Viscosity Modeling of MES and SLS Using Machine Learning Method

Viscosity is crucial to improve the efficiency of injected fluids for oil displacement in reservoirs. Traditionally, research has focused on polymers that help reduce the mobility of injected fluids, while surfactant viscosity has received less consideration. This research investigated the viscosity behavior of methyl ester sulfonate (MES) and sodium lauryl sulfate (SLS) surfactant solutions using a machine learning method—adaptive neurofuzzy inference system (ANFIS). This study aimed to predict the viscosity of surfactant solutions. Experimental data included viscosity measurements of 36 MES and SLS samples at various concentrations and temperatures, obtained by digitizing viscosity curves. These data served as input and validation for the ANN and ANFIS models. The results showed that ANFIS predicted viscosity values ​​reliably, yielding only 1.33% and 0.43% differences for MES and SLS, respectively. Comparison of viscosity prediction with Artificial Neural Network (ANN) showed that ANFIS prediction was better, because ANN yielded two deviating predictions.