Short Title: Int. J. Mech. Eng. Robot. Res.
Frequency: Bimonthly
Professor of School of Engineering, Design and Built Environment, Western Sydney University, Australia. His research interests cover Industry 4.0, Additive Manufacturing, Advanced Engineering Materials and Structures (Metals and Composites), Multi-scale Modelling of Materials and Structures, Metal Forming and Metal Surface Treatment.
2024-10-25
2024-09-24
Abstract—Process monitoring is necessary in machining operation to increase productivity, improve surface quality and reduce unscheduled downtime. Tool wear and breakage are important and common sources of machining problems due to high temperatures and forces of machining process. Therefore, it is highly beneficial to develop an online tool condition monitoring system. This paper investigates a robust tool wear monitoring system for milling operation. Spindle current is employed as the fault indicator due to its cost-effectiveness and ease of use in an industrial environment. Wavelet time-frequency transform is used as a superior tool to simultaneously investigate time-varying characteristics of the signal and its frequency components. After the time-frequency step, spectral subtraction algorithm is employed to intensify the effect of tool wear in the signal and reduce the effect of other cutting parameters. Based on this method, the average signal spectrum of a healthy case is subtracted from all the signals with the same cutting parameters. After further processing and noise reduction, fault features and indicators are extracted from the results of the processed signal. Finally, five advanced machine-learning algorithms are implemented for modeling the system. Gaussian process regression, support vector regression, Bayesian rigid regression, nearest neighbor regression and decision tree methods are compared. The methods are validated based on the experimental data. Results show a high accuracy for the tool wear estimation while decision tree method was superior to others with accuracy of 91.6%.