The Q Estimation Method Study Of Sparse Bayesian Reflection Coefficient Inversion And Generalized Linear Inversion

In the development process of seismic exploration,seismic data are used for oil and gas reservoir detailed description and produce instantaneous changes of reservoir monitoring,so the high precision and high resolution seismic data acquisition requirements more and more high.The absorption and attenuation of formation is an important factor affecting the resolution of seismic data.The study of absorption and attenuation of seismic wave in underground media has become an important topic in seismic exploration field.The quality factor Q is a physical quantity describing the attenuation of seismic wave propagation in underground viscoelastic media.This paper first introduces the basic parameters of the attenuation,attenuation mechanism and main factors affecting the absorption of seismic wave attenuation of seismic waves,and then introduced the VSP(Vertical Seismic Profile)data quality factor Q value theory of three kinds of extraction methods,and the numerical model for these classification methods were compared and analyzed.Then,a new Q estimation method based on generalized linear inversion is put forward.This method fully combines the advantages of the existing methods,not only can effectively use the full band information,but also reduce the impact of noise on the accuracy of Q estimation,so as to achieve accurate estimation of Q value.The validity and feasibility of this method are verified by the numerical model,the physical model and the actual data.Finally,the development of a method for simultaneous estimation of Q and sparse Bayesian multi reflection coefficient inversion,weaken the assumptions the non sparse inversion of reflection coefficient,and Q filtering factor,frequency dispersion operator formation integrated into conventional steady-state convolution model,design reflection coefficient inversion results of the L_P norm the evaluation criteria of formation quality factor as products output from the side,then gives a new Q estimation method.The test of synthetic data and real data shows that the frequency band of the reflected reflection coincides with the frequency of seismic data,and at the same time,the correct high-frequency components are greatly restored,and the influence of Q value on the inversion result is reduced.