Noncontact direct temperature and concentration profiles measurement of soot and metal-oxide nanoparticles in optically thin/thick nanofluid fuel flames

A novel inverse reconstruction model was presented to simultaneously estimate temperature and concentration distributions of soot and metal-oxide nanoparticles in optically thin/thick nanofluid fuel flames from the knowledge of flame emission radiation intensities received by a CCD camera. The flame self-absorption effect was considered and the additional help with thermophoretic sampling particle diagnostics (TSPD) technique in the previous model was no longer needed. The monochromatic radiation intensity at the central wavelength of B channel was used to search the volume fraction ratio of metal oxide nanoparticles to soot, and those at the central wavelengths of R, G channels were used to calculate the unknown parameters profiles. Hybrid algorithms including one-dimensional searching, least-square QR decomposition (LSQR) and iterative algorithms were introduced to deal with the inverse problem. The radiation intensities were simulated based on the line-of-sight method, which were regarded as the input values in the inverse problem. The effects of input metal-oxide nanoparticles concentration fields, optical thicknesses and measurement errors on the reconstruction accuracy were discussed in details. Accurate reconstructed results can be obtained from the ideal radiation intensity even on the conditions of the extremely low distribution of metal-oxide nanoparticles concentration and the large optical thickness. The temperature and soot concentration fields can be successfully retrieved even with the noisy data, while the reconstruction of metal-oxide nanoparticles concentration profile with the random distribution at the locations close to the flame center was susceptible to the measurement error. However, in the flame with the continuous distribution in the metal-oxide nanoparticles concentration, the reconstructed profiles were still satisfying with the high level of noise. (C) 2019 Elsevier Ltd. All rights reserved.