A model of silica colloid growth, stability and transport used to predict geothermal reinjection lifetime

Abstract

When cooled geothermal brine is reinjected, deposited silica may block the fluid pathways in the aquifer, reducing injectivity. Previous authors have used TOUGHREACT to successfully model silica deposition, using an empirically determined reaction rate to predict the rate of deposition. An a-priori model of hydrodynamic effects on the transport of colloidal silica is desirable. This paper describes a new model of fluid flow and silica particle nucleation, growth and transport. 1D transport PDEs are solved for heat and mass transfer. Silica particle growth is modelled by considering homogeneous nucleation and growth due to molecular deposition (Weres et al. 1981), Ostwald ripening (Wagner 1961), and aggregation (Elimelech et al. 2013, S?kvarla 2013).
The deposition rate is sensitive to the stability (resistance to aggregation) of the colloidal particles. DLVO theory predicts stabilities which, under certain conditions, differ from experimental observations. The model implements empirical stability relationships fitted to experimental data. The model replicates bench observations of silica growth and stability (Weres et al. 1981, S?kvarla 2013, Tobler and Benning 2013). It is used to model deposition in the fracture networks of a simplified geothermal reservoir. The results are compared to the injectivity predictions of Xu et al. (2004).