Obtaining a reduced kinetic mechanism for methyl decanoate using layerless neural networks

Major efforts in the search for techniques for the development of reduced kinetic mechanisms for biodiesel have been observed, since these mechanisms may have thousands of species. This paper proposes a reduction strategy and presents the development of a reduced kinetic mechanism for piloted jet diffusion flame of methyl decanoate (MD). The strategy consists of applying the DRG, Directed Relation Graph, technique for initial reduction, and the use of Layerless Neural Network (LNN) to define the main chain and obtain a skeletal mechanism. Hence the hypotheses of steady-state and partial equilibrium are applied, and the assumptions are justified by an asymptotic analysis. The main advantage of the strategy is to reduce the work required to solve the system of chemical equations by at least two orders of magnitude for MD, since the number of species is decreased in the same order.