Design of combustion experiments using differential entropy

Abstract

The aim of several combustion experiments is the determination of the rate coefficients of important elementary reactions. The experimental conditions are usually selected on the basis of local sensitivity analysis. Shock tube and tubular flow reactor experiments are often designed in such a way that only one reaction step is important at the investigated conditions. Sheen and Manion (J. Phys. Chem. A, 118 (2014) 4929–4941) suggested a method for the design of shock tube experiments based on differential entropy. Their method was modified and extended in this work. In the extended method, both the experimental and residual errors of the measurements are considered at the calculation of the posterior uncertainty of the determined rate parameters, the differential entropy matrix is calculated in an analytical way and the net information flux value is calculated for each suggested experimental point. In an iterative procedure, all investigated experimental points with negative net information flux values are discarded and the remaining experimental conditions are recommended for the measurements. The most valuable candidate experimental points can be determined based on the net information flux values. The method was used for the selection of experimental conditions for the determination of the rate coefficient of reaction NO2 + H = NO + OH at conditions similar to the tubular flow reactor experiments of Alzueta et al. (Energy Fuels, 15 (2001) 724–729). In these experiments, the oxidation of methanol was investigated with and without NO addition. Our method suggested a range of temperature, equivalence ratio and initial NO concentration, where the experimental data carry the most information on the rate coefficient of this elementary reaction.