This paper reports on the first modeling of sandstone matrix acidization using a comprehensive geochemical simulator based on the partial local equilibrium assumption. This new model allows any combination of kinetic and equilibrium reactions involving any number of chemical species, which greatly increases the simulator's predictive abilities compared to existing sandstone acidizing models. The new simulator, KGEOFLOW, which is applicable to a variety of reactive transport processes in petroleum reservoirs, is used to predict optimal acid injection rates based on the amount of mineral dissolution and precipitation.
Sandstone matrix acidizing, particularly with mixtures of HF and HCl, has long been the primary near-wellbore stimulation technique. Nevertheless, many questions of both fundamental and practical importance remain: the depth of penetration for both live acid and spent acid, the importance of the initial mineral assemblage, the consequences of precipitation of reaction products, and the optimal injection rate and volume.
The primary difficulty in developing an effective acidizing design is most certainly the complexity of naturally-occurring sandstones and of the aqueous chemistry. Laboratory experiments are usually carried out in relatively short cores that do not normally exhibit the extensive mineral precipitation which can cause permeability reduction. Thus, short-core experiments cannot be readily translated to field treatments. There have been several experiments conducted in long cores, where precipitation is evident. However, because of the difficulty and expense of such studies, an important alternative approach for designing effective acidizing treatments is to incorporate comprehensive reactive transport modeling in the design.
"Two-parameter" kinetic models represent well the effluent acid concentrations from short cores at modest flow rates and temperatures.They do not, however, include precipitation of product minerals and their subsequent reaction with acid, nor do they consider the reaction of spent acid deep in the formation. Four-parameter models indicate that the failure of the two-parameter model to fit laboratory results reflects the lack of mineral precipitation reactions in the model. However, both two- and four-parameter models use only a small number of chemical reactions from the enormous number possible, and this type of simplification limits the applicability of these models.