Control, estimation and prognostic for partially observable deteriorating systems with multivariate observations

通过多变量观测对部分可观测的恶化系统进行控制、估计和预测

• 批准号: 121384-2011
• 负责人: Makis, Viliam
• 金额: $ 1.97万
• 依托单位: University of Toronto
• 依托单位国家: 加拿大
• 项目类别: Discovery Grants Program - Individual
• 财政年份: 2012
• 资助国家: 加拿大
• 起止时间: 2012-01-01 至 2013-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=509215
• 关键词: Control; estimation; prognostic; partially; observable

The objective of the proposed research is to develop the optimal control, estimation, and prognostic procedures for models of partially observable deteriorating systems subject to random failure, considering both single failure mode as well as a combination of degradation and sudden failure (competing risks). The models will be formulated and developed with practical applications in mind, to represent real systems as well as possible. Condition monitoring is currently widely used both in industries and medicine to prevent the occurrence of high degradation or failure, both of which have serious consequences. In technical applications, these are mainly the cost or safety consequences, in medical applications, the high cost is combined with health issues. In most cases, diagnosis of a disease in an early stage results in a better prognosis of successful cure at lower cost. Condition monitoring and screening procedures currently in use have not been designed optimally or even sub-optimally. They are widely used and more such procedures will soon be implemented especially in the healthcare sector worldwide, but also in many larger companies mainly in industries where stopping, full inspection, and testing is very expensive. Proper system modeling and the optimal design of such procedures, which is the focus of the proposed research is expected to bring huge economic as well as health benefits. The proposed research is highly novel and unique in the world. The results obtained so far using simpler system modeling are very promising and considerably better results are expected with the proposed more advanced modeling and better utilization of multivariate data for estimation, prognostic, as well as optimal decision-making. The Principal Investigator has extensive experience with modeling, estimation and control of deteriorating systems as well as practical experience acquired through collaboration with the University of Toronto CBM Consortium companies and many other companies for many years. He will continue publishing the unique results of his research in the top international journals and it is expected that results obtained in the proposed research will contribute very significantly to the further development of the discipline.

本研究的目的是为遭受随机故障的部分可观察退化系统模型开发最优控制、估计和预测程序，同时考虑单一故障模式以及退化和突然故障（竞争风险）的组合。这些模型的制定和开发将考虑到实际应用，以尽可能地代表真实的系统。状态监测目前广泛应用于工业和医学领域，以防止严重退化或失效的发生，这两种情况都会带来严重的后果。在技​​术应用中，这些主要是成本或安全后果，在医疗应用中，高成本与健康问题相结合。在大多数情况下，疾病的早期诊断可以带来更好的预后，并以较低的成本成功治愈。 目前使用的状态监测和筛选程序的设计不是最佳的，甚至是次优的。它们被广泛使用，并且很快将实施更多此类程序，特别是在全球医疗保健行业，而且在许多大型公司（主要是停止、全面检查和测试非常昂贵的行业）中也是如此。 适当的系统建模和此类程序的优化设计是本研究的重点，预计将带来巨大的经济和健康效益。拟议的研究在世界上非常新颖和独特。迄今为止，使用更简单的系统建模获得的结果非常有希望，并且通过提出的更先进的建模和更好地利用多变量数据进行估计、预后以及最佳决策，预计会获得更好的结果。首席研究员在恶化系统的建模、估计和控制方面拥有丰富的经验，以及通过与多伦多大学煤层气联盟公司和许多其他公司多年合作获得的实践经验。他将继续在国际顶级期刊上发表其独特的研究成果，预计拟议研究中取得的成果将为该学科的进一步发展做出重大贡献。