基于神经网络PID的纤维板调施胶控制系统的设计  

Design of Control System for Dosing and Applying Glue to MDF Based on Neural Network PID

作　　者：谢永华[1] 周宏威[1] 

机构地区：[1]东北林业大学,哈尔滨150040

出　　处：《东北林业大学学报》2011年第6期125-127,共3页Journal of Northeast Forestry University

基　　金：黑龙江省自然科学基金项目(F200920);中央高校基本科研业务专项基金(09AB09)

摘　　要：基于神经网络的优点,设计了神经网络作为主控制器的对系统进行辨识的纤维板调施胶控制系统和神经网络作为辅助控制器来修正PID控制器的纤维板调施胶控制系统。分析了两种神经网络控制器的性能优劣。仿真表明,神经网络作为辅助控制器调节PID比神经网络作为主控制器的超调小,系统稳定时间更短。Based on the advantages of neural network,two control systems for dosing and applying glue to Medium-density fiberboard（MDF） were designed using neural network as the main controller as well as using neural network control system as the auxiliary control unit to revise the main controller.Advantages and disadvantages of performances of the two neural network controllers were analyzed.Simulation results show that the neural network control system using the auxiliary control unit to revise the main controller has a more favorable effect than the control system using neural network as the main controller.

关 键 词：调施胶 神经网络 系统辨识 PID控制 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]