集气管压力系统的神经网络逆解耦复合控制

Neural Network Inverse Decoupling Compound Control of the Gas Collector Pressure System

针对集气管压力系统具有强干扰、强耦合、非线性、多参数等特性,依据逆系统解耦原理,分析了集气管压力系统数学模型的可逆性;采用对非线性具有较强逼近能力的BP神经网络,逼近集气管压力系统的逆系统;神经网络逆系统与原系统构成伪线性解耦复合系统,实现集气管压力系统的神经网络逆解耦复合控制。仿真结果表明该方法实现了系统解耦,具有一定的应用性。

For collectors＇pressure system strong interference,coupled,nonlinear,multi-parameter and oth-er characteristics,based on the inverse system decoupling principle,the reversibility of the mathematical model of the gas collectors＇pressure system is analyzed.BP neural network which has strong nonlinear ap-proximation ability is applied,to approximate inverse system of gas collectors＇pressure system.Neural net-work inverse system with the original system composes of the pseudo linear decoupling composite system. The neural network inverse decoupling control of gas collectors＆#39;pressure system is implemented.The simu-lation results show that this method realizes decoupling,has a certain application.