Fault diagnosis in HVAC chillers using data-driven techniques

Failures in HVAC systems occur frequently and lead to loss of comfort, degradation in operational efficiency, and increased wear and tear on the system equipment. Faulty HVAC systems seriously affect the energy efficiency of commercial buildings; they are oftentimes the causes for exceeding the allocated demand margins resulting in steep monetary penalties. A real-time fault detection and isolation (FDI) system can ensure uninterrupted and energy-efficient operation of the HVAC systems, and thus enhance the quality of service in modern buildings. In this paper, we propose a data-driven approach for real-time fault detection and isolation (FDI) in the chillers in HVAC systems. Our techniques diagnose a number of faults belonging to both gradual degradation and abrupt fault classes.