Through the years, a number of methods have been used to assess the significance of crack-like defects in pipelines. Some methods have been developed from first principles and fracture mechanics, but other, more empirical methods have been developed to better account for observable complex mechanisms.
This article briefly examines the different methods currently available and outlines the underlying assumptions upon which each method relies, all with the aim of helping readers select the most appropriate method to assess their pipeline.
When crack-like defects are discovered in a pipeline, the first questions operators usually ask are related to whether the pipeline can still be operated safely, how many repairs (if any) will be required and when those repairs need to be completed.
A number of methods have been developed to help provide answers to these questions. Some models are based on first principles and fracture mechanics and were initially developed for generic applications (not specific to pipelines). These models can sometimes be very conservative, as they do not take into account the intricacies of a particular situation. Other methods have been developed by the pipeline industry that are heavily based on observation and experimentation, giving rise to semi-empirical methods that can be simple to use and give accurate results when used within the range of their original calibration. This article aims to give an overview of these different assessment methods and models to provide operators with a better understanding and assist integrity engineers in selecting an appropriate assessment method for their specific case.
Two of the most widely used assessment methods for pipeline defects are the British Standard BS 7910:2019 “Guide to methods for assessing the acceptability of flaws in metallic structures”i and the American Petroleum Institute/American Society of Mechanical Engineers’ API579-1/ASME FFS-1 – 2021 “Fitness for Service”ii. These documents are referenced in a number of pipeline-industry standards and becoming practices, bringing benefits to the industry due to a consistent and robust approach that is widely understood. These methods have gained a level of popularity in the pipeline industry for the assessment of crack-like defects, and there is now extensive experience in applying them to pipelinesiii,iv.
BS 7910 is concerned with all types of metallic structures and components, while API 579-1/ASME FFS-1 (more commonly known as API 579) is specific to pressurized equipment used in the petrochemical and allied industries. Assessment of flaws is carried out based on the concept of fitness for purpose (FFP). This FFP concept deems a structure “fit to carry out its purpose,” provided that conditions for failure are not reached. Hence, by carrying out a critical engineering assessment (or simply “engineering assessment”), known or hypothetical flaws are assessed against this criterion.
To determine acceptability of crack-like flaws, both BS 7910 and API 579 use a failure assessment diagram (FAD) that considers failure by fracture (either brittle or ductile) and plastic collapse. A safe/not safe boundary is calculated using material properties and in some cases the geometry. Points are then plotted on this domain based on the loads and the flaw geometry and subsequently deemed acceptable or not. The figure shows a typical FAD with an assessment point inside the boundary, which means this particular flaw would be considered safe.
When the field of fracture mechanics was being developed, it was for an idealized, simplified linear-elastic regime. This regime was difficult to apply in the pipeline industry, where the structures have thin sections, sometimes low strength and high ductility, and definitely display non-linear-elastic behaviors. After much work in this area, the NG-18 equations (produced by the NG-18 committee of the American Gas Association) were developed, based on both simple theory and full-scale experiments. This experimental program helped overcome a number of challenges presented by phenomena such as high ductility, crack blunting and material constraints that are observed in the real world but not easily accounted for in the generic methods. Since the initial development of the NG-18 semi-empirical equations in the 1960s, there have been a number of refinements made to them to take into account the controlling mechanisms for failure (fracture or flow controlled) and different flaw shapes.
Other assessment methods exist that include the ductile-flaw growth model (DFGM) and the closely related pipe axial-flaw failure criterion (PAFFC), which are based on elastic-plastic fracture mechanics theory. Corrosion life assessment software (CorLAS) based on the Electric Power Research Institute (EPRI) J estimation scheme was originally developed for assessing stress corrosion cracking (SCC) flaws in pipelines but has been applied to cracks and crack-like flaws of different origin.
FRACTURE TOUGHNESS INPUT
A critical input into both the generic and the pipeline-specific models is the fracture toughness of the pipeline steel. Traditionally, the “toughness” of line pipe has been quantified by the impact energy measured using the Charpy V-notch (CVN) test. However, this CVN method is a qualitative measure of toughness, and these require further validation and calibration in full-scale tests in order to use them as a proxy for quantitative values. Quantitative methods for directly measuring a material’s toughness, such as crack tip opening displacement [CTOD] or the J-integral, are not as popular because they tend to be more expensive than CVN tests. Hence, the majority of test results available (historical and current) are for CVN tests only.
This lack of quantitative data has significant effects when trying to apply the assessment methods mentioned before; “generic” methods rely on this quantitative data, whereas the semi-empirical methods use the CVN values directly. To attempt to use the CVN results in generic methods, a number of correlations have been formulated, which are greatly dependent on a number of factors, such as fracture mechanism, temperature, etc. Choosing the correct correlation becomes vital, as selecting the incorrect correlation can mean that the results are meaningless and do not accurately represent the asset’s condition. Also, as a general rule, generic methods tend to give much more conservative results than semi-empirical methods, again stemming from the conservative nature of the toughness correlations.
No matter which type of assessment method is used, another dimension that needs to be taken into account is time. Are the anomalies likely to be growing over time, or are they static and stable? Are they actually cracks or just planar features with blunt tips? When considering crack changes over time, many parameters can have a very large impact on the assessment. These include assumed initial size, assumed growth mechanism and rate, estimated driving forces/loads, and the estimated final critical size.
SO, WHICH METHOD TO USE?
In general, the NG-18 equations tend to be more accurate for ductile pipe steels (as long as the steel is in the CVN “upper-shelf” region, noting that most good quality pipe steels exhibit upper-shelf behavior at typical pipeline operating temperatures). The generic methods tend to be more conservative, but they can be used to give safe results when material properties are uncertain or there is known low ductility and/or toughness. It should be noted that generic methods can give non-conservative results when calculating fatigue life. Of course, these statements are generalizations made for clarity in this introductory overview. It is good practice to always seek expert advice prior to assessing crack-like anomalies.
When it comes to assessing crack-like defects in pipelines, one method is not necessarily “better” than another. Each method has its advantages and disadvantages, which may depend on the particular material and type of crack. Constrained by the data available and the set of assumptions made, a decision choosing one method over the other is eventually made. If time-dependent growth is to be considered, further complexities and considerations are required for the assessment of crack-like anomalies in a pipeline. It is always better to seek the advice of subject-matter experts to ensure that the most appropriate method is chosen and the inputs used are the most suitable for that method, thus avoiding falling into the age-old “garbage in, garbage out” trap.
Figure 1: Failure assessment diagrams (FAD) are used by both BS 7910 and API 579 to determine acceptability of crack-like flaws.
i BS 7910:2019, “Guide to methods for assessing the acceptability of flaws in metallic structures,” 2019, BSI Standards Publication.
ii API579-1/ASME FFS-1 – 2021, “Fitness for Service,” 2021, American Petroleum Institute/American Society of Mechanical Engineers.
iii “Application of BS 7910 to high pressure pipelines,” Andrews, R., Cosham, A. and Macdonald, K. 2018, International Journal of Pressure Vessels and Piping, 168, pp. 148-155.
iv “Crack-Like Defects in Pipelines: The Relevance of Pipeline-Specific Methods and Standards,” IPC2012-90459. Cosham, A., Hopkins, P. and Leis, B. 2012, International Pipeline Conference IPC2012.