Milling tool wear monitoring through time-frequency analysis of sensory signals

Abstract

The states of milling tool are closely related to the quality of the workpieces under machining. A high quality product often implies high quality surface finish and dimensional accuracy. Therefore, tool wear has to be controlled. However, the tool wear cannot be measured continuously while the machine is still in operation. Thus an alterative condition monitoring approach should be adopted. The condition parameters, e.g. electric current, vibrations, acoustic emissions, are considered as indirect data in data-driven health management technology as they are not directly related with the machine health states. The sensory signals acquired during the operational process are generally time varying (TV) and non-stationary. The features will be lost if the signals are analyzed from just the time domain or frequency domain. The combination of time and frequency analysis (TFA) of the signals is very useful to extract the features hidden in the signals.