气垫式流浆箱解耦控制的分析及应用

The Analysis and Application of Decoupling Control for Aircushion Head Box

关键词:  气垫式流浆箱  神经网络  解耦控制

Key Words:head box  the neural network  decoupling control

摘要:在建立气垫式流浆箱数学模型的基础上,利用神经网络原理设计了神经网络解耦控制器。根据总压和浆位给定值以及其输出值,通过自学习、调整网络权值来实现闭环控制的神经网络解耦控制思路,将强耦合的总压和浆位分解为两个单回路PID闭环控制系统。采用PID控制算法对解耦后的两个单回路闭环系统进行常规控制。实际运行效果表明，该控制系统可使总压和浆位的调整互不影响，且系统稳定。

Abstract:On the basis of the established mathematical model of aircushion head box neural network decoupling controller was designed based on neural network principle. According to the total pressure, a given value of pulp suspensions level and its output value, through self-study, adjusting the network weights to realize closed loop control of neural network decoupling control, the strong coupled total pressure and pulp suspensions level was decomposed into two single loop PID closed-loop control systems. It successfully solved the control difficult of strong coupling of total pressure and pulp suspensions level.