Variable flow heating control system based on neural network decoupling

The diagonal recurrent neural network (DRNN) has been brought up and applied to the variable flow heating system with features of being nonlinear, time-varying, uncertainty and serious coupling among variations. This application of the network can overcome and eliminate the coupling among the variations, when the control process of heating system is either static or dynamic. What is more, the design of a neural network decoupling controller has also been given. After the Matlab simulation, it shows that the purposes of shortening the dynamic time and completely static decoupling can be achieved and to our satisfaction.