An approach based on total load current for lathes working state recognition

Current monitoring method has become a new method. In this paper, we took the total load current signal of lathes as the main object, used wavelet analysis and FFT to process this signal respectively, and extracted signal features to indicate the working state. Then by constructing the Support Vector Machine (SVM) “One against One” clustering structure model the recognition of the working state can be implemented correctly. The results of experiment shows that, based on load current and SVM, this model could realize the condition monitoring of two or more lathes and instead of the method of measurement and state diagnosis in the past that can only works on individual device.