单晶Si电火花线切割工艺的多目标优化研究
Multi-objective Optimization for WEDM of Single-crystal Silicon

由于电火花线切割放电过程很难通过理论分析得到有效的切割机理数学模型。引入中心复合实验设计（Central composite design，CCD）实验方法，建立四因素三水平的单晶Si电火花线切割的实验方案。采用响应曲面法（Response surface methodology，RSM）建立单晶Si表面粗糙度和材料去除率与空载电压、脉冲宽度、脉冲间隔和运丝速度等主要工艺参数的二阶数学模型并进行分析。分析结果表明：预测模型具有较好的拟合度和适应性。并以提高单晶Si表面的加工质量和材料去除率为目标建立工艺参数优化模型；设计非支配快速排序遗传算法（NSGA-Ⅱ）进行优化问题求解，得到了最优的Pareto解集。实验表明，优化模型对表面粗糙度和材料去除率的预测是准确的，并且能实现相应的半导体材料的线切割过程中的粗糙度和材料去除率的预测。
It is difficult to obtain an effective mathematical model for cutting mechanism by using theoretical analysis. This paper introduces the experimental method of central composite design (CCD), and establishes the experimental scheme of four factors and three levels for single-crystal Silicon wire cutting. The second order mathematical model for single-crystal Silicon surface roughness and material removal rate, such as no-load voltage, pulse width, pulse interval and wire speed, was established by using response surface methodology (RSM). The results show that the prediction model has good fittness and adaptability. Non-dominated sorting genetic algorithm Ⅱ (NSGA-Ⅱ) is used to solve the optimization problem, and the optimal Pareto solution set is obtained. The optimization of the process parameters is carried out to improve the processing quality and material removal rate of the surface of single-crystal Silicon. Experiments show that the present model is accurate for predicting the surface roughness and material removal rate, and which can realize the prediction of roughness and material removal rate during the wire cutting of the corresponding semiconductor materials