Optimal Tool Path Planning in Five-Axis Flank Milling for Cylindrical Cutters using Surrogate Models and Multi-Level Space Reduction Techniques

Abstract

Precise error control in 5-axis flank milling of complex surfaces through optimal tool path planning is a challenging task. The solution process usually involves a large number of decision variables in a highly nonlinear search space. The performance of previous meta-heuristic algorithms on this problem has been unsatisfactory, largely due to the curse of dimensionality. This paper proposes a computational scheme for generating optimal tool paths based on multi-level space reduction techniques integrated with surrogate models. A global model is first constructed to offer an approximation to the actual solution landscape that facilitates quick identification of a potential global optimal area. Next, a local model is created around the optimal area to estimate more accurate solutions. This two-step process can repeat for multiple iterations until a satisfactory solution is achieved. We compare how different surrogate modeling methods and sampling techniques influence the optimization process based on the scheme. Test results on representative surfaces show that it outperforms other methods in terms of both solution quality and search efficiency. This work provides an effective approach for enhancing geometric accuracy in five-axis flank milling.