等离子喷涂过程的BP神经网络自适应控制*

Adaptive Control of Plasma Spraying Process Using BP Neural Network

中文关键词  等离子喷涂  自适应控制  ＢＰ神经网络

英文关键词  plasma spraying  adaptive control  BP Neural Network

中文摘要

      分析了ＢＰ网络辨识器和控制器的常规训练方法存在的不足之处及其限制，提出了一种通过ＢＰ网格辨识器的灵敏度网络反传对象的误差来训练ＢＰ网络控制器的新方法.根据ＢＰ网络和自适应控制的相容性，提出了一种新的ＢＰ网络自适应控制结构，并利用ＢＰ网络辩识器的灵敏度网络训练ＢＰ网络控制器的思想，来调整ＢＰ网络控制器的参数.等离子喷涂过程控制模型的训练研究表明其效果良好.

英文摘要

      The drawback and limit of the general methods for training BP network identifier and control are analyzed，a new method for training BP network controller by error back propagation through sensitivity network of BP network identifier is put forward. On basis of resemblance between BP network and adaptive control，a new adaptive control structure using BP network is put fir ward.The parameters of BP network controller by the idea of training BP network controller with error back propagation through the sensitivity network of BP network identifier.The training results of plasma spraying process control model clearly show its excellent effect.