Assessing Risk in Long-Term CO2 Storage Under Uncertainty via Survival Analysis-Based Surrogates

We develop a survival analysis-based framework for assessing long-term CO2 storage risk under geological and operational uncertainty. Shortened reservoir simulations produce censored observations when leakage has not occurred before monitoring ends. Our approach accounts for these observations to construct computationally inexpensive and interpretable surrogate models.

Using methods ranging from Kaplan–Meier estimation to random survival forests, we predict leakage risk in new scenarios with shorter simulations or without additional simulations. A saline aquifer case study evaluates the framework under varied injection and production rates and uncertain geological properties, demonstrating its potential to reduce the computational burden of storage risk assessment.

Recommended citation: Gurwicz, A., Chen, J., Gutman, D. H., & Gildin, E. (2024). Assessing risk in long-term CO2 storage under uncertainty via survival analysis-based surrogates. SPE Annual Technical Conference and Exhibition, New Orleans, Louisiana, USA, September 23–25. Paper SPE-220737-MS. https://doi.org/10.2118/220737-MS