Title: Artificial Intelligence Across Disciplines: Research, Innovation, and Intelligent Solutions
Editors: Dr. Subita Bhagat, Dr. A. Balamurugan, Dr. P. Krishna Kumar, and Mrs. S. Nandhini Devi
ISBN: 978-81-69857-83-3
Chapter: 11
DOI: https://doi.org/10.59646/815/11
Authors: A. Balamurugan, M. Bhuvaneswari, and M. Sudha
Abstract
Modern planetary challenges necessitate computational paradigms capable of resolving non-linear, multi-scale physical dynamics across coupled energetic and climatic systems. This chapter investigates the convergence of physical law and deep learning to model, forecast, and optimize atmospheric-energy interactions. By integrating physics-informed neural networks (PINNs) with Fourier neural operators (FNOs), we formulate an end-to-end framework that addresses high-dimensional, turbulence-governed environmental dynamics. Applied to an empirical offshore renewable energy zone, the architecture directly internalizes Navier-Stokes continuity and thermodynamic conservation equations within the neural loss topology. This formulation minimizes spectral dispersion errors, preserves spatio-temporal coherence, and yields physically consistent wind-wake and local boundary-layer microclimate forecasts. Numerical experiments demonstrate that our physics-informed operator method achieves near-numerical precision while reducing computational inference latencies by over three orders of magnitude relative to traditional large-eddy simulations. The chapter concludes with rigorous operational guidelines for integrating these models into real-time grid balancing and resilient climate mitigation infrastructures.
Keywords: Physics-Informed Neural Networks, Fourier Neural Operators, Climate Dynamics, Environmental Physics, Renewable Energy Optimization, Navier-Stokes Constraints