Bayesian corrosion assessment of unpiggable pipelines

点击翻译按钮获取中文摘要

Corrosion assessment of unpiggable oil and gas pipelines is a challenging task because such pipelines are inaccessible to inspection tools. Bayesian methodologies have been used to carry out corrosion assessment of unpiggable pipelines based on limited number of field survey data. However, the complexity of the formulation and computational techniques involved greatly hinders the application of these methodologies. The proposed project will employ the Bayesian network (BN) to develop methodologies and tools for updating the depth, length, density and growth of corrosion defects on unpiggable pipelines based on limited number of field survey data or inspection data from similar neighbouring pipelines. The focus of the research is to ensure that the developed methodologies and tools are theoretically sound as well as amenable for practical application. The simple visualization and ease to incorporate both observed data and expert judgements make BN particularly suited for the proposed research. The research will be carried out in collaboration with the industry partner, Lloyd's Register in Halifax, Nova Scotia, and will train one postdoctoral fellow for a duration of six months. The research outcome will facilitate the integrity management of unpiggable oil and gas pipelines and improve the safety, reliability, and economical viability of the vast oil and gas pipeline networks in Canada.