Challenges in Simulation of Salt Clogging

Abstract

Abstract Carbon sequestration in depleted reservoirs or aquifers is highly demanded but still faced with technical challenges in many aspects. Among them, losing well injectivity during the storage process is a major concern. This can be caused by salt deposited in the reservoir, particularly near the injection well, which may sometimes creep into the injection well. Therefore, it is desirable to estimate the amount and distribution of salt precipitation at the injection conditions for a smooth implementation of CO2 sequestration. In this paper, we investigate how much commercial software CMG-GEM can help the evaluation of salt precipitation. We first review the critical mechanisms involved in salt precipitation and then analyze the challenges in simulating these mechanisms. According to the literature, water saturation and saturation index are the two most influential parameters that control the amount and pattern of salt precipitation and clogging due to water vaporization. Their values are determined by the complex interplay between viscous force, gravity, the evaporation of water into the CO2 stream, the molecular diffusion of dissolved salt in the brine, and surface phenomena such as the spreading of a thin water film on the rock surface, the Marangoni convection, and disjoining suction. Here we investigate the challenges of simulating the aforementioned mechanisms as well as salt precipitation due to the backflow of brine toward the injection well. The surface-related phenomena are difficult to account for in simulation. However, the extent of the CO2 plume can be significantly underestimated if they are neglected. Although water vaporization, salt diffusion, and capillary pressure can be formally included in the simulation, it is arguable whether they always describe the actual phenomena adequately. In most cases of CO2 injection into an aquifer, water spreads all over the rock surface, which increases the rate of vaporization and surface-related phenomena, such as the Marangoni effect, dramatically. Marangoni turbulent fluxes originating from the unbalanced shear stresses on the interface can accelerate the mixing effect in homogenizing the ions composition, which results in self-enhanced salt precipitation via the thin brine film spreading on the rock surface. We examine different simulation techniques as remedies to mimic those phenomena.