【摘要】 针对薄壁筒切削加工过程中的多频颤振复杂现象，提出一种在线监测声压信号小波包能量相对熵的薄壁筒切削状态识别方法。以不同状态对应主振动频率以及能量分布不同作为颤振依据，首先，通过小波包分解得到切削过程中声压信号能量概率分布；其次，以机床空载时的信号能量概率分布作为参考分布；最后，建立小波包能量分布相对熵指标来监测薄壁筒零件的切削状态。切削振动试验结果表明：随着加工状态的恶化，小波包能量相对熵值明显增大；在主振动频率发生跳跃的位置，相对熵值也会发生突变；相比于传统的频谱监测方法，所提出方法表现出更快的响应速度，能够对多频颤振状态进行准确判别。更多还原

【Abstract】 During the cutting process, the vibration characteristics of the thin-walled cylinder workpiece are sensitive to material removal. The vibration response of the workpiece shows strong time-varying characteristics, and the chatter frequency exhibits multi-frequency and jump phenomena. This paper aims at detecting the complex multi-frequency chatter phenomena when machining thin-walled cylinder workpieces. A thin-walled cylinder cutting state identification method is proposed based on the on-line monitoring of wavelet packet energy relative entropy of sound pressure signals. Firstly, based on the main vibration frequency and energy distribution corresponding to different states, the energy probability distribution of acoustic signals during the cutting process is obtained by wavelet packet decomposition, and then taking the signal energy probability distribution of machine tool without load as the reference distribution. Finally, the relative entropy index of wavelet packet energy distribution is established to monitor the cutting state of thin-walled cylinder. The results of the machining tests show that the wavelet packet energy relative entropy is increased obviously with the deterioration of the machining state. At the position where the main vibration frequency jumps, the relative entropy value could also change quickly. Compared with the traditional spectrum monitoring methods, the proposed method exhibits faster response speed and accurate identification of the multi-frequency chatter state.更多还原