| dc.description.abstract | Water +n-alkane systems at high pressures and temperatures, including the critical region, have been studied and modeled. All fluids exhibit fascinating behaviour near the critical point. Thus, the thesis begins with a description of supercritical fluids, including recent industrial applications. Special attention is given to the phase diagrams. The systems of our interests, water+n-alkanes, belong to the type III pf phase diagrams according to van Konynenburg and Scott, with three phase equilibrium and interrupted critical line, of which the lower branch connects the critical point of n-alkane and the upper critical end point and the upper branch runs from the critical point of pure water to high pressures (valid to water – n – hexacosane). As this work has utilized only published measurements, a review of vapour-liquid and liquid-liquid equilibria of water+n-alkanes systems at high pressures and temperatures has been made.
A classical cubic equation of state has been used to model such systems in the first place, both pure fluids and binary mixtures. As expected, a failure in the critical region was observed, especially for longer chain alkanes. Therefore, an alternative solution has been tested. This combines the traditional equation of state with a so-called crossover function (taken from the literature and modified), which leads to correct behaviour (correct critical exponents) near the critical point.
Both approaches have been tested for pure water and n-alkanes. In addition to the fact that the crossover function improves the behaviour in the critical region, the following was obtained: The modified crossover EOS gave better results in simultaneous description of the vapour pressure (within 0.2 – 1%), saturated liquids (within 2.4 – 3.9%), saturated gases (within 0.7 – 11.2%) and single phase properties(within 3.8 – 6.1%) compared with the classical EOS (vapour pressure with 0.2-0.9% accuracy, saturated liquids with 1-3.6% accuracy, saturated gases with 4-11% accuracy and single phase properties with 10-30% accuracy). The biggest improvement was observed in the single phase properties while keeping the accuracies of other properties comparable with the classical EOS. The modification of the crossover EOS brought an overall improvement compared to the crossover EOS from literature. The greatest advantage of the crossover EOS as modified here is the ability to predict the volumetric properties of both two and single phase regions of higher n-alkanes from the vapour pressure fit.
The application of crossover EOS to binary mixtures has been done in two ways. First, the method of isomorphism, which keeps the correct critical exponents along the critical lines, has been utilized. However, this method is suitable only for binary mixtures with continuous critical lines. Therefore, a new methods of applying crossover EOS to mixtures bas been introduced in this work.
The results for binary mixtures are divided into two parts. In the first part, the simple mixtures of methane-ethane and carbon dioxide-n-butane have been modelled in order to verify our calculations. Results comparable of crossover EOS to mixtures brought improvement in predicting the critical line for the methane-ethane system. However, the carbon dioxide-n-butane comparable accuracy was not obtained.
In the second part, water-n-alkane systems up to water-n-heptane have been modelled by classical EOS with and without association and with the crossover EOS as introduced here. The goal was to correlate vapour-liquid-equilibria (VLE) and/or liquid-liquid-equilibria (LLE) data in the range of 300 – 650 K in temperature and 1 – 40 M Pa in pressure for hydrocarbons in water and to predict the critical line. The new crossover model was suitable for water-methane and water ethane. The model brought improvement in certain regions, however, overall improvement was not observed. Mean square deviation (MSD) was MSD(T) = 1.58 K, MSD(P)=2.11%, MSD(x)=20.33%. MSD(y)=8.56% for mater-methane and MSD(T)=3.03 K, MSD(P)=18.60%, MSD(x)=47.35%, MSD(y)=10.00% for water-ethane. Water-n-alkanes higher than water-ethane were not described well with such a model. From the EOS tested, the best was the classical EOS without association. The best results were obtained for water-methane, with MSD(T)= 1.09 K, MSD(P)=1.73%, MSD(x)=21.90%, MSD(y)=4.37%. The worst results were obtained for water-hexane, with MSD(T)=1.52 K, MSD(P)= 3.00%, MSD(x)=63.7%, all with simple mixing rules. Water-propane and water-n-heptane lack good experimental data at high temperatures and pressures. Association din not seem to play a major role in the conditions considered. The results were comparable with more sophisticated models as the statistical association fluid theory (SAFT) or cubic plus association (CPS). The best mixing rules for correlation of VLE/LLE were not in general the best for predicting the critical lines. The prediction were very good for water-methane and water-hexane, the critical lines of the other systems were predicted with worse, but sufficient accuracy. The best results were for water-methane MSD(Tc) = 0.38% and MSD(xc)=14% up to 150 M Pa and the worst for water-pentane MSD(Tc)=3.34% and MSD(xc)=27.6% up to 80 M Pa, with critical pressure equals to the experimental one for all systems. Simultaneous fit to VLE/LLE and the critical points gave unacceptable poor fit to VLE/LLE with only a small gain for the critical line description. | nb_NO |
