Research On Life Prediction Of Avionics Air Conditioning System Based On Multi-feature

According to the actual demand of avionics air conditioning system of a certain airline A320 fleet,a method of life prediction of avionics system based on multi-feature is proposed.Through the investigation of the existing methods of life prediction of avionics air conditioning system,this study proposed the reason why existing methods cannot be used on a large-scale is that there is lack of prior knowledge and abnormal data classification difficult problem.In view of the existing problems,a degradation assessment algorithm for air conditioning system of passenger aircraft based on improved Fuzzy-C Means algorithm was proposed.This algorithm evaluates the gap between normal and abnormal state of the aircraft air conditioning data through the distance-based method to obtain the state degradation quantity,which effectively solves the problem of insufficient prior knowledge in other multi-feature datadriven methods.The algorithm uses the optimal fuzzy unsupervised clustering algorithm.It solves the problem that it is difficult to classify abnormal data into the same category when dealing with complex systems such as aircraft air conditioning systems.The introduction of contrast about air conditioning components as one of FCM algorithm distance evaluation parameters,effectively improve the sensitivity of the model,timely find air conditioning system's early performance degradation.Finally,the feasibility of the algorithm is verified by comparison.This algorithm greatly improves the effect of the model on the "descending state" in the process of system performance degradation,improves the rationality of the model and the evaluation effect,greatly increases the application value of the algorithm in the actual maintenance work,and provides technical guidance for the realization of predictive maintenance.