Dynamic Cross-Sectional Temperature Imaging From LAS Labeled Electrical Tomography

A laser absorption spectroscopy (LAS) fused electrical tomography (ET) is introduced to non-invasively measure the admittivity and temperature distributions of Bunsen burner flames. All the operations are carried out at the same region and height at multiple fuel-rich combustion states in the same cross section. The closer the fuel-to-air ratio of the combustion state is to one, the greater the values of admittivity and temperature. Random forest (RF) regression method is used to construct a mapping model for admittivity and combustion state to temperature. The maximum absolute error and relative error between mean temperatures of flame region estimated from reconstructed admittivity distributions and constructed mapping model with those measured by LAS are 60.45 K and 4.8%. The proposed method is used to monitor dynamic acoustically excited Bunsen burner flame, and the flame frequency measured in the spectrums of admittivity and temperature matches the measurements of a visible light photodetector. Experiments demonstrate the ability of ET to monitor the combustion states and temperature distribution of dynamic combustion fields.