Research On Intelligent Fault Diagnosis Method Of Rudders

During the movements of missiles,aircrafts and ships,the rudder is the executive mechanism of is navigation control system,and plays an important role in the heading control.The fault state and performance of the rudder directly determine the dynamic quality of navigation.With the continuous development of aviation technology and artificial intelligence technology,there is a need for intelligent development of rudders’ precise test.Therefore,the intelligent fault diagnosis method of rudders is an important research content which conforms to the development trend and has practical significance.As early as last century,some scholars began to study the test methods and test equipment of the rudders.At present,the traditional automatic test methods of the rudders have been able to test the performance parameters accurately.However,the performance parameters after the test need to be judged and decided by manpower.The data analysis and decision-making process are time-consuming and laborious.With the progress of science and technology,the test equipment of rudders began to develop in the direction of intelligence.This article is based on the practical application of engineering background whose main purpose is to promote the intelligent process of the rudders’ test equipment.The performance parameters of the traditional test equipment are trained and learned by using the machine learning technology.By building and optimizing the learning model,the fault location and defective product screening of the rudders can be realized.This paper first introduces the structure and test method of the traditional test equipment of the rudders,and puts forward the design idea and common algorithm of the intelligent test equipment.Secondly,the data is preprocessed to make the data clear and complete and more universal for all kinds of classification algorithms.Finally,the fault location and defective product screening of the steering gear are realized respectively,in which,fault location uses the abundant rudders data set with multi types,while,the random forest algorithm is used to learn the training set data and the shuffled frog leaping algorithm is used to carry out multi-weight weighted optimization to improve the performance of the classifier.Finally,the performance of the classifier is comprehensively evaluated in the test set by the cross validation,accuracy,recall and kappa coefficient based on confusion matrix.Besides,defective product screening uses the one-class data set with only qualified product data,and the One-Class Support Vector Machine classification algorithm is used to train the data of the rudders to achieve the defective screening.The performance of the classifier is evaluated by the composite accuracy(composed of the accuracy of the training set,the accuracy of the positive test set and the accuracy of the negative test set).The intelligent fault diagnosis and defect screening of the rudders integrate the fields of artificial intelligence machine learning technology,computer algorithm,instrument and testing science,promote the intelligence of the traditional rudders testing equipment,reduce the repeated and complicated artificial observation,and save manpower and time.The accuracy of the intelligent fault screening method can meet the basic test requirements,and has a certain practical application value.