Premium
A multiscale multilayer vertically integrated model with vertical dynamics for CO 2 sequestration in layered geological formations
Author(s) -
Guo Bo,
Bandilla Karl W.,
Nordbotten Jan M.,
Celia Michael A.,
Keilegavlen Eirik,
Doster Florian
Publication year - 2016
Publication title -
water resources research
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.863
H-Index - 217
eISSN - 1944-7973
pISSN - 0043-1397
DOI - 10.1002/2016wr018714
Subject(s) - geology , buoyancy , scale (ratio) , vertical integration , plume , aquifer , geotechnical engineering , mechanics , groundwater , meteorology , physics , quantum mechanics , political science , law
Efficient computational models are desirable for simulation of large‐scale geological CO 2 sequestration. Vertically integrated models, which take advantage of dimension reduction, offer one type of computationally efficient model. The dimension reduction is usually achieved by vertical integration based on the vertical equilibrium (VE) assumption, which assumes that CO 2 and brine segregate rapidly in the vertical due to strong buoyancy and quickly reach pressure equilibrium. However, the validity of the VE assumption requires small time scales of fluid segregation, which may not always be fulfilled, especially for heterogeneous geological formations with low vertical permeability. Recently, Guo et al. (2014a) developed a multiscale vertically integrated model, referred to as the dynamic reconstruction (DR) model, that relaxes the VE assumption by including the vertical two‐phase flow dynamics of CO 2 and brine as fine‐scale one‐dimensional problems in the vertical direction. Although the VE assumption can be relaxed, that model was limited to homogeneous geological formations. Here we extend the dynamic reconstruction model for layered heterogeneous formations, which is of much more practical interest for saline aquifers in sedimentary basins. We develop a new coarse‐scale pressure equation to couple the different coarse‐scale (vertically integrated) layers, and use the fine‐scale dynamic reconstruction algorithm in Guo et al. (2014a) within each individual layer. Together, these form a multiscale multilayer dynamic reconstruction algorithm. Simulation results of the CO 2 plume from the new model are in excellent agreement with full three‐dimensional models, with the new algorithm being much more computationally efficient than conventional full three‐dimensional models.