Predicting plume spreading during CO2 geo-sequestration: benchmarking a new hybrid finite element-finite volume compositional simulator with asynchronous time marching

In this paper, we present the results of benchmark simulations for plume spreading during CO2 geo-sequestration conducted with the newly developed Australian CO2 Geo-Sequestration Simulator (ACGSS). The simulator uses a hybrid finite element-finite volume (FEFVM) simulation framework, integrating an asynchronous local time stepping method for multi-phase multi-component transport and a novel non-iterative flash calculation approach for the phase equilibrium. The benchmark investigates four standard CO2 storage test cases that are widely used to assess the performance of simulation tools for carbon geo-sequestration: (A) radial flow from a CO2 injection well; (B) CO2 discharge along a fault zone; (C) CO2 injection into a layered brine formation; and (D) leakage through an abandoned well. For these applications, ACGSS gives results similar to well-established compositional simulators. Minor discrepancies can be rationalised in terms of the alternative, spatially adaptive discretisation and the treatment of NaCl solubility. While these benchmarks cover issues related to compositional simulation, they do not address the accurate representation of geologically challenging features of CO2 storage sites. An additional 3D application scenario of a complexly faulted storage site demonstrates the advantages of the FEFVM discretisation used in the ACGSS for resolving the geometric complexity of geologic storage sites. This example also highlights the significant computational benefits gained from the use of the asynchronous time marching scheme.

Automated summary

The study presents benchmark simulations for plume spreading during CO2 geo-sequestration using ACGSS with results similar to established compositional simulators. Minor discrepancies can be attributed to alternative discretisation and NaCl solubility treatment. While the benchmarks did not cover geologically challenging features of CO2 storage sites, they demonstrated the advantages of FEFVM discretisation and asynchronous time stepping.