基于NSGA-Ⅱ的钛合金铣削工艺参数多目标支持向量机优化

Multi-objective SVM optimization of milling process parameters of titanium alloy based on NSGA-Ⅱ

通过实验测试了钛合金铣削加工性能,采用正交试验方法分析了切削参数组合形式引起的切削力变化情况。从材料去除效果与切削载荷方面进行综合评价,建立了包含多目标元的启发算法,由此确定Pareto前沿参数组成的集合,对比了各算法对上述问题的处理效果。研究结果表明:采用设定切削参数时,切削步距对切削力造成最明显影响,每齿进给量次之,主轴转速影响程度最低。沿径向观察切削宽度发现,4.5 mm位置达到最低切削力,7.5 mm位置达到最高切削力。支持向量机(SVM)模型获得了最佳性能,能够实现精确预测的目标,防止发生过拟合的问题。初始参照支配解各参数都呈现正增益状态,表明采用此方法可以确保各方面都比初期参照更优,符合可靠性。

The milling performance of titanium alloy was tested experimentally.The cutting force variation caused by the combination of cutting parameters was analyzed by orthogonal test.Based on the comprehensive evaluation of material removal effect and cutting load,the heuristic algorithm containing multi-objective element was established,and the set of Pareto front parameters was determined,and the processing effects of each algorithm on the above problems were compared.The results show that the cutting step has the most obvious influence on the cutting force,followed by the feed per tooth,and the spindle speed has the least influence.The lowest cutting force was found at 4.5 mm and the highest was found at 7.5 mm.The support vector machine(SVM)model achieves the best performance and can achieve the goal of accurate prediction and prevent the problem of over-fitting.Each parameter of the initial reference dominant solution presents a positive gain state,indicating that the method in this paper can ensure that all aspects are better than the initial reference and conform to the reliability.