Deep Reinforcement Learning Optimizing for Geological CO2 Storage Considering Geomechanical and Leakage Risks
Bibliographic record
Abstract
Optimizing well placement and injection rates to maximize CO2 storage while limiting leakage and geomechanical risks is a critical challenge for scaling up geological carbon storage (GCS). This study develops a deep reinforcement learning (DRL) framework in which a convolutional neural network (CNN) policy maps surrogate-predicted CO2 saturation and pressure fields to well locations and dynamic injection rates. Policy parameters are trained using Proximal Policy Optimization (PPO) with a multi-objective reward that promotes CO2 storage while penalizing pressure buildup and boundary leakage. High-fidelity surrogate models, trained on coupled flow-geomechanics simulations, enable efficient agent–environment interactions at reduced cost, supporting high-resolution evaluation across diverse geological realizations. Geological uncertainty is addressed through domain randomization, where 1,000 permeability realizations are embedded via multidimensional scaling and grouped into 20 clusters for streamlined training. The framework is applied to a conditional geostatistical model of the Aquistore storage site in Saskatchewan, Canada. Performance is benchmarked against fixed-rate injection schedules and the covariance matrix adaptation evolution strategy (CMA-ES). Results show that PPO consistently identifies superior well configurations and dynamic injection strategies, achieving greater stored volumes while maintaining operational safety across all validation samples.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.003 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".