基于WPD与功率谱熵的铣削颤振仿真识别

Simulation Identification of Milling Chatter Based on WPD and Power Spectrum Entropy

摘要: 颤振是高速铣削加工过程中的一种不稳定振动，它会导致工件表面质量差、刀具磨损严重、噪声严重。为了避免铣削颤振带来的这些负面影响，需要对铣削颤振进行早期检测。本文提出了一种基于小波包分解和功率谱熵的轻微颤振识别方法。首先，对加速度传感器采集到的原始振动信号进行小波包分解，选取含有丰富颤振信息的小波包进行重构仿真。然后，计算重构仿真信号的功率谱熵作为识别指标。实验结果表明，该方法能够有效地检测出颤振。

Abstract: Chatter is a kind of unstable vibration during high-speed milling machining, which leads to the poor surface quality of the workpiece, serious tool wear and noise. In order to avoid these negative effects of milling chatter, early detection of milling chatter is needed. In this paper, a slight chatter identi-fication method based on wavelet packet decomposition (WPD) and power spectrum entropy is proposed. Firstly, wavelet packet decomposition is performed on the original vibration signals col-lected by the acceleration sensor, and the wavelet packets containing rich chattering information are selected for reconstruction simulation. Then, the power spectral entropy of the reconstructed simulated signals is calculated as the identification index. The experimental results show that the method can effectively detect the chattering vibration.