RBF-based reconstruction method for tomographic imaging of temperature and water vapor concentration in flames

Tunable diode laser absorption spectroscopy (TDLAS) tomography can reconstruct the local integral absorbance distributions at two different center wavelength lines to obtain the characteristic parameter distributions of the component to be measured. For the reconstruction at each wavelength, a proper algorithm was required due to restricted projections. In this paper, a linear combination of radial basis function (RBF) was used to approximate the continuous distribution in the region of interest, then the compact support function was introduced to reduce the error and computing resources. The distributions of local integral absorbance were obtained by solving the coefficient vector of RBF. Numerical simulated data at different levels of Gaussian noise were fabricated for both temperature and concentration distributions and local integral absorbances at two different wavelengths were reconstructed. Reconstructed distributions of temperature and concentration using proposed RBF-based methods agreed well with real distributions.