The effectiveness of various chatter detection methods under noisy conditions

Abstract

Unmanned operations are sought after in manufacturing processes such as milling and lathing. During these processes, the detection and mitigation of machine tool chatter is critical. The veracity of these methods under noise conditions that would be found in a live factory environment is not well understood. This study aims to evaluate the performance of various classification methods for the detection of chatter under periodic and white noise. Different training methods and artificial noise injection are used to highlight the benefits and pitfalls of the different methods for chatter detection. It is found that machine learning models like Support Vector Machines have a significant ability to classify noisy data even when untrained on noise.