A Study On Seismic Prediction Of Fractured-Vuggy Carbonate Reservoirs

Deep carbonate reservoirs are the key and difficult issues for future oil and gas exploration,and fracture-vuggy carbonate reservoirs are one of the important types.The current exploration practice shows that due to the large buried depth,complex surface conditions,and strong heterogeneity within the fractured-vuggy body,the reflections on the seismic data are often characterized by scattered and chaotic.Due to the weak seismic response characteristics,the detection results of faults and geological boundaries are fragmented and discontinuous,it is difficult to predict the oil and gas content of the reservoir.Aiming at the key and difficult targets of carbonate fractured-vuggy reservoir prediction,this paper discusses the prediction methods from the perspectives of high-precision time-frequency analysis and optimized semblance detection methods.The main research contents of the paper are as follows:First of all,from the perspective of high-precision time-frequency analysis,this paper focuses on the method of hydrocarbon-bearing reservoir prediction based on Hilbert-Huang transform(HHT)and its improved methods.On the basis of detailed discussion on the theoretical basis and existing problems of the HHT method,a high-accuracy time-frequency analysis method combined with Variational Mode Decomposition(VMD)and Envelope Derivative Operator(EDO)is proposed.The VMD method can adaptively and non-recursively decompose the original seismic signal into a series of band-limited quasi-orthogonal IMFs,then the EDO operator with good non-negative characteristics and anti-noise ability is introduced to calculate instantaneous amplitude and frequency of the selected IMF which contains the effective information of each frequency band instead of Hilbert transform method,combined with the characteristics of "strong energy in low-frequency and high-frequency attenuation",the final obtained time-frequency distribution can track the energy change to predict the fracturedvuggy reservoir.Application of the VMD-EDO method in real seismic data verified its effectiveness,this method has higher time-frequency resolution to effectively characterize energy anomalies hidden in broadband seismic data.The VMD-EDO can be used as a valid time-frequency analysis method for oil and gas detection.Then,a workflow has been developed to improve the quality of semblance coherence algorithms for clearly describing the faults and fracture-vuggy boundaries.By adding structureoriented weighting factors on the basis of semblance coherence algorithm to reject random noise and improve the stratigraphic and structural discontinuities,and further using the Laplacian of Gaussian operator to the improved semblance attribute that sharpens the steeply dipping faults,attenuates the interference information.The effectiveness of this workflow is verified by real seismic data of western China,the final results highlights the discontinuities and fracture-vuggy boundaries more clearly and precisely,the workflow shows better at tracking discontinuities and has a higher resolution.Finally,combining the two methods,we obtain the mode of ‘Energy attributes describes area and improve semblance coherence defines the boundary’ of the recognition of the deep carbonate fractured-vuggy reservoir.Based on this recognition mode,We validate the effectiveness of our methods by applying it to real seismic data acquired from western China,the result shows that this method can effectively track the energy anomaly and describe the boundary information.The combination of the two methods for reservoir prediction has achieved good application results,and can provide important references for similar types of work.