(598l) Model Predictive Control of Solar-Thermal Reactors

Saade Saade, E. - Presenter, University of Colorado
Clough, D. E., University of Colorado
Weimer, A. W., University of Colorado at Boulder

A transport tube reactor uses concentrated solar irradiation in order to drive the endothermic biomass gasification reaction for the production of synthesis gas (H2, CO and CO2). This process is extremely dependent on the available sunlight and on the absence of clouds. Transient clouds constitute a significant disturbance for the steady operation of the reactor, causing unnecessary shut-downs and start-ups as well as complications in the separation processes downstream of the reactor. In order to make the process more feasible, flow manipulation is required to ensure steady conversion and a more efficient operation. Thus, a robust control system that allows continuous high performance operation of the reactor is required.

A model predictive control system was developed for the process. A digital charge coupled device (CCD) camera captures images of the sky at a sampling interval of one minute. These images are analyzed and used to anticipate the incoming solar irradiation at future instances. The predicted solar irradiation is used as an input for the model predictive controller (MPC), which ensures steady conversion despite transients in solar irradiation through the manipulation of the flow rates into the reactor. The control algorithm was developed based on linear time series models for different operating points of the reactor. These linear models were fitted to simulation results of a dynamic model of the process that was developed based on unsteady mass and energy balances. The control system was tested using SIMULINK and it was implemented at the High Flux Solar Furnace (HFSF) at the National Renewable Energy Laboratory (NREL).  This poster describes the development of the control algorithm and presents the results obtained after its implementation.