论文部分内容阅读
针对车用等速万向节滑套在实际冷精整后杯壁端部出现不平整的问题,基于Deform有限元分析软件,提出Kriging代理模型与粒子群算法相结合的方法,以杯壁高度差为优化目标,使用拉丁超立方抽样法对预成形件尺寸参数进行抽样、完成实验设计,通过数值模拟得到仿真结果,并拟合成相应的Kriging模型。采用MATLAB将Kriging模型与粒子群算法相结合,对因素进行全局寻优,得到最小杯壁高度差的预成形件尺寸。产品试制后,杯壁高度差由原来工艺的12 mm缩小为3 mm,验证了此优化方法的实用性和正确性,提高了锻件的材料利用率。
Aiming at the problem of unevenness in the end of cup wall after the actual cold finishing of the automotive universal joint sleeve, the method of combining Kriging agent model and particle swarm optimization is proposed based on Deform finite element analysis software, The difference is the optimization objective. The dimensions of the preform are sampled using Latin hypercube sampling method. The experimental design is completed. The simulation results are obtained by numerical simulation and the corresponding Kriging model is fitted. Using MATLAB to combine Kriging model with particle swarm optimization algorithm, the factors are globally optimized to obtain the minimum preform size of the cup height difference. After trial production, the height difference of the cup wall was reduced from 12 mm to 3 mm in the original process, which verified the practicability and correctness of this optimization method and improved the material utilization rate of forgings.