Monitoring Rotating Tool Wear Using Higher-Order Spectral Features

Abstract

A framework for detecting incipient wear in rotating machinery is proposed. In this paper, statistical techniques that combine power spectrum estimates with higher-order spectrum (HOS) estimates for feature development are applied to discriminate and classify vibration signals from new and slightly used drill bits in a drill wear study. Results from experimental data obtained when drilling composite circuit cards reveal that the performance of a power spectrum-based tool wear monitoring system can be enhanced by complementing the power spectrum information with HOS information on the accelerometer signal. Evidence presented supports the proposition that a HOS approach provides better signal features to a pattern classifier which allows better decisions on the state of rotating tool wear.