Optimization models for evidence-based reliability and maintenance decisions

石油和天然气行业是加拿大经济的主要驱动力。加拿大目前是世界第六大石油生产国，并有望在未来20年内成为世界前三大石油生产国之一。然而，对于石油和天然气管道等地下资产来说，糟糕的安全措施可能会对环境、经济和社会产生重大的负面影响。例如，2010年密歇根恩布里奇石油泄漏事件是美国中西部历史上最严重的环境灾难。这些引人注目的事故的主要原因是几十年前建造的管道老化，没有适当维护以延长其设计使用寿命，或者预防性更换。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。该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，同时又不会过于复杂而无法在实践中应用。更具体地说，我们将开发定量风险评估模型，这可以帮助管道管理者优先考虑他们的管道段，并将他们的维护资源应用到最需要的地方。不必要和过度的预防性维修也可能造成损失，同样，维修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管道暴露在环境多样性和变化负荷下的可靠性，并预测其剩余使用寿命。这些模型有助于在管道段发生故障之前进行预测，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，我们将构建管道的检测和维修优化模型，以最低的运维费用将故障风险最小化。这些模型允许确定最佳更换/维修策略，从而平衡检查和维护成本与故障成本和风险。这项研究计划的结果有望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.