Abstract
Acoustic emission (AE) is considered one of the main methods of on-line detection of catastrophic tool failure (CTF). Some strategies have claimed a subsequent increase of the root mean square value of the AE signal (AERMS) which in turn has been used as a measure of the CTF. However this measure was found to be not always sensitive to CTF. The aim of this paper is to present a method of catastrophic tool failure detection which uses symptoms other than the direct AERMS signal. The method is based on the statistical analysis of the distributions of the AERMS signal. The b distribution which was assumed in this study has been used with a density function of two parameters. The skews and kurtosis of the b distribution were the main measures employed. They were found to be highly sensitive to changes in tool condition and have given promising results with regard to chipping and tool breakage detection.