A Generative-AI Modeling Framework for Explainable Decision Support in Complex Geosteering Scenarios
cs.LG, math.OC, physics.geo-ph, stat.AP
Submitted: 2025-03-11
Updated: 2026-09-17
Comments: The conference version of this paper is published in EAGE ECMOR 2024 proceedings: https://doi.org/10.3997/2214-4609.202437018
License: http://creativecommons.org/licenses/by/4.0/
The gist: The real-time process of directional changes while drilling, known as geosteering, is crucial for hydrocarbon extraction and emerging directional drilling applications such as geothermal energy,
Terminology
Abstract
The real-time process of directional changes while drilling, known as geosteering, is crucial for hydrocarbon extraction and emerging directional drilling applications such as geothermal energy, civil infrastructure, and CO2 storage. The geo-energy industry seeks an automatic geosteering workflow that continually updates subsurface uncertainties and captures the latest geological understanding, informed by real-time observations. We propose a real-time, AI-driven geosteering workflow that integrates Generative Adversarial Networks (GANs) for geological parameterization, ensemble methods for model updating, and global discrete dynamic programming (DDP) optimization for complex decision-making during directional drilling operations. Our framework relies on offline training of a GAN model to reproduce relevant geology realizations and a Forward Neural Network (FNN) to model the response of Logging-While-Drilling (LWD) tools for a given geomodel. This paper introduces a first-of-its-kind workflow that progressively reduces GAN-geomodel uncertainty around and ahead of the drilling bit and adjusts the well plan accordingly. The workflow automatically integrates real-time around-bit LWD, which, through learned geological correlations, reduces uncertainty in predicted geology ahead of drilling. A DDP-based decision support system leverages probabilistic look-ahead predictions to suggest better steering strategies. We test the workflow prototype on a small yet challenging low-net-to-gross drilling scenario with several possible targets. The results show that the workflow produces meaningful steering recommendations and, through its probabilistic updates, automatically maps formation boundaries along the drilled well.
Sources
- Parametrization and generation of geological models with generative adversarial networks
- Latent diffusion models for parameterization and data assimilation of facies-based geomodels
- Latent Diffusion Model for Conditional Reservoir Facies Generation
- Optimal Sequential Decision-Making in Geosteering: A Reinforcement Learning Approach
- High-Precision Geosteering via Reinforcement Learning and Particle Filters
- Strategic Geosteering Workflow with Uncertainty Quantification and Deep Learning: A Case Study on the Goliat Field
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: Local-Consistency Optimisation for Bayesian Neural Networks
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks