Dynamic Simulation and Direct Adaptive Control of a Fixed-Bed Reactor

摘要

This paper considers the conversion control problem of a fixed-bed reactor. A finite element method using the software of COMSOL Multiphysics is utilized to investigate the system dynamics of the reactor which are described by a system of partial differential equations (PDE). To circumvent the runaway phenomena of the hot spot temperature and to achieve the stable conversion control, a direct adaptive control strategy is suggested for a fixed-bed reactor having highly exothermic reactions. A bounded single neuron controller is implemented in the control system to regulate the exit conversion rate. In associated with a simple yet stable parameter tuning algorithm, the controller is able to learn to control the fixed-bed reactor adaptively with using the information of process output errors. The proposed scheme is directly applicable and easy to implement; it does not rely on a complex process PDE model and the only a priori process knowledge is the system output response direction. The control performance of the proposed scheme is compared with an IMC-PI controller. Extensive simulation results show that the direct adaptive control scheme outperforms the conventional linear controller, especially when facing with unexpected system variations.