Pollution formation and the response of premixed hydrogen blending flame to the external excitation using large eddy simulation: The application of an artificial neural network and genetic algorithm

Abstract

BackgroundThe regulations for combustion pollution are getting more stringent every day. To decrease the level of pollution, serious attention to the unstable operating conditions of the flame is essential, as any flame may suffer from a level of instability. Further, the application of hydrogen-hydrocarbon blending fuels is developing in the industries. MethodsTherefore, the current study investigates the pollution produced by unstable combustion. An artificial neural network applying 4-15-1 neurons is also used by the input parameters of excitation frequency, amplitude, equivalence ratio, and hydrogen content. It is shown that the prediction is well for emissions. Using a genetic algorithm, the operating conditions corresponding to a minimum pollution emission are determined. Additionally, the importance of the governing parameters is evaluated in the pollution formation mechanism, and the excitation frequency is set in the first rank. Significant findingsHydrogen addition makes the NOx to be increased by a factor of 2.5 and CO to be mitigated by a rate of 2. This factor for increasing the excitation amplitude from 0.05 to 0.45 is around 5. Furthermore, the flame response to the inlet excitation is completely assessed using the heat release fluctuation, transfer function magnitude, and phase. Adding more hydrogen was shown to shift the cut-off frequency.