A wavelet model on reconstructing complex aerodynamic field in furnace with acoustic tomography

Abstract

Acoustic method is a non-intrusive and promising full-field measurement for aerodynamic fields in furnace. A novel wavelet model, containing not only a wavelet skeleton but wavelet basis, was established for the acoustic measurement. It focuses adaptively to the optimal scale of basis function built as a 2-D ‘Mexican Hat’ wavelet. Since the inverse problem uses scarce data, regularization method was introduced. Defined with a pseudo Shannon entropy, an objective function was proposed for the determination of optimal regularization factor and further optimal scale of the basis, respectively. The model was essentially validated by numerical experiments with four typical scenarios of flow patterns. The wavelet basis was compared to Gaussian basis, with the same wavelet model skeleton, and gave inherently better results. Nevertheless, a configuration with more acoustic sensors proved to give more accurate reconstructions. Overall, the wavelet model turned out reliable to provide satisfactory reconstruction to random complex flow fields.