Modeling In-Situ Scale Deposition: The Impact of Reservoir and Well Geometries and Kinetic Reaction Rates

Abstract

Summary Previous work has demonstrated how and where the mixing of incompatible brines occurs in waterflooded reservoirs and what the impact on scale prevention strategies is in terms of timing and placing squeeze treatments. This paper extends this work by modeling the resulting in-situ deposition process. The location of maximum scale deposition and the resulting brine compositions at the production well are calculated for a range of sensitivities, including reservoir geometry (1D, 2D areal and vertical, and 3D), well geometry (location and orientation within the field and with respect to other wells and the aquifer), and the reaction rate (ranging from no precipitation to equilibrium). In conventional systems with no aquifer, it is demonstrated that maximum scale deposition occurs in the immediate vicinity of the production wellbore; therefore, low produced-cation concentrations indicate inadequate squeeze treatments. In systems in which water injection is into the aquifer, low cation concentrations may also result from deposition deeper within the reservoir. Maximum scale dropout still occurs as the fluids approach the production well but is sufficiently far from the wellbore to be unaffected by squeeze treatments or to have any major impact on productivity. The reaction rate is critical in determining the amount of scale deposition; however, even under equilibrium conditions, sufficient concentrations of scaling ions are delivered to the production well to necessitate squeezing it but with lower inhibitor volumes. Once cation concentrations have been reduced, it is predicted that they will never increase again. This paper also discusses some of the limitations of modeling such systems, including determination of kinetic reaction rates, mixing zone size, and impact on permeability. Although the thermodynamics are fairly well understood, the kinetics are much more difficult. The size of the mixing zone is affected by numerical dispersion, and computationally intensive techniques are required to overcome this problem. Previous experience shows that formation damage factors are very difficult to extrapolate from coreflood data because there is a great difference between the dimensions of the mixing zone in the reservoir and the core plug.