Experimental investigation and optimization of process parameters in micro-milling LF21 thin-walled parts

The micro-milling of thin-walled aluminum alloy parts is often constrained by the inherent trade-off between surface quality and machining efficiency, particularly for low-stiffness structures such as LF21 aluminum alloy. In this study, a systematic investigation is conducted to quantitatively characterize and optimize this trade-off. A central composite circumscribed experimental design based on response surface methodology (RSM) is employed to analyze the individual and interactive effects of spindle speed, feed per tooth, axial depth of cut, and radial depth of cut on surface roughness. A predictive model for surface roughness is established and experimentally validated, exhibiting a maximum relative error of 13.25%. To simultaneously improve surface quality and material removal rate, an improved NSGA-II algorithm is developed. The proposed algorithm incorporates Latin Hypercube Sampling (LHS) for population initialization, an adaptive hybrid crowding distance mechanism, and a Pareto-front-shape-aware environmental selection strategy to enhance convergence stability and solution diversity. Based on the developed models and optimization framework, a multi-objective optimization of micro-milling parameters for LF21 thin-walled structures is performed with the dual objectives of minimizing surface roughness and maximizing material removal rate. The optimized parameter combinations provide quantitative guidance for achieving high-efficiency and high-quality micro-milling of LF21 aluminum alloy thin-walled parts.