Study On Gaussian Beam Migration Methods In Complex Medium

As an important approach to accurately image the subsurface, pre-stack depth migration(PSDM) is playing a more and more important role during the process of oil and gas exploration since 90th last centry. However, the conventional PSDM can't compromise imaging accuracy and computational efficiency. For example, reverse-time migration is very accurate but with low efficiency, on the contrary, Kirchhoff migration is very efficient but with low accuracy. Gaussian beam migration(GBM) is an imaging method with computational efficiency, imaging accuracy and flexibility, this unique advantage offer it high research value and broad applicate prospect. In this doctoral dissertation, i take a study on GBM methods in complex medium with two purpose, one is to intensify its advantage in order to make it more practical, the other is to extend its range of application so as to be well applied in oil and gas exploration.In this dissertation, i futherly develop the existing GBM methods, including the following several aspects: (1). With an unique approach, i derive GBM formulas suitable for different seismic data gather, and verify those validitiy with numerical tests. (2). I present an amplitude-preserved GBM method which can be used in complex surface conditions. By decompose seismic data into local plane-waves directly from the rugged surface, this method not only have better imaging quality especially the near surface part, but also get an amplitude preserved image which could measure angle-dependent reflection coefficients. (3). I present pre-stack elastic GBM which images the multi-component seismic data. With the P and S wave velocity, this method decompose multi-component data into local plane-waves with different initial wavemode and different direction, and downward-continue and image local plane-waves using elastic Gaussian beams. Due to decoupling the coupled elastic wavefield during the downward-continuation process ,this method can get P and S wave image without wavefield separation. Common Image Gathers(CIGs) are important outputs during the process of PSDM, it can be used to refine velocity field and provide AVO/AVA informations. In this dissertation, i study CIGs of GBM, and present ways to extract ODCIGs, ADCIGs and PWDCIGs. Because GBM naturely have ray angle imformations, it can directly extract ADCIGs without complicate mapping process. I present methods to calculate 2D/3D reflection angle, 3D reflction azimuth angle and reflection angle of converted wave, then verify the exactness of those method with extraction tests of ADCIGs.Parameters such as initial half-width, reference frequency and imaging angle range are important factors which control the imaging quality and computational efficiency of GBM. I analysis the effects and characters of those parameters, and give optimized seletion criterions. Concerning the aspects of efficiency enhancement, i give high effiective algorithm presented by Hill(1990), analysis its fundamental principle and use a series of optimized algorithms during programming.