Research On Fault Diagnosis Of Radar Servo System Based On Data

As an important subsystem of the radar system,the radar servo system directly affects the operating accuracy of the radar system.Various types of failures will occur during the operation of the radar servo system.Fast and accurate fault diagnosis of the radar servo system is the basis to ensure its stable operation.At present,the fault diagnosis of radar servo system is usually completed by manual or traditional mechanism models,and the efficiency and accuracy of fault diagnosis are difficult to be guaranteed.The historical fault data of radar servo system has the law of fault generation.Based on the data,fault diagnosis research is helpful to improve the efficiency and accuracy of fault diagnosis,and has high theoretical significance and application value.This subject is a project in cooperation with a research institute in Nanjing,and a certain type of radar servo system is used as the research object to conduct related fault diagnosis research.In this regard,the main work of this article is as follows:First,the overall design of the radar servo fault diagnosis system is given,the overall architecture design of the fault diagnosis system and the design of each functional module are completed,and the development tools and programming languages suitable for the program are selected.Secondly,based on the historical switching fault data of the radar servo system,a switching fault diagnosis model is established.Before establishing a switch fault diagnosis model,a multi-signal flow graph model and an improved genetic algorithm are used to optimize the selection of test points,so that the test cost can be reduced while ensuring the testability index.After optimizing the test points,this paper selects the Bayesian network model as the switch fault diagnosis model.At the same time,in order to solve the problems of Bayesian network model structure learning difficulties and accuracy bottlenecks,a structure fusion learning method based on K2 algorithm,MHS algorithm and MMHC algorithm is proposed.Then,based on the analog historical fault data of the radar servo system,an analog fault diagnosis model is designed.Before establishing the analog fault diagnosis model,remove the outliers in the sample through the isolated forest model,add time series characteristics to the data through targeted feature engineering,and use the SMOTE algorithm to solve the sample imbalance problem to ensure the accuracy of the analog fault data,Making the data more complete description of the fault.In order to solve the bottleneck problem of single learner diagnosis accuracy in traditional analog fault diagnosis,a fault diagnosis model based on Stacking integration method is proposed.Next,considering that the switch fault data and the analog fault data have different descriptions of the faults of the radar servo system,there are certain limitations in performing fault diagnosis based on the switch fault data or the analog fault data.Therefore,this paper finally proposes an information fusion fault diagnosis model,which integrates the output results of the digital fault diagnosis and the output results of the analog fault diagnosis according to the D-S evidence theory.Experimental results show that the method optimizes the fault diagnosis results of the radar servo system,greatly improves the accuracy of fault diagnosis,and has high practical value.Finally,based on the radar servo system fault diagnosis model designed in this paper,the development of the radar servo system fault diagnosis software is completed.Through the analysis of the actual demand,the overall design of the radar servo system fault diagnosis software and the design of each functional module are given.At the same time,the operation process of the software is explained with examples,which embodies the intelligence and humanization of the software.