Non-Luminous Flame Temperature Determination Method Based on CO2 Radiation Intensity

A new method for estimating the temperature of non-luminous flames is presented. The spectral radiation intensity emitted from uniform or temperature-axisymmetric CO2 gas is simulated. A new inversion scheme that utilizes intensity ratios instead of directly spectral intensities is applied. This technique eliminates the errors caused by inaccuracies in the absolute radiation intensity that result from both calibration and on-site measurement. For example, this situation can occur when a spectrometer acquires radiance within only a part of its field of view. Neural networks are used as inverse models. The proposed inversion approach shows good results with simulated data. The analysis of example measured emission spectra from practical flames is also demonstrated.