Current regulations for prediction and management of potential delayed failures from existing pipeline dents rely primarily on depth and conservative assumptions related to threat interactions, which have shown limited correlation with industry failures. Such miscorrelation can lead to challenges in managing effectiveness and efficiency of pipeline integrity programs. Leading integrity techniques that entail detailed assessment of complex dent features rely on the use of finite element analysis, which tends to be inefficient for managing large pipeline systems due to prohibitively complex modeling and analysis procedures. While efforts are underway to improve dent assessment models across the industry, these often require significant detailed information that might not be available to operators; moreover, they suffer scattered model error which makes them susceptible to unclear levels of conservatism (or non-conservatism). Nevertheless, most techniques/models are deterministic in nature and neglect the effect of both aleatory and epistemic uncertainties. Operators typically utilize conservative assumptions based on subject matter experts’ opinions when planning mitigation programs in order to account for different types of uncertainties associated with the problem. This leads to inefficient dig programs (associated with significant costs) while potentially leaving dents on the pipeline which cannot be quantitatively risk assessed using current approaches. To address these concerns, the problem calls for a dent assessment framework that balances accuracy with the ability to assess dent and threat integration features at a system-wide level with available information in a practical timeframe that aligns with other integrity programs.
This paper expands upon the authors’ previously published work regarding a fully quantitative reliability-based methodology for the assessment of dents interacting with stress risers. The proposed semi-quantitative reliability model leverages a strain-based limit state for plain dents (including uncertainty) with semi-quantitative factors used to account for complex geometry, stress riser interactions, and operating conditions. These factors are calibrated to reliability results from more detailed analysis and/or field findings in order to provide a simple, conservative, analytical-based ranking tool which can be used to identify features that may require more detailed assessment prior to mitigation. Initial validation results are provided alongside areas for continued development. The proposed model provides sufficient flexibility to allow it to be tailored/calibrated to reflect specific operator’s experience. The model allows for a consistent analysis of all types of dent features in a pipeline system in a short period of time to support prioritization of features while providing a base-level likelihood assessment to support calculation of risk. This novel development supports a dent management framework which includes multiple levels of analysis, using both deterministic and probabilistic techniques, to manage the threat of dents associated with stress risers across a pipeline system.