Geophysical control of geologic carbon storage using deep reinforcement learning: Sensitivity to multigeophysical noise and to the uncertainty of digital twins
Bibliographic record
Abstract
ABSTRACT Geologic carbon storage (GCS) must be safe and profitable. To achieve these goals for gigaton-scale GCS operations, decision-making in the presence of uncertainty is required. Geophysical monitoring methods can inform such decisions, given their sensitivity to the spatiotemporal changes in the subsurface during and after injection. We investigate a novel framework for the optimal control of GCS operations using geophysical monitoring. We refer to this decision-making tool as “geophysical control” and develop sequential decision-making models trained using digital twins of GCS operations and the corresponding geophysical monitoring signals. In particular, we obtain these models via deep reinforcement learning (DRL) and specifically focus on two types of uncertainty: geophysical noise and uncertainty in the subsurface petrophysical model. Our objective is to demonstrate how each source of stochasticity affects the decision-making process when one seeks to maximize profit while minimizing the risk of induced seismicity through an optimal policy that determines the annual target CO2 injection rate. We train a suite of DRL agents with different geophysical observations (surface time-lapse gravity, surface seismic amplitude-variation-with-offset (AVO), and combined gravity and AVO surveys), different signal-to-noise ratio levels, and with/without petrophysical model uncertainties. A comparison of the learning behavior of these independent DRL agents shows that (1) the DRL framework has the capacity to learn optimal CO2 injection policies; (2) training performance degrades with increasing geophysical noise (especially more in the seismic AVO case); and (3) the combination of AVO and gravity enhances decision-making, especially in the presence of geophysical noise. Our results show that the use of multigeophysical measurements and the incorporation of subsurface model uncertainties are critical in developing robust injection control agents using DRL.
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 distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 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 teacher head, 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".