Since 2003, I have concerned myself with thinking strategically about pipeline failures. I spent 8 years developing software for pipeline integrity management. I quickly became dissatisfied with the lack of predictive power in integrity models. Failures, it appears, are due to outliers. The best models will penalize the general case while missing these extreme outliers. Failures will occur no matter how conservative you make your integrity models or how conservatively you calculate your response times after an inspection. I am convinced that the best way to minimize seemingly random pipeline failures is direct and continuous monitoring.
Important Pages
These links point to pages that will orient you to the concepts and motivation for this site.
acK for pipeline integrity 
Leave a comment