Research on the PCA-based Intelligent Fault Detection Methodology for Sewage Source Heat Pump System

Abstract As an energy-saving and environmentally friendly technology, sewage source heat pump is rapidly developing to be widely implemented in urban distributed energy system. However, in the lifecycle, the faulty operation of sewage source heat pump system (SSHPs) results in energy wastage, low efficiency, system unreliability and shorter equipment life. A novel PCA-based fault detection methodology is proposed for fault detection of SSHPs. Analyzing of the square prediction error (SPE) and Hotelling’s T2 principal component analysis (PCA) is a valid intelligent detection method to figure out faulty operation. Approximately 4,800 records of running state parameters was investigated from a 500,000 m2 sewage source heat pump central heating system of community building in Xi’an, China. The PCA-based fault detection model of the SSHPs was established, and the SPE and Hotelling’s T2 was calculated and analyzed. The analysis results presented that the PCA-based fault detection methodology can distinguish anomalies from normal operation, and availably detect the faults of SSHPs. Hence, the PCA-based fault detection methodology can ensure the healthy and efficient operation of SSHPs, and improve the application prospect of the system.