No abstract

Dehydration of natural gas (NG) by absorption is a common industrial procedure implemented in order to avoid flow blockage and equipment breakdown. Despite the widespread use of dehydration units, few experimental data are available in the literature and most engineering practices are based on empirical correlations for the design and the determination of the optimum operational parameters. In this paper, an accurate thermodynamic model for NG mixtures, the UMR-PRU, is further extended to mixtures containing triethylene glycol (TEG) and is then used to simulate a typical NG dehydration unit using TEG by incorporating it in commercial simulators through the CAPE-OPEN standard. The results are compared with those obtained by the recommended by Aspen Hysys, TST/NRTL model. The two models calculate similar lean TEG purity, TEG circulation rate, and stripping gas rate in order to obtain the same level of dehydration, ca. 30 ppm water in the dry gas, while some differences are observed in the component distribution in the vapor and liquid phases. In addition, different reboiler duties are calculated by the two models, with those of UMR-PRU to be considered more realistic due to better prediction of the heat capacities of aqueous TEG mixtures.

No abstract

No abstract

Simultaneous chemical and phase equilibrium (CPE) calculations are essential in chemical engineering, finding numerous applications in industrial processes, such as reactive distillation. CPE calculations can be challenging in systems involving polar and associating components, which also exhibit azeotropes or even reactive azeotropes, with two common examples being methyl tert-butyl ether and isopropyl acetate synthesis systems. To tackle the problem, a robust algorithm for solving CPE is required, coupled with an accurate thermodynamic model for describing system nonideality. In this work, a Gibbs energy minimization algorithm based on the method of Lagrange multipliers is employed for performing CPE calculations in the aforementioned systems. The algorithm is coupled with traditional activity coefficient models UNIQUAC and UNIFAC, as well as the UMR-PRU model. The results indicate that all models can successfully describe the CPE in the studied systems. UMR-PRU yields very satisfactory results, despite being utilized as a purely predictive tool.