Spectral Interference Rejection Algorithm for Beat Frequency Estimation in Machining Chatter

doi:10.1109/TIM.2023.3329216

2023-01-01

PRE

DOI：10.1109/TIM.2023.3329216

原文链接

原文求助

摘要

En 中文

Chatter detection is an important approach to reduce detrimental effects induced by machining chatter on part quality and machine tool components. High-accuracy beat frequency estimation is crucial to effectively detect chatter with the beat effect phenomenon. The beat frequency can be obtained from the modulated instantaneous amplitude (IA) of the chatter vibration component. However, the estimation accuracy is severely impacted by the spectral leakage induced by negative parts and harmonics, since the IA is generally a nonsinusoidal periodical waveform. In this article, a novel spectral interference rejection method is developed for accurate beat frequency estimation during machining processes. This algorithm operates within the frequency domain and utilizes a method of accurately aligning a theoretical spectrum of a signal with multiple harmonics to the collected spectrum of the sampled signal. Through computer simulations and simulated machining signals, the performance of our algorithm is evaluated comprehensively in comparison to other estimators. Performance comparisons between different algorithms are carried out under different levels of noise and harmonic interference. Our algorithm is 15-25 dB more accurate than other algorithms, when the harmonic interference is severe. When dealing with simulation machining signals, the mean square error (MSE) of our algorithm is 15 dB less than others in three cases. Moreover, our algorithm is evaluated through two machining tests, and the results demonstrate that the proposed algorithm provides more accurate beat frequency estimation.

Keyword:

Beat effect

chatter detection

frequency estimation

machining chatter

spectral inference