Reactive Flow Simulation
Master thesis
Submitted version
View/ Open
Date
2019-06-15Metadata
Show full item recordCollections
- Studentoppgaver (TN-IER) [166]
Abstract
Sea water-injection in carbonate formations leads to reactive processes that are linked to affecting oil recovery via wettability alteration and chemical compaction. The concentrations of divalent ions, such as Ca2+, Mg2+, and SO42 –, have proved to affect the stability of the carbonate matrix and the oil recovery. These effects are essential for chalk fields such as Ekofisk and Valhall on the Norwegian Continental Shelf (NCS).
This study considers history matching of recently performed brine injection experiments of three Mons Belgium chalk cores, with specific ion composition at reservoir (Ekofisk) conditions.
A 1D advection-dispersion-reaction (ADR) geochemical model is developed in PHREEQC, to capture the geochemical effects of Ba2+ , Ca2+ , Mg2+ , Sr2+ , and SO4 2 – in the injection brine. The model considers steady-state dissolution and precipitation reactions of anhydrite (CaSO4 ), calcite (CaCO3 ), celestite (SrSO4 ), magnesite (MgCO3 ), strontianite (SrCO3 ), and witherite (BaCO3 ). The minerals are selected based on a static model, experimental findings, and literature. Literature reaction rate kinetics give too high dissolution and precipitation rates, hence direct application does not match experimental results. To match experiments tuning parameters are introduced to the reaction rate equation, to reduce the literature reaction kinetic rates.
The model produces suitable calcite precipitation and magnesite precipitation, both considering effluent concentrations and post-flooding mineral distribution. The behaviour of witherite was captured, but its precipitating rate seems to have a higher meta-stable saturation, hence require a higher super-saturation for precipitation initiation. Moreover, at super-saturations beyond the meta-stable level, the precipitation rate accelerates faster, compared to calcite and magnesite. The lack of reaction kinetic data for strontianite introduces great uncertainty to the simulation. Consequently, simulations of Sr2+ injection sequences were adjusted to match effluents, but mineral distributions were not matched. Transient effluent behaviour during sulphate-bearing mineral precipitation was not matched.
Description
Master's thesis in Petroleum Engineering