Answers are grounded in this paper and cite their sources.
Mini-review explores proxy modeling techniques for CO2 storage in karstified carbonate reservoirs, indicating significant improvement in efficiency.
Proxy modeling provides a computationally efficient alternative to conventional numerical simulations in petroleum engineering, enabling rapid uncertainty quantification, optimization, risk assessment, and forecasting. This review examines the evolution of proxy models from traditional statistical methods to advanced artificial intelligence/machine learning (AI/ML)-based techniques, with a particular focus on their applications in CO2 storage. The general methodology for proxy model development is outlined, including the objective definition, sensitivity analysis, sampling strategies, model construction, and validation. Applications in field development planning, history matching, and probabilistic forecasting are discussed through relevant cases. Recent advances in AI/ML-based proxy modeling for CO2 injection and storage, primarily in sandstone saline aquifers, are critically evaluated. While proxy models significantly reduce the computational burden of full-scale simulations and deliver results with acceptable accuracy, their application to complex carbonate reservoirs, especially those with karst features, remains limited. Quantifying the uncertainty on plume migration within and around karstified regions during and after CO2 injection is essential for designing cost-effective monitoring strategies and verifying the conformance of the predicted reservoir behavior. This gap underscores the need for targeted research to develop robust proxy modeling approaches for these highly heterogeneous systems.
No takes yet. Share an insight, caveat, or question.
Jalil et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: