Quantitative imaging of the elastic properties of the subsurface is essential for reservoir characterization. We assess with a realistic synthetic example the potentialities of 2D elastic full-waveform inversion for imaging complex onshore structures. Full-waveform inversion of land data is challenging because of the increased non-linearity introduced by free surface effects such as surface waves. To mitigate theses non- linearities, different multiscale strategies are assessed. Numerical optimization relies on the L-BFGS Quasi-Newton method which outperforms more classic preconditioned conjugate-gradient one. Sequential inversions of increasing frequencies define the most natural level of hierarchy in the multiscale imaging. We show that this regularization is not enough for adequate convergence in the case of land data.Asecond level of hierarchy over aperture angles implemented with complex-valued frequencies is necessary and allows convergence of the inversion towards acceptable velocity models.Among the possible strategies for sampling frequencies in the inversion, successive inversion of slightly-overlapping frequency groups has proven to be the most reliable one when compared with more standard sequential inversion of single frequencies. This suggests that simultaneous inversion of multiple frequencies is critical when considering complex wave phenomena such as surface wave propagation.
展开▼
