SBIR-STTR Award

In recent years, there have been a number of pipeline failures that have made national headlines. One such failure occurred in San Bruno, California in 2010 and resulted in 8 lives lost, 51 people injured and 38 homes destroyed. The catastrophic rupture of the pipeline created a 72 ft by 26 ft crater, and a 28 ft long, 3,000 lb piece of pipe was thrown 100 feet. The National Transportation Safety Board (NTSB) issued an incident report that determined the probable cause of the failure was the owner’s inadequate integrity management program. As of 2018, there were 301,227 miles of gas transmission pipelines in the United States, 20,435 miles of which are designated as being in high consequence areas. According to the Pipeline and Hazardous Materials Safety Administration (PHMSA), 56% of pipelines in the US were installed prior to the Minimum Federal Safety Standards being finalized in 1970. Pipelines constructed prior to 1970 were awarded grandfather status and were deemed exempt from the tighter controls placed on more modern pipelines. The pipeline that failed in San Bruno was one of these grandfathered pipelines. A crucial component of any effective integrity management program is the inspection and assessment of threats to the pipeline’s integrity. A typical pipeline inspection involves the use of an in-line inspection device that travels along inside the pipeline and takes a high resolution (millimeter-scale) ultrasonic scan of the pipe wall. This high resolution scan produces terabytes of data that is then interrogated to determine critical regions (threats) where damage may be present. These threats often number in the thousands. Each threat then undergoes an engineering critical assessment (ECA), which involves the numerical modeling and analysis of the damaged region. The large volume of data, combined with the potential for a large number of detected threats, directly lends itself to high-performance computing (HPC). The R&D group at The Equity Engineering Group will produce a user-friendly, HPC-accelerated platform for assessing the aging pipeline infrastructure in the United States. The platform will be web-based and will leverage high-performance cloud-computing resources. The proposed platform will contain the following five modules: (1) Data Import and Management – Upload, store and retrieve data from ILI inspection devices, along with other sources of data (e.g., streaming data from SCADA systems and field inspection data); (2) Screening Data Analysis – Process data using clustering algorithms combined with low level engineering assessments to identify and rank critical threats. Perform calculations using highly parallelized processes; (3) Advanced Engineering Analysis – Create advanced numerical models of the critical threats identified during the screening analysis and assess them using high performance cloud-computing resources; (4) Predictive Maintenance – Use the results from the advanced engineering analysis to enhance a high- performance Bayesian Decision Network to determine the timing of inspection, repair and replacement activities; and (5) Data and Results Visualization – Deliver the results in the web-based PipeSight platform, optimizing the presentation of the data based on the role of the user. Use principles from the business intelligence community to present analysis results in a hierarch