论文部分内容阅读
对304不锈钢试样进行了激光打孔试验,使用形貌仪测得了孔截面粗糙度参数,并通过反向传播神经网络,建立了基于激光功率、脉冲频率和离焦量三个工艺参数与孔表面粗糙度之间关系的神经网络预测模型。利用大量试验数据对样本进行网络训练,证实了该人工神经网络模型预测精度高,预测误差控制在6%左右,最大误差不超过8.08%。该模型可以准确地预测激光打孔表面的粗糙度和有效地缩短激光打孔作业的准备周期。
The laser drilling test was carried out on the 304 stainless steel sample. The roughness parameters of the hole section were measured by using a profilometer. The three parameters of the laser power, the pulse frequency and the defocus amount were established through the back propagation neural network. Neural Network Prediction Model of Relationship Between Surface Roughness. Using a large amount of experimental data to conduct network training on the samples, it is proved that the artificial neural network model has high prediction accuracy, the prediction error is controlled at about 6% and the maximum error is no more than 8.08%. The model can accurately predict the roughness of laser drilling surface and effectively shorten the preparation period of laser drilling.