Wavelet representation of Acoustic Emission in turning process

This paper deals with the representational and analysis issues of Acoustic Emission (AE) signals in turning processes. In earlier works, the power spectral density of AE signals is computed from Fourier transform based techniques. This paper investigates the superiority of a wavelet transform over the Fourier transform in analyzing rapidly changing signals such as AE, in which high frequency components are to be studied with sharper time resolution than low frequency components. The effectiveness of the wavelet representation of AE signals is studied in the context of flank wear estimation problem in turning processes. A set of turning experiments are conducted in which the flank wear is monitored through AE signals. A specially designed neural network architecture is used to relate AE features to flank wear. The accurate flank wear estimation results indicate that the wavelet transform representation of AE signals is very effective in extracting the AE features sensitive to gradually increasing flank wear.