InterPore2025

Geochemical Modelling CO2 Mineralisation in Peridotite Formations, History Matching of a Pilot Test

22 May 2025, 09:35

15m

Oral Presentation (MS25) Advances in Carbon Mineralization: Unveiling Multiscale Geo-processes and Coupled Mechanisms MS25

Speaker

Mr Juerg Matter (44.01)

Description

Efficient implementation of CO2 mineralization can be a fast and safe method for the long-term disposal of anthropogenic CO2. However, quantifying mineralized CO2 can be a challenge in assessing CO2 mineralization projects. Modelling can facilitate the long-term and large-scale assessment of mineralisation if a reliable model can be developed. Herein, a series of history-matching exercises was executed to calibrate the reaction kinetics behind CO2 mineralization in fractured peridotite, using the results of a successful pilot test carried out in the peridotite of the Samail ophiolite in Oman. The pilot test, known as the Chalk project, was conducted in an inject-soak-retrieve mode. A fully coupled thermo-hydro-chemical reservoir model was developed to history match the pilot test using well-logging, core data, and well-testing analyses. Two realizations were constructed to investigate the role of grid size (coarse and fine grid size) in the history matching of reaction kinetics behind CO2 mineralization. Additionally, two cases for the primary minerals of the host formation were considered: (i) Mg-rich olivine and (ii) serpentinized olivine.

For history-matching, the tracer concentration profile was matched as the baseline where no reaction takes place. Then, the reaction rate and surface area of primary and secondary minerals are tuned to reproduce the field observation. CMG-GEM package was used to perform the coupled simulations. The reaction rates of lizardite and pyroxene minerals were increased to enhance the dissolution, which helped match pH of the produced water. Also, dolomite and calcite were identified as the main carbonate minerals controlling the CO2 mineralisation as they could match the trends of Ca and Mg, where the reaction rate and activation energy of precipitation for these two minerals were increased in the model. The dominance of dolomite and calcite in the modelling results matches field observations that reveal calcite and dolomite as the main carbonate minerals in the shallow subsurface (<200 m depth).
Also, overly coarse grids ease the progress of numerical simulation as the mixing of resident and injected fluid is better controlled in large grid volumes; however, large grid blocks fail to differentiate regions of the reservoir dominantly under dissolution or precipitation. Furthermore, using different primary minerals impacts the numerical stability of the reservoir model as well as the reaction kinetics. Serpentinized minerals such as lizardite could adversely affect numerical stability due to high reactivity.
From the various history-matching cases, it was concluded that the area of dissolution is very localized around the injection well, whereas precipitation takes place variably depending on whether it is calcium or magnesium-rich carbonates. Therefore, the gridding scheme of a reservoir model should differentiate the areas with predominant precipitation and dissolution. Additionally, the history-matched model was used for a long-term assessment of CO2 mineralization in peridotite formations, where it demonstrated promising longevity for dissolution and precipitation projects. This work puts forward a robust methodology for modelling CO2 mineralization in fractured formations. This work also demonstrates that using batch-based reaction kinetics may require further tuning for reservoir conditions, where dynamic dissolution and precipitation processes are in play.