Investigations Into Characteristics Of Burner Flames Using Spectral Analysis And Neural Network Techniques

Characteristics of flame radicals play an important role in an in-depth understanding of a combustion process including fuel and air inputs and pollutant emissions. This research programme is concerned with a method for the on-line prediction of the equivalence ratio of a methane-air premixed flame using spectroscopic and neural network techniques. A portable spectrometer-(USB2000+) is used to acquire the spectroscopic information of the flame radicals including OH*, CN*, CH* and C2* of a flame. The spectral characteristics of flames are investigated. A back propagation neural network model for the prediction of the equivalence ratio of a flame is established based on the flame radical radiative intensities. Experimental results from a laboratory-scale gas-fired combustion rig show the effectiveness of the proposed method. It is also found that the characteristics of radicals in premixed air and methane flame, vary with the air and fuel supplies, particularly, spectral intensities of four radicals, i.e., OH*、CN*、CH*、C2*, increase with the volume flow rate of methane. The prediction of the equivalence ratio of air and methane is also performed using the spectroscopic and neural network techniques. The research lays the foundation for the experimental study on the combustion characteristics of gas-and biomass-fired flames, which is of great significance on optimizing the combustion processes.