Reliability Assurance Methodology Incorporating Advanced Diagnostics and Prognostics

今天的社会依赖于电网、电信网络、管道以及空中、地面和水上交通系统等庞大工程系统的可靠、不间断运行。尽管可靠性保证方法在过去70年中取得了重要进展，但影响数百万人的事故和重大服务中断仍然频繁发生。随着设备的使用而发生劣化，材料性能和退化行为存在固有的不确定性，使得传统的基于时间的可靠性保证方法显得力不从心。可靠性研究界最近的努力是将状态监测、高级诊断和预测纳入到复杂工程系统的在线可靠性保证中。本研究旨在通过在齿轮箱试验台上进行动态模拟和实验室实验，开发有效的设备劣化健康指标，开发信息融合方法来推断设备隐藏的健康状态，利用健康指标和推断的健康状态信息估计运行系统的剩余使用寿命分布，根据部件状态分布和部件之间的依赖关系评估系统健康状态分布，并开发动态维护优化方法，以确保工程系统的安全、可靠和低成本运行。这项研究将为探测和识别潜在故障模式、评估多状态系统的可靠性、预测运行中设备的剩余使用寿命和工程系统的动态运维决策提供基本原理、有效算法和实用模型。将产生的结果在管道、石化炼油厂、发电机、电信网络、电网和运输车辆等现代工程系统的设计、制造和运营中具有极好的应用潜力。这项研究的结果可以应用于工业环境中，直接造福于加拿大的工业。将为加拿大的工业培训高素质的人员。

Today’s society depends on the reliable, uninterrupted operation of colossal engineering systems such as power grids, telecommunication networks, pipelines, and air, ground, and water transportation systems. Despite important advancements in reliability assurance methodologies over the past 70 years, accidents and major service disruptions affecting millions of people still occur frequently. The fact that deterioration occurs as a device is used and inherent uncertainty exists in material properties and degradation behaviors makes traditional time-based reliability assurance methods inadequate. Recent thrusts of the reliability research community are to incorporate condition monitoring and advanced diagnostics and prognostics into online reliability assurance of sophisticated engineering systems. This research program aims to develop effective health indicators of equipment deterioration by conducting dynamic simulation and laboratory experiments on a gearbox test rig, develop information fusion approaches to make inferences on the hidden health state of the equipment, estimate the remaining useful life distribution of a running system utilizing information on health indicators and inferred health states, evaluate system health state distribution based on component state distributions and dependency relationships among components, and develop dynamic maintenance optimization approaches to ensure safe, reliable and cost-effective operation of engineering systems. The proposed research will generate fundamental theorems, efficient algorithms, and practical models for detection and identification of impending failure modes, for reliability evaluation of multi-state systems, for prediction of the remaining useful life of running equipment, and for dynamic operation and maintenance decision making of engineering systems. The results to be generated have excellent potential for applications in design, manufacture, and operation of modern engineering systems such as pipelines, petro-chemical refineries, power generators, telecommunication networks, power grids, and transportation vehicles. The results of this research can be applied in industrial settings to benefit the Canadian industries directly. Highly qualified personnel will be trained for the Canadian industries.