Thèse de doctorat en Génie chimique
Sous la direction de Jean-Louis Chevalier.
Soutenue en 1998
à Aix-Marseille 3 .
Ce mémoire décrit la mise au point et l'utilisation d'un modèle de four de vapocraquage dans le but d'optimiser les conditions de marche industrielle. Le formalisme développé est un modèle de connaissance basé sur les lois de transfert thermique et d'hydrodynamique et sur un schéma cinétique simplifié des réactions de craquage. Il intègre également tous les paramètres liés au cokage des tubes, phénomène subi qui limite la durée des runs et pénalise la sélectivité de l'opération. Le premier chapitre fait le point sur la bibliographie du sujet et les outils actuellement existants. Une première version du modèle avait été mise au point pour un ancien type de four. . .
Kinetic and thermal modelling of liquid petroleum gases in modern furnaces
This memo describes the development of a steam-cracking furnace model to optimise furnace operating conditions. The model is a knowledge model using heat transfer laws, pressure drop equation and simplified kinetic reactions scheme. The coking model and its effects on furnace operations are taken into account in model. Coke is the principal constraint which reduces run length and decreases the selectivity of steam cracking operations. The first part of this report describes the literature and the existing tools. It presents also the version developed for Lavera old horizontal gas furnaces cracking. The second part is development, tuning and validation of the model for new vertical furnaces. The kinetic scheme from the previous version has been tuned by little adjustment of the kinetic constants. Heat transfer, pressure drop and coking models are updated with new laws to improve the results of the models and to take into account of the new technology of the furnace. Then, the model has been tuned on a large set of industrial data and validated following a sever procedure before use in simulation. Applications of the model, which are the justification of the work progressed, are described in the third part. The first application is to follow actual runs to check the running conditions of the furnace and to detect as soon as possible any unexpected event. The second application is optimisation of turn operating conditions to improve the benefit of the plant. . .