Acta Chimica Slovaca, Vol.2, No.2, 2009, 21 - 36 Neural Network Predictive Control of a Chemical Reactor Anna Vasičkaninová*, Monika Bakošová Institute of Information Engineering, Automation and Mathematics, Faculty of Chemical and Food Technology, Slovak University of Technology, Radlinského 9, 81237 Bratislava, Slovakia monika.bakosova@stuba.sk Abstract Model Predictive Control (MPC) refers to a class of algorithms that compute a sequence of manipulated variable adjustments in order to optimize the future behaviour of a plant. MPC technology can now be found in a wide variety of application areas. The neural network predictive controller that is discussed in this paper uses a neural network model of a nonlinear plant to predict future plant performance. The controller calculates the control input that will optimize plant performance over a specified future time horizon. In the paper, simulation of the neural network based predictive control of the continuous stirred tank reactor is presented. The simulation results are compared with fuzzy and PID control. Keywords: model predictive control, fuzzy control, PID control, neural network, continuous stirred tank reactor Introduction Conventional process control systems utilize linear dynamic models. For highly nonlinear systems, control techniques directly based on nonlinear models can be expected to provide significantly improved performance. Model Predictive Control (MPC) concept has been extensively studied and widely accepted in industrial applications. The main reasons for such popularity of the predictive control strategies are the intuitiveness and the explicit constraint handling. The predictive controllers are used in many areas, where high-quality control is required, see e.g. Qin and Badgwell (1996), Qin and Badgwell (2000), Rawlings (2000). Model-based predictive control refers to a class of control algorithms, which are based on a process model. MBPC can be applied to such systems, as e.g. multivariable, non- minimum-phase, open-loop unstable, non-linear, or systems with long time delays.