the document,
3. making the Annex more user friendly for the assessment engineer to navigate.

6. Acknowledgements

The authors wish to acknowledge with thanks, the advice and information provided by other members of the BS WEE/37 Sub-Committee on Residual Stresses. These are Professor John Bouchard, Dr Ali Sisan, Ms Anne Teughels and Mr Paul Hurrell.

The paper is published by permission of Serco Ltd, and TWI Ltd.

7. References

1. BS 7910. Guidance on methods for assessing the acceptability of flaws in metallic structures. Chapter 7. British Standards Institution. London. UK. 2005.
2. R6: Assessment of the Integrity of Structures containing Defects, Revision 4, (latest revision , 2010).
3. FITNET Fitness-for-Service (FFS) - Procedure (Volume 1) ISBN 978-3-940923-00-4, Koçak, M., Webster, S., Janosch, J.J., Ainsworth, R.A., Koers, R and FITNET Fitness-for-Service (FFS) - Annex (Volume 2) ISBN 978-3-940923-01-1, Koçak, M., Hadley, I., Szavai, S., Tkach, Y., Taylor, N., both printed by GKSS Research Center, Geesthacht, 2008.
4. API: 579-1/ASME FFS-1 2007, 'Fitness-for-service'.
5. Bate S. K., Green D. and Buttle D. J., 'A review of residual stress distributions in welded joints for the defect assessment of offshore structures', OTH 482, HMSO (1997).