Determination of effective kinetic laws for cement alteration due to the contact with a CO2-rich solution
2 Géosciences Montpellier, CNRS-Université de Montpellier, 34095, Montpellier, France
O 12.2 in Reactive transport in heterogeneous aquifers
24.03.2018, 09:15-09:30, 3
Reactive transport modelling is growing as a necessary tool to study the fate of reactive solutes in the subsurface for a variety of applications including geo-energy related applications such as CO2 injection in deep geological formations. One of the challenges posed by these models is the choice of the values for the kinetic parameters characterizing the simulated processes which can not be considered in equilibrium. Kinetic rates and reactive surface area values are of particular importance for very reactive systems, such as injection of acidic CO2-rich solutions in deep aquifers, where dissolution processes can reduce the sealing properties of the well cement, potentially causing CO2 to leak from the reservoir to other permeable layers and therefore having a negative impact on the environment and also on human health.
The interaction between a CO2-rich solution and the cement is known to cause different reaction fronts to appear. The objective of this study is to determine the effective kinetic and diffusion laws of cement alteration combining an experimental and a numerical study.
To that end we first performed diffusion experiments of CO2-rich solutions using Class G Portland cement samples at P = 12 MPa and T = 60 °C. Measurement of aqueous chemistry by ICP-MS enabled the determination of net reaction rates. SEM images of the reacted cement samples were taken to identify the reactions occurring during the alteration process and also to measure the position and the thickness of the different reaction fronts. Afterwards, the experimental results were used in an automatic calibration process whose objective is to infer the optimal set of effective reaction rate coefficients and effective surface area of the solid phases and the parameter linking diffusion to porosity changes. For this purpose we coupled a diffusion-reaction modelling tool to PEST. The statistics provided in the estimation procedure allow to quantify the effect of each parameter on the numerical results, and therefore to identify those with a consistent impact on the processes identified in the experiments.
Export as iCal: