低频缺失数据的弹性波全波形反演策略

弹性波全波形反演对地震数据的低频成分具有很强的依赖性。当多分量地震数据缺失低频成分时,混合域频率多尺度全波反演方法容易陷入局部极值,得到错误的反演结果。为降低弹性波波形反演的对初始模型的依赖,提出了Laplace-Fourier域反演与混合域反演相结合的反演策略,其基本思路为:先应用Laplace-Fourier域反演对低频缺失的地震数据指数衰减并反演,重建出模型的中、长波成分,然后将该模型作为后续混合域反演的初始模型,利用多尺度反演逐步恢复模型的短波长成分。模型数据的反演结果显示,在数据缺失低频时利用Laplace-Fourier域反演可获得一个更加准确的长波长纵、横波速度模型,将此模型作为混合域频率多尺度反演的初始模型能够更好地重建地下介质的弹性参数,从而得到更准确的结果。

Full Waveform Inversion Strategy of Elastic Wave with Low Frequency Missing Data

The full waveform inversion of elastic wave has a strong dependence on the low frequency components of seismic data.The mixed domain frequency multi-scale full wave inversion method easily fall into the local extremum and result in the wrong inversion results,when the low frequency component is absent from the multi-component seismic data.The inversion strategy of combining Laplace Fourier domain inversion with mixed domain inversion is proposed in order to reduce the dependence of elastic wave shape inversion on the initial model.We use the inversion of Laplace Fourier domin to attenuate and inverse the index of low-frequency missing seismic data.We reconstruct the medium and long wave components of the model.Taking the model as the initial model of the subsquent mixed domin inversion,we gradually restore the short wavelength components of the model by using multi-scale inversion.The inversion results of the model data show that a more accurate long wave longitudinal and shear wave velocity model can be obtained by using the Laplace Fourier domain inversion when the data is missing in low frequency.Taking this model as the initial model of multi-scale frequency inversion in the mixed domain,the elastic parameters of the underground medium can be reconstructed better and the more accurate results can be obtained.