Abstract
Abstract. Fluid extraction from producing hydrocarbon reservoirs can cause anthropogenic land subsidence. In this work, a 3-D finite-element (FE) geomechanical model is used to predict the land surface displacements above a gas field where displacement observations are available. An ensemble-based data assimilation (DA) algorithm is implemented that incorporates these observations into the response of the FE geomechanical model, thus reducing the uncertainty on the geomechanical parameters of the sedimentary basin embedding the reservoir. The calibration focuses on the uniaxial vertical compressibility cM, which is often the geomechanical parameter to which the model response is most sensitive. The partition of the reservoir into blocks delimited by faults motivates the assumption of a heterogeneous spatial distribution of cM within the reservoir. A preliminary synthetic test case is here used to evaluate the effectiveness of the DA algorithm in reducing the parameter uncertainty associated with a heterogeneous cM distribution. A significant improvement in matching the observed data is obtained with respect to the case in which a homogeneous cM is hypothesized. These preliminary results are quite encouraging and call for the application of the procedure to real gas fields.
Highlights
Fluid extraction from aquifer systems and hydrocarbon reservoirs are among the most frequent causes of anthropogenic land subsidence
The calibration focuses on the uniaxial vertical compressibility cM, which mostly influences the occurrence of land subsidence
The Ensemble Smoother (ES) is deemed adequate for this purpose because its implementation (i) avoids the simulation restart necessary to apply other more common data assimilation techniques, such as the Ensemble Kalman Filter (EnKF); (ii) significantly reduces the overall computational cost required by geomechanical model and the inversion procedure
Summary
Fluid extraction from aquifer systems and hydrocarbon reservoirs are among the most frequent causes of anthropogenic land subsidence. An ensemble-based DA method is used to infer the geomechanical parameters characterizing the rock formation of a deep gas reservoir, reducing the prior uncertainties of the geomechanical model response. A 3-D finite-element (FE) geomechanical model is used to predict the land surface displacements above a gas field where displacement observations are available. An Ensemble Smoother (ES) is the DA algorithm used to estimate the compressibility by inversion of land surface displacement data. The ES is deemed adequate for this purpose because its implementation (i) avoids the simulation restart necessary to apply other more common data assimilation techniques, such as the Ensemble Kalman Filter (EnKF); (ii) significantly reduces the overall computational cost required by geomechanical model and the inversion procedure.
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call