Detection of change in processes using wavelets

The paper presents some applications for which the wavelet transformation can be used to detect defects in manufacturing processes. This is accomplished by characterizing process parameters using various forms of wavelet coefficients. The identification of process defects is demonstrated using detection indices that are derived from these coefficients. It has been found that these indices are able to detect obscure defects/changes in the process parameters, thus providing an early indication of possible failure. To illustrate the effectiveness of those indices, different signals from various processes, namely, vibration signals from bearings, force signals from end milling, and acoustic emission signals from grinding, are used