The prediction of nitric oxide conversion by dielectric barrier discharge using an artificial neural network model

Abstract

NO conversion to NO2 by DBD (Dielectric Barrier Discharge) was investigated in the N2/O2/NO system. Since NO2 could be generated from NO oxidation by O species, the increase of specific energy input (SEI), function of discharge power and residence time, facilitated the production of O radicals, promoting NO2 production. However, high temperature accomplished by DBD inhibit the NO2 generation to some extent. Also, the conversion is affected by stoichiometric ratio, i.e., oxygen content and inlet NO concentration, which makes a challenge to characterize and predict the products using traditional methods, such as chemical kinetic model. To address above problem, an artificial neural network (ANN) model was developed to predict NO conversion by DBD in the N2/O2/NO system. The experimental data was adopted to train the proposed ANN model in order to simulate and predict the concentrations of NO, NO2, N2O and NOx during reactions. Good agreement was observed between the simulated results and the validated tests. The ANN model showed that inlet NO concentration plays a dominant role in NO2 generation, accounting for 36.22%, followed by residence time and discharge power, which were 26.25% and 23.52%, respectively. O2 content has marginal effect, taking 14.01%. As a byproduct, N2O was much more affected by stoichiometric ratio, which accounted for 64.75%, compared to 35.25%, belonging to discharge power and residence time.