Analysis and prediction model of vibration strength in milling wood–plastic composites

This study designs multiple experimental schemes and conducts high-speed milling experiments using hard alloy cutting tools in order to solve the vibration problem encountered during high-speed milling of wood–plastic composites. A single-factor experiment is conducted to investigate the effects of milling parameters on vibration acceleration using time-domain analysis. The frequency-domain characteristics of vibration acceleration are then analyzed, and energy variations across different frequency bands under different milling conditions are examined using wavelet packet decomposition. An orthogonal experiment based on the vibration acceleration test results is then designed, and a milling vibration acceleration prediction model is established using a PSO-BP neural network. Results indicate that during high-speed milling, the vibration acceleration amplitude initially increases and then decreases as the cutting speed increases. In contrast, the vibration acceleration amplitude increases as the feed rate, axial cutting depth, and radial cutting depth increase. The observed milling vibration is forced vibration, with the energy distribution in the medium- and high-frequency range serving as a key indicator for vibration signal recognition. The PSO-BP neural network prediction model can effectively predict vibration acceleration signals during high-speed milling of wood–plastic composites.