Improving Optimization of Tool Path Planning in 5-Axis Flank Milling by Integrating Statistical Techniques

Abstract

Optimization of the tool path planning in 5-axis flank milling of ruled surfaces involves search in an extremely high-dimensional solution space. The solutions obtained in previous studies suffer from lengthy computational time and suboptimal results. This paper proposes an optimization scheme by integrating statistical techniques to overcome these problems. The scheme first identified significant factors in the tool path planning that influence the machining error of a machined surface by a first sampling plan. We then conducted a series of simulation experiments designed by the two-level fractional factorial method to generate experimental data with various settings. A regression model was constructed with Response Surface Methodology (RSM) that approximates the machining error in terms of those identified factors. This simplified model accelerates estimation of the objective function, computed as a black-box function in previous studies, with less computation. Test results show that the proposed scheme outperforms PSO in both the computational efficiency and the solution quality.