INTELLIGENT CONTROL SYSTEMS, MPC SOLAR TRACKER

KOM YTU Department of Control and Automation Engineering MPC WITH DUAL AXIS SOLAR TRACKER USING MATLAB By Muntasser MOSLEH

MPC WITH DUAL AXIS SOLAR TRACKER USING MATLAB

This work focuses on the simulation of a photo voltaic (PV) application technology in harvesting renewable energy from solar radiation, and the efforts to improve its efficiency through various technological developments. One such development is solar tracking, which can be achieved through single axis or dual axis models. The simulation proposes using model predictive control (MPC) applied to a dual axis solar tracker using Matlab/Simulink, which is capable of producing continuous results and handling various systematic process control limits. MPC is chosen due to its ability to predict future process outputs, consider existing constraints, and control various levels of the process structure, including multivariable cases. Overall, the simulation showcases the potential benefits of applying MPC to improve the efficiency and performance of dual axis solar trackers.

Solar trajectory is determined by two celestial angles, altitude and azimuth, which form the basis of the ecliptic coordinate system. The Sun's position is expressed as the ecliptic longitude, ranging from 0° to 360°. Photo voltaic (PV) panels are driven by a dual axis motor system that follows the Sun's altitude and azimuth. The system's input is voltage and its output is angle. A state space model is necessary for model predictive control (MPC) controllers. The model is used to predict system outputs based on current and previous values, with an optimizer minimizing an objective function to obtain optimal input. The MPC block uses reference angles to determine the optimal input. The simulation shows that the MPC controller performs well in tracking the solar trajectory, as evidenced by the similarity between the predicted and actual curves.

Acknowledgments

I would like to express my gratitude to the students of the Intelligent Control Systems course of the YTÜ Control and Automation Engineering department, Class of Fall 2022, whose dedication and hard work made this project possible. I am also deeply thankful to Doctors Marco Rossi, Julia Hoerner, and Melda Ulusoy for their invaluable contributions.