In the process of optimizing petroleum refining, it is essential to predict the distribution of products and their quality. Petroleum is a very complex substance and it is practically impossible to characterize it in the usual manner, by analyzing the quantity of each single component. Therefore, conventionally, it is characterized by methods based on the analysis of distillation, such as the True Boiling Point (TBP) distillation test and the ASTM D86 test. In order to predict the distribution

more »

... f products and their relationship with the operating conditions, stationary state simulations based on primary principals, in terms of fluctuations of mass, energy and relations of balance, are extremely important. As the number of components in petroleum is very high, the composition of the mixture cannot be used directly. Conventionally, dozens of pseudocomponents are created, in order to simulate petroleum. More recently the approximation of petroleum through a reduced set of real components has been proposed. This is the approach studied in this dissertation. The main challenge of this method is the determination of the real components chosen to simulate the original mixture, and their quantities. It is necessary to simulate the available distillation curves experimentally in the methodology for the selection of the substitute mixture. In this study, we simulated the TBP and ASTM D86 distillation curves in an Aspentech BatchSep environment. Applications of these simulations are presented in six different examples. Of these examples, five are TBP simulations and one is an ASTM D86 simulation. Some examples are based on mixtures with known compositions, and in others there are comparisons with experimental data from petroleum samples. The methodology of the substitute mixture and that of the pseudo-components were compared in the simulation of a distillation column. In general, the methodology proposed in literature is suitable when the only simulation that is required is a temperature curve. When density data is available, the method becomes limited by the availability of components in the data base and by the hypothesis that the components emerge pure, one by one, which is what happens when conventional methodology is adopted. A methodology that takes into account the fact that the samples are a mixture of components must be implemented so that the method is technically appealing