[H-SC07] CCUS (Carbon Dioxide Capture, Utilization, and Storage) for Climate Mitigation

[HSC07-P02] Application of Dimensionality Reduction Methods for the Classification of Porous Rock Structures

*Masashige Shiga¹, Masao Sorai¹ (1.Research Institute for Geo-Resources and Environment, National Institute of Advanced Industrial Science and Technology)

Keywords:CO2 geological storage, CT image, dimensionality reduction, machine learning

Understanding and predicting the behavior of CO₂ and formation water during the process of CO₂ underground storage is significantly facilitated by rock CT image data. It is not straightforward to analyze these multidimensional data, but one effective solution is dimensionality reduction through machine learning. In this study, various dimensionality reduction methods were applied to actual rock CT images and similar model structures to attempt classification of the structures. The extraction of features through dimensionality reduction not only deepens our understanding of rock structures but also plays a crucial role in improving the accuracy and efficiency of CT image reconstruction using deep generative models, which has been actively pursued in recent years. We plan to report results including those from images reconstructed using deep generative models.