Application of non-parametric methods for monitoring of tool chipping

Abstract

Cutting tool wear is a consequence of a normal wearing process, which is considerably slow, or of a chipping process, which happens instantly and could often lead to tool breakage.The paper presents a methodology for monitoring the tool chipping process.The reliability of monitoring the tool chipping, in particular its initialization, depends on the sensor signal acquisition, processing, and interpretation.Tool chipping is a very fast nonlinear process, and too complex to be described by deterministic models.Not only the cutting process itself, but also the mechanical link (from motors to toolworkpiece interface) is a significant cause of signal deviation and noise.Therefore, knowing the signal power spectra and analyzing them are of great importance for the monitoring of stochastic processes.This paper presents an analysis of the tool chipping process by means of a nonparametric method for the spectral analysis of force, current, and acoustic signals.