Methods Research On The Key System Of Aeroengine Based On Intelligent Fault Diagnosis

Aero-engine has been considered as the "heart" of airplane for a long time,whether it works directly affects the safety of aircraft flight.But because of the influence of the aero-engine working environment(high temperature,high pressure and large stress),it is difficult to ensure the reliability when using it.In order to diagnose aero-engine fault timely,accurately and efficiently for ensuring the safety of the aircraft flight,this paper puts forward a kind of intelligent fault diagnosis method.It takes the lubrication system which is one of key systems of aero-engine as the research object to study the intelligent fault diagnosis method.Firstly,we obtained the valuable experimental data from the aero-engine health diagnostic experimental platform.The paper adopts the consensus data fusion algorithm to obtain data from preprocessing the raw data to make the processed data not only keep the details of data but also remove noise and redundant information.At the same time,on the basis of thorough analysis of health characterization parameters,using dynamic principal component analysis to extract the characteristic information which has high fault recognition resolution.Then,in view of aero-engine lubrication system failure modes and the nonlinear characteristics of data,the paper combines with Support Vector Machine(SVM)theory respectively with radial basis function kernel,polynomial kernel and linear kernel,maps the characteristics of the sample vectors from low dimensional space to high-dimensional space,then to make the samples linearly separable in high dimensional space.Doing with probability voting strategy theory to further improve the established diagnosis model/algorithm,the presented method can achieve the better classification ability and resolve the unclassifiable problems.Experiment research shows that the improved support vector machine fault diagnosis method can effectively avoid the unclassifiable problems,so as to further improve the accuracy of fault diagnosis.Finally,this paper utilizes evidence theory of the multi-source information fusion method to do decision fusion for the fault diagnosis results,which is based on different kernel function of improved support vector machine.The experimental results show that the decision-making level fusion can effectively improve the accuracy of aero-engine lubrication system fault diagnosis.The fault diagnosis method presented in this paper can accurately diagnose the fault of aero-engine and achieve the desired research goal.