基于主成分分析与BP神经网络的激光拼焊板力学性能预测
Mechanical Property Forecasting of Laser Tailored Welded Blank based on Principal Component Analysis and BP Neural Network
陈楼
点击:2342次 下载:0次
作者单位:江苏大学先进成型研究所
中文关键字:激光拼焊板; 力学性能预测; 主成分分析; BP神经网络
英文关键字:Laser tailor-welded blanks; Mechanical property prediction;PCA; BP ;
中文摘要:为了预测并控制激光拼焊板的力学性能,本文通过对0.8~1.5 mm的St12板及其镀锌板进行差厚、等厚拼焊,在此基础上建立了以焊接工艺参数为输入变量的基于主成分分析的BP神经网络拼焊板力学性能预测模型。通过实例验证表明,本文所建预测模型对拼焊板抗拉强度及延伸率的预测精度均达91%以上。充分表明该模型与试验结果吻合良好,验证了该预测模型的合理性及适用性。
英文摘要:For prediction and control of mechanical property of laser tailor-welded blanks, On the basis of the research of 0.8~1.5 mm St12 steel and galvanized steel welding, prediction models of mechanical properties of tailor-welded blanks trained with welding processes as input variables were put forward based on Principal Component Analysis and improved BP Neural Network. According to testing result, the forecasting accuracy of tensile strength and elongation was above 91%. Therefore,the proposed models are reasonable and applicable to examine deep-drawing steel for demonstrative purposes.