An effective numerical strategy for retrieving all characteristic parameters of an elastic scatterer from its FFP measurements
Autores: Hélène Barucq Rabai Djellouli
Fecha: 15.10.2020
Journal of Computational Physics
Abstract
A new computational strategy is proposed for determining all elastic scatterer characteristics including the shape, the material properties (Lamé coefficients and density), and the location from the knowledge of far-field pattern (FFP) measurements. The proposed numerical approach is a multi-stage procedure in which a carefully designed regularized iterative method plays a central role. The adopted approach is critical for recognizing that the different nature and scales of the sought-after parameters as well as the frequency regime have different effects on the scattering observability. Identification results for two-dimensional elastic configurations highlight the performance of the designed solution methodology.
BIB_text
title = {An effective numerical strategy for retrieving all characteristic parameters of an elastic scatterer from its FFP measurements},
journal = {Journal of Computational Physics},
pages = {109683},
volume = {419},
keywds = {
Helmholtz equation, Elastic scatterer, Inverse scattering problem, Ill-posedness, Regularization, Newton-type method
}
abstract = {
A new computational strategy is proposed for determining all elastic scatterer characteristics including the shape, the material properties (Lamé coefficients and density), and the location from the knowledge of far-field pattern (FFP) measurements. The proposed numerical approach is a multi-stage procedure in which a carefully designed regularized iterative method plays a central role. The adopted approach is critical for recognizing that the different nature and scales of the sought-after parameters as well as the frequency regime have different effects on the scattering observability. Identification results for two-dimensional elastic configurations highlight the performance of the designed solution methodology.
}
doi = {10.1016/j.jcp.2020.109683},
date = {2020-10-15},
}