基于模糊神经网络的梭式窑自适应控制系统应用研究

摘要

针对梭式窑温度控制过程存在非线性、时变性和不确定性等情况,设计了基于模糊神经网络的自适应控制(MRAC)系统并运用于梭式窑中,即采用RBF径向基网络作为梭式窑温度对象的辨识网络,它产生的Jacobin信息与控制误差一起提供给模糊神经网络控制器作为其学习信号,经过MATLAB仿真调整初始参数.最后运用VC++软件结合数据采集卡对燃气梭式窑进行控制.该控制系统能辨识出梭式窑时变性、非线性、长滞后性的对象特性,实时修正控制器参数,以适应对象特性的变化,在实际应用中取得了良好的控制效果.%After analyzing the non-linearity,time -varying and uncertainty during the temperature control of shuttle kilns,an adaptive controller(MRAC)based on Fuzzy Neural Network is designed and used in shuttle kilns.That is,by adopting radial basis RBF network as a shuttle kiln identify network,which pro-duces the Jacobin information and the control error together with the information provided to be a study sig-nal of the fuzzy neural network controller,and then the initial parameters is adjusted through MATLAB simulation.Eventually the gas shuttle kiln shall be controlled by applying YC++,combining with data acqui-sition card.In order to adapt to the change of object properties ,the system can identify the Object properties such as the nonlinearity, time -varying and uncertainly accurately, revise the controller parameters in real lime ,which control effect was verified to be good in practical application.