Subscribe to our newsletter to receive the latest news and events from TWI:

Subscribe >
Skip to content

Engineering critical analyses to BS 7910 - the UK guide on methods for assessing the acceptability of flaws in metallic structure

   
By: C S Wiesner (1), S J Maddox (1), W Xu (1), G A Webster (2), F M Burdekin (3), R M Andrews (4) and J D Harrison (1)
(1)TWI, (2)Imperial College, (3)UMIST, (4)BG Technology

Paper published in International Journal of Pressure Vessels and Piping, vol.77, no.14-15. December 2000. pp.883-893

Abstract

The well-known fitness-for-purpose standard PD 6493:1991 has been revised and is now published as the British Standards Guide BS 7910. The revisions include modifications to the fracture and fatigue assessment clauses and many additional appendices have been added to complete the guidance provided. The PD 6539:1994 procedures for high temperature assessment have been incorporated as clause 10. This paper presents a brief historical review of the UK developments of fitness-for-purpose methods and outlines the modifications to BS 7910. In addition, the new BS 7910 fracture assessment clause is validated using a series of wide plate tests.

1. Introduction

The fracture mechanics based fitness-for-purpose (FFP) approach, also referred to as Engineering Critical Analysis (ECA), enables the significance of flaws to be assessed in terms of structural integrity. The ECA concept has undergone extensive developments in the past 30 years or so and the widely used PD6493 [1] procedure has been produced in the UK. The document has recently been revised and is now published as BS 7910 'Guide on methods for assessing the acceptability of flaws in metallic structures' [2] .

This paper provides a brief historical overview of engineering critical assessment methods; outlines the additions and changes which have been made to the procedures; and presents wide plate validation results of the BS 7910 fracture/plastic collapse clauses compared with results obtained using PD6493:1991 procedures.

2. Historical development of engineering fitness-for-service procedures and standards

2.1. 1960s to 1970s and PD6493:1980

In 1968 a paper was published [3] proposing that it was possible, at that time, to write an acceptance standard for weld imperfections based on fitness-for-purpose. Such a standard could, it was argued, replace the arbitrary or 'workmanship' standards contained in most design and fabrication codes. Workmanship standards give maximum permissible levels for imperfections based on the most widespread form of non-destructive testing (NDT) used for inspecting new fabrications - radiography. Unfortunately, whilst this NDT method is good at finding relatively harmless volumetric flaws such as slag inclusions and porosity, it is not always good at finding, and is even less satisfactory for sizing, potentially more harmful planar flaws such as cracks, lack of sidewall fusion, etc. Because workmanship standards lay down quite specific rules for permissible lengths of slag inclusion and density of porosity, a large amount of repair work is carried out for these relatively innocuous flaws. It has been estimated that such unnecessary repairs may add as much as 10% to construction costs. Often these added costs would be far outweighed by the consequential costs of lost production. Furthermore, repair of innocuous imperfections has been known to introduce more deleterious defects which have led to structural failure.

This was the background to the publication of reference 3. Following its publication, the British Standards Institution set up a committee to determine whether it was indeed possible to draft a rational acceptance standard. Once it had reported in the affirmative, the work of drafting began with the establishment of the British Standards Institution's (BSI) WEE/37 Committee in 1970. The Committee deliberated for the next ten years. During this time much research was undertaken throughout the world and early experience was being gained in the application of the techniques to industrial problems.

This research and experience increased the confidence of the Committee in its ability to produce a recommendation which would lead to structures which were both safer and more economical than those where the arbitrary acceptance standards were rigorously imposed. In 1980, PD6493 [4] was published. It gave methods which could be adopted by agreement between contracting parties whereby imperfections could be assessed on a fitness-for-purpose basis. PD6493:1980 was, in British Standards terminology, a Published Document. This was in recognition of the fact that much research was still on-going. Published Documents have less force than Standards or Guides.

PD6493:1980 concentrated on the assessment of imperfections with regard to their possible effects on failure by brittle fracture and fatigue. The treatment for brittle fracture was based on measurements of fracture toughness in terms of K Ic or CTOD (crack tip opening displacement) and utilised the CTOD design curve proposed by Burdekin and Dawes [5] .

Failure by plastic collapse of the remaining net section was treated only very briefly and as an entirely separate failure mode from brittle fracture. General and simplified procedures were given for making fatigue assessments using an integration of the Paris fatigue crack growth law to predict whether the given imperfection would grow to failure within the design life. The simplified procedure gave a series of S-N curves representing 'quality categories'. Flaw sizes were given for each category which would be acceptable for a component which was required to meet the associated S-N curve as a fatigue design requirement. This approach was recommended for the treatment of non-planar flaws (slag inclusions and porosity). Acceptable sizes of planar flaw were also given and, for these, the Paris law integration had been carried out for the user. Other failure modes - leakage, corrosion fatigue, stress corrosion, buckling and creep - were all treated very briefly.

The publication of PD6493:1980 was a milestone in that, for the first time, a standardised framework was laid out, which could form the basis of agreement between contracting parties and licensing bodies for assessing imperfections found during fabrication.

In parallel with the development of PD6493:1980, the UK Central Electricity Generating Board (CEGB) was developing its own approach to the assessment of imperfections with regard to static ductile and brittle fracture. This was being driven primarily by the need to demonstrate the integrity of nuclear pressure vessels and of large rotor forgings. These developments culminated in the publication, in 1976, of Revision 1 of the so-called R6 procedure, now in its third revision [6] . The fourth revision is due to be published in 2000. This combined the assessment of brittle and ductile fracture by using the so-called 'two parameter approach'. It utilised a Failure Assessment Diagram (FAD) in which the vertical and horizontal axes were the ratios of the fracture driving force to the fracture toughness and of the applied load to the plastic collapse loads respectively. Failure was predicted when either of these ratios exceeded unity. Interaction between brittle behaviour and plastic collapse was allowed for by a curve derived from a strip yield analysis. The advantage of this approach over that in PD6493:1980 was that the two static failure modes were treated explicitly in one operation. Also, the FAD could be used to assess how the assessment point approached the failure locus as stress or flaw size increased. This is a useful feature, since it indicates the extent to which failure would be dominated by brittle fracture or by ductile instability.

R6 quickly gained international recognition in the power generation industry, whilst PD6493:1980 was more widely used in other industries - oil and gas in particular.

2.2. 1980s and PD6493:1991

The existence in Britain of standardised methods of flaw assessment was very advantageous to industry. However, the fact that there were two such methods available was a possible source of confusion. During the 1980s, the BSI WEE/37 Committee worked towards a harmonisation of these approaches. As a result, PD6493 was reissued in 1991 [1] .

The treatment of fatigue remained very similar to that in the 1980 edition. An enhancement was that the Quality Categories were related to the design S-N curves for particular weld details. Thus, it was possible to determine the permissible levels of imperfection in a butt weld in a structure also containing fillet welds subjected to the same cyclic stress ranges, such that there would be an equal probability of failure from the imperfect butt weld as from the (nominally sound) fillet weld.

The treatment for ductile and brittle fracture was changed in PD6493:1991. Three levels of assessment were given. Level 1, termed the 'Preliminary Assessment' method, was similar to that in the 1980 edition, except that a failure analysis diagram was used, so that the treatment for plastic collapse was explicit rather than implicit, as in the 1980 edition. Level 2, the 'Normal Assessment' method was similar to the 1980 edition (Rev.2) of the R6 method. Level 3, the 'Advanced Assessment' method was based on revision 3 of R6 and allowed the user to take account of the resistance to ductile crack extension. At all these levels, the fracture toughness could be expressed in terms of CTOD or K Ic.

Other changes in the 1991 edition were that the treatments from some of the other failure modes (corrosion/erosion, environmentally assisted cracking, buckling and creep) were enhanced, although still much less detailed than those for fracture and fatigue. The creep section gave exemption criteria. If these were met, creep could be ignored as a potential failure mode. However, the document gave no direct guidance on how to deal with situations when they were not met.

3. Outline of revisions to PD6493 leading to BS 7910

3.1. General

Following the publication of PD6493:1991 and continuing throughout the 1990s, the WEE/37 Committee continued to work on improvements to the document and an extensively revised edition has been first published in 1999 as BS 7910. Following feedback from users in late 1999/early 2000, several typographical errors and technical inconsistencies were corrected and a revised version appeared in 2000 [2] .

BS 7910 includes extensive modifications to the fracture assessment procedures (Clause 7), together with modifications to the fatigue assessment procedures (Clause 8). A completely new chapter (Clause 9) has also been added, covering assessment of flaws in plant operating at high temperature, see Fig.1.

Fig.1. BS 7910 overall structure
Fig.1. BS 7910 overall structure

 

The document has been completely re-written to improve clarity and usability, and to incorporate modern flaw assessment technology, including a number of methods originally published in revision 3 of the R6 procedure. Since feedback from users of PD6493 indicated that the section dealing with stresses caused confusion, in BS 7910 special attention was paid to the description of the stresses to be determined and how they are actually used in an assessment. The changes include a new definition of peak stress, to mean the highest stress at a structural discontinuity rather than the elevation in stress it causes. The relevant parts of the fracture and fatigue assessment clauses have also been revised to reflect the changes introduced into BS 7910.

3.2. Fracture assessments

As in PD6493:1991, there are three assessment procedures, or levels, available for a fracture assessment. Flow diagrams have been included to guide users through them. The Level 1 fracture assessment procedure in BS 7910 is unchanged, but is renamed the 'Simplified Assessment' procedure. The Level 1 procedure is based on a conservative Failure Analysis Diagram (FAD). The Level 1 FAD has K r (or √ δ r), S r co-ordinates, where K r (or √ δ r) is the ratio of applied crack driving force to fracture toughness and S r the ratio of applied stress to flow strength (where flow strength is mean of yield and tensile strength hence incorporating some plasticity). The graphical procedure (the CTOD design curve approach, originally presented in PD6493:1980) becomes Annex N in BS 7910.

Level 2 remains the 'Normal Assessment' method for cases where single-value measurements of fracture toughness (e.g. K Ic, δ mat) are available. The PD6493:1991 Level 2 failure assessment method (strip yield model) is now superseded by Levels 2A and 2B; the choice between the two depends on the type of stress-strain data available for the material in which the flaw is situated. Level 2B is used if the relevant full stress-strain curve is available. Guidance has been included regarding the derivation of Level 2A FAD in cases of discontinuous yielding developed in the recently completed European project SINTAP [7] (Structural Integrity Assessment Procedures for European Industry), as follows:

For materials which exhibit a yield discontinuity (often referred to as Lüders plateau) in the stress/strain curve (i.e. any curve which is not monotonically increasing), or for which it cannot be assumed with confidence that no discontinuities exist, either a cut-off value for L r (the ratio of applied stress to yield strength) of 1.0 should be applied or Level 2B should be used. If it is impractical to determine a Level 2B FAD, the Level 2A FAD at and beyond L r = 1.0 can be estimated [7] using:

√ δ r(L r=1) or K r(L r= ) =
{1 + E ε L/ σ u Y + 1/[2 (1 + E ε L/ σ u Y)]} -0.5

where

ε L = 0.0375 (1 σ u Y/1000) is the estimated length of the Lüders plateau (this relation is restricted to σ u Y < 800N/mm 2), (Note)
(Note: There is a typographical error in BS7910: 1999 (Amendment 1) which states a validity limit of 976N.mm 2; users of BS7910 should correct this limit to 800N/mm 2)

σ u Y is the upper yield strength (if this is unavailable, it is conservative to use the yield or 0.2% proof strength),

and

√ δ r(L r>1) = √ δ r(L r= ) L r (N-1)/2N

or K r(L r>1) = K r(L r= ) L r (N-1)/2N

where

N = 0.3 (1 - σ Y/ σ u) is the lower bound strain hardening exponent estimated [7] from the yield to tensile strength ratio, σ Y/ σ u.

For continuous yielding, the failure assessment line is similar to that used in PD6493:1991, Level 3. L r is now used in place of S r for plastic collapse predictions at Level 2, see Fig.2.

Fig.2. BS7910 Level 2B, material-specific failure assessment diagram
Fig.2. BS7910 Level 2B, material-specific failure assessment diagram

 

Level 3 of BS 7910 (ductile tearing instability assessment) remains unchanged with respect to PD6493:1991, with Level 3A and 3B dependent on the type of stress-strain data available as for Level 2. A novelty of Level 3 options is the addition of the R6 Option 3 method, which becomes Level 3C in BS 7910. In this approach, the FAD and driving force may be derived from elastic-plastic finite element analysis to give more accurate predictions of structural behaviour.

BS 7910 includes 21 Annexes, several of which originate from the R6 procedure. Those which are particularly relevant to fracture assessments include flaw re-characterisation rules, a leak-before-break analysis procedure, advice on calculating reserve factors and performing sensitivity analyses, and consideration of mixed mode loading. Others provide a more extensive collection of reference stress (limit load) and stress intensity factor solutions, including solutions from three-dimensional finite element analyses (FEA) for weld toe cracks, guidance on the treatment of weld metal/parent material strength mismatch and on the fracture toughness testing of different areas of weldments. Profiles of residual stress distributions for common joint configurations are given and new guidance has been written, again based on work in SINTAP, on correlations between Charpy energy and fracture toughness including incorporation of the so-called Master Curve concept [8,9] . There is also improved consideration of proof testing and warm pre-stressing and guidance on reporting the results of flaw assessments. The lengthwise flaw interaction criteria for fracture assessment have been relaxed compared to PD6493:1991 based on the findings that there is almost no crack driving force enhancement of adjacent flaws in this direction.

3.3. Fatigue assessments

For BS 7910, the fatigue assessment clauses in PD6493:1991 were reviewed in the light of new information and experience gained from their use in practice. The main change was the introduction of new fatigue crack growth laws, based on an extensive review and analysis of published data for steels [10] . These include more precise two-branch Paris laws and allowance for applied stress ratio, R (the ratio of minimum to maximum applied stress during fatigue loading). More attention is paid to environmental influences and the new recommendations cover marine corrosion, with and without cathodic protection, and fatigue crack growth at elevated temperature. The specific cases covered include ferritic steels in air, freely-corroding in seawater and in seawater with cathodic protection (-850mV and -1100mV Ag/AgC1), see Fig.3.


     

Fig.3. Recommended fatigue crack growth laws in BS7910

 

New, simplified conservative (upper bound) single-branch Paris laws are also provided, for convenience. They relate to high R-values (R ≥ 0.5) in order to give conservative estimates of fatigue crack growth in welded structures. As in fatigue design, these are assumed to contain high tensile residual stresses and hence to experience a high effective R under any fatigue loading. The basic law for ferritic steels in air gives a slightly higher crack growth rates than the corresponding law in PD6493:1991, based on more recent experimental data obtained at R = 0.5. Apart from air, the recommended laws also cover ferritic steels in seawater and at elevated temperature.

Austenitic steels can be treated using the simplified law for ferritic steels in air, but no advice is given for other environments. New data did not justify any changes to the recommended stress intensity factor threshold values in PD6493:1991, except that a value of zero is now recommended for steels in freely corroding seawater. However, it is now strongly recommended that crack growth rate and threshold values for high R values are used when assessing a flaw in a welded structure, to allow for the influence of high tensile residual stresses.

Advice on the derivation of fatigue crack growth laws and threshold values for non-ferrous metals is also given, using correlations based on relative Young's modulus values.

Allowance has been made for the extensive evidence now available which indicates that there is no need to impose the flaw interaction criteria in PD6493:1991 in a fatigue assessment. Thus, multiple flaws are assessed separately without any consideration of flaw interaction [11,12] .

The fatigue assessment method referred to in PD6493:1991 as the 'Simplified Procedure', which relates the required and actual fatigue performance of a flaw to a grid of quality category (stress versus endurance) S-N curves, is retained, but it is now referred to as 'Assessment Using Quality Categories'. For consistency with Eurocode design S-N curves, the reference stress range associated with each quality category curve now corresponds to N = 2 x 10 6, rather than 10 5, cycles. New graphs for assessing planar flaws using the simplified fracture mechanics method have been introduced, based on the new upper bound Paris fatigue crack growth law. The acceptance levels for non-planar flaws (slag inclusions, porosity and undercut) are still consistent with available experimental data, and apart from extending the thickness range over which the undercut acceptance limits apply, they are unchanged in BS 7910.

3.4. Procedures for the assessment of flaws in high temperature plant

The high temperature flaw assessment procedure [13] is described in Clause 9 of BS 7910 with further information included in Annex T on how to make the calculations and a worked example in Annex U. Both failure by net section creep rupture and creep crack growth are considered. The procedures are based on PD6539:1994 [14] guidance which has been updated for BS 7910.

The BS 7910 calculation procedure follows a similar format to that used for making fracture and fatigue assessments. The plant operating conditions and material properties are determined. Any flaws present are characterised and evaluated for fatigue and fracture. If the flaw is acceptable with respect to these failure modes, the creep damage in the uncracked ligament is assessed via a ductility exhaustion criterion using the appropriate reference stress values. In addition, the amount of creep crack growth is calculated for the conditions of concern. The whole process is repeated for successive time steps until the failure condition is reached or the desired lifetime is achieved.

Calculations can be carried out at the initial design stage for postulated defects or after a defect has been found during an inspection to determine the remaining lifetime. It is recommended that a sensitivity study is performed to give added confidence in the predictions. The step-by-step details are given in Annex T and a worked example, showing how secondary stresses and combined creep and fatigue are dealt with, in Annex U. Protection against creep rupture is achieved by limiting the ductility exhaustion in the uncracked ligament to a suitable fraction and prevention of fracture by restricting the amount of cracking allowed.

3.5. Partial safety factors in BS 7910

3.5.1. Background

For general structural assessment purposes, a comparison of load and resistance is used to predict failure. When there are uncertainties in the input variables, or scatter in the materials data, reliability analysis methods can be employed to determine the probability of failure, i.e. the probability that the load effects will exceed the resistance effects. Partial safety factors (PSFs) can be applied to individual input variables to give a target reliability without having to carry out probabilistic calculations. The values of partial safety factors depend on the target reliability, the coefficients of variation (COV) and the number of standard deviations from the mean taken to represent the load and resistance distributions.

3.5.2 Existing partial safety factor specifications

The first published document giving recommendations for partial safety factors in connection with fracture mechanics-based structural integrity assessments was PD6493:1991. The recommendations were derived as part of work by Glasgow University/UMIST [15] on defect assessment methodology for offshore structures in the late 1980s and were included as part of an optional appendix in PD6493. The procedure in use at the time involved the then Level 2 assessment curve based on the flow strength parameter S r, applied to results of wide plate tests. In view of changes to the PD6493 fracture assessment clause and also because of developments in structural Eurocodes it was appropriate that a reassessment of partial safety factors should be carried out.

There is a general target reliability index value adopted in Eurocodes for ultimate limit state conditions in structures for which failure would have major consequences which corresponds to a failure probability of about 7 x 10 -5. Since this value has been derived to deal with the appropriate uncertainties in loading for plastic collapse failure, the same PSFs were included for fracture/plastic collapse failure to ensure consistency with existing procedures.

3.5.3. Derivation of BS 7910 partial safety factors

Recommendations were produced to cover different requirements for target reliability and different degrees of variability of the input data. The target reliability levels corresponded to those used previously in PD6493 with the addition of the standard level adopted in Eurocode 3, namely failure probabilities of 2.3 x 10 -1, 10 -3, 7 x 10 -5 and 10 -5.

The input variables considered for these assessments were stress, flaw size, fracture toughness and yield strength. The recommendations are intended to cover the assessment of a single flaw rather than a full distribution of flaw sizes. For the purposes of determining partial safety factors the results are derived in terms of different COVs.

It is important to recognise that there is no unique solution for partial safety factors and even when a preliminary separation is made into load and resistance groups, there are still many alternative combinations of factors which could be applied to the separate input variables to give the same required target reliability. The most appropriate solutions are those for which the partial safety factors remain approximately constant over a wide range of input values. For each of the data groups, values of partial safety factors have been selected for each target reliability and COV to cover for all the other cases.

The partial safety factors in PD6493 for fracture toughness and flaw size are significantly lower than those recommended in BS 7910 for target reliabilities of 10 -3 and 10 -5. On the other hand, the partial safety factors on stress in PD6493 are somewhat higher than those in BS 7910. An important difference is that the calculations for the BS 7910 factors have been carried out assuming that 'failure' occurs in accordance with the failure assessment diagram, whereas in practice it is often found that the diagram gives safe predictions rather than critical ones. In contrast, the original PD6493 partial safety factors were calibrated against wide plate test results, rather than the failure assessment curve and this explains some of the differences.

3.5.4. Concluding remarks

New recommendations have been produced for BS 7910 partial safety factors for use in structural integrity assessments where the primary modes of failure are fracture and plastic collapse. The recommendations are more comprehensive and have been designed to be compatible with relevant codes for the design of steel structures. The new factors are higher for fracture toughness and flaw size than PD6493 values, but this is compensated to some extent by the new values being lower for stress levels. A more detailed description and discussion of the development of BS 7910 partial safety factors has been given by Burdekin et al [16] .

3.6. Assessment of general corrosion in pipes and pipelines

3.6.1. Introductory remarks

General corrosion damage in pipelines reduces wall thickness, either locally or globally, resulting in a reduction of load bearing capacity and/or stiffness of the pipeline structure compared with the design conditions. The presence of corrosion damage has therefore safety and cost implications with respect to the operation of high pressure transportation and storage systems.

Whilst there are a number of established engineering methods which are suitable for the assessment of corroded pipes, pipelines and cylindrical vessels, significant developments in the areas of flaw detection and fitness-for-purpose assessment techniques have been made which enable pipeline companies to determine the remaining strength of corroded pipelines more accurately and with higher confidence. More specifically, a large group sponsored project has been managed by BG Technology, the results of which have been incorporated in BS 7910 Annex G.

3.6.2. Corrosion assessment method

3.6.2.1. Procedure development

The project combined an extensive programme of full-scale burst tests on pipe samples containing machined corrosion models, with numerical modelling methods incorporating three-dimensional non-linear finite element analysis (FEA).

Isolated, grouped (interacting) and combined (complex-shaped) corrosion defect models were considered. Defect shapes included pits, grooves (corrosion bands) and patches (general corrosion) with defect depths ranging from 20% to 80% of wall thickness. The pipe grades tested ranged from Grade X52 to X65. Various pipe diameters from 203mm (8in) to 914mm (36in) and diameter/thickness ratios from 8 to 64 were included. The pipe materials and pipe sizes considered in this programme are complementary to the existing AGA (American Gas Association) database. In total, 81 full-scale pipe burst tests and 52 ring expansion tests were completed in the experimental programme; these included six burst test results and 16 ring expansion test results.

The results from the guidance development studies were incorporated in BS 7910 Annex G, prepared in a suitable format, following feedback by practising pipeline engineers.

3.6.2.2. Outline of BS 7910 Annex G procedure

General procedure

The general methodology for assessing corroded pipeline is illustrated by a flow chart, as shown in Fig.4. The methods recommended in this procedure are classified into three levels of assessment, depending on required accuracy of the assessment and the level of information available. Limitations of the methods and safety factors recommended are included in Annex G of BS 7910.
Fig.4. BS7910 Annex G assessment procedure for assessment of corroded pipe and pipelines
Fig.4. BS7910 Annex G assessment procedure for assessment of corroded pipe and pipelines

Level 1 Corrosion assessment

A set of rules identifying longitudinal and circumferential flaw interaction is used for grouping corrosion flaws.

The flaw groups are then treated as single isolated flaws and are assessed using a simple equation, which only requires limited information on material properties and defect sizes. A relation is presented in Annex G to determine safe working pressures using actual or specified minimum tensile strength values for the pipe material and the total projected axial length and maximum depth of the flaw. This relation is based a large parametric database of failure pressures, predicted using non-linear various pipe sizes, pipe grades, defect configurations and defect sizes, and validated against over 100 full-scale burst tests. The estimated safe working pressure is correct only if there is no flaw interaction.

Level 2 Corrosion assessment

Two engineering methods have been developed for the Level 2 assessment dealing with corrosion interaction (Level 2a) and with complex-shaped defects (Level 2b). As further validation of Level 2b is required, it has not been included in BS 7910 Annex G.

With respect to Level 2a, a flaw group may include a number of separate but adjacent corroded areas which interact. The failure pressure is higher than that for a single flaw configuration of the same dimensions.

A safe working pressure is determined by the lowest pressure value predicted from assessments of all combinations of adjacent flaws. Each of the assessments assumes the total length of the defect group but an equivalent depth. This assessment requires additional information on flaw spacing. The method has been validated against a large number of failure predictions (using FEA) for equally sized, axially-separated flaw groups and some pipe burst tests.

Level 3 Corrosion assessment

As the most advanced method, included in BS 7910 Annex G, non-linear FEA together with a validated criterion is recommended for detailed assessment if full material information and flaw configuration are available. The failure pressure of a corrosion model is deemed to have been reached at the load where the von Mises equivalent stress value throughout the remaining ligament reaches the tensile strength of a material. This failure criterion has been well validated. Such analyses generally give accurate failure predictions (within ±5%).

 

3.6.3. Concluding remarks

New guidelines have been developed by a BG Technology-led group sponsored project enabling pipeline engineers to indicate more accurate and less over-conservative assessments of pipeline corrosion, particularly for complex-shaped and interacting groups of corrosion. The guidelines have been incorporated in BS 7910 Annex G and have the potential to significantly reduce the costs associated with continued operation or repair of pipelines in which in-service corrosion is found. More details of the corrosion assessment procedures have been given by Fu and Andrews [17] .

 

3.7. Summary of changes in BS 7910

The principal changes are as follows:
  • The procedures now have the enhanced status of a 'Guide'. This unfortunately, means that the well known numbering 'PD6493' will change to 'BS 7910'. The title has been changed to 'Guide on methods for assessing flaws in metallicstructures'. This is because the committee considered that the methods were as applicable to flaws in castings, forgings, etc, as to flaws in welds.
  • The scope now formalises existing practice so that the document can be used not only to assess initial fabrication flaws, but also for those found as a result of in-service inspection. Thus a major application of the document willbe to enable decisions to be made on life extension of ageing plant.
  • The treatment for fracture has been even more closely aligned with the R6 approach.
  • In 1994, another BS Published Document was issued, PD6539 [7] , giving methods for assessing flaws in high temperature plant. This has now been integrated into BS 7910, so that creep crack growth can now be assessed.
  • A number of annexes (twenty-one in total) have been added, providing state-of-the-art methods for engineering critical analysis and increasing the scope of the document.
  • Guidance on the assessment of generalised corrosion in pipelines and vessels has been introduced in Annex G.

4. Validation of BS 7910 fracture assessment procedures using TWI wide plate data

Validation of the BS 7910 fracture and collapse assessment procedures has been carried out against TWI experimental data of large-scale wide plate fracture tests. Relevant small-scale fracture toughness specimens were used to provide appropriate input data. Eighty-two wide plate specimens have been assessed at Level 2A. The TWI software Crackwise 3 which automates the BS 7910 assessment procedures has been used to carry out the appraisal. The wide plate tests cover a wide range of materials, flaw type/location, load configurations and test temperatures. Materials used were pressure vessel steels, C-Mn structural steels, pipeline steels, aluminium alloys and type 316 stainless steels and their weldments. Flaw types included through thickness cracks, semi-elliptical surface cracks, extended long surface cracks. The flaw location was in parent material, weld and HAZ. Applied loads included externally applied tension, bending or combined tension and bending and welding residual stresses. The test temperatures covered the entire transition range from the lower to the upper shelf.

The failure points of all eighty-two wide plate specimens were predicted correctly by BS 7910 Level 2A assessment procedures, and both K-based and CTOD based assessment routes produced conservative predictions. The K-based results by BS 7910 are presented graphically in Fig.5a. In comparison, assessments to PD6493 Level 2 led to some non-conservative failure predictions, see Fig.5b. It was well known that the PD6493 Level 2 FAD can result in marginal failure predictions in the 'knee' region of the FAD for high work hardening materials. This was one of the reasons to discontinue its use in BS 7910.

(a) to BS7910 Level 2A;
(a) to BS7910 Level 2A;
(b) to PD6493 Level 2
(b) to PD6493 Level 2

Fig.5. Fracture assessment of TWI wide plate data results based on K or J fracture toughness values: (a) to BS7910 Level 2A; (b) to PD6493 Level 2.


Figure 5a shows that the degree of conservatism in the BS 7910 fracture assessment procedure varies and can be quite small. This implies that to ensure a safe deterministic fracture assessment, conservative approaches must be adopted. The current assessment results have been obtained under the following conditions: fracture toughness data were obtained from high-constraint SENB specimens; the minimum value of three tests or the second lowest value of six tests were employed. If the population of fracture toughness tests is large enough to enable a statistical analysis, the mean minus one standard deviation value was used. Caution is advised when using maximum load fracture toughness data for high work hardening materials; rather, it is recommended to apply Level 3 tearing assessments.

In the as-welded condition, residual stresses must be considered not only for cracks located in weld metal and HAZ, but also for cracks close to the weld. The reduction of welding residual stresses after proof tests should be according to BS 7910 Annex O. As recommended in BS 7910 for cracks in the weld metal, a conservative approach is to use the lower properties of parent and weld metals for calculating the collapse parameter L r, even if the weld metal over-matches the parent material as recommended in BS 7910 Annex I. It is recommended to use only validated and well-established stress intensity factor and collapse load solutions, and to be aware of the validity range of solutions for unusual geometries given in BS 7910 Annex M and P.

5. Outlook

The British Standards Institution has recently submitted BS 7910 to the European Committee for Standardisation (CEN). Because of differences in the methods of treating fracture in various European countries, the CEN committee (TC121/WG14) responsible for this topic at present considers that it would not be possible to publish a universally acceptable standard immediately. It was therefore decided to issue BS 7910 as a CEN Technical Report with a view to commence drafting a CEN Standard in the near future.

In the meantime, the major three-year European collaborative research project, Structural Integrity Assessment Procedures for European Industry (SINTAP), has been completed. The project consortium consists of seventeen partners from nine different European countries. The project aimed at resolving national differences in fracture assessment methods and at arriving at a consensus approach. Another large European collaborative project to resolve national differences in assessment methods for High Temperature Defect Assessment (HIDA) is also nearing completion.

Important developments are also taking place in the United States. The American Petroleum Institute has just published a document, API 579 [12] , giving recommendations for fitness-for-purpose evaluation of pressurised equipment in the refinery and chemical industry. Fitness-for-purpose is defined as the ability to demonstrate the structural integrity of a component containing a flaw. API 579 is mainly targeted at the assessment of ageing plant, a subject of increasing importance. Some of the methods are similar to those contained in BS 7910. However, API 579 devotes more attention to specific situations which arise in ageing petrochemical plant. Examples are: general metal loss and locally thinned areas, blisters and laminations and fire damage. Valuable guidance is also given on the possible remedial measures to be taken for each of the flaw types and failure mechanisms covered by the document. In addition, the American Society of Mechanical Engineers is working on a document dealing with post-construction inspection. Details of this are not yet available.

The outcome of the European collaborative projects mentioned above will be combined with input from the mentioned international procedures and the existing BS 7910 methods to begin drafting of a CEN or international procedure in due course.

6. References

  1. PD6493:1991: 'Guidance on methods for assessing the acceptability of flaws in fusion welded structures'. British Standards Institution, London, 1991.
  2. BS 7910:1999: (incorporating Amendment No.1) 'Guide on methods for assessing the acceptability of flaws in metallic structures', British Standards Institution, London, 2000.
  3. Harrison J D, Burdekin F M and Young J G: 'A proposed acceptance standard for weld defects based upon suitability for service', In: Procs. 2nd conference on the significance of defects in welds, London, May 1968. Abington, Cambs: The Welding Institute. Paper 1, 65-79.
  4. British Standards Institution. 1980. PD 6493: 'Guidance on methods for assessing the acceptability of flaws in fusion welded structures'. 1st ed. London, British Standards Institution.
  5. Burdekin F M and Dawes M G: 'Practical use of yielding and linear elastic fracture mechanics with particular reference to pressure vessels'. Conf. On Practical Applications of Fracture Mechanics to Pressure Vessel Technology, London, May 1971, Mechanical Engineering Publications.
  6. Milne I, Ainsworth R A, Dowling A R and Stewart A T: 'Assessment of the integrity of structures containing defects'. CEGB report R/H/R6 - Rev.3. Barnwood, Glos, British Energy Generation Ltd, 1987.
  7. BRITE EURAM Project: Structural Integrity Assessment Procedure for European Industry - SINTAP, Procedure Document, British Steel (now Corus Group) Swinden Technology centre, Rotherham, UK, November 1999.
  8. Wallin K: 'A simple theoretical Charpy V - K IC correlation for irradiation embrittlement'. Proc Conf ASME PVP 1989. (D L Mariott, T R Mager, W H Bamford Eds.), American Society for Mechanical Engineers, 1989.
  9. Wallin K: 'Fracture toughness transition curve shape for ferritic structural steels'. Proc. Conf. Fracture of Engineering Materials and Structures, 61-79, Elsevier Applied Science, 1991.
  10. King R: 'A review of fatigue crack growth rates in air and seawater', HSE Report OTH 511. Health and Safety Executive Books, London, 1998.
  11. Morgan H G: 'Interaction of multiple fatigue cracks'. Proc. Conf. Fatigue of Welded Structures, TWI, Cambridge, UK, 1987.
  12. Soboyejo W O: 'On the prediction of the fatigue propagation of semi-elliptical defects', ASTM STP 1122, Advances in Fatigue life prediction techniques, American Society of Testing and Materials, 1997.
  13. Webster G A and Ainsworth R A: 'High temperature component life assessment'. Chapman and Hall, London, 1994.
  14. PD6539:1994: 'Guide on methods for the assessment of the influence of crack growth on the significance of defects in components operating at high temperatures'. British Standards Institution, London,1994.
  15. Plane C A, Cowling M J, Nwegbu V K and Burdekin F M: 'The determination of safety factors for defect assessment using reliability analysis methods', Third International Symposium on Integrity of OffshoreStructures, September 1987.
  16. Burdekin F M, Hamour W, Pisarski H G and Muhammed A: 'Derivation of partial safety factors for BS 7910'. Proc. Conf. IMechE seminar: 'Flaw assessment in pressure equipment and weld structures - PD6493 toBS 7910'. IMechE Publications, London, UK, 2000.
  17. Fu B F and Andrews R M: 'Assessment of general corrosion in pipes and pipelines'. Proc. Conf. IMechE seminar: 'Flaw assessment in pressure equipment and weld structures - PD6493 to BS 7910'. IMechEPublications, London, UK, 2000.
  18. High-Temperature Defect Assessment - HIDA. Final report will be available in due course: enquiries to Dr I A Shibli, European Technology Development Ltd, 2 Warwick Gardens, Ashtead, (KT21 2HR), Surrey,UK.
  19. American Petroleum Institute: API 579. Recommended practice for fitness-for-service, API, Washington DC, 2000.

 

Presented with permission from Elsevier

For more information please email:


contactus@twi.co.uk