An investigation into the integrity of wellbore cement in CO2 storage wells: Core flooding experiments and simulations

Abstract

An important issue for geological storage of CO2 is the potential for wellbore cements to degrade in contact with the acidic formation waters resulting from CO2 dissolution. Cement degradation is a two stage process; cement carbonation occurs as various cement phases react to form calcium carbonate. The key second stage is the potential for erosion of the cement as this calcium carbonate dissolves into the formation water. For significant erosion to occur there would need to be a flow of water, under-saturated in calcium and carbonate ions, across the cement to remove dissolved calcium carbonate. This paper, presents a program of work that investigates cement degradation at the cement-formation interface. Two core flooding experiments were conducted at pressures and temperatures representative of storage conditions using composite cement–sandstone core plugs using CO2 saturated waters with chemistries representative of formation waters. The relatively high permeability of the sandstone allowed sufficient water flow rates for regular water samples to be collected and the chemistry analysed. As the sandstone simply provided a flow path for water, and did not impart any substantial chemical effect, the observations are applicable to a range of situations involving water flow in contact with cement. As the experiments, were structured such that the inflow water flowed across the cement plug surface before passing through the sandstone, each experiment provided two sets of observations with significantly different water flow velocities and chemistries. The measurements of water chemistry were combined with the flow rate observations to calculate the cumulative dissolution of the calcium carbonate and thus estimate the erosion of the cement. This compared well with direct estimates of the volume eroded by the flow across the cement plug surface. Using μXRD it was found that where the cement came into contact with the water it reacted to form calcium carbonate with none of the original cement phases detected. The erosion rate of the cement, when normalized by the water flow rate, had a clear relationship with respect to the difference between the inflow and outflow calcium concentrations. An empirical relationship was used to fit this data, thus providing a mathematical description of the cement erosion rate with respect to water flow velocity and the calcium solubility deficit. This was applied in a simulation model to a series of hypothetical case studies to investigate cement erosion at the cement-formation interface of a well, where there was an initial flow channel, across the geological seal in a CO2 storage formation.