Integrity reliability analysis is becoming an important component of effective pipeline integrity management systems. It aims at utilizing reliability engineering to address integrity uncertainties and check pipeline reliability measures against safety objectives/targets. In current practice, pipeline safety is typically verified using simplified deterministic procedures based on a safety factor approach that is tailored to the design of new pipes. A more realistic verification of actual safety performance of existing pipelines can be achieved by probabilistic methods where uncertainties of basic random variables are considered and the impact on the reliability of the system is analyzed. To enable such an approach, specification of integrity target reliability levels is required in order to benchmark the safety level of an existing pipeline system. The probability of failure (PoF) per pipeline segment or unit length is quantified and then checked against an integrity permissible probability of failure (PoFp) or integrity target reliability (1-PoFp). This check against a specified reliability target allows the operator to confidently determine whether a segment of pipe is safe at current operating conditions while considering identified uncertainties. However, the main challenge around reliability targets is choosing such targets to begin with.
This paper presents a semi-quantitative validation approach for estimating integrity reliability targets based on calibrating past failure incidents and evaluating PoF at the time of failure. Accounting for both aleatory and epistemic uncertainties in assigning the integrity targets, pipeline operators can gauge how to choose such targets and how to be flexible in terms of customizing integrity targets based on their asset performance and adopted integrity programs. A brief summary of published reliability targets in pipeline and non-pipeline industries is presented herein.
