New interaction parameters from VLE data for group contribution (GC-NRTL) model

Seyf, Jaber Yousefi; Asgari, Mohammad; Heidaryan, Ehsan

doi:10.1007/s43153-022-00290-y

Cited by 1 publication

(1 citation statement)

References 50 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Local composition models are recommended when experimental data is available, but for n number of components, they require n × ( n – 1)/2 binary interaction parameters and also the experimental pair data. In lack of experimental data, predictive models such as analytical solutions of groups (ASOG), UNIQUAC functional group activity coefficients (UNIFAC), UNIFAC-DMD, − NIST-modified UNIFAC, and NRTL functional activity coefficient (NRTL-FAC) , are used. Although these models are called predictive, the group–group interaction parameters, first, are fitted to many experimental data.…”

Section: Introductionmentioning

confidence: 99%

Phase Equilibrium Calculation of Bio-Oil-Related Molecules Using Predictive Thermodynamic Models

Jalalinejad,

Yousefi Seyf,

Funke

et al. 2024

Energy Fuels

View full text Add to dashboard Cite

In the present study, predictive thermodynamic models including original UNIFAC, Dortmund-modified UNIFAC (UNIFAC-DMD), NIST-modified UNIFAC (NIST-UNIFAC), and the COSMO-segment activity coefficient (COSMA-SAC) were used to predict the phase equilibrium of binary and ternary mixtures relevant for the description of fast pyrolysis bio-oils. A total of 3371 binary vapor−liquid equilibrium (VLE) isothermal or isobaric data sets were used to study the predictive power of the investigated models. Based on the obtained deviation for VLE of binary mixtures, the NIST-UNIFAC is recommended for class I mixtures (including non-bio oil), while the COSMO-SAC model is suggested for class II mixtures (bio-oil molecules). In sequence, 62 available ternary vapor−liquid equilibrium (VLE) isobaric data including bio-oil-related and non-bio-oil molecules were used to investigate the models. Results showed that both the UNIFAC-DMD and NIST-UNIFAC provide the lowest deviation compared to UNIFAC and COSMO-SAC. Also, the COSMO-SAC model requires a large CPU time (51 min) for 62 ternary systems, while NIST-UNIFAC, with a CPU time of 13.2 s, allows the fastest model calculations. To further evaluate the ability of the models, 125 binary LLE systems with 850 binary data sets and 2543 data points were collected from the literature. In terms of CPU time, the group contribution models demand run times in the order of seconds, and those of COSMO-SAC require run times in the order of hours (1.91 h). Also, the NIST-UNIFAC model provides the lowest deviation (with a slight difference from the UNIFAC-DMD model) compared with other models. In many cases, the COSMO-SAC model cannot predict the phase separation of the binary LLE data. Finally, 157 ternary combinations with 276 experimental ternary LLE data were collected, and the results show that UNIFAC-DMD and NIST-UNIFAC are the models with the lowest deviation. In summary, according to the obtained results in VLE systems for the bio-oil-related molecules with polar and complex structures, the group contribution models should be used with more investigation and the COSMO-SAC model is not recommended for LLE systems at all.

show abstract