铝合金板快速加热弯曲的参数预测

Parameter prediction of bending of aluminum alloy sheet induced by laser prompt heating

基于BP神经网络平台,建立了铝合金板快速加热弯曲的角度预测BP网络模型,实现了脉冲激光加工工艺的参数控制与优化。通过试验获得样本数据,将试验样本数据用于BP网络的训练,利用训练好的BP网络对非线性的样本数据规律进行拟合,对脉冲激光弯曲角度和工艺参数进行准确的预测,预测误差范围可控制在<5～8%,研究结果为实际生产中精密成形提供了有效的理论与试验依据。

Based on the basic platform of BP neural network, a BP network model was founded to predict the bending angle in the process of laser bending of aluminum alloy sheet to optimize laser bending parameter control. The sample data obtained from experiment were used to train BP network. The nonlinear regularities of sample data were fitted through trained BP network, the predicted results in- clude laser bending angles and laser bending parameters. Experimental results indicate the prediction allowance is controlled less than 5～8%, it can provide effective foundation both theory and experiment for industry purpose.