Control of a selective catalytic reduction system based on NARMA-L2 model

Abstract

The plant of the selective catalytic reduction (SCR) system is characterized by significant nonlinearity, time delay and temperature sensitivity. In order to control the urea injection accurately, the (nonlinear auto regressive moving average) NARMA-L2 model based control is applied to the SCR system. In this paper, a data-based technique is taken and a model of the plant is identified on the basis of input-output data. Then the identified model is used to the design of a NARMA-L2 controller. Simulation of the NARMA-L2 model based control for the SCR system is presented to demonstrate the effectiveness and superiority. The comparison results show better performance over the traditional PID control.