Performance analysis of a predictive control strategy for a simulated central tower CSP plant using real-time solar radiation measurements

Central tower Concentrated Solar Power (CSP) technology is a renewable electricity generation method that uses heliostats to direct solar thermal radiation toward a central receiver. While it is considered a promising technology it still faces several challenges. One key issue is the implementation of a control strategy that can effectively manage the short-term variability of solar irradiance caused by cloud cover. This study presents the design and implementation of a cost-effective real-time data acquisition and processing system that uses a commercial fisheye camera and solar radiation sensors. The system allows the real-time creation of shadow maps and short-term forecasts, achieving prediction accuracies of 64.4% and 63.3%, respectively. Additionally, a simplified methodology is introduced for the dynamic identification of the thermal-optical model of the central receiver, making it suitable for real-time control applications. A control strategy is also developed, which combines a generalized predictive controller with a complementary PID. This is to maintain the fluid’s outlet temperature and heat flux at their reference values. It is done while integrating spatially distributed solar radiation forecasts to enhance responsiveness to disturbances, improve operational stability, and protect the central receiver. The system’s performance is validated using actual measured data across various operating scenarios. The implemented strategy results in a significant reduction, up to 93%, in temperature fluctuations at the receiver when compared to conventional methods, while maintaining safe levels of heat flux. These results show the practical feasibility and strong performance of the proposed hybrid predictive control system, highlighting its potential to improve operational reliability, extend equipment longevity, and improve overall efficiency in central-tower CSP plants.