Mechanical and Civil Engineering Seminar: Special Seminar
Mechanical and Civil Engineering Seminar Series
Title: Can computers be trained as porous media experts?
Abstract: Transport phenomena in porous media occur within tiny pore spaces, and pore-scale imaging and modelling can elucidate the associated physical and chemical mechanisms. Micro-CT imaging and pore-scale modelling have developed rapidly over the last two decades, bridging the disciplines of geology, chemical engineering, image processing, and computational fluid dynamics. They have provided new pathways for understanding complex transport phenomena in underground geological formations and other porous media. However, several steps in this framework are time-consuming and subject to user bias. Machine learning and Convolutional Neural Networks (CNNs), as part of the broader field of Artificial Intelligence (AI), can be integrated into the framework of pore-scale modelling and imaging. The trade-off between sample size and image resolution, image segmentation, as well as the expensive computational cost associated with numerical simulations of fluid flow in the pore spaces, can be addressed using CNNs. I will demonstrate how we can recreate porous media images at super-resolution and explore porous media transport phenomena. Additionally, I will demonstrate the reliability and accuracy of CNNs in determining rock properties based on images of pore spaces. Finally, I will discuss challenges and opportunities for the development of machine-learning approaches in porous media applications.
Bio: Peyman Mostaghimi is a Professor of Minerals and Energy Resources at the University of New South Wales (UNSW), Australia, where he leads a multidisciplinary research group on Multiscale Transport in Porous Systems (MUTRIS). He is currently a Senior Fulbright Scholar and the Chair of the Council of the International Society for Porous Media. His research is focused on fluid dynamics and transport phenomena in porous media with application to geological carbon dioxide storage, subsurface hydrology, hydrogen storage, minerals and hydrocarbon recovery and groundwater modelling. He performs theoretical, numerical, and experimental research into the characterization of heterogeneous porous materials at different scales. He obtained his PhD from Imperial College London on transport phenomena in porous media.
NOTE: At this time, in-person Mechanical and Civil Engineering Lectures are open to all Caltech students/staff/faculty/visitors.