Direct Depth Domain Seismic Inversion Method And Its Application

The seismic exploration target of oil and gas resources has gradually shifted to deep and complex subtle oil and gas reservoirs.As a result,the corresponding seismic data processing and interpretation methods have evolved from simple to complex.Among these methods,because the prestack seismic depth migration technology can effectively carry out high-precision seismic imaging in complex structural areas,it has been widely used in the processing of three-dimensional seismic data.However,due to the lack of depth-domain seismic wavelet estimation methods and seismic record synthesis methods,the inversion and interpretation of depth-domain seismic data are usually implemented in the time domain.It is then necessary to convert the results back to the depth domain.Such a process is complex and time-consuming.Meanwhile,the conversion of data between different domains leads to the loss of useful information and the introduction of errors.This dissertation aims to form a practical and efficient direct depth domain seismic inversion workflow to provide a foundation for the direct inversion and interpretation of depth-domain seismic data.The seismic wavelet estimation,seismic record synthesis,well-to-seismic ties,elastic parameter inversion,and fluid detection in the depth domain are studied in this dissertation to achieve this goal.The main research work and innovation of this dissertation are as follows:(1)Direct depth domain seismic record synthesis.Combining the characteristics of depth-domain seismic wavelet varying with medium velocity with non-stationary convolution theory,a depth domain convolution model which can synthesize seismic records directly in the depth domain is proposed.This method replaces the stationary seismic wavelet in the conventional convolution model with the velocity variant depthdomain seismic wavelet.In addition,a depth-domain phase-shift seismic record synthesis method in matrix form is proposed based on the wave equation migration theory and phase shift method.The model test results show that the accuracy of the two direct depth domain seismic record synthesis methods is the same and is higher than that of the indirect depth domain seismic record synthesis method.(2)Direct depth domain seismic wavelet estimation.The depth-domain generalized seismic wavelet is derived based on the time-domain generalized seismic wavelet.Then,a statistical seismic wavelet estimation method and a deterministic seismic wavelet estimation method are proposed,respectively.The statistical seismic wavelet estimation method is used for the initial well-to-seismic ties.The deterministic seismic wavelet estimation method is used for subsequent iterative well-to-seismic ties and seismic inversion.Because the depth-domain seismic wavelet model is determined by the order and the reference wave number,only these two parameters need to be estimated when extracting the wavelet.The corresponding wavelet extraction method is efficient and straightforward and can be used to realize efficient well-to-seismic ties in the depth domain.(3)Direct depth domain well-to-seismic ties.An efficient direct depth domain well-to-seismic ties method is proposed based on the direct depth domain seismic record synthesis method and seismic wavelet estimation method.This method directly ties the synthetic record with the actual seismic trace at the well location in the depth domain and can introduce the dynamic regularization method to improve calibration efficiency.Real-data examples show that the method can output a reliable well-toseismic relationship after 2-4 calibrations.(4)Direct depth domain elastic inversion.A model-driven depth-domain elastic inversion method is proposed based on the stack impedance inversion theory and the depth domain convolution seismic record synthesis method.This method is solved by the linear method,which has high calculation efficiency.The inversion results obtained by this method can be directly compared with depth domain logging and geological data.The proposed method has been successfully applied to predicting favorable reservoirs in a field area of the South China Sea.The results verify the effectiveness and reliability of the proposed method.(5)Fluid detection based on depth-wavenumber spectrum feature extraction.Based on an asymmetric analysis window,a depth-domain generalized hyperbolic Stransform method is proposed.The method is combined with the fluid mobility calculation method to realize the fluid detection of depth-domain seismic data.Synthetic and real-data examples demonstrate that the asymmetric analysis window could highlight the low wavenumber components of depth-domain seismic signals and provide high resolution,which is helpful to identify low-wavenumber seismic anomalies accurately.