Optimization models for evidence-based reliability and maintenance decisions

石油和天然气行业是加拿大经济的关键驱动力。加拿大目前是世界第六大石油生产国，并有望在未来20年跻身世界前三名。然而，石油和天然气管道等埋在地下的资产的安全做法不佳，可能会对环境、经济和社会产生重大负面影响。例如，2010年密歇根州的Enbridge漏油事件是美国中西部地区历史上最严重的环境灾难。这类备受瞩目的事故的主要原因是几十年前修建的管道老化，没有得到适当的维护以延长设计的使用寿命，或者没有进行预防性更换。管道基础设施的主动完整性管理对于防止死亡和对环境的巨大破坏至关重要。这一研究项目将通过基于现实假设的循证科学方法的发展，为管道完整性管理做出贡献，但又不会过于复杂，无法在实践中应用。更具体地说，我们将开发量化风险评估模型，帮助管道管理人员确定管道段的优先顺序，并将维护资源应用于最需要的地方。不必要和过度的预防性维护也可能造成损失，同样，维护水平不足也可能造成损失。此外，我们将开发概率方法来估计暴露在环境多样性和不同负荷下的管道的可靠性，并预测其剩余使用寿命。这些模型有助于在发生之前预测管道段的失效，以便采取维修或更换等预防措施，从而显著降低对环境的负面影响和维护成本。最后，我们将构建管道检测和维护优化模型，以最低的运行维护费用最小化故障风险。这些模型允许确定最佳的更换/维修策略，该策略平衡了检查和维护的成本与故障的成本和风险。这项研究提案的结果预计将为加拿大社会带来环境、经济和社会效益。

The oil and gas industry is a key driver of Canada's economy. Canada is currently the world's 6th largest oil producer and is poised to become one of the top three in the world over the next two decades. However, poor safety practices with regard to buried assets such as oil and gas pipelines can have a significant negative impact on the environment, economy, and society. For example, the 2010 Enbridge oil spill in Michigan was the worst environmental disaster in the history of the American Midwest. The main causes of such high-profile accidents are aging pipelines constructed decades ago, that are not properly maintained to extend their designed useful life, or preventively replaced. Proactive integrity management of pipeline infrastructure is essential to the prevention of fatalities and immense damage to the environment. This research program will contribute to pipeline integrity management through the development of evidence-based scientific methods based on realistic assumption, yet be not too complicated to be applied in practice. More specifically, we will develop quantitative risk assessment models, which can help pipeline managers prioritize their pipeline segments and apply their maintenance resources where they are most needed. Unnecessary and excessive preventive maintenance could be also loss-making likewise inadequate level of maintenance. Moreover, we will develop probabilistic methods to estimate the reliability of pipelines exposed to environmental diversity and varying loads and to predict their remaining useful life. These models help to predict the failure of pipeline segments before occurrence, so that a preventive action such as repair or replacement can be taken, which should result in significantly lower negative environmental impact and maintenance cost. Finally, we will construct inspection and maintenance optimization models for pipelines to minimize failure risk at the lowest operating and maintenance expenses. These models allow the determination of an optimal replacement/repair policy, which balances the cost of inspections and maintenance with the cost and risk of failures. The results of this research proposal is expected to deliver environmental, economic, and social benefits to Canadian society.