The Research On Thin Reservoir Prediction Under Strong Shield

F159 well block is located in the Midwest Fandong area,its main oil layer is E₂s₄₁C.The depth of the reservoir is 3000³400 meters.The reservoir thickness of F159 is about 20 meters,single layer thickness about 10 meters.In this area,the oil shale on the top of the reservoir form a strong reflection phase on the top of the target layer,and the reservoir signal is suppressed to make the reservoir prediction accuracy less than the requirement.In this area,the thickness of single layer is thin,the sand mudstone is mutual,and the lateral variation is large.Besides,the reservoir is buried deep,the accuracy of seismic data less than the requirements,and the main frequency is low,the frequency band is narrow.Therefore,to eliminate the effect of strong reflection layer on thin reservoir seismic reflection information and enhance thin reservoir weak reflection signal has important significance on improve the thin reservoir prediction accuracy,determine the favorable spatial distribution of thin reservoirs,improve the producing rate and tight oil reservoir development effect.After literature research home and abroad,according to the problem in this research area,,this paper makes detailed interpretation of seismic data,studies the geophysical characteristics of reservoir,which provide the basis data for forward modeling and data processing.Through forward modeling and analysis,we can find that improve the main frequency can improve the resolution of seismic data when there is no strong reflection layer.However,it is necessary to consider the influence of signal-to-noise ratio in practical application.In this paper,in order to improve the uniqueness of the solution and the spatial continuity of the decomposition,we study the strong shielding stripping technology based on multi-channel matching and horizon constraints.By the way,this paper achieves two methods of improving the resolution.The first one is spectral whitening processing based on Generalized S–transform（GST）.The second one is the spectral inversion based on Moore-Penrose algorithm.The former can improve the resolution of seismic data by expanding the seismic frequency band.The latter can improve the resolution by inverting the reflection coefficient of seismic section,it can achieve great effect to improve the resolution in model test,but the real seismic data processing needs further analysis and research.After removing the strong reflection layer,through the attribute analysis and comparison,we can find that the arc length attribute and the frequency attribute are not match with the well data accurately.Obviously,the RMS attribute has the best effect.Through the comparison,it can be seen that the absorption attenuation attribute based on area difference method,and we use this characteristic to predict reservoir range.Because of the strong reflection layer,the constrained sparse spike inversion is affected greatly.After striping the strong reflection layer,the inversion gets better.