Reducing Pipeline Failures in a Complex System Using Statistical Methods

In 2006, pipeline failures in the Swan Hill Unit #1 field, in Northern Alberta, Canada, had increased to a point where they were negatively impacting operating costs, production, and the company’s standing with the provincial regulator.  The size and complexity of the field infrastructure, as well as the continuous reactive response to failures, made it difficult to focus efforts and prioritize work in the field.  With over 600 pipeline segments, and 1000 km of pipe in the ground, a field specific assessment, inspection, and replacement strategy was necessary.  The initial step in this process was a systematic review of the past 40 years of failure and operational data.  Statistical methods were then used to determine which risk factors should be used to prioritize the pipelines for immediate and future attention.  A field specific risk assessment process was developed which allowed for targeted inspections and pipeline replacements.

This paper presents the results of the statistical analysis, how this information was fed into a dynamic risk assessment model, the ongoing efforts to improve the accuracy of the model, and the steps that were taken to reduce the number of failures to a third of their 2006 magnitude.