Fault Diagnosis And Prediction System Of A Radar Servo System

Radar servo system is an important part of radar,which plays an important role in target discovery,target tracking and accurate measurement of target position.However,it is found that the frequency of servo system failure is very high.At present,the maintenance activities of radar servo system mainly rely on professional and technical personnel to use instruments and meters,and rely on experience to carry out regular maintenance,after-the-fact maintenance,situation-based maintenance and other traditional maintenance methods.Traditional maintenance methods are inefficient,and sometimes even other failures may occur due to lack of experience.In this thesis,a fault diagnosis and prediction system of radar servo system is designed,which can improve the support efficiency and the accuracy of fault diagnosis and prediction of radar servo system,so as to give full play to the effectiveness of radar.The main work is as follows:Firstly,the selection of sensitive parameters of radar servo system and data acquisition system have been studied.On the basis of analyzing the working principle and composition of radar servo system,the types,characteristics and selection methods of measuring points,a sensitive parameter selection method of radar servo system based on fault tree and fault mode,influence and hazard analysis is proposed.The data acquisition system of radar servo system is realized based on NI data acquisition card.Secondly,aiming at the problem that signal noise is difficult to remove,an improved empirical wavelet transform denoising algorithm has been studied.According to the hierarchical structure characteristics of radar servo system,the three-level fault diagnosis method of radar servo system has been studied.They are as follows:(1)Module-level fault diagnosis method,based on fault tree and expert system,forms a rule base stored in dictionary,and then traverses the rule base for fault diagnosis.(2)Board-level fault diagnosis method,which analyzes the correlation between the input and output signals collected by the circuit board and the normal input and output signals,and then obtains the fault diagnosis results.(3)Component-level fault diagnosis method,based on feature extraction of signals,the state diagram is constructed,and finally input into convolutional neural network for training,and then the fault diagnosis results are obtained.Then,the fault prediction method of radar servo system based on SOM and LSTM has been studied.On the basis of extracting signal features,the state vector is constructed.Taking this as input,it is input into SOM network for feature fusion.According to the distance between each state vector and the weight vector corresponding to the optimal neuron,the HI degradation curve is constructed,and then the time series corresponding to the degradation curve is used as the input of LSTM,which is input into LSTM network for training,and then the fault prediction result is obtained.Finally,on the basis of analyzing the system software architecture and software requirements,the software for fault diagnosis and prediction is compiled by Labview.