InterPore2021

Dynamic coarsening for efficient simulation of geothermal energy applications

1 Jun 2021, 15:40

15m

Oral Presentation (MS7) Mathematical and numerical methods for multi-scale multi-physics, nonlinear coupled processes MS7

Speaker

Dr Øystein Klemetsdal (SINTEF Digital)

Description

Whereas extraction of hydrocarbons from the subsurface typically involves transport phenomena over large distances (e.g, in-between injection and production wells), transport of geothermal heat during extraction and storage is chiefly confined to the proximity of wells. This indicates significantly reduced computational efforts by considering only a region of interest around the wells. In this work, we investigate a dynamic coarsening approach that uses aggregation of cells in an underlying fine-scale geomodel to form a coarse representation with refinement only in regions of interest. The dynamic grid is constructed from a predefined set of nested partitions of the underlying fine grid, and constructed at each timestep based on a set of suitable indicators. The approach is highly flexible, since it applies to any underlying fine grid consisting of non-overlapping cells, and efficient, since properties of the dynamic grids can be computed in a preprocessing step.

We assess the method on a number of test cases, and compare and contrast the results to fine-scale simulations using different refinement indicators and coarsening strategies. All numerical experiments will be conducted using the newly developed geothermal module of the open-source MATLAB Reservoir Simulation Toolbox.