Analysis of Tool Chatter in Terms of Chatter Index and Severity Using a New Adaptive Signal Processing Technique

Abstract

Regenerative chatter is a predominant phenomenon in the turning process. In machining of metals, identification of tool chatter is essential in order to improve productivity and thus enhance the tool life. In spite of the immense work done within this domain, still many facets are yet to be explored. For detection of chatter many researchers have used sensors, but usually measured chatter signals from sensors are contaminated by background noise and other disturbances. Hence, it is essential to develop an efficient signal processing technique by the aid of which induced noise can be removed and onset of chatter can be detected at the earliest. In present work the scholar has explored four objectives. Firstly, turning process has been simulated using simulink tool Matlab and signals have been recorded at different combinations of cutting parameters. The simulation has been validated by comparing the simulated and experimentally recorded signals. Secondly, wavelet denoising technique has been implemented for denoising the noisy signals. Moreover, the denoising technique have been validated by comparing the results with simulated noise free signals. Thirdly, the peak frequency have been determined at different combination of cutting parameter by power spectral density analysis. Lastly, a new output parameter chatter index (CI) has been calculated. CI makes it convenient to analysis the severity of chatter.