基于GR神经网络的汽车U形纵梁多工步冲压成形回弹预测分析
Springback Prediction in the U-shaped Auto Longeron multi-steps Stamping Process Based on the GR Neural Network
王梦寒,刘文,赖啸
点击:4840次 下载:0次
作者单位:重庆大学,宜宾职业技术学院
中文关键字:汽车纵梁;多工步;神经网络;回弹预测
英文关键字:auto longeron; multi-operation; neural network; Springback prediction
中文摘要:针对纵梁件单工步成形回弹预测误差较大的缺陷,本文提出对汽车纵梁件进行多工步成形回弹控制分析,在数值模拟的基础上,建立了基于广义回归神经网络(GRNN)回弹预测模型,并运用该模型对不同凹模圆角与压边力等重要成形工艺参数下的回弹值进行了模拟预测。结果表明,论文采用的广义回归神经网络模型的预测值与模拟试验值有较好的吻合度,说明广义回归神经网络模型能够准确的预测纵梁多工序后回弹分布。
英文摘要:According to the drawback that the prediction error is large for the forming spring-back of longeron part after single operation, this paper puts forward the control analysis for multi-steps forming springback of automobile longeron part, established generalized regression neural network (GRNN) springback prediction model based on numerical simulation, and used the model to do some simulated prediction for springback value under some important technological parameters, such as different fillet of dies and blank-holder force .The results show that prediction value of generalized regression neural network model and simulated test value have good inosculation, this proves that the generalized regression neural network model can accurately predict the springback distribution of longeron after multistep .