Abstract

In this seminar, we present an innovative deep-learning-based generative method as an alternative to traditional subsurface earth model (geomodel) generation procedures. This method, utilizing advanced autoencoder and autoregressor techniques, enables the rapid generation of thousands of geomodels similar to a user-defined source geomodel in seconds. This breakthrough approach facilitates unprecedented control and manipulation of the generated geomodels. Furthermore, we explore the application of neural operator and transfer learning to rapidly forecast spatiotemporal evolutions of saturation and pressure in large, heterogeneous reservoirs. Demonstrated on the field-scale SACROC geomodel with permeability-porosity heterogeneity, this novel workflow significantly reduces computational costs and simulation time for reservoirs with varying injector-producer locations and injection rates. This efficiency is crucial for large-scale, probabilistic assessments of carbon storage and containment, offering the potential for low-cost adaptive monitoring over extended periods.

Bio

Dr. Sid Misra, an Associate Professor at Texas A&M University, specializes in subsurface monitoring and forecasting for the exploration and production of earth resources. With an undergraduate degree in electrical engineering from the Indian Institute of Technology, Bombay, and a Ph.D. in petroleum and geosystems engineering from The University of Texas at Austin, Dr. Misra has authored two books and spearheaded

References or Related Papers

Falola, Y., Misra, S., & Nunez, A. C. (2023, October). Rapid High-Fidelity Forecasting for Geological Carbon Storage Using Neural Operator and Transfer Learning. In Abu Dhabi International Petroleum Exhibition and Conference (p. D011S020R003). SPE.

Misra, S., Chen, J., Falola, Y., Churilova, P., Huang, C. K., & Delgado, J. (2023, June). Massive Geomodel Compression and Rapid Geomodel Generation Using Advanced Autoencoders and Autoregressive Neural Networks. In SPE EuropEC-Europe Energy Conference featured at the 84th EAGE Annual Conference & Exhibition. OnePetro.

Chen, J., Huang, C. K., Delgado, J. F., & Misra, S. (2023). Generating Subsurface Earth Models using Discrete Representation Learning and Deep Autoregressive Network. arXiv preprint arXiv:2302.02594.