Neural network based flow controller

Abstract

Presents an application of a backpropagation neural network to control flow rate. The backpropagation neural network is trained to learn an inverse dynamics model of a flow control system and then configured as a direct controller of the process. The ability of the neural network to learn the inverse dynamics model of the process plant is based on input vectors with no a priori knowledge regarding process dynamics. For interfacing a flow transmitter to a personal computer and also for giving controller output to the valve from a personal computer, an add on card and signal conditioning cards are designed. The process is tested by using the neural network controller. Experimental results show that the neural network gives satisfactory results for different set points and also exhibits better performance for load disturbance.