The correct representation of a fuel, in terms of its physical and chemical properties and its combustion kinetics poses, a challenge to modern engine development when state-of-the-art simulation technology is used. In this context, a promising approach is the use of surrogates that emulate the properties of real fuels, where the surrogates are made up of a significantly lower number of components than the original fuels. The goal of this paper is to present an algorithm that can be used to generate surrogates composed of real chemical components, as opposed to pseudo components.The algorithm was developed by simultaneously fitting the True Boiling Point (TBP) curve, the liquid density at 15 ℃ and the cetane number. To illustrate the algorithm, surrogates for four different fuels were generated: a commercially available European diesel and three research diesel proposed by the FACE (Fuels for Advanced Combustion Engines) CRC Research Group. Two of the resulting surrogates were produced on a lab-scale and subjected to laboratory examination. For validation, the experimental data for these two surrogates were compared to those for the target fuels and to data generated by thermodynamic models on the basis of the surrogates' compositions.Both the fitted properties and additional properties, which were not used for fitting, were compared with experimental properties such as the ASTM D86 boiling curve, content of aromatics, flash point, heating value, cloud point, viscosity, and tempera-

For characterization of crude oil and its primary fractions, the generation of substitute mixtures (surrogates) containing only real chemical components is a promising approach. The abandonment of pseudo-components, except for the utmost high-boiling fractions, allows for rigorous application of standard thermodynamic models (e.g., activity coefficients and equations of states), increasing reliability of phase-equilibrium calculations and predictive capabilities using process simulators. In this paper, an improved algorithm for characterization of petroleum fractions with real components is developed and applied to characterization of crude oil and its products through generation of substitute mixtures. The capabilities of emulating the separation behavior of crude oil are verified through a comprehensive analysis of a simulation conducted with real components by comparison to real plant data of an operating crude oil distillation unit (CDU). Additionally, a simulation based on the traditional pseudo-component approach is used for comparison. ■ MOTIVATIONCrude oil is a hydrocarbon mixture containing thousands of individual components ranging from light gases to very heavy, high-boiling components. 1 This mixture of a vast number of components with unknown chemical composition has to be processed in the refineries. Because of the increased need for efficiency, a much deeper understanding of the chemical specificity of refinery streams will be necessary for optimization. 2 A molecular-based characterization of the refinery streams can help to achieve this task. 3 A state-of-the-art approach for crude oil characterization is the pseudo-component approach, which is readily available in commercial simulation programs. Pseudo-components are generated on the basis of measured bulk properties, and all further calculations are based on these artificial components. Within the generation of pseudo-components, especially the estimation of their critical data and the acentric factor is arguable and no single commonly accepted method has been established thus far. 3 Another approach to characterize complex hydrocarbon mixtures is the use of real chemical components instead of pseudo-components. One advantage of using real components is the applicability of rigorous thermodynamic models instead of mainly empirical correlations for property estimation, which might be prone to errors. 4 This allows also for the modeling of fractions and mixtures with non-traditional components, e.g., bio-based products, which are not captured in the original pseudo-component approach. Furthermore, such an approach enables use of reaction kinetics or inclusion of key components within the simulation. 3 Moreover, it is possible to define key components within simulations, where calculations can rely on measured pure component data and interaction parameters. Additionally, the real component approach allows for experimental validation of predictions and applied models. ■ INTRODUCTIONPioneering activities in the generation of substitute mixtures for ...

In this paper a precedently developed surrogate optimization algorithm for fossil fuels, which originally allowed simultaneous fitting of the true boiling point (TBP) curve, the liquid density at 15 °C, and the cetane number, is refined toward its application to biodiesel and its mixtures with fossil diesel. For this purpose, the algorithm is extended (1) to also include fitting of the kinematic viscosity at 40 °C and (2) to account for peculiarities of biodiesel concerning its narrow boiling range and compensation of systematic errors of measured boiling curves. To illustrate these improvements, first, the algorithm is applied to property estimation and surrogate optimization of three different biodiesel fuels, for which surrogates consisting of one to three components are proposed. Second, a surrogate for a commercial European fossil diesel is calculated and produced in lab-scale. Finally, the algorithm is used for surrogate optimization and property estimation of mixtures of biodiesel and fossil diesel, considering fractions of biodiesel of 7% and 20% per volume. It is shown that the improved algorithm is capable of reliably optimizing surrogates for fuels containing both biogenic and fossil components.