Research On Fault Diagnosis Of Artillery Motor Based On Empirical Wavelet Transform

Artillery servo system is one of the most stringent systems for artillery operation requirements and operating conditions.If the artillery servo system is like a " body ",the artillery drive motor on the servo system is the " heart " that maintains the " body " normal operation.In other words,the drive motor is the core component of the artillery servo system.The working environment of the motor in the artillery servo system is bad,which not only affects the working life of the motor,but also increases the failure rate of the motor.Mastering a suitable fault diagnosis method of motor,and taking countermeasures to eliminate the fault in time before the failure of the equipment component.It is a great value for improving the operational performance of the artillery.This topic takes the artillery motor as an example to conduct in.depth research on common fault types and diagnostic schemes.The main contents are as follows:(1)In order to overcome the shortcomings of the soft threshold function which will cause signal edge distortion,and make full use of the hard threshold function to preserve the signal advantage of the signal edge local characteristics,the subject tries to improve based on those threshold scheme.Aiming at the lack of experience in choosing decomposition layers for classical wavelet threshold de.noising method,the empirical mode decomposition has defects such as false mode and modal aliasing.In this paper,a denoising scheme based on empirical wavelet transform and improved threshold is proposed.It tries to avoid blindly pursuing the improvement of signal.to.noise ratio,which denoising too much affects the original signal.It also tries to avoid insufficient noise reduction in order to retaining the original signal.(2)A rolling bearing fault diagnosis scheme based on energy difference and genetic algorithm is proposed.Firstly,the energy difference characteristic index is proposed.When the energy operator is used as the signal fault characteristic index,it can only represent the defect of each fault scale and cannot display the fault characteristic transition trend between adjacent scales.Secondly,in view of the appropriateness of the selection of SVM parameters,the genetic algorithm with strong iterative search function is used to optimize the parameters.It solves the difficulty that the commonly K-CV mesh search method can only find the local highest classification accuracy,and K-CV method is a large amount of calculation and time.consuming and difficult to operate.(3)A method for detecting the state of artillery motor based on Hankel matrix transformation and random forest is proposed.Firstly,in order to utilize the motor signal information more efficiently,the Hankel matrix conversion is performed before the singular value calculation of the motor noise reduction signal.Then,for the irregularity and complexity reality of the signal singular value feature,the singular sample entropy and singular energy value are used to assist the analysis again.Aiming at the current situation of the motor in this subject,we choose a random forest classifier with strong binary ability and determine the decision tree with great influence on the classification result by local search method.Finally,the feasibility of the motor fault diagnosis scheme proposed is verified by the motor fault test and the field motor experiment of the Institute.(4)For the solution proposed in the above section,a set of user operating system for artillery motor fault diagnosis based on the research results of this paper was compiled,it makes all the research results in this paper will be practical and operable.At the same time,it provides simple guidance for the user's real fault diagnosis operation,and lays a foundation for further improvement of the system.