University of Wisconsin AI model enhances CSP operations at Solana

Researchers from the University of Wisconsin-Madison developed a high-fidelity simulation model for parabolic trough solar fields to address operational challenges in Concentrating Solar Power (CSP) plants. The model, based on a neural network, computed heat absorption with a 0.3% error compared to a detailed model while increasing simulation speed by a factor of 100. It was validated using data from the Solana Solar Generating Station in Arizona, demonstrating accuracy in temperature computation and net optical efficiency. The model accounted for time-varying collector defocusing and was tested in case studies for training and operational optimization. It served as both an operator training simulator and a tool for improving solar field control strategies.