Prediction of dead oil viscosity using mathematical simulation in wells of the Ecuadorian Amazon Basin

Abstract

The reliable determination of dead oil viscosity directly influences transportation planning and reserve estimation within the Oriente Basin. Based on this need, the objective was to develop a mathematical simulation model to predict dead oil viscosity in wells of the Ecuadorian Amazon Basin. The study was conducted using an experimental, quantitative, and cross-sectional approach. PVT tests from ten Amazonian wells were analyzed, Pearson correlation coefficients were used to identify relevant predictor variables, and a multiple linear regression model was fitted using R software. The results were then compared with laboratory measurements. The findings show that reservoir temperature and API gravity exhibit significant inverse correlations of −0.76 and −0.86, respectively. These variables were integrated into the proposed model, which achieved a coefficient of determination (R2) of 0.8 and a reduction in mean error to 19.8%. In contrast, other evaluated models showed deviations of 81.6%, 40.1%, and 94.7%. In specific cases, such as the SUSHUFINDI-51 and YNNA-009 wells, the discrepancy was reduced to 6.8%, while traditional methods exceeded 500% errors under high-temperature conditions, highlighting the need to adjust predictive expressions to local contexts. Finally, it is concluded that the developed equation facilitates faster diagnostics and contributes to optimizing the design of pipelines and surface equipment.