COMPUTATIONAL INTELLIGENCE IN DIAGNOSTIC AND PROGNOSTIC APPLICATIONS

Abstract

The main focus of this thesis is to develop advanced diagnostics and prognostics algorithms and framework for industrial applications. Both diagnostics and prognostics are the key enablers of Condition-Based Maintenance (CBM) and health monitoring system which has been gaining in popularity. The benefits and motivations for CBM and health monitoring system, diagnostics and prognostics have been well documented in the literature. Although considerable research has been devoted to the development and improvement of diagnostics and prognostics, there exist several open issues in fault diagnosis and remaining useful life estimation problems. This thesis, therefore, aims to develop new diagnostic and prognostic algorithms and framework to overcome limitations in existing approaches. In this thesis, novel solutions addressing the gaps in imbalance learning, deep learning, ensemble learning, and multi-state framework are proposed. The proposed approaches are analyzed and evaluated its suitability for not only benchmark datasets but also real-world applications.