论文部分内容阅读
基于有限元数值模拟软件LS-DYNAFORM,对拼焊板方盒形件拉深成形进行模拟研究。通过改变拉深成形过程中压边力这一最重要且易于控制的工艺参数,寻求拼焊板方盒形件拉深成形时较优的变压边力曲线加载形式。为预测不同工艺参数下拼焊板方盒形件拉深成形时的较优压边力加载曲线,建立了变压边力的BP神经网络预测模型,并将该模型预测的结果与数值模拟得到的结果进行对比分析。研究结果表明,拼焊板薄板采用变压边力、厚板采用恒定压边力、且薄板压边力不小于厚板压边力的加载形式,拼焊板成形件整体质量较好,焊缝移动量较小;神经网络预测模型能较好的预测拼焊板方盒形件拉深成形时的变压边力,与数值模拟结果的最大相对误差在12.3%以内。
Based on the finite element numerical simulation software LS-DYNAFORM, a simulation study on the blank box-forming of the tailor welded plate was carried out. By changing the most important and easy-to-control process parameters of blank holder force in the process of deep drawing, the better loading condition of blank holder edge curve during blank-forming of box blank is sought out. In order to predict the loading curve of the blank holder under different process parameters, the BP neural network prediction model of variable blank holder force was established and the results of the model prediction and the numerical simulation were obtained The results of comparative analysis. The results show that the tailor-welded blank plate adopts the variable blank holder force, the thick blank holder adopts the constant blank holder force, and the blank holder force is not less than the blank holder blank load. The overall quality of the tailor-welded blank holder is better. And the moving amount is small. The neural network prediction model can better predict the percussion force of blank box during forming, and the maximum relative error between numerical simulation results and 12.3%.