Optimization of the Arrhenius Parameters in a Pseudo-detailed Mechanism for Jet Fuel Thermal Oxidation Using Genetic and Simplex Algorithms

• Published in: Energy & fuels 2004-11, Vol.18 (6), p.1896-1908
• Main Authors: Wade, Andrew S, Kyne, Adrian G, Mera, Nicolae S, Pourkashanian, Mohammed, Ingham, Derek B, Whittaker, Sean
• Format: Article
• Language: English
• Subjects: Applied sciences; Energy; Energy. Thermal use of fuels; Exact sciences and technology; Fuels
• ISSN: 0887-0624; 1520-5029
• DOI: 10.1021/ef049814h

In this paper we present a novel way to determine new Arrhenius parameters and unknown initial reactant concentrations in a pseudo-detailed reaction mechanism for the autoxidation of aviation fuel. The technique employed is a specialized optimization procedure known as a genetic algorithm (GA), which utilizes an abstraction of the Darwinian principle of survival of the fittest to “breed” good solutions over a predefined number of “generations”. Temporal O2 consumption profiles for a given fuel or paraffin blend, which have been measured experimentally over a range of conditions, are reproduced by integrating the governing set of differential equations that result from the chemical kinetic mechanism and the optimized set of rate constants and initial reactant concentrations obtained using a GA inversion process. A simplex algorithm is then applied to further improve the GA-optimized parameters. The new set of rate constants lie within chemically reasonable predefined boundaries based upon values found for analogous reactions in the literature. This powerful method also offers the ability to optimize unknown reactant concentrations in real fuels which cannot be measured with current experimental techniques simultaneously with the reaction rate parameters. Such a process can lead to the development of reaction mechanisms whose newly optimized rate constants and reactant concentrations reproduce closely all the experimental data available, enabling a greater confidence in their predictive capabilities. The process is also shown to be an effective tool to facilitate the elucidation of areas of particular success or failure of the mechanism to simulate the autoxidation characteristics of jet fuels. A GA-optimized pseudo-detailed oxidation reaction mechanism for a “clean” jet fuel, approximated by paraffin blends virtually devoid of antioxidants, is presented, along with optimized values of reactant concentrations for several real fuels. This mechanism offers a remarkable improvement over previous mechanisms in its generality, i.e., its ability to simulate autoxidation behavior for a range of fuels and their blends. In addition, the new mechanism is capable of predicting autoxidation characteristics for fuel blends that were not part of the optimization process.