FOURIER AND WAVELET TRANSFORM FOR FLANK WEAR ESTIMATION — A COMPARISON

Abstract This article presents potential sensor data representation schemes for force and vibration signals in the context of flank wear estimation in turning processes. In particular, the performances of methods based on fast Fourier transforms (FFTs) and fast wavelet transforms (FWTs) are compared using data from turning experiments. This research, for the first time, studies the performance of these modern sensor data representation schemes for flank wear estimation on a common platform and provides a useful insight into their merits and drawbacks. The flank wear estimates are computed continually from the features extracted through each representation scheme by using a simple recurrent neural network architecture. The results can be used for selecting correct data representation schemes for flank wear estimation.