Comparison of Fatigue Models in API 1183 for Predicting Fatigue Life of Pipeline Dents

Abstract

Pipeline dents are mechanical damage caused by a third party in buried pipelines during construction or maintenance repair. Since 2008, the Pipeline Research Council International (PRCI) has sponsored a large pipeline dent program on experimental tests, integrity assessment, and inline inspection of pipeline dents with and without interfacing other anomalies. For more than ten years, many experimental data, numerical results and assessment models have been developed. On this basis, American Institute of Petroleum (API) recently published a recommended practice (RP) 1183 – Assessment and Management of Pipeline Dents. This API code provides the pipeline industry “standard methods” for evaluating severity and predicting fatigue life of pipeline dents with a single peak. This paper discusses fatigue methods prescribed in API RP 1183 for predicting fatigue life of pipeline dents. This API standard provides two types of dent fatigue criteria. One is screening methods based on dent fatigue life in Sections 7.4.1, 7.4.2, and 7.4.3 of API 1183. The other is assessment methods for predicting fatigue life of dents in Section 8.3.4. Because of complexity of fatigue methods and math equations given in the content, Annexes A.1 to A.5 in API 1183 provide a set of standard examples for calculating fatigue life, determining restraint condition, and evaluating the screening and assessment methods for plain dents. However, after careful examination, it was found that many issues and errors exist in this new API standard. In particular, comparisons using these Annex examples in API 1183 showed self-inconsistence between the screening methods and the assessment method. This paper presents the detailed fatigue life calculations, comparisons, and self-inconsistent results of the dent screening and assessment methods given in API 1183. On this basis, recommendations are made on how operators should use API RP 1183, and what actions API should take to improve this code.