Improving the Performance of a Pipeline Crack Detection Embedded System

点击翻译按钮获取中文摘要

Petroleum products produced in Alberta need to be safely and efficiently transported to other markets for purposes such as refining. Pipelines are becoming an increasingly preferred solution since they pose lesser direct risk to human lives than other modes such as rail. It is crucial however that pipeline operators detect problems such as leaks proactively and fix them before they cause environmental damage. Typically, pipeline operators use embedded computer systems called smart pigs to detect minute imperfections, which could potentially lead to leaks. Encompass Inspections Canada is a company that has developed a state-of-the-art pig system and successfully deployed it for major clients such as Shell and Enbridge. Typically, the pig travels a section of the pipeline collecting information about the contours of the segment. This data is stored internally in the pig's memory. After traversing the segment, the device is collected and its data is transferred to a field laptop. Visualizing software on the laptop is used to provide graphical views of the location and dimensions of pipeline imperfections. Several technical challenges need to be addressed in the long-term to improve the efficacy of the system. In particular, the visualization software currently does not provide real-time renderings of the data for long pipeline segments. Analysts have to wait for a long time thereby compromising the agility of the process. This proposal aims to initiate research that focuses on optimizing the performance of the entire system to redress such problems. For example, one solution is to compress the amount of data transmitted by the pig since typically most of the data pertains to portions of the pipeline that are unlikely to develop serious cracks. This can reduce the time needed to transfer the data to the laptop and also the amount of data handled by the visualization software. Another complementary solution is to use parallel processing techniques, e.g., processing using the laptop's Graphic Processor Units (GPUs), to speed up data visualization. Outcomes of long term research in this area is likely to improve pipeline safety thereby benefitting the environment as well as the economy of Canada.