模糊神经网络在非线性电站制粉控制系统中的应用

火电厂制粉系统是一个相互关联强耦合的复杂多输入多输出控制对象,运行时具有存滞后、大惯性和非线性的特点,其动态特性非常复杂,数学模型难以建立,并且数学模型随煤质、外部环境等因素变化较大.因此,提出一种基于模糊神经网络的2级控制方案,第一级采用神经网络控制,第二级采用模糊控制.该方案很好地解决了在多变量模糊控制系统中的控制规则多、维数灾难问题,在理论上为多变量、高维数非线性模糊控制系统提出了一种新的控制方案.实验室测试结果表明,如果将其应用到中储式制粉系统中,将提高电站制粉系统的自动化水平及经济运行指标.与传统的手动及PID定值控制比较,该方法大大改善了电站制粉系统的控制性能,是一种很有实用价值的控制方法.

Application of fuzzy neural network on nonlinear pulverizing control system of power plant

Pulverizing system of power plant is a sort of complicated multiple inputs and outputs control object,it takes on the distinct characteristics of strong coupling,interrelation,pure delay,big inertia,nonlinear during operation.At the same time,its dynamic characteristic is very complicated,it is very difficult to build its mathematics model,furthermore,mathematics model varies with the factors of coal quality,external environment,etc..According to above characteristics,puts forward a two-stage control project based on fuzzy neural network,first-stage adopts neural network control,second-stage adopts fuzzy control.This project resolves well the questions of many control rules and curse of dimensionality in multi-variable fuzzy control system,provides a new control project for multi-variable,nonlinear fuzzy control system on the theory.Experiment results show applying it on the pulverizing system of power plant will advance the automatization level and economic index.Comparing with traditional hand control and PID fixed value control,this project improves the control performance of pulverizing system of power plant,it is a control way with practical value.