Seismic Prediction Method Of Tight Limestone Fault Controlled Reservoir In FuMan Oilfield

The ultra-deep(7500～9500m)strike-slip fault(SSF)fractured zone(SSFFZ)is one of the most important oil and gas enrichment zones in Tarim Basin.The formation of ultra-deep heterogeneous reservoirs is closely related to the multi-stage activities of SSFs and corrosion transformation.Seismic data provide the main reference information for the prediction of fractured heterogeneous reservoirs in ultra-deep SSFs,and play an important role in the identification of SSFFZs,physical parameters of fracture zones and lithofacies / fluid prediction.However,due to the limitations of effective frequency band of seismic data,noise distribution,modeling error and other factors,there is great uncertainty in the identification of ultra-deep SSFFZs and the prediction results of its heterogeneous reservoir,which challenges the accurate support of reservoir characterization.At present,seismic exploration technology is one of the most common and effective techniques used to characterize SSFFZs in oil and gas field exploration and development,especially for ultra-deep fault fractured zones.The seismic response of different energy and disordered reflection caused by SSFFZ is the main feature of Ordovician carbonate fractured vuggy reservoir.It is controlled by SSF.Some of them are mixed strong "beaded" reflection groups,while others are banded disordered reflections of different widths,lengths and energy levels.The difficulty is that the latter is difficult to identify directly from seismic data.It is difficult to distinguish such reflection features effectively by using conventional attribute analysis.This paper presents a new combined method for seismic identification of ultra-deep SSFFZ.By introducing the scanning results of formation dip angle,a horizontal smoothing filter based on seismic dip angle data is constructed,and the horizontal smoothing filter is applied to the original seismic data to obtain the horizontal trend data with local formation background characteristics.Then,calculate the residual between the horizontal trend data and the original seismic data,and finally obtain the data to enhance the seismic response of the strike slip fault fracture zone,so as to guide the identification of the SSFFZs.After locking the fault fractured zone,according to the characteristics of its threedimensional structure with width in the transverse direction,thickness in the longitudinal direction and subsection extension,two kinds of seismic identification ideas are proposed,namely,the main fault fracture zone and the micro fault system between the main faults.For the main fault fractured zones,the artificial intelligence(AI)method is used to identify the internal configuration characteristics of the fracture zone based on the seismic feature interpretation framework of the fractured zone and the label of three-dimensional forward training samples.For the micro fault systems,a weak information extraction and enhanced tracking method based on wavelet decomposition and energy sorting optimization and reconstruction is used to identify the configuration characteristics of the micro fault systems between the main fault fractured zones,which further solves the problem of identifying the internal configuration characteristics of the ultra-deep SSFFZs.However,the prediction of physical property difference in fracture zone is another bottleneck and difficulty in reservoir description.The inversion method based on traditional isochronous sediment related low-frequency impedance model is difficult to describe heterogeneous reservoirs.In order to solve this bottleneck problem,this study puts forward the view that the fracture of SSF is geological isochronous,and fully considers this factor when establishing the low-frequency impedance model.A multi iteration strategy of facies controlled post stack geostatistics inversion is adopted to improve the accuracy of fault controlled heterogeneous reservoir description.This method extracts the information of strike-slip fault zone wrapped in tight limestone,and combines the prior information to establish the "double isochronous" framework of heterogeneous strata.Then,the iterative deterministic inversion is implemented,and the absolute impedance data consistent with the drilling and geological understanding is obtained by combining the relative impedance and the heterogeneous low-frequency impedance model.Considering the reservoir resolution of seismic inversion,post stack statistical inversion is implemented to maximize the inversion of underground heterogeneous reservoir characteristics.At the same time,in order to solve the problem that some reservoirs are filled with solid materials,a new method is proposed to jointly identify the characteristics of ultra deep solid filled caverns by using supervised LDA waveform component analysis based on local unsteady state and automatic reconstruction method based on adaptive waveform component features.Firstly,based on the supervised component analysis of local data,combined with the waveform and component combination characteristics of filled,partially filled and unfilled caverns confirmed by drilling,the characteristics of ultra deep solid filled caverns are identified through efficient automatic reconstruction method,and the seismic characteristic data of unfilled heterogeneous bodies are obtained to further improve the prediction accuracy of heterogeneous reservoirs in fractured zones.Finally,the prediction results of heterogeneous reservoirs are verified by model data and actual data to guide the quantitative characterization of fractured vuggy reservoirs.The effectiveness of this method is verified by actual seismic data,drilling,logging and logging data.The application shows that the technical system in this paper has achieved remarkable results in Fu Man oilfield and in the surrounding blocks,indicating that the method has good practicability and popularization.