The research objective of this award is to develop methods for improving the diagnosis and prognosis of complex engineered systems. Real world systems are inherently subject to faults, and information on abnormal behavior of components and processes can be obtained through measured signals. The aim of this research is to exploit the information provided by sensor signals together with deep knowledge of the physical behavior of a complex system to create a unified framework for the systematic design of diagnostic and prognostic algorithms and for the prediction of system aging. Today, there are few, if any, general results concerning system prognosis and aging; the research, if successful, will result in methods that are general enough to be applied to engineered systems in different domains, for example automotive, aerospace, chemical processes and manufacturing. Deliverables for this project include software for the design and analysis of diagnostic algorithms, experimental data, mathematical models, and educational materials for engineering students.
This GOALI project is conducted in close collaboration with a major automotive manufacturer, General Motors, and aims to apply the new methodology to the diagnosis of automotive electrical systems and to the prognosis and life prediction of automotive batteries. If successful, the research will result in significantly greater reliability of future automotive electrical and energy storage systems. This may eventually result in reduced warranty costs and increased customer satisfaction, with the potential of touching millions of automobile owners around the world. The research may also have a positive impact on accelerating the introduction of advanced electrical and energy storage systems in future electric and hybrid automobiles by increasing their reliability and serviceability. Further, the educational outcomes of the project will affect undergraduate and graduate engineering students at two major universities as well as practicing engineers in the automotive industry through distance education programs.
The supervision of complex systems and processes requires the presence of a variety of diagnostic and prognostic functions to insure three important system properties: 1) availability – that is the ability to retain critical system functionality at all times; 2) safety – that is the guarantee that under faulty conditions the system will continue to operate in a safe manner, or in compliance with existing regulations; and 3) serviceability, that is the assurance that the system can be rapidly serviced or re-configured in the event of one or more malfunctions; this latter property is intimately connected to the subject of system prognosis. This project, a GOALI collaboration between the PIs at The Ohio State University (OSU) and Clemson University (CU) and General Motors Corporation (GM), has resulted in two major contributions: i) the ability to improve the detection of faults in uncertain systems through the use of adaptive detection thresholds that account for uncertainty in system model parameters; and ii) a formal understanding of aging that allow for model-based prognosis of system health. The results of the research have been applied to two automotive systems. 1. The adaptive detection threshold ideas were applied to an automotive electrical power generation and storage system, and in particular to an automotive alternator. Experimental validation of the approach was made possible by the collaboration with GM. 2. Aging models were developed for high voltage lithium-ion batteries such as are used in hybrid and electric vehicles, and it was experimentally demonstrated that such aging models can assist in state-of-health determination and in model-based prognosis of advanced batteries. The results of this research will benefit service technicians at the repair shops and dealerships by improving early diagnosis of problems and their resolution, and will also benefit automotive OEMs in improving product quality, reducing warranty, increasing revenue, and ultimately enhancing customer satisfaction. In summary, the outcome of this research has the potential of having a positive impact on millions of consumers, and also of improving the competitiveness of the largest U.S. automotive manufacturer. Further, this project has had a significant educational impact: during the course of the project one PhD student at OSU completed his degree, a second PhD student at OSU (a Hispanic female) as well as a third PhD student at CU reached the PhD candidacy stage, two MS student completed a thesis, an undergraduate student completed a BS Honors thesis thanks to a Research Experience for Undergraduates supplement to the grant, and three visiting scholars (graduate students) from Italy participated in various aspects of the project. Finally, the project also served as an opportunity to further the training of a former post-doctoral scholar (a female), who during the course of this project, in which she served as a senior investigator, was awarded principal investigator status by OSU and received a promotion to the position of Research Scientist.
