An approach to improve the performance of floating photovoltaic (FPV) systems by incorporating the complex interplay among platform motion, hydrodynamic loads, solar irradiance and other factors has been devised by researchers from Beijing University of Posts and Telecommunications (China) and Cranfield University (U.K.).

The artificial intelligence (AI)-driven digital twin framework for these systems uses a physical FPV twin deployed on a water surface, with sensors transmitting data to the cloud and to the digital version of the installation. The digital twin system was trained on data from 155 physical experiments, using a two-tier artificial neural network with a high-fidelity model and a reduced-order model.

The dataset generated, which reflects different incidence angles, wave amplitudes and wavelengths, was used to train the digital twin. The modeling scheme described in Solar Energy achieved coefficient of determination values of 0.9996 for photovoltaic surface temperature and 0.9189 for power output. It accurately captured oscillatory behaviors in surge and pitch motions, and reproduced rapid variations in mooring forces and transient power fluctuations.

“Three-dimensional performance maps generated by the trained models revealed strong nonlinear interactions between environmental inputs and system behavior,” the researchers explained. “For instance, heave motion peaked under wave lengths of 2.5–3.5 m and higher amplitudes (~0.0375 m), indicating resonant conditions. Power output was maximized when solar irradiance exceeded 340 W/m² at a 90° incidence angle, and PV temperature exceeded 75° C under the same conditions. These insights enable predictive optimization and enhance understanding of FPV performance under variable sea states.”