TC4钛合金电火花小孔加工多目标优化试验研究  

Experimental Study on Multi-objective Optimization of EDM Small Hole Machining for TC4 Titanium Alloy

作　　者：张文超 王帅[1] ZHANG Wenchao;WANG Shuai(School of Mechanical Engineering and Automation,Dalian Polytechnic University,Dalian 116034,Liaoning,China)

机构地区：[1]大连工业大学机械工程与自动化学院,辽宁大连116034

出　　处：《机械科学与技术》2023年第1期113-118,共6页Mechanical Science and Technology for Aerospace Engineering

摘　　要：为提升电火花加工TC4钛合金的表面加工质量和加工效率,选取紫铜圆柱电极开展TC4钛合金电火花小孔加工试验,采用正交试验法,以电极相对损耗率、表面粗糙度、工件材料去除体积为工艺指标,分析峰值电流、维持电压、放电脉宽对工艺指标的影响重要性。采用RBF(Radial basis function)神经网络对已有试验数据进行训练,建立放电参数与工艺指标之间的数学预测模型。以该预测模型为适应度函数,将遗传算法与Skyline选择算法结合进行多目标优化仿真,得到最佳工艺指标,最后开展多目标优化验证试验。结果表明:当峰值电流为14 A、维持电压39 V/42 V、放电脉宽102μs/108μs时能够取得最优的加工结果,优化值与试验值误差较小。To improve the surface machining quality and machining efficiency of TC4 titanium alloy in EDM(Electrical discharge machining),the copper cylindrical electrode was selected to carry out EDM small hole machining experiment of TC4 titanium alloy.The orthogonal experiment method was adopted.Taking the relative electrode wear rate,surface roughness and material removal volume of workpiece as optimization objectives,the influences of the peak current,discharge voltage and discharge pulse width on the optimization objectives were analyzed.RBF(Radial basis function)neural network was used to train with the experimental data,and the prediction model between the discharge parameters and the optimization objectives was established.Taking the prediction model as the fitness function,the multi-objective optimization simulation was carried out by combining the genetic algorithm with the Skyline selection algorithm,and the optimal technical index was obtained.Finally,the multi-objective optimization verification experiment was carried out.The results show that when the peak current is 14 A,the maintenance voltage is 39 V/42 V,and the discharge pulse width is 102μs/108μs,the optimal machining results can be obtained,and the error between the optimal value and the experimental value is small.

关 键 词：电火花加工 RBF神经网络 遗传算法 Skyline选择算法 多目标优化 

分 类 号：TG661[金属学及工艺—金属切削加工及机床]