Development of vapour liquid equilibrium calculation methods for chemical engineering design

Abstract

This thesis deals with the development of computational methods for vapour liquid equilibrium (VLE) and volumetric properties. The VLE in this thesis can be divided into the low- and medium-pressure VLE with an experimental part and into the high-pressure VLE with a modelling and simulation part. The volumetric properties in this thesis deal with the extension of the model for compressed liquid densities.

At low-pressure VLE, the emphasis was on the optimisation of model parameters. Two apparatus were built, a circulation still and an automated total pressure apparatus for the vapour liquid equilibrium measurements. The measurements were correlated with activity coefficient models for the liquid phase and with equations of state for the vapour phase. A program for correlating the vapour liquid equilibrium was developed. The measurements and VLE models optimised were needed in developing gasoline additives to replace methyl tertiary-butyl ether (MTBE).

At near-critical VLE, the emphasis was on the robustness of the VLE and simulation routines. There was a need for a simulator to find out the dynamics of several vessels and buffer tanks when vessels were in a runaway condition, exposed to fire and imbalance of flows, or all of these events simultaneously. In addition, the operation point near the VLE critical point was of special interest. A dynamic simulator where the vapour and liquid phases were assumed to be in equilibrium was developed. The pressure relieving devices were assumed to be the only devices to control the flow of material. The effect of the pipe network was not included in the simulator.

The temperature range of the model for the compressed liquid density of mixture was extended. The rigorous bubble point pressure and the critical point computed from the cubic equation of state were more consistent with the experimental data than the pseudo-bubble point and pseudo critical point of the original model. The application range of the model was extended at the expense of accuracy, but the extended model was better than a cubic equation of state.