Design of nonlinear model predictive control with application to regenerative thermal oxidizer system

Abstract

A regenerative thermal oxidizer (RTO) is a treatment process of exhaust air. The RTO system absorbs heat from the exhaust gas and uses the captured heat to preheat the incoming gas stream and destroy air pollutants emitted from process exhaust streams at high temperatures. In order to regulate the gas stream output, temperature control loop is an important requirement. This paper presents an application of nonlinear model predictive control (NMPC) designed for RTO system. Applying the Wiener and Hammerstein model to NMPC can transform the design criterion into a quadratic function. The performance of NMPC based on Wiener and Hammerstein model is compared to linear MPC. Effects of each control parameter are also obtained. Simulation results indicate that NMPC provides better transient performance while error converges to zero.