Fault Diagnosis Research Of Automatic Ammunition Loading System Based On Current Signal And Wavelet-SVM

The artilleries equipped or researched in recent years have been equipped with theadvanced automatic ammunition loading system.Although the degree of automation has beengreatly improved, the system is more complex and has high malfunction occurrence rate.Based on current signal and in terms of theoretical research, simulation and test on theDC motor ammunition loading mechanism fault diagnosis, this article studies thecharacteristics of fault feature among the current signal and tests the feasibility and validity ofcurrent signal from theory to practical application, which is applied to the DC motorammunition loading mechanism fault diagnosis. Concrete methods are: On the basis ofwavelet theory and support vectors machine theory(SVM), this paper uses goodtime-frequency analysis capability of wavelet algorithm, and selects an appropriate waveletfunction to improve the effect of the fault diagnosis according to the actualsituation,literature reviews indicate that wavelet function selection greatly affects the qualityof the analysis of transient signal. This article chooses the sym8wavelet function to detect thefault of ammunition loading mechanism and obtain the time and location of the mutation ofcurrent signal. Then, this paper combines wavelet packet algorithm with SVM algorithm,classifies and recognizes the current fault, extracts the energy eigenvalues of the signal afterdecomposing the current signal into three layers with wavelet packet algorithm. As the inputvector of SVM, energy eigenvalues can be trained to enable SVM to diagnose and classify.The result of signal analysis test indicates that SVM can efficiently identify the normal stateand failure state of artillery autoloader.The results of this study expand the fault diagnosis method of automatic ammunitionloading system, provide the automatic ammunition loading system with an effective methodof mechanical fault diagnosis, and have some academic values.