Sensor Fault Diagnosis of Air Conditioning System Based on Multi-dimension Clustering Algorithm

With the development of modern architecture, indoor refrigeration and air conditioning system has become more and more complex. In these complex systems, the ability to accurately and timely detect and diagnose multiple faults is beyond human ability, so fault detection and diagnosis becomes more and more important. The purpose of this paper is to study sensor fault diagnosis of air conditioning system based on multi - dimensional clustering algorithm. This paper describes the basic theory of clustering algorithm and the basic process of sensor fault diagnosis in air conditioning system. Taking the laboratory air conditioning system as the experimental data object, multi-condition modeling data were identified and classified by multi-dimension Gaussian mixture model EM clustering algorithm. By selecting TPR statistics as the test measure, the feasibility and rationality of the strategy are verified.