Thermodynamic software

This article follows on from the articles "Thermodynamic diagrams – General to Thermodynamic diagrams – Wet mixtures and combustion", in which diagrams for pure bodies and mixtures were presented. It deals with the software packages used in thermodynamics.

Diagrams provide a visual representation of a fluid's properties in its field of use, and are very well suited to qualitative analysis of the changes that this fluid may undergo in industrial operations. However, they are relatively inaccurate to read, and of little use for dimensioning, where a continuous representation is often required (distillation calculations, etc.).

For quantitative analyses, we now prefer to supplement diagrams with the use of software packages (a contraction of the two terms "software product" characterizing standard software developed to meet generic specifications not specific to a single user). Thermodynamics software packages are software tools that enable the thermodynamic state of a fluid to be calculated over a wide range of its state variables. To do this, they use mathematical formulations of their properties, based on equations whose form and parameters have been adjusted or are adjustable by comparison with experimental values recorded by research laboratories and published in the literature or in databases. The main equations of state are fitted using numerical techniques, and the differential equations of thermodynamics are used to derive the derivative functions.

When the same formulation does not provide sufficient accuracy throughout the fluid's operating range, it is subdivided into several regions, each described by an appropriate formulation (liquid, liquid-vapor equilibrium, vapor zones, etc.). At the boundaries of these regions, small differences in value may exist. An example, concerning water, is given in this article.

Since the input variables of interest to a given user are not necessarily those which appear in the mathematical formulations adopted, it is often necessary to make changes to the variables, by inverting the software package's state equations in a more or less complex way. In some cases, convergence difficulties may arise.