One of the methods of reducing the amount of CO2 in the air is CO2 sequestration by dissolving the gas underground in an aquifer. In practice,there will be impurity gases in the injected CO2 stream or in the aquifer, which influence the dissolution rate of the CO2 stream. This res
...
One of the methods of reducing the amount of CO2 in the air is CO2 sequestration by dissolving the gas underground in an aquifer. In practice,there will be impurity gases in the injected CO2 stream or in the aquifer, which influence the dissolution rate of the CO2 stream. This research aims at finding a method to evaluate and calculate this influence. We focused on H2S and CH4. We found that given the specific circumstances underground and with added impurity gases, it is advisable to use an Equation of State, for which we chose Peng-Robinson. For the first simulation with a large aquifer model, we were able to see the difference between a stream of pure CO2 and mixtures of CO2 with H2S and CH4 respectively. We observed that the gas stream with CH4 moved considerably faster along the aquifer than the streams of pure CO2 or CO2 mixed with H2S. On the other hand, we found that the dissolution rate was higher when H2S was present in the mixture. In the dissolution process, we see a large influence of instabilities on the dissolution rate as a result of density differences in the aquifer. However, the resolution of the large aquifer model did not allow to evaluate this influence sufficiently. In two consequent steps, we adapted our simulation model. First, we used a small scale model using only a fraction of the large aquifer model, thus considerably reducing the calculation time. Next, we increase the resolution of the small scale model. This higher resolution gave considerably more accurate results, reflecting the differences between the three cases and the influence of the instabilities in the aquifer. We concluded that the simulation based on the small scale and high resolution model yields accurate and reliable results. However, the calculated outcomes have to be corrected for the difference in boundary conditions between the large and small scale models.