Modeling the viscosity of (water + methanol), (water + ethanol), (water + 1-propanol) and (water + 2-propanol) mixture up to the high pressures

The viscosity of alcohol-water systems has been modeled using the Eyring model. Four alcohols, including methanol, ethanol, 1-propanol, and 2-propanol are employed for this modeling. The UNIFAC-DMD activity coefficient and the cubic plus association equation of state are applied to the excess activation free energy within the Eyring model. Furthermore, a new expression has been proposed for the binary interaction parameter of the Eyring model. The overall results demonstrate a good agreement with the experimental data, yielding an average absolute deviation (AAD) of 4.52% and 5.36% for the two strategies, respectively. However, the Eyring model exhibits significant deviations at high pressures and low temperatures. Subsequently, the data sets are utilized to develop an artificial neural network for accurately computing the viscosity of alcohol-water systems. The results align well with the experimental data, achieving AAD% of 1.37. Moreover, the neural network model effectively addresses the limitations inherent in the Eyring model.