In-process Evaluation of the Overall Machining Performance in Finish-Turning via Single Data Source

Abstract

Sensor fusion often uses multiple sensors to evaluate a single quantity. The work presented in this paper attempts to use information from a single sensor to estimate overall machining performance (characterized by cutting forces, chip breakability, surface roughness, and dimensional deviation due to tool wear). In particular, the performance is aimed at reflecting the in-process changes of the above-named quantities with respect to tool wear progression (major flank, crater and minor flank wear). 3-D cutting force measured by a tool dynamometer is fully utilized by aggregating multivariate time series models and neural network techniques. Dispersion analysis is used to extract signal features which correlate well with progressive tool wear. The results have shown the effectiveness of the proposed method which also has the obvious merit of simplicity.