Development of a Computational Fluid Dynamics Compositional Wellbore Simulator for Modeling of Asphaltene Deposition

The asphaltene problem is a two-step process: (1) asphaltene precipitation, as predicted by the thermodynamic model, and (2) asphaltene deposition, the amount of which is estimated by the kinetic model. Asphaltene precipitation is a prerequisite but not a sufficient condition for deposition. Deposition is dependent on other factors such as surface properties, phase behavior, rheology, and flow patterns. As a result, in addition to understanding thermodynamic and kinetic models, it is critical to also understand flow models. In fact, multiphase flow modeling is at the core of simulation, and it must be coupled with thermodynamic and kinetic models. Numerous studies on modeling asphaltene deposition on pipe walls have been performed theoretically and experimentally, but a comprehensive theory to properly understand this phenomenon has not yet been presented. In thermodynamic modeling, the perturbed chain statistical associating fluid theory (PC-SAFT) equation of state is used to predict the asphaltene phase behavior. In this study, we show that the proposed PC-SAFT model is more accurate than the solid model used in commercial software. Unlike prior research that neglected flow patterns or used empirical relations to model multiphase flow, this study simulates multiphase flow using separate momentum equations for each phase. Among the existing kinetic models, the Kurup model has been used to predict the asphaltene deposition profile in the wellbore due to its greater compatibility for computational fluid dynamics application. The results of the proposed model show good agreement with field case data of asphaltene deposition thicknesses along the wellbore tubing.

Automated summary

The asphaltene problem involves two steps: precipitation and deposition, where deposition is influenced by factors such as surface properties, phase behavior, rheology, and flow patterns. Understanding flow models is critical, and the PC-SAFT equation is shown to be more accurate than solid models in commercial software.