GOALI: Adaptive Degradation-Based Prognosis with Application to Vehicular Electrical Systems

这项学术与工业联络资助机会（GOALI）奖的研究目标是使用基于实时性能/状态的传感器信号来表征多组件系统中组件之间的退化相互作用，并使用这种表征来改进复杂工程系统的基于传感器的预测。传统上，用于在给定系统中对组件相互依赖进行建模的一般方法侧重于调查组件故障对剩余幸存组件的影响。相比之下，这个项目的方法通过关注系统组件逐渐和部分退化的影响，而不是它们失败的影响，在更基本的层面上解决了这个挑战。这将通过结合随机和统计建模框架来实现，该框架将用于开发具有相互依存降解过程的组件的自适应预测模型。该项目的可交付成果包括具有预测算法的软件、研究结果文档、工业平台验证和工程学生教育。这个项目的成功将对提高人类安全，降低美国汽车工业的维护和保修成本产生直接影响。具体来说，GOALI项目是与通用汽车公司密切合作进行的。目的是将预测方法应用于车辆发电与储能系统的关键部件。这项研究的结果也将使许多其他工业部门受益，包括航空业、发电行业、制造业和服务业领域。该研究议程将作为两名博士生博士论文的基础，为他们提供丰富的理论与实际工作相结合的培训经验和实习机会。研究结果将被纳入预测研究生课程。传播将包括向学术界的会议报告以及工业研讨会和讲习班。

The research objective of this Grant Opportunity for Academic Liaison with Industry (GOALI) award is to use real-time performance-/condition-based sensor signals to characterize component-to-component degradation interactions in multi-component systems, and use this characterization to improve sensor-based prognostics of complex engineering systems. The general approach used for modeling component interdependencies within a given system has traditionally focused on investigating the effects that a component's failure has on the remaining surviving components. In contrast, this project's approach addresses this challenge at a much more fundamental level by focusing on the effects of gradual and partial degradation of system components rather than the effects of their failures. This will be achieved through a combined stochastic and statistical modeling framework, which will be used to develop adaptive prognostic models for components with interdependent degradation processes. Deliverables for this project include a software with prognostic algorithms, documentation of research results, validation on industrial platform, and engineering student education. The success of this project will have a direct impact on improving human safety, and reducing maintenance and warranty costs of the American automotive industry. Specifically, this GOALI project is conducted in close collaboration with General Motors. It aims to apply the prognostic methods on key components of the Vehicular Electric Power Generation and Storage (EPGS) system. Findings of this research will also benefit many other industrial sectors, including the airline industry, the power generation industry, manufacturing sector, and domains of the service sector. The research agenda will serve as the foundation for the doctoral dissertations of two Ph.D. students, providing them with a rich training experience that combines theoretical and industrial work as well as internship opportunities. Research findings will be incorporated in a Prognostics graduate course. Dissemination will include conference presentations to the academic communities as well as industrial seminars and workshops.