Select thermodynamic models for process simulation
A Practical Guide to a Three Steps Methodology

Chapter 3 – Abstract

This chapter will discuss how the fluid composition affects the thermodynamic calculations. As already pointed out in the first chapter, discussion brings up both the issue of models as that of their parameters i.e. the fluid description. These two issues cannot be dissociated.

In table 1.1, it has been proposed to subdivide this analysis criterion into three sub-questions:

In order to help answering these questions, we shall stress that:

In one way or another, the model-parameter combination must always be compared (either for regression or for validation) to experimental data.

The first two sections are therefore devoted to the description of the type of data that may exist, the way they are presented and some discussions on how to evaluate their quality (pure components and mixtures). Figure 3.21 illustrates the regression procedure. The third section is entirely focused on the determination of model parameters by fitting on experimental data. The fourth section will discuss the thermodynamic models, focusing on their molecular construction, and hence how they provide the species behaviour in a mixture. Table 3.20 provides a summary for the activity coefficient models and table 3.26 for equations of state. Finally, in the fifth section, the concept of “key component” will be discussed and the mixture data to be evaluated will be first identified (it is often almost impossible to consider all binary data in a multicomponent mixture).

Figure 3.21 Procedure for regressing and validating a model
Table 3.20 Overview of the most important activity coefficent models
Predictive Correlative
Enthalpic Hildebrand NRTL
Van Laar
Entropic Flory
Enthalpic + Entropic UNIFAC

* The Flory Huggins model is presented here as a predictive, but it is very often used with a correlated parameter for solvent-polymer mixtures.

Table 3.26 Summary of important equations of state and their use.
Correlative Corresponding states Group contribution
Hydrocarbons Cubic +α(T)
Slightly non-ideal Cubic + kij mixing rule Cubic + Jaubert kij
PC-SAFT (Tihic)
Strongly non-ideal
Cubic + GE-type mixing rule
CPA (water, alcohols, gases)
VR-γ-SAFT (Lymperiadis)
GCA (bio-applications)
+ entropic non-ideality
NRHB GC-Lattice Fluid
GC-PPC-SAFT (Passarello)
Lee Kesler
Starling - BWR