PR244-243907-Z01 Analytical Prediction of Leak Events and Detection Using Historical Data

Unintentional methane releases from natural gas pipelines introduce environmental, health and safety, financial, and reputational risks for pipeline operators. The need for swift leak detection is essential to mitigate these risks. This work builds actionable insights to improve gas pipeline leak detection through the analysis of historical leak incident information. A review of commercially available leak detection techniques was performed to identify candidates to fill potential technology gaps. Leak incident records were collected across various sources and then collated, cleaned, and evaluated for analysis. The collated dataset was analyzed to identify relationships among pipeline, leak, and detectability characteristics for the development of key insights. A predictive model based on these insights was developed to quantitively assess various leak detection approaches. This document aims to provide operators with techniques and insights to optimize leak detection resources through an increased understanding of the contributing factors that lead to high priority leaks. The collated dataset along with insights and the predictive model is delivered for operator use. The zipped deliverable contains the following two items: 1. PR244-243907-R01 Analytical Prediction of Leak Events and Detection Using Historical Data Final Report 2. PR244-243907-E01 Collated Data Deliverable