基于神经网络的板料冲压回弹预测系统的研究

Research on ANN-based Model Used to Predict the Springback of Stamping

文中以NUMISHEET’93标准考题中U形板料作为研究对象,将正交试验法和数值模拟法结合,产生各工艺参数组合下的冲压回弹量,以这些仿真试验结果作为神经网络的训练样本,在NeuroSolution软件环境下建立输入为工艺参数、输出为冲压回弹量的神经网络模型预测模型,以测试样本验证模型的预测精度作为反馈信息,构成闭环系统。该方法为板料冲压工艺参数的优化提供了有效的工具。

In this paper, U-shape sheet metal of Numisheet＇93 test is chosen to be a test object and suitable orthogonal test table is chosen to arrange the stamping experiments in CAE environment, The CAE results of simulating are used as teh training data to obtain a neural network model, which is implemented using the NeuroSolution software, for predicting the springback values, when different stamping process parameters are given.Test samples are used to verify the precision of the model and as feedback information. All of these make up of the closed loop system. This method provides an effective way for the process parameters optimization in stamping.