Research Of Sensor Fault Diagnosis Method In Constant Air-volume Air-conditioning Unit Based On PCA

Sensor fault is one of the typical faults in Air-conditioning system. The fault sensor will offer the system inaccurate measurements, which would mislead the control system to do wrong actions, thus causing the energy waste, cost increase and decline of indoor environment quality. Therefore, the precision and reliability of sensors is very important for control and monitoring of Air-conditioning system. Researches were focused on sensor fault detection in constant air-volume (CAV) Air-conditioning system and the main contents of the research include:1. Summarizing the existent methods of Air-conditioning system's sensor fault diagnosis and analyzing the superiority and deficiency of each method. Based on the close relativity among the sensors of CAV Air-conditioning system, principal component analysis (PCA) method was proposed to detect faults of temperature sensors and humidity sensors.2. Making study on all types of Air-conditioning system sensor faults, especially the precision grade decrease fault.3. Deeply analyzing the advantages of PCA method and discussing the principles of PCA to improve the method by selecting more approporiate data matrix and more suitable way of unitizing data dimensions.4. Setting up the simplified model of CAV Air-conditioning system to explore its energy relations based on enthalpy analysis and establishing PCA model according to such balance of the system.5. Colleting the actual running data from Tianjin museum building control system, analyzing and pre-disposing the data.6. Validating the fault diagnosis and recovery ability of established PCA model in the CAV Air-conditioning system. The simulation testified that PCA model based on the enthalpy analysis had good fault detection and recovery effect.In the paper, PCA model was set up according to the enthalpy analysis in the constant air volume Air-conditioning system to solve the problem of sensor fault diagnosis and the simulation results proved that the model has satisfying fault diagnosis ability. The subject is very meaningful to perfect sensor fault diagnosis methods in Air-conditioning system.