Multi-objective modelling and optimal parameter selection of a multi-pass milling process considering uncertain milling stability constraint

Optimizing machining parameters and pass numbers under machining stability constraint is important for manufacture engineers planning a multi-pass milling operation. Considering that milling stability varies with the machine tool dynamics when the tool and workpiece position changes, a generalized regression neural network (GRNN) is adopted to obtain the position-dependent stability constraint. Then a multi-objective optimization model of a multi-pass milling is established, where the objectives are the total cutting time (T_c) and surface roughness (R_a) and the variables are the machining position and cutting parameters for each pass. A non-dominated sorting genetic algorithm (NSGA-II) is introduced to solve the multi-objective optimization model for providing an optimal Pareto front, from which one satisfactory solution well balancing the T_c and R_a is selected by combining the entropy-weighted algorithm (EWA) and technique for ordering preferences by similarity to ideal solution (TOPSIS). A case study was carried out to establish a multi-objective optimization model after the milling stability constraint and surface roughness were predicted with the aid of GRNN and back-propagation neural network (BPNN), respectively. An ideal solution containing optimal machining position and cutting parameters was solved using NSGA-II, EWA, and TOPSIS, which was compared with solutions of two mono-objective models for optimizing T_c and R_a, respectively, to validate its feasibility. The observed stability and acceptable R_a value of the milling test under the ideal solution also indicated that the proposed optimization method can realize a trade-off between two conflict objectives and fascinate the process planning of a multi-pass milling.