Investigation on the motion fault of main spindle using the wavelet neural network

【摘要】：正The purpose of this work is to put forward a new method using neural network of harmonic wavelet.Fault of workpiece is separated from examined signal by adopting technology in separating fault form,thus the denoising method of learning ability of threshold value in harmonic wavelet and neural net- work is put forward.In this method the optimal value may be found for lathe main spindle so as to denoise the motion fault of lathe main spindle,which overcomes the shortcoming of traditional hard threshold value and soft threshold value.The experiment shows the validity and applicability of the proposed method.