Sequential independent component analysis for cutting forces de-noising in micro-machining tool condition monitoring

摘要

In this paper, we applied a sequential independent component analysis (ICA) algorithm for cutting forces de-noising, as a preprocessor in micro milling condition monitoring. The sources can be extracted from the instantaneous mixtures (sensor outputs) with ICA based on kurtosis. We can extract the sources one by one with deflation until as expected with the algorithm. Such methods are attractive in micromachining monitoring since the expected cutting forces are contaminated by relatively large noises and the sources and noises are identified non-Gaussian. The results were illustrated in time domain.