Fault detection for systems with multiple unknown modes and similar units

This dissertation considers fault detection for large-scale practical systems with many nearly identical units operating in a shared environment.;A special class of hybrid system model is introduced to describe such multi-unit systems, and a general approach for estimation and change detection is proposed. A novel fault detection algorithm is developed based on estimating a common Gaussian-mixture distribution for unit parameters whereby observations are mapped into a common parameter-space and clusters are then identified corresponding to different modes of operation via the Expectation-Maximization algorithm. The estimated common distribution incorporates and generalizes information from all units and is utilized for fault detection in each individual unit.;The proposed algorithm takes into account unit mode switching, parameter drift, and can handle sudden, incipient, and preexisting faults. It can be applied to fault detection in various industrial, chemical, or manufacturing processes, sensor networks, and others. Several illustrative examples are presented, and a discussion on the pros and cons of the proposed methodology is provided.;The proposed algorithm is applied specifically to fault detection in Heating Ventilation and Air Conditioning (HVAC) systems. Reliable and timely fault detection is a significant (and still open) practical problem in the HVAC industry---commercial buildings waste an estimated 15% to 30% (;Results are presented from an extensive performance study based on both Monte Carlo simulations as well as real data collected from three operational large HVAC systems. The results demonstrate the capabilities of the new methodology in a more realistic setting and provide insights that can facilitate the design and implementation of practical fault detection for systems of similar type in other industrial applications.