Physics-Guided Deep Learning for Prediction of Energy Production from Geothermal Reservoirs

Deep learning in CO2 geological utilization and storage: Recent advances and perspectives

“…By combining physical laws (Navier-Stokes equations, multiphase fluid seepage equations, water-rock reactions) with deep neural networks, Physics-informed deep learning (PIDL) can learn more generalizable models with fewer data samples and can accelerate the training rate (Wang et al, 2024). PIDL is particularly well suited to scientific computational and engineering problems that can be described by explicit physical laws (Qin et al, 2024). In summary, PIDL incorporates data-and physics-driven synergies that offer promising directions for advancing the science of CO 2 utilization and storage modeling.…”