Currently, there is no existing methodology within commercially available software packages for accurately simulating the gradual evaporation of the aqueous phase in batch thermomechanical dehydration processes involving highly stable water-hydrocarbon emulsions. This limitation constitutes a significant obstacle to the widespread industrial implementation of a promising approach for liquid hydrocarbon waste disposal, which relies on the evaporation of the aqueous phase under intensive stirring conditions, ultimately producing a hydrocarbon product with residual water content. In this study, the widely used Aspen HYSYS V12 software was employed to model these processes. The primary objective was to identify the most appropriate thermodynamic model accurately describing vapor–liquid phase transitions during the boiling of the aqueous phase in highly stable water–hydrocarbon emulsions, with water content ranging from 2 to 60% by weight. The modeling of the gradual boiling process was divided into several sequential stages, each representing a single evaporation step. The initial feedstock temperature was set at 90 °C, with subsequent stages involving temperature increments of 5 °C until the residual water content in the product fell below 0.5% by weight. Four thermodynamic models were evaluated for their ability to predict phase equilibria: Peng–Robinson, Wilson, UNIQUAC, and NRTL. It was observed that the Peng–Robinson model poorly describes the dehydration process, as it predicts water evaporation only at 100 °C, which contradicts experimental evidence indicating that evaporation occurs over a broader temperature range. The Wilson model significantly overestimates boiling points, reaching values up to 290 °C. Although the UNIQUAC model accurately reflects the process at higher water contents, it results in elevated energy consumption, necessitating substantial superheating of the feedstock up to 220 °C. The NRTL model provided the best correlation (among studied thermodynamic models) with experimental data, providing an average relative deviation of 3.68% and effectively capturing the two-stage evaporation mechanism: initial removal of free water at 100–110 °C, followed by bound moisture evaporation at temperatures approaching 160 °C. Vaporization rates were also examined across all models. The Peng–Robinson approach predicted the highest vaporization peaks but was the least representative of actual process conditions. Notably, in the NRTL model, the peak vaporization rates were 1.9 to 2.7 times higher than those estimated using the UNIQUAC and Wilson models. This parameter is critical for the optimal selection and design of subsequent condensation equipment. Based on these findings, the NRTL thermodynamic model is recommended for the industrial-scale implementation of thermomechanical dehydration processes involving heavy hydrocarbon feedstocks, given its accuracy in modeling phase transitions and the temperature-dependent vapor generation rates derived from sequential equilibrium flash calculations.
Safiulina et al. (Thu,) studied this question.