|dc.description.abstract||This dissertation consists of four published or accepted journal articles that address some of the key problems in prognostics and health management area (PHM). Effective health diagnostics and prognostics provide multifarious benefits such as improved safety, improved reliability and reduced costs for operation and maintenance (O&M) of complex engineered systems. Extensive literature reviews on PHM for diagnostics of system health conditions and O&M decision-making for complex engineered systems have identified important challenge problems for this dissertation as follows:
- Effective diagnostics of current health states based on heterogeneous sensory data from multiple sensors is an intricate problem for condition monitoring techniques to be applied on complex engineered systems, mainly due to high system complexity and sensory data heterogeneity;
- With an increasing system complexity, it is extraordinarily difficult to have a well-tested system so that all potential faulty states can be realized and studied at product testing stage. Thus, real time health diagnostics requires automatic detection of unexampled system faulty states based upon sensory data to avoid sudden catastrophic system failures;
- Despite successful applications of different diagnostic algorithms in various engineering fields, a challenge for health diagnostics is that an implicit relationship between different system health states and features of sensory signals makes it difficult to develop a generally applicable health diagnostics technique.
- Although diagnostics and prognostics can provide valuable information for proactive actions in preventing system failures, their benefits have not been fully utilized for the O&M decision-making process.
To carefully address these important research problems, this dissertation proposes four research solutions: a multi-sensor health diagnostics technique using deep belief network, a tri-fold hybrid classification approach for diagnostics with unexampled faulty states, a multi-attribute classification fusion technique to develop a generally applicable health diagnostics framework and a generic prognostics-informed O&M decision-making framework by utilizing failure prediction information in the O&M decision-making process. In this dissertation, different practical engineering applications will be employed as case studies to demonstrate the efficacy of proposed research solutions.||