Chiller Unit Fault Detection and Diagnosis Based on Fuzzy Inference System

A serials of adverse effects will occur after the chillers units´ production of faults. The chiller units are strong non-linear and long-time delay system, so there are existences of fault detection and diagnosis (FDD) limitation if adoption regular fault judgment method. Fuzzy inference system (FIS) has an outstanding control effect because it is close to human thoughts. Moreover, the result of FIS is easy to be understood by operators. Firstly, fuzzy inference system of chillers unit was modelled in the paper, and then the FIS was trained and checked by the operation data obtained from a real building chiller unit under normal and unhealthy conditions respectively. The results indicated the chillers unit FDD based on FIS is feasible and prompt. Furthermore, this method is simple and prone to be automated.