CNC Corner Milling Parameters Optimization Based on Variable-Fidelity Metamodel and Improved MOPSO Regarding Energy Consumption

DOI: https://doi.org/10.1007/s40684-021-00338-3

2021-05-18

Abstract:In the corner milling process, processing energy consumption is a very important objective, since the energy efficiency of CNC machine is barely above 14.8%. Meanwhile, the excessive processing temperature will increase the thermal deformation of the product and leads to quality decline. Improper process parameters will lead to unnecessary high temperature and energy consumption. By optimizing the process parameters, the appropriate temperature and Specific Energy Consumption can be obtained. This study investigated into modeling Specific Energy Consumption and temperature in corner milling process using variable-fidelity metamodels. The adopted variable-fidelity metamodels are constructed by Hierarchical Kriging, in which 48 sets of low-fidelity data obtained from the AdvantEdge software simulation are used to reflect the trends of the metamodels, and 16 sets of high-precision data obtained from physical experiments are used to calibrate the trends. The experimental cost is reduced and the prediction accuracy is increased by making full use of both sets of data. An improved K-means Multi-objective Particle Swarm Optimization algorithm was adopted and applied on the multi-objective corner milling parameters optimization problem to find satisfactory specific energy consumption and temperature. The obtained Pareto solutions can provide guidance for selecting process parameters according to different requirements, such as reducing energy consumption or temperature.