Full factorial sensitivity analysis of asphaltene precipitation and deposition in CO2 and CH4 coreflooding

Modelling of asphaltene precipitation and deposition in petroleum reservoirs requires adjustable parameters and introduces uncertainty sources. To address the main sources of uncertainty, a two-level full factorial sensitivity analysis was applied to asphaltene precipitation and deposition. Six factors were critically evaluated in the models, three concerning the precipitation phenomenon and the others related to the asphaltene deposition: asphaltene molar content, binary interaction parameters, reference pressure, asphaltene molar volume, surface deposition rate coefficient and flocculation parameters. The effect of these factors on the average of permeability reduction is assessed and their relative importance and relevant interactions were analyzed. Curves of precipitated, flocculated and deposited asphaltene showed that asphaltene deposition caused permeability reduction. Results showed that thermodynamic sources of uncertainty have a more significant effect on the model performance compared to the kinetics parameters ones. This suggests that fitting deposition parameters using experimental permeability reduction data may jeopardize the applicability of these parameters in real field conditions. The aggregation of asphaltene precipitates to form larger particles is modelled using the Kohse and Nghiem flocculation model, one of the most used. Simulations for different conditions were also performed to improve the reliability of the findings. We conclude that reestimation of these model parameters using experimental data of kinetics of asphaltene aggregation is required to provide acceptable description of the aggregation phenomenon.

Automated summary

A study on asphaltene precipitation and deposition in petroleum reservoirs conducted a two-level full factorial sensitivity analysis, evaluating the impact of six key factors on permeability reduction and found that thermodynamic uncertainty sources had a more significant effect on model performance.