Select thermodynamic models for process simulation
A Practical Guide to a Three Steps Methodology
Jean-Charles de Hemptinne, Jean-Marie Ledanois, Pascal Mougin and Alain Barreau
The choice of a modelling approach is very often guided by the experience on other systems, or by the availability of parameter databases. This is acceptable in the sense that most industrial computations are based on previous cases that worked correctly. Yet, it may also be dangerous, and it is then recommended to submit the problem to a thermodynamic analysis.
Our purpose is to provide a vademecum that should help the practicing engineer in finding the right questions to answer when faced with a novel type of thermodynamic problem: when the questions are correctly stated, the answer is half on its way.
The construction of the book, that may seem awkward at first, is designed for this purpose. It is constructed on three pillars (chapters 2, 3 and 4) that each represent a different point of view converging to the same goal: the development of an adequate thermodynamic set of models for an industrial problem. We believe that in order to analyse completely a physical modelling problem, these three pillars must be correctly mastered (figure 0.1):
- understanding the fundamental principles,
- use of the available mathematical models, and
- knowledge of the system physical (phase) behaviour.
The domain of application that is aimed is petroleum and energy process design. Yet, readers working in other fields of chemical engineering may also find interesting topics.
The first chapter goes into the details of the philosophy of the approach. The main messages we want to carry here is:
- A thermodynamic method may contain many different models (as in the image of the Russian doll); each of which has to be parameterised. The origin of the parameters is at least as important as the choice of the model.
- The three questions that will help the engineer solve his problem concern the relevant properties of his process, the type of mixture he has and the phases he might encounter.
The second chapter summarises some of the basic principles of thermodynamics (what O’Connell  calls ‘Always True’). It is written so that most concepts and equations that the chemical engineer may need are provided in a very concise form. It contains two main sections:
- How to read and understand a phase diagram, using Gibbs phase rule?
- How are the fundamental thermodynamic principles used for calculating engineering properties? This results in a rather short presentation of residual properties, excess properties, and a few algorithmic aspects on phase equilibrium calculation. Chemical equilibrium is also touched upon.
The third chapter is the heart of this book. Rather than providing an exhaustive list of thermodynamic models, it emphasises the use of these models, starting from the concept of the fluid composition. In fact, this composition must be regarded as a set of parameters to be used with the chosen mathematical model. The link between this parameter + equation combination and the true physical behaviour is performed through a comparison with experimental data. This should explain the construction of the third chapter that has three parts:
- It starts with the description of experimental information: section 1 discusses pure components; section 2 discusses mixtures. The third section illustrates how this experimental information is used for parameter estimation using data regression.
- In section 4, the actual thermodynamic models are summarised. No effort was made to be exhaustive (the user is in principle limited to what his process simulator provides), but it was rather attempted to help the reader make the link between the molecular structure of the fluid phase and the significance of the parameters values.
- In a final, short but important section, the concept of key component is introduced. The sensitivity of the requested type of information to some parameters may be much larger than to others. Some guidelines are provided here to help the engineer focus his attention on the most important parameters (i.e. key components or key binaries).
The fourth chapter puts the modelling issue in the perspective of the true physical behaviour of a physical system. The phase behaviour depends greatly on the type of mixture that is considered. This is why a number of important industrial mixtures are discussed. For each system, some model recommendations are suggested. The discussion is organised in two main parts:
- First, the property behaviour is discussed through some thermodynamic diagrams. As phase properties are considered qualitatively independent of the fluid composition, the example of pure CO2 is used.
- Second, the phase diagrams are analysed in some details. Here, the varying complexity of several types of mixtures is shown.
The true originality of our approach is presented in the final, fifth chapter. It both concludes the theoretical part of this endeavour (which is the book), and introduces the practical part that will be proposed on the web. The intention is to publish electronically a number of case studies that illustrate industrial examples that have been treated at IFP Energies nouvelles. The solution pattern will follow as closely as possible the approach suggested in this fifth chapter. The same final chapter also lists some important industrial processes in order to guide the engineer in his problem-analysis.
This user guide is not designed as a thermodynamics textbook. Many very good thermodynamic handbooks exist for helping teachers in designing a course in a logical, linear fashion (Elliot & Lira , Prausnitz et al. , Smith & Van Ness , O’Connell , Vidal …). Teachers can use this document as suggested below, but will be disappointed by the lack of demonstrations, and the non-linear conception of the logic.
This manual is not either a review of the existing thermodynamic models, that would help developers in understanding and code the most adapted model to their situation. Other authors have made a significant effort in summarizing the models in a logical fashion, stressing their strengths and weaknesses and providing indications on how they should be coded, if needed (Poling et al. ; Kontogeorgis & Folas ; Mollerup & Michelsen ; Riazi …). Thermodynamic experts will be disappointed by the lack of completeness, of numerical parameters and of algorithmic analysis.