Modeling and control of nonlinear nitrogen oxide decomposition process

This paper presents a modeling and predictive control approach for a non-stationary nonlinear nitrogen oxide (NO_x) decomposition process whose dynamics depend on the time-varying working-points and may be locally linearized. An off-line identified hybrid pseudo-linear ARX model (RBF-ARX model), which is composed of Gaussian radial basis function (RBF) networks and linear ARX model structure, is utilized to describe the process behavior. On the basis of the RBF-ARX model, a long range predictive control (RBF-ARX-MPC) strategy that does not require on-line parameter estimation is investigated for this kind of nonlinear process. Stability of the controller proposed under certain condition is discussed. Particularly, an industrial experiment result is also given to show satisfactory modeling precision and control performance obtained by the proposed approach in real industrial application.