Diagonal recurrent neural network based integrated control system of an industrial boiler

An integrated control system based on diagonal recurrent neural network is proposed in this paper to control an industrial boiler. The process inverse dynamic modelling technique is applied to design the proposed controller. A multilayer diagonal recurrent neural network is trained to identify the unknown inverse dynamic model of the boiler plant by a dynamic back propagation algorithm with an adaptive learning rate. The adaptive learning rate was developed by introducing a Lyapunov function. The new software controller is then developed for the integrated control system of the boiler using the weights of the trained network. The developed controller is tested by using the boiler output data that are not used as the training data. The output response of the developed controller is compared with the existing PID controller response of the plant. Test results and comparisons show gratifying performance of the developed controller