【摘要】 利用Taguchi方法对多工位锻造工艺参数进行显著性筛选,采用随机抽样的拉丁超立方(LHS)设计并进行数值模拟实验,运用最小二乘法(LS)和移动最小二乘法(MLS)建立抽样点的响应面模型,应用遗传算法进行迭代优化,并对2种建模方法加以对比.结果表明,MLS模型的误差比LS模型平均降低36.7%.多工位锻造零件经过近似模型方法优化,最大成形载荷从918 kN降至569 kN,成形质量显著提高.更多还原

【Abstract】 Taguchi method was used to select the most important factors of multistage forging process and a kind of random sampling method,Latin hypercube sampling was applied to design and execute the numerical simulation experiment.Furthermore,based on the samples,the moving least squares(MLS) and least square(LS) methods were adopted to establish the response surface model,which was used for the iterative optimization process with genetic algorithm.Comparison was carried out between the MLS and LS modeling methods,and the result shows 36.7% decrease of MLS model error compared with LS.The loads of multi-station forging part decrease from 918 kN to 569 kN after optimized by the approximation model optimization method,and the forming quality is improved significantly.更多还原