基于模型参考神经网络实现锅炉水位自适应控制  

The Water Level Control System of Boiler Based on Reference Neural Network

作　　者：朱翚[1] 芮延年[1] 刘开强[1] 

机构地区：[1]苏州大学机电学院,江苏苏州215021

出　　处：《中国制造业信息化（学术版）》2007年第8期51-54,58,共5页

摘　　要：锅炉水位是影响锅炉安全运行的重要参数,而水位控制系统是一个复杂的非线性系统。设计了一个稳定的参考模型来表达所期望的汽包水位特性,再通过构建神经网络控制器,控制锅炉水位输出曲线匹配期望的锅炉水位特性。神经网络控制器由2个神经网络组成:被控对象辨识神经网络和控制神经网络,采用反向传播学习算法作为神经网络训练算法。仿真结果表明,采用该神经网络控制器的系统输出性能明显优于采用传统控制器的系统。The boiler water level control system is a complex nonlinear system. It designs a neural network controller to improve the boiler water level control system. This system comprises of two neural networks： the plant model network and the controller network. The former is used to identify the water level control system while the latter to control the output of the plant to follow the reference model. Back - propagation algorithm is applied to train these two networks. The result shows that the water level control system is satisfied control performance.

关 键 词：锅炉水位 自适应控制 辨识神经网络 模型参考神经网络控制器 

分 类 号：TP202[自动化与计算机技术—检测技术与自动化装置]