论文部分内容阅读
成形极限图(FLD)是评价金属板材成形能力的重要工具。为快速的建立拼焊板(TWB)成形极限图,建立基于人工神经网络(ANN)拼焊板FLD的预测模型。采用试验设计和有限元法获得训练样本,L-M算法对样本数据进行训练,建立了FLD预测模型并与物理试验结果对比。基于预测模型,分析了摩擦系数对拼焊板最小极限应变的影响。结果表明,基于ANN预测的拼焊板FLD与试验结果吻合,主应变的相对误差最大为8.71%。摩擦系数f对最小极限应变影响较大,f从0增大到0.12时,最小极限应变先增大后减小,并在摩擦系数f=0.06附近出现极小值。
Forming limit map (FLD) is an important tool to evaluate the forming ability of sheet metal. In order to quickly establish the TWB forming limit map, a prediction model based on Artificial Neural Network (ANN) tailored welding (FLD) was established. The training samples were obtained by experimental design and finite element method. The L-M algorithm was used to train the sample data. The FLD prediction model was established and compared with the physical experiment results. Based on the prediction model, the influence of the friction coefficient on the minimum ultimate strain of the tailor welded plate was analyzed. The results show that the FLD of tailor-welded blanks based on ANN is in good agreement with the experimental results, and the maximum relative error of the main strain is 8.71%. The friction coefficient f has a great influence on the minimum ultimate strain. When f increases from 0 to 0.12, the minimum ultimate strain first increases and then decreases, and the minimum value appears near the friction coefficient f = 0.06.