Multi-Scale Full Waveform Inversion Based on Normalized Energy Spectrum Misfit Function
Y. Hu, L. Han, P. Zhang and Y. Yin
Event name: 81st EAGE Conference and Exhibition 2019
Session: Poster: Full Waveform Inversion B
Publication date: 03 June 2019
Info: Extended abstract, PDF ( 1.5Mb )
Price: € 20
The misfit function of the conventional Full Waveform Inversion (FWI) based on the L2 norm is a strong nonlinear optimization problem and it may easily fall into a local minimum. In addition, the conventional FWI misfit function also needs an accurate source wavelet and a high signal-to-noise ratio (SNR) seismic data. These restrictions constraint the application of FWI in the real field seismic data processing. In this abstract, the misfit function of normalized energy spectrum is proposed to alleviate the strong nonlinear problem for the FWI method and reduce the dependence of high SNR seismic data and source wavelet. It is to match the normalized energy spectrum of recorded and synthetic data with the frequency distribution, which ignores the waveform and phase oscillation characteristics. Numerical examples show that the normalized energy spectrum FWI (NESFWI) results are better than the normalized FWI (NFWI) results in some areas, such as constructing a good initial velocity model, noise immunity, and source wavelet independence.