Features and Capabilities
This page outlines the main features and capabilities of MRST-co2lab, which is a module in the open-source MATLAB Reservoir Simulation Toolbox dedicated to modeling large-scale storage of CO2 in saline aquifers and other geological formations that are limited upward by a low-permeable caprock.
Data input

Extraction of surface grid from 3D grids having an areal IJ numbering (rectilinear grids, corner-point grids, etc)   Construction of volumetric grid models and top surface grids from depth and thickness maps given on GIS format   Interface to public data sets including e.g., geological formations from the CO2 Storage Atlas of the Norwegian Continental Shelf
Trap and spill-point analysis
Computation of traps and their associated catchment regions, i.e., regions that lie deeper than the trap and from which a tricling flow of CO2 would eventually end up in the trap.    Computation of spill-point paths, i.e., the migration path of a trickling flow of CO2 as it gradually fills up a succession of shallower traps.   Capacity estimates: upper theoretical bounds on structural, residual, and solubility trapping assuming that the entire volume of the aquifer is filled with in some trapped form.
Vertical equilibrium simulations: fast simulation of large-scale, long-term trapping
Vertial equilibrium models assuming a sharp interface between each of the four different fluid states: free CO2, residually trapped CO2, brine with dissolved CO2, and brine. The flow models include structural and residual effects, buoyancy, compressibility, dissolution, and sub-scale trapping.   Vertical equilibrium models assuming a smooth transition from CO2 to brine (capillary fringe). These models include fine-scale capillarity forces and fine-scale hysteresis effects in addition to the effects accounted for in the sharp-interface simulators.   Detailed CO2 inventories delineating how the injected carbon volumes have been trapped by different mechanisms, stacked according to increasing risk of leakage: from dissolved and residually trapped CO2 (green) to volumes that are still movable (yellow / orange) or have already exited the domain (red).


MRST-co2lab offers two types of vertical-equilibrium simulators:

  • fully-implicit solvers based on industry-standard input description; these solvers are based on automatic differentiation and offer simple computation of adjoints and sensitivities for all kinds of vertical-equilibrium models
  • sequential solvers (flow and transport separated) formulated with either the height (h) or the vertically averaged CO2 saturation (S) as primary variable; these solvers only work for simple sharp-interface models

Both types of solvers can handle industry-standard grid formats and give detailed inventories delinating how the injected carbon volumes have been trapped by different mechanisms.  For the fully-implicit solvers, support for thermal and geomechanical effects is in progress.

For those seeking a more conventional simulation approach, CO2 sequestration can be simulated using MRST's standard reservoir simulation capabilities on structured and unstructured grid:

We also have several examples of compositional flow capabilities inhouse (these may be available upon request).

Graphical user interfaces
Identification of traps and catchment areas, spill-point analysis, and launching of simplified VE simulation.
Read more..
  Estimation of trapping capacities for geological formations from the Norwegian Continental Shelf.
Read more or watch video tutorial
  Full vertical equilibrium simulation for geological formations from the Norwegian Continental Shelf.
Read more or watch video tutorial
Optimizing aquifer utilization
Identification of optimal injection points based on a greedy algorithm that places each new injection point so that it maximizes the volume of structural traps that can be reached by a trickling flow of CO2 in the upslope direction.   Rigorous mathematical optimization of injection rates (or pressures). The objective function is set to maximize structural, residual, and solubility trapping and penalize pressure buildup and/or volumes that exit the simulated domain.      

Published January 28, 2016