Behnam Jafarpour

Research Summary

Dr. Jafarpour's research focuses on integrating machine learning with subsurface energy science to develop physics-informed deep learning models for complex energy systems. His work addresses challenges such as high-dimensionality, limited data, and multiphysics coupling in applications including carbon storage, geothermal energy, and hydrocarbon reservoirs. Central to his approach are hybrid models that embed domain knowledge and physical constraints into deep learning frameworks, enabling predictions that are both computationally efficient and physically consistent. His group develops surrogate and reduced-order models, latent-variable representations, and transformer-based architectures to capture spatiotemporal dynamics in heterogeneous subsurface environments. These models are applied to forward prediction, real-time subsurface monitoring, characterization, and decision-making, providing scalable solutions for energy operations.His work bridges methodological innovation and practical impact, producing tools that are deployable in industry and training students to operate at the interface of engineering, geoscience, and data science.