<p>Knowing the physical properties of hydrocarbons and petroleum fractions is essential for designing most crude oil production and refining processes. Several correlations, called group contribution methods, have been proposed in the literature to assess these parameters and have been used extensively. The majority frequent correlations reported in the literature, where it is generally accepted in the petroleum industry, are used to describe unspecified fractions of oil. The only input parameters required are specific gravity and normal boiling point or molecular weight. Calculated properties include: normal boiling point (T<sub>b</sub>), Molecular weight (Mw), critical properties for instance critical pressure (P<sub>c</sub>), critical temperature (T<sub>c</sub>), critical volume (V<sub>c</sub>) and acentric factor (ω), as well as other parameters including Watson factor (K), and compressibility factor (Z<sub>c</sub>). In this approach, x samples of petroleum fractions of Messla and Sarir crude oils of Arabian Gulf Oil Company, Libya have been collected. A quantity of characterization technique of untainted and unknown petroleum fractions have been presented to foresee the physical properties of these petroleum fractions. Correlations for characterizing non-specific petroleum fractions suppose precise seriousness and boiling point as input parameters. These correlations are Twu Correlations, Cavett Correlations, Kesler-Lee Correlations and Riazi-Daubert Correlations. The physical properties of the compounds in terms of specific gravity and (T<sub>b</sub>) were shown on the algorithm and the correlation models for the auxiliary acids were corrected in this study. The considerations addressed revealed that there is no significant difference between the correlation models and the results obtained and it appears to be very close to the similar published data of the cited authors.</p>
Prediction of the Viscosity of Heavy Petroleum Fractions and Crude Oils by Neural Networks.
