Session: 09-08-03 Floating Solar: Loads and Responses

Submission Number: 180130

Efficient Uncertainty Quantification of Extreme Loads in Floating Photovoltaic Arrays 

Floating photovoltaic (FPV) arrays are emerging as promising renewable energy technology, but their survival under extreme environmental conditions remains uncertain in light of recent failures. Response uncertainty is driven largely by nonlinear multi-body interactions that induce nonstationary behavior, broadened response spectra, and significant variability across ensembles and load cases. On the other hand, wind–wave coupling constrains the feasibility of conducting direct high-fidelity simulations for large-scale reliability studies because accurately resolving the coupled air–sea processes necessitates costly mesh refinements, long integration windows, and tightly synchronized solvers. To address these challenges, this work presents a framework for uncertainty quantification (UQ) of FPV arrays using a multi-fidelity surrogate model. High-fidelity hydrodynamic simulations are combined with reduced-order models to train a single machine-learning surrogate, based on Gaussian Process Regression (GPR), that maps environmental inputs (wind speed, wave height, and wave period) to multiple correlated structural responses, including platform motions, mooring loads, and connector forces. Environmental uncertainty is introduced through probabilistic wind–wave models and propagated efficiently through the surrogate to estimate variability in key structural responses under typhoon-like conditions. Case studies demonstrate that the proposed approach significantly reduces computational cost while maintaining predictive accuracy, providing a practical tool for the reliability-based design of FPV systems. The results highlight the critical influence of joint wind–wave loading uncertainty on FPV multi-body dynamics and offer guidance for resilience-oriented design criteria in large-scale deployments.

Presenting Author: Ding Peng Liu Independent Researcher

Presenting Author Biography: Ding Peng Liu, graduated from The University of Texas at Austin with a PhD in Civil Engineering. His study focuses on structural reliability and surrogate modeling of offshore marine structures. He just finished one year of postdoctoral research in National University of Singapore and now started his career as an assistant professor in National Taiwan University.

Authors:

Ding Peng Liu Independent Researcher
Shun-Wen Zheng Department of Hydraulic and Ocean Engineering, National Cheng Kung University
Taemin Heo Hongik University
Wen-Huai Tsao McNeese State University
Nai-Chi Chen Department of Hydraulic and Ocean Engineering, National Cheng Kung University
Lance Manuel The University of Texas at Austin