Research And Implementation Of Intelligent Flight Control Fault Diagnosis System

In the current flight control fault diagnosis system,,an expert system is generally used to diagnose the fault.When using traditional expert systems to diagnose flight control faults,there are rules conflicts,slow matching,and problems such as new faults that cannot be accurately located.This article addresses the deficiencies of expert systems,applying machine learning algorithms to the fault diagnosis of flight control systems,Improve the diagnostic accuracy of the device,it is of great significance to the development of flight control fault diagnosis.The main contents of this article are as follows:First,the current fault diagnosis method is analyzed,study the characteristics of flight control fault data,according to the characteristics of flight control data samples,two machine learning algorithms,support vector machine and random forest,were selected as the algorithm of the system and the model building and parameter tuning of the two algorithms were carried out separately.Using different data sets to test the classification performance of support vector machine model and random forest model respectively,according to the test results,the random forest with better classification performance is selected as the machine learning algorithm of the system.Then,in order to achieve better classification accuracy,the random forest was optimized in two aspects: decision tree node splitting and decision tree integration.In splitting optimization of decision tree nodes,a splitting algorithm based on linear combination of minimum Gini index and maximum information gain rate is proposed;In decision tree integration optimization,a decision tree integration optimization algorithm based on classification accuracy is proposed,then using different data sets to verify the two algorithms experimentally.On this basis,this paper also comprehensively uses two algorithms,integrated optimization of random forest generated by node splitting and experimented with the data set on the algorithm.Experimental results show that the algorithm improves the classification accuracy of random forest the most and finally chooses to use the comprehensive algorithm as the algorithm in this paper.Finally,according to actual needs,the overall frame design and logical frame design of the system and used python to develop software for the system.After the software is implemented,the function of the flight control fault diagnosis software and the accuracy of fault diagnosis are tested experimentally.