Optimization And Application Of Pre-stack PP And PS Joint Inversion

With the rapid development of the national economy, the traditional oil and gas reservoirs(structural reservoirs, lithological reservoirs, etc.) have been difficult to meet the future demand for oil and gas resources, so we need to find new oil and gas reservoir types. Therefore, our current and future oil and gas exploration need to shift the focus to the unconventional oil and gas exploration, such as tight gas, shale oil, and fractured reservoirs. Meanwhile, with the rapid development of the computer industry, depth-domain pre-stack depth migration have been used widely, and depth-domain seismic data processing advantages become more evident, so depth-domain lithological parameters inversion have attracted wide attention of Geophysicists.Integrated application of pre-stack PP and PS angle gathers information in Joint-inversion, not only reduces the number of inversion matrix conditions, but also improves inversion accuracy, especially the shear wave velocity and density of inversion results can provide more accurate information on reservoir properties so that eliminate false. Compared with traditional approximate formula, Generalized Linear Inversion(GLI) method can reflect the variation of Zoeppritz equation more accurately, thus the inversion results are more accurate.In this paper, the author does a complete derivation of the GLI-related formulas and gives the definition of depth-domain wavelets according to the definition of time-domain wavelets. This paper uses deconvolution method on the Angle Domain Common Image Gathers(ADCIG) in Pseudo-depth domain based on Pseudo-speed conversion technology used in previous production of depth-domain synthetic seismogram. Experiments show that using L1 norm sparse spike deconvolution before pre-stack AVA GLI can eliminate the impact of wavelet, make the seismic data have higher resolution, and achieve the optimization of Pre-stack PP and PS joint inversion.In addition, reference to the low-frequency modeling techniques in time-domain inversion, this paper using Kriging techniques to establish the depth-domain low-frequency model, which used as the low frequency compensation of inversion results, making the P-wave velocity, the S-wave velocity, and density information got through inversion more accurately, thereby effectively guide lithology and fluid prediction, oil analysis. The author tests the Pre-stack sparse spike AVA GLI using on the model and the seismic data, which achieved good results.