Measurements and modeling of physical properties for oil and biomaterial refining

Abstract

The aim of this thesis was to investigate a set of binary systems for developing necessary thermodynamic models for oil and biofuel industries. Extensive experimental work was performed for supplying the necessary vapor – liquid equilibria (VLE) and excess enthalpy data for the selected systems. Binary systems with C4 hydrocarbons + alkenes, alcohols and ketones were measured due to their importance in production of fuel additives. Static total pressure apparatus was utilized for these isothermal measurements. Important for fuel purification, systems with sulfur containing compounds were measured using recirculation still apparatus at isothermal conditions. Several experimental techniques for VLE measurements were applied for investigation of furfural containing binary systems and for the 2-methoxy-2-methylpropane (MIBK) + alcohol systems. Furfural can serve as one of the precursor in production of biofuel. Calorimetric measurements were also made for these systems using flow mixing calorimeter. Consistency of all experimental data was analyzed with four different consistency tests. The obtained experimental data were used for optimization of Wilson, NRTL an UNIQUAC model parameters and for validation of predictive group contribution model UNIFAC(Dortmund) and predictive COSMO-RS model. A method for improving group contribution methods is suggested. The distance weighting technique takes into account similarity of the compounds for optimization of the group contribution parameters. The suggested method is applicable to linearized group contribution methods. Normal boiling points of the pure compounds were predicted applying new technique to Joback – Reid and Marrero – Gani group contribution methods. Average improvement of the normal boiling point predictions was 6 K for the Joback – Reid method and 4 K for the Marrero – Gani method.