Abstract:Limiting cutting depth for evaluating the milling stability is dependent on the machining position. The consequence is that the stability constraint of the process parameters optimization model has uncertain. To solve this problem, the tool tip frequency response functions at different machining positions are combined with the milling stability theory. Firstly, a general regression neural network (GRNN) is formulated for predicting the positiondependent limiting cutting depth, which can be used to determine the milling stability constraint. Then, a process parameters optimization model of multipasses milling for minimizing cutting time is established. Displacements of the machine tool moving parts and cutting parameters for rough and finish milling processes are taken as variables. The particle swarm optimization algorithm (PSO) is utilized to solve this optimization model. A case study is implemented on a vertical machining center. The optimal combination of machining position and cutting parameters can be obtained, including the spindle speed, cutting depth, cutting width and feed rate per tooth. The total cutting time of the rough and finish processes decreases 2247% after the optimization. There is no chatter during the milling process, which verifies the feasibility of the proposed optimization model.