Rapid Forecasting of Carbon Storage and Migration using Neural Operators and Transfer Learning

Explore innovative techniques for predicting CO2 migration in storage reservoirs through this 51-minute lecture by Dr. Siddharth Misra, Associate Professor at Texas A&M University. Delve into the powerful combination of Fourier Neural Operator (FNO) networks and Transfer Learning, which significantly reduces computational time and data requirements for 3D CO2 plume migration predictions. Discover how this approach enables efficient and accurate analysis of CO2 saturation and pressure evolution in large reservoirs with diverse geological and engineering conditions. Examine the effectiveness of this technique through experiments on the 3D SACROC geomodel, showcasing a 75% reduction in training time and 50% decrease in data needs compared to conventional methods. Learn how this innovative approach can reduce traditional forecasting times from 40 minutes to just 1 minute, revolutionizing carbon storage and migration predictions in the field of geoscience and petroleum engineering.

Rapid Forecasting of Carbon Storage and Migration using Neural Operators and Transfer Learning