Porosity and permeability prediction from petrographic point-counting data using machine learning

2026-05-05 · Source: A Geodyssey – Geoscience Text Analytics and Enterprise Search Research · Field: Science & Research — Environmental Science & Earth Systems, Artificial Intelligence & Machine Learning · Depth: Advanced, quick

Summary

A new study from the Karlsruhe Institute of Technology demonstrates that machine learning models can accurately predict reservoir porosity and permeability. Researchers trained these models using petrographic point-counting data, which are detailed mineral-by-mineral descriptions of rock composition. This data has been routinely collected by geoscientists for decades but has been underutilized for predictive modeling. The approach offers a low-cost method to re-analyze extensive legacy petrographic archives from the oil and gas industry, generating new insights for reservoir characterization. This technique is particularly relevant for the growing geothermal energy and underground gas storage sectors. The study serves as a proof of concept, with blind-well validation identified as the crucial next step.

Key takeaway

For geoscientists and reservoir engineers evaluating new energy projects or optimizing existing assets, this study suggests that your legacy petrographic point-counting data holds significant untapped value. You should consider integrating machine learning techniques to extract predictive insights into porosity and permeability, potentially reducing exploration costs and enhancing reservoir understanding for geothermal, oil, and gas applications.

Key insights

Machine learning can predict reservoir properties from historical petrographic point-counting data.

Principles

• Legacy data holds untapped predictive value.
• Mineral composition dictates reservoir quality.

Method

Machine learning models are trained on petrographic point-counting data to predict porosity and permeability, potentially using SHAP values for interpretability.

In practice

• Re-analyze existing petrographic archives.
• Evaluate reservoir properties from cuttings.
• Apply to geothermal and gas storage.

Topics

• Machine Learning
• Porosity Prediction
• Permeability Prediction
• Petrographic Point-Counting
• Reservoir Characterization

Best for: AI Scientist, Research Scientist, Domain Expert

Related on AIssential

• Machine learning for sustainable geoenergy — Machine learning can enhance sustainable geoenergy projects by addressing data scarcity and uncertainty through physics-informed methods.
• How AI and Advanced Tech are Reshaping Mineral Exploration — AI and advanced imaging transform mineral exploration by integrating diverse data, enhancing subsurface understanding, and improving safety.
• Physics-informed convolutional neural networks for fluid flow through porous media — Physics-informed CNNs accurately predict fluid flow in porous media, significantly accelerating traditional LBM simulations.
• Kaggle Winners Walkthroughs: Yale/UNC-CH - Geophysical Waveform Inversion with Team greySnow — Combining physics-based models with machine learning enhances Full Waveform Inversion for subsurface imaging.
• Machine-Learning Emulation of Satellite Greenhouse Gas Retrievals: Stability over Time — Machine-learning emulators for GHG retrievals require time-aware training to maintain accuracy and stability over extended periods.
• Prediction of Thermal Breakthrough Probability and Spatial Optimization Geothermal Injection Well — Machine learning can predict geothermal thermal breakthrough and optimize well placement, replacing costly simulations.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by A Geodyssey – Geoscience Text Analytics and Enterprise Search Research.