Multi-objective feed rate optimization of three-axis rough milling based on artificial neural network

Feed rate in the computerized numerical control (CNC) milling is an essential parameter that could affect both the machining efficiency and the working conditions of the machine tool. In this paper, we proposed a multi-objective feed rate optimization method of three-axis rough milling. The in-process machining data as generated in the machining process is calculated and aligned for building the data-based model of spindle power, and an artificial neural network (ANN)-based modeling approach is proposed for the spindle power. Based on the proposed model, a multi-objective optimization framework is presented to optimize the feed rate for the objectives of increasing the machining efficiency and the loading stability of the spindle. To validate the feasibility and advantage of the proposed methods, a set of machining experiments are conducted, showing that our proposed ANN-based model has good accuracy in terms of predicting the spindle power, as well as that the feed rate optimization framework as solved based on the multi-objective evolutionary algorithm based on decomposition (MOEA/D) can effectively improve the machining efficiency and reduce the fluctuation of spindle power.