Fault diagnosis in HVAC chillers

In this article, we consider a data-driven approach for fault detection and isolation (FDI) of chillers in HVAC systems. To diagnose the faults of interest in the chiller, we employ multiway dynamic principal component analysis (MPCA), multiway partial least squares (MPLS), and support vector machines (SVMs). The simulation of a chiller under various fault conditions is conducted using a standard chiller simulator from the American Society of Heating, Refrigerating, and Air-Conditioning Engineers (ASHRAE). We validated our FDI scheme using experimental data obtained from different types of chiller faults.