Fault Detection And Diagnosis Method For Air Handling Unit Of Variable Air Volume Air-conditioning System

Variable air volume(VAV)air-conditioning system is an emerging type of air-conditioning system which is widely used in large and medium-sized construction projects in China.Compared with constant air volume(CAV)system,VAV system has higher fault possibility.Therefore,research on fault detection and diagnosis(FDD)of VAV air-conditioning system can discover the fault in time,which has practical significance to ensure the stable operation of system and save energy.Most modern buildings are equipped with energy management system(EMS),which can collect real-time data of the air-conditioning system.This provides the rich information base and convenient conditions for the research of FDD.The air handling unit(AHU)of VAV air-conditioning system is the research object of this paper.Firstly,the working principle and every functional unit of AHU are briefly described.The full-scale intelligent experimental platform of central air-conditioning system of Chang'an University is introduced.The running data of the normal operation is collected based on the experimental platform.Then,the model is established by principle algorithm of principal component analysis(PCA)with the measurement of normal operation,and four groups of tests of common faults of AHU are designed.The method of AHU fault detection and diagnosis is presented with the SPE statistics and Q contribution rate as the test standards.The results prove the validity of PCA method and show that the detection efficiency and the accuracy of FDD are higher for serious faults.However,the FDD capability of small faults is insufficient.In order to improve the results of FDD,the VAV air-conditioning system is modeled in TRNSYS,and four typical faults of AHU are simulated.The comparative analysis of the respective under normal operation and fault operations have been made,and the corresponding fault qualitative characteristics are obtained.On the base of these characteristics,the identification patterns of AHU fault are formed,and these patterns can be used to diagnose faults.Applying this pattern to diagnose and analyze,the correct diagnosis results were obtained.It was proved that pattern recognition is a better method for diagnosing faults.The results of study prove the FDD method combined with PCA and pattern recognition reliable.It can detect and diagnose possible faults of AHU.Further more,it can perform the prejudgment of faults and guide the monitoring personnel to repair the fault.Finally it makes AHU run in the normal condition which contributes to ensure the thermal comfort and reduce the energy consumption.