Chatter detection and spindle speed selection by scanned periodic sampling frequency

Abstract

This paper proposes and evaluates an approach for chatter detection using periodic sampling of time domain milling signals. In this approach, the periodic sampling frequency is scanned and the sampled values of the time domain milling signal are used to calculate a stability metric based on the normalized sum of the absolute value of differences between adjacent samples for each sampling frequency. It is shown that local minima in the stability metric values occur at the tooth passing frequency (and its subharmonics) for stable behavior, while the local minima occur at/near the chatter frequency (and its subharmonics) for chatter. It is also demonstrated that the chatter frequency identified using the metric value local minimum can be used to update the spindle speed and determine stable operating parameters. Both numerical and experimental results are provided to validate the approach.