Online monitoring and multi-objective optimisation of technological parameters in high-speed milling process

Online monitoring and optimisation of technological parameters are very effective methods of improving productivity and machining surface quality, especially in high-speed milling. During high-speed milling processes, cutting tools wear fast, leading to increased cutting forces and vibrations and decreased surface quality with increased power consumption. To investigate the effect of cutting forces and vibrations on high-speed milling processes, models for determining cutting forces and vibrations are presented in this paper. Stochastic tool wear was obtained from a probabilistic approach based on the combination of cutting force and systematic single-point vibration analyses. The singularity obtained from the vibration sensor signal is determined by the holder exponent (HE) through the wavelet transform maximum module. In addition, the nonlinear processes caused by the deformation and geometry of the cutting blade, the basis of selecting HE as an indicator to estimate the singularity points of the vibration signal, are also considered. To provide a model for predicting and optimising cutting forces, tool wear, vibrations, surface quality and power consumption, a new hybrid algorithm, i.e. back-propagation neural network and multi-objective particle swarm optimisation, was developed to determine the optimal cutting parameters to minimise the total power consumption, improve surface quality and increase tool life. High-speed milling experiments were conducted to confirm the accuracy and availability of the proposed multi-objective prediction and optimisation model. The improved optimisation method based on the proposed model can increase the surface quality and tool life by 5.95% and 9.87%, respectively. The power consumption can be reduced by 10.49% compared to empirical selection.