Fault Detection of an Air Handling Unit Using PCA Method

This paper uses principal component analysis (PCA) method for detection of faults in a five zone air handling unit (AHU). PCA approach is used to extract the correlation of variables in air handling unit and reduce the dimension of measured data. A PCA model is built to determine the thresholds of statistics and calculate square prediction errors (SPE) of new observations, which are used to check if a fault occurs in AHU. The fault detection system consists of a PCAbased fault detection code, a building management system (BMS) with three temperature sensors, control valve and output air controller signals. Allexperiments were carried out on air conditioning systemof the RAZAVI hospital in Mashhad. Five faults artificially introduced to the system, are as follows: (1) Hot air temperature,(2) Room temperature sensors,(3) Return air damper stuck, (4) Outdoor damper stuck, (5) Control valve damper stuck, signals. The results show that the PCA-based fault detection method is applicable and effective for AHU.