基于神经网络和SQP算法的焊接飞溅优化控制
OPTIMIZING CONTROL OF WELDING PROCESS BASED ON NEURAL NETWORKS AND SQP ALGORITHM
-
摘要: 提出了神经网络和逐次二次规划法的在线优化控制方法,在神经网络建模的基础上,把SQP算法和直接搜索两种寻优方法作了比较;用NN-SQP方法对焊接过程飞溅进行了控制,取得较好的效果.并提出了用电弧声表征焊接飞溅量的软测量方法.Abstract: An on-line optimizing control approach based on neural networks and SQP algorithm is proposed, artificial neural network is used for buiding a model based on which optimization methods of SQP algorithm and direct search are compared. The control of welding process spatter using NN-SQP algorithm is studied and better effect is grained. The welding spatter is showed by the electricity arc sound in this work.