Long term assessment of massive CO_2 sequestration. Reactive transport modelling in deep saline aquifer according to natural and incidental scenario evolutions

摘要

The deep underground sequestration of CO_2 in natural reservoirs may provide the possibility of a good containment for the greenhouse gas emissions. But on long term evolution, some natural or incidental leakages of the sequestration may occur, bringing about the potentially release of significant amounts of CO_2. Coupled reactive transport modelling can provide useful information at large time and space scales. Two kind of deep saline aquifers have been chosen: a carbonated and a sandstone aquifers, to illustrate the main expected phenomena: CO_2 migration and dissolution in a carbonated aquifer and mineral precipitation in a sandstone aquifer. According to this contrasting behaviour, some natural or incidental leakages of the CO_2 have been taken into account. For the carbonated aquifer: 1) short-circuit with higher aquifer: a disruption of the cover brings a CO_2 leakage which is limited by the hydrodynamic flow. It can be observed a chemical disturbance of the water composition due to the acid plume. The modelling indicates a possible rock dissolution containing some pollutants which then migrate. 2) The CO_2 migration in fractured aquifers: in presence of fractures, the CO_2 migration and dissolution are quicker. 3) Leakage by natural submarine exutory in the far field: an acid plume in water disrupts significantly the local pH. For the sandstone aquifer, the progressive mineral precipitation, into carbonate, forms a protective layer. Consequently, natural or incidental evolutions do not bring about significant modification on the long term assessment. The simulations point out the performance of the reactive transport code for taking into account geochemical processes and consequences in the far field of the initial sequestration. CO_2 has exhibited a non negligible chemical reactivity. This approach offers the possibility of providing some qualitative features rather than risk assessment.