Effect of thermodiffusion on compositional grading in hydrocarbon reservoirs: Insights from field cases and simulations

Abstract

Heterogeneity of fluid composition in vertical and lateral directions is observed in many reservoirs, known as compositional grading. It is necessary to evaluate the diffusion forces for modeling this phenomenon. In addition to the chemical and gravity forces, the temperature gradient causes mass diffusion. In the present investigation, the accuracy and reliability of the Haase theory and the Shukla-Firoozabadi method in thermodiffusion modeling are studied comprehensively using field samples from light to heavy-oil reservoirs. On the one hand, the Shukla-Firoozabadi model results in compositional grading reduction by thermodiffusion; on the other hand, the Haase theory reveals enhancement and attenuation effects of thermodiffusion depending on reference enthalpy quantification. A new approach, which employs the reference enthalpies of pseudocomponents as tuning parameters, is proposed to improve the results of the Haase theory. The proposed procedure is also validated using field cases. In addition, the experimental field information compared with isothermal and nonisothermal models are studied to answer the question of "how" and to "what extent" thermodiffusion contributes to compositional grading. Results demonstrated the limited attenuation effect of thermodiffusion in light oils, while this phenomenon leads to severe enhancement of compositional grading in the medium and heavy-oil reservoirs. In heavy oils, the thermodiffusion effect is more pronounced on heavy pseudocomponents rather than light paraffinic species. Finally, the practice of compositional gradient model optimization, which is of particular interest to reduce the uncertainty in fluid initialization and to address reservoir compartmentalization, is presented.