Reproduce “Multi-scale Dynamic Wake Modeling and Prediction of Floating Offshore Wind Turbines via Physics-Informed Neural Networks and Fourier Neural Operators” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Guodan Dong, Jianhua Qin, Chang Xu
Using high-fidelity LES data for a floating offshore wind turbine undergoing prescribed surge and pitch, the study compares physics-informed neural networks (PINNs) with Fourier neural operators (FNOs) for reconstructing and predicting unsteady wakes. Under the tested conditions, the FNO better retained small-scale flow structure and future-time accuracy, while training in about 15 minutes versus more than two hours for the PINN.
Key insights
- Both methods reproduced the broad wake motion, but the PINN produced smoother fields and tended to suppress small structures and velocity fluctuations.
- The PINN’s simplified two-dimensional physical constraints were applied to data from three-dimensional LES, while the FNO learned field mappings directly from the simulation data.
- Good agreement in overall wake shape did not guarantee accurate fluctuation or spectral predictions; these measures exposed shortcomings in the PINN’s representation of unsteady detail.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.