Model predictive control of a solar-thermal reactor

This paper presents the application of a Model Predictive Control (MPC) system to a solar-thermal reactor for carbon-steam gasification. The controller aimed at rejecting the disturbances in solar irradiation, caused by the presence of clouds, through the manipulation of the flow rates into the reactor. Simulations of a first-principles dynamic model were used to fit a linear, time-invariant model, to be used directly in the MPC algorithm. The MPC was tuned through simulations, until robust control was achieved. The proposed control system was compared to a fine-tuned multi-loop feedback control strategy. Simulations of the controlled reactor system showed that the linear MPC was capable of minimizing the effect of the disturbances within the entire operating region, outperforming a more conventional multi-loop feedback strategy. The results obtained suggest that the proposed controller is suitable for a real-world implementation. (c) 2014 Elsevier Ltd. All rights reserved.