Structural Integrity Technology Group, TWI Ltd, Cambridge, UK.
Paper presented at ESIA9 - 9th International Conference on Engineering Structural Integrity Assessment - 15-19 October 2007, Beijing, China.
TWI Ltd has an ongoing research program aimed at validating and improving Fitness-for-Service assessment procedures for Environment Assisted Cracking. Initial studies have focused on the shallow crack phenomena and this paper reviews current assessment procedures, highlighting one area where further experimental work is required.
Setting conditions for the avoidance of in-service crack growth in aggressive corroding environments has long been a major challenge due to the number of variables that have a significant effect on material behaviour. Under static loading conditions, shallow stress corrosion cracks may grow faster or slower than deeper cracks, depending on the material-environment system. There are several reasons why shallow cracks might behave differently to deep cracks. For example, a crack's size relative to microstructural features, environmental effects within the crack and the size of the crack tip plastic zone can all influence behaviour, Jones and Simonen. 
Review of EAC assessment procedures
It is fair to say that none of the established Fitness-for-Service (FFS) standards contain comprehensive assessment procedures for environment assisted cracking (EAC), although all highlight the importance of only using datarelevant to the actual environment and loading conditions. Methods for evaluating EAC within current integrity assessment procedures are usually based on avoiding the phenomena by limiting the stress ( σ< σ SCC ) for crack free components, or limiting the stress intensity factor (Kt;K ISCC ) where a crack or flaw already exists. For relatively deep cracks, K ISCC is an appropriate characterising parameter, as suggested in BS 7910, and standard test techniques can be used to determine material-environment specific data. Similarly, for defect free components, where theinitiation of cracks is the dominant factor on life, a threshold stress ( σ SCC ) is appropriate, and again test techniques exist for generating suitable data. Shallow cracks fall between these two extremes, and it is currently unclear how behaviour in this regime should be characterised.In some cases the extrapolation of 'deep-crack' test data to predict the performance of a component containing a shallow crack may be non-conservative, ie the threshold for avoiding cracking may be lower, or the crack growth rate may be higher. Only R6 makes specific reference to the limitations associated with the use of deep-crack test data, and only in reference to fatigueor corrosion fatigue behaviour.
Currently the European Fitness-for-Service Thematic Network's recently released final draft (FITNET Mk7)  is the only standard to contain a specific section that deals with EAC, although API RP 579 may incorporate something similar in the next revision. FITNET Mk7 highlights a number of weaknesses in current materials testing and integrity assessment procedures, and specifically mentions difficulties associated with the assessment of shallow cracks (  , Turnbull et al.  ). It also introduces the concept of a two-parameter approach to the assessment of EAC ( Figure 1). This diagram describes a transition between K-controlled and stress-controlled behavior as the crack size is reduced, and emphasizes the need for care in the shallow-crack regime as the critical stress may be lower than that obtained by extrapolating the deep crack data. Unfortunately, although the diagram in Figure 1 describes this aspect of the shallow-crack problem very well, it does not identify a clear methodology for the assessment of shallow flaws. The form of the curve in the shallow-crack regime has no experimental justification and further work in this area is needed before a robust procedure for a quantitative assessment of shallow flaws could be developed.
Regimes of behaviour
There are a number of ways that the early stages of EAC may be modelled, not covered in this paper. The diagram in Figure 1 describes one aspect of the shallow crack problem, in that the appropriate characterising parameter changes from K to σ as the crack size decreases. In the transitional regime, models developed to explain the nature of this transition will rely on the availability of suitable shallow crack test data, and work is currently under way at TWI to generate data of this kind for a number of material-environment systems. In Figure 1 the shallow crack regime is essentially defined by the relative magnitude of K ISCC and σ SCC . However, crack size dependency associated with differences in crack tip chemistry may occur over a different scale and the potential influence that this has on observed thresholds or crack growth rates should not be ignored. With respect to material response to static loading, there is a need to define the extent over which deep-crack K ISCC or da/dt data can be used to assess real, potentially shallower, flaws.
In defining the applicable limits of deep-crack data ( ie K ISCC and/or da/dt), it can initially be assumed that the stress state can still be characterised in terms of an applied stress intensity factor (K), and that we are considering the possibility that cracks may be 'chemically small' due to differences in crack tip chemistry. If flaws are still large with respect to the extent of crack tip plasticity or microstructure, K remains an appropriate characterising parameter. Under these conditions, conventional assessments (using K ISCC and/or da/dt data) may be carried out, so long as the input data are conservative, ie test data are associated with specimens containing flaws of a comparable size to those being assessed. In many instances this will be an adequate approach, as the minimum crack size that can be reliably detected by NDT, may be significantly larger than that needed for the stress field to be characterised by K. Under these circumstances there is no need to consider the behavior of smaller ('mechanically small' or 'microstructurally small') flaws as any FFS assessment will be limited to a minimum assumed flaw size. Only if such flaws are 'chemically small' is there a danger of non-conservatism. It is likely that conventional fracture mechanics test techniques can be adequately modified for exploring material behaviour in this regime.
In cases where there is a need to model the behaviour of mechanically or microstructurally small cracks, the conventional criterion for avoiding subcritical crack growth (K<K ISCC ) is no longer appropriate. As illustrated in Figure 1, the situation that arises is similar to that described by Kitagawa and Takahashi  for fatigue loading. In essence, it is known that in the absence of a crack, there is a critical stress to avoid crack growth ( σ SCC ), and in the shallow crack regime some transitional behaviour might be expected. Leaving aside the precise details of this transitional regime, a dual criterion for avoiding stress corrosion cracking could be envisaged as follows:
σ/ σ SCC < 1 and K/K ISCC < 1 (1)
as described by the two dashed lines in Figure 1. Of course, for static loading conditions, the conventional way of describing the interaction between two failure mechanisms (one K-controlled and one stress-controlled) is a failure assessment diagram (FAD). Indeed for the simplest type of FAD ( ie BS 7910 level 1)  the criterion for avoiding failure can be expressed as:
S r = σ ref / σ f < 0.8 and K r = K l / K mat < 0.707. (2)
Safety factors aside, the similarities between the two sets of criteria are apparent. Exploring the validity of either the straight line construction in Figure 1, or some form of postulated transitional behaviour requires the availability of shallow crack test data. In this case test data are required over a wide range of crack sizes, although some estimate of the crack size of interest can be determined by examining the values of K ISCC and σ SCC for the particular material-environment system of interest. The two straight lines in Figure 1 will intersect at a point corresponding to:
where Y is a geometry dependent term. It is clear that the ratio K ISCC
/ σ SCC
is critical in determining the crack size over which the anticipated transition from K-controlled to stress-controlled behaviour occurs and this will depend on the material-environment system of interest. For example, for pipeline steels in a sour environment, where K ISCC
≈ 30 MPa√m and σ SCC
≈ 220-440 MPa (Contreras et al
, Pargeter et al
, Sponseller 
), this transition is expected when the crack depth is approximately 3-6 mm. By contrast, for high strength aluminum alloys in seawater, where K ISCC
≈ 2-7 MPa √m and σ SCC
≈ 550 MPa, (Bayoumi 
, Ohsaki 
), the critical regime is 4-50 µm. It is clear that specimen geometry and testing procedures for examining the anticipated transition in these two cases would differ considerably.
For applications where FFS assessments are based on NDT inspection limits, it is unlikely that flaws smaller than 1mm will be of practical interest. Therefore slow strain rate and/or constant load tests should be carried out to investigate the effect of crack depth on the measured value of K ISCC to examine the possibility that differences in crack tip environment have an influence on material behaviour. For other applications, where the design philosophy may be different, there may be an interest in characterizing the behaviour of smaller flaws. Under these circumstances there may be a different transition from K-controlled to stress-controlled behaviour, and shallow crack data are needed to develop models for material behaviour in this regime.
- Bayoumi M R, 1996: 'The mechanics and mechanisms of fracture in stress corrosion cracking of aluminium alloys', Engineering Fracture Mechanics 54, No. 6, pp879-889.
- BS 7910, 2005: 'Guide to methods for assessing the acceptability of flaws in metallic structures', British Standards Institution, London.
- Contreras A, Albiter A, Salazar M, Perez R, 2005: 'Slow strain rate corrosion and fracture characteristics of X-52 and X-70 pipeline steels', Materials Science and Engineering A407, pp45-52.
- FITNET, 2006, FITNET Fitness-for-Service Procedure Final Draft MK7, Prepared by European Fitness-for-Service Thematic Network FITNET.
- Jones R H and Simonen E P, 1994: 'Early stages in the development of stress corrosion cracks', Materials Science and Engineering, A176 211-218.
- Kitagawa H and Takahashi S, 1979: 'Applicability of fracture mechanics to very small cracks of the cracks in the early stage,' 2nd Int. Conf. on Mechanical Behaviour of Materials, pp627-631.
- Ohsaki S, Kobayashi K, Iino M, Sakamoto T, 1996 : 'Fracture toughness and stress corrosion cracking of aluminium-lithium alloys 2090 and 2091', Corrosion Science, 38, No. 5, pp793-802.
- Pargeter R J, Gooch TG and Bailey N, 1990 : 'The effect of environment on threshold hardness for hydrogen induced stress corrosion cracking of C-Mn steel welds', Conference Proceedings 'Advanced Technology in Welding, Materials, Processing and Evaluation', Japan Welding Soc, Tokyo, April 1990.
- Sponseller D L, 1992 : 'Interlaboratory testing of seven alloys for SSC resistance by the double cantilever beam method', Corrosion, 48, No. 2, pp.159-171.
- Turnbull A, Koers R W J, Gutierrez-Solana F and Alvarez J A, 2005: 'Environment induced cracking - A fitness-for-service perspective', Proceedings of OMAE 2005, 24th International Conference on Offshore Mechanics and Arctic Engineering, OMAE2005-67566.