Reducing the uncertainty of risk assessment in geologic CO2 sequestration by conditional simulation of permeability fields and monitoring data assimilation
Abstract
Research on risk assessment at geologic CO2 sequestration sites has been widely conducted in the past decade. Risk metrics, such as CO2 plume area, plume mobility and leakage rate at legacy wells have been used for risk analysis to predict CO2 plume location, identify potentially leaky locations, determine when a storage site can be closed, etc. However, substantial uncertainties are usually involved in the above risk metrics if no direct measurements (such as permeability estimates from pumping tests) or monitoring data is incorporated into the geological models.
In this study, we propose to reduce the uncertainty in risk assessments by sequentially performing conditional simulation and data assimilation. An ensemble of prior geological models (denoted as model 1) are first generated based on the prior geological information of a storage site. Then, reservoir models (denoted as model 2) which are constrained by direct measurements are generated by conditional simulation. When monitoring data from the storage site become available, they are assimilated into models by a recently developed data assimilation method, Ensemble Smoother with Multiple Data Assimilation (ES-MDA, a superior method over ensemble Kalman filter). After the data assimilation process, the reservoir models (denoted as model 3) are calibrated and then can be used for future risk assessment. The uncertainty in risk metrics obtained from models 1, 2, and 3 are quantified and compared. The effect of the number of direct measurements and the number of monitoring wells on uncertainty reduction of risk metrics are also investigated in this study. The results show that a combination of conditional simulation and data assimilation approaches can substantially reduce the uncertainty of risk assessment compared to the risk assessment based on prior models (model 1) or based on models generated from only conditional simulation or only data assimilation.- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2018
- Bibcode:
- 2018AGUFM.H23M2140C
- Keywords:
-
- 1816 Estimation and forecasting;
- HYDROLOGYDE: 1846 Model calibration;
- HYDROLOGYDE: 1847 Modeling;
- HYDROLOGYDE: 1873 Uncertainty assessment;
- HYDROLOGY