Statistical Reconstruction of Gas Oil Cuts

Gas oil cuts are extremely complex mixtures of several thousands of different chemical species. Consequently, conventional petroleum analyses do not allow to obtain the molecular detail that is required for the development of robust and predictive kinetic models. Recently, two-dimensional Gas Chromatographic techniques (GC2D) have greatly improved the knowledge in the field of characterization of gas oils. However, they remain R&D tools and are hardly utilized in the refining industry. Hence, the goal of the statistical reconstruction of gas oils is to provide a surrogate for this GC2D analysis. To this aim, the gas oil cuts are characterized by means of matrices of molar fractions of pseudo-compounds, which are classified by chemical family and by carbon atom number. The input analyses are the Fitzgerald mass spectrometry, the sulfur speciation (one-dimensional gas chromatography coupled to a specific sulfur chemiluminescence detector) and the total nitrogen and basic nitrogen contents, and allow to quantify the proportions of all the chemical families present in the matrix. The simulated distillation is also used in order to introduce information on the volatility of the gas oil cut. The reconstruction method proposed in this paper is mainly based on a reference statistical distribution of the number of carbon atoms for the side chains connected to the naphtheno-aromatic cores. For each chemical family, the knowledge of the number of potential side chains and the estimation of the maximum length of these alkyl chains allow to determine the carbon number distribution by adjusting of the reference distribution. After reconstruction, the properties of the resulting molar fractions matrix are very close to the analyses used for the reconstruction. Moreover, the method allows to predict, with a high precision, complementary analyses such as the hydrogen content, the aromatic carbon content and the density at 15 ˚C. Finally, the matrix can be efficiently used to develop kinetic models like those employed at IFP to predict the performances of gas oil hydrotreating units