Unconventional/complex reservoirs

Machine-Learning Techniques Characterize Source-Rock Images at the Pore Scale

The paper demonstrates the ability of deep-learning generative models to enable new shale-characterization methods.

Image-based characterization and deep-learning work flow
Fig. 1—Image-based characterization and deep-learning work flow. Images are first acquired for a source-rock sample from one or multiple modalities. The image data then are curated, including alignment, normalization, and subsampling, to create a data set suitable for machine-learning model training. After training the image-processing model, new images are generated or predicted. The synthesized images are post-processed, including segmentation into pore and mineral phases, and the morphological and petrophysical properties of the sample evaluated.

Nanoimaging techniques for characterizing pore-scale structure of shales trade off between high-resolution/high-contrast sample-destructive imaging modalities and low-contrast/low-resolution sample-preserving modalities. Acquisition of nanoscale images also is often time-consuming and expensive. In the complete paper, the author introduces methods for overcoming these challenges in image-based characterization of the fabric of shale source rocks using deep-learning models.


A promising application of data-driven scale-translation techniques is nanoscale imaging. This application is important for studying shales because of the importance of nanoporosity in shale gas storage.

Continue Reading with SPE Membership
SPE Members: Please sign in at the top of the page for access to this member-exclusive content. If you are not a member and you find JPT content valuable, we encourage you to become a part of the SPE member community to gain full access.