Pre-stack seismic inversion is an effective way to investigate the characteristics of hydrocarbon-bearing reservoirs.Multi-parameter application is the key to identifying reservoir lithology and fluid in pre-stack inv...Pre-stack seismic inversion is an effective way to investigate the characteristics of hydrocarbon-bearing reservoirs.Multi-parameter application is the key to identifying reservoir lithology and fluid in pre-stack inversion.However,multi-parameter inversion may bring coupling effects on the parameters and destabilize the inversion.In addition,the lateral recognition accuracy of geological structures receives great attention.To address these challenges,a multi-task learning network considering the angle-gather difference is proposed in this work.The deep learning network is usually assumed as a black box and it is unclear what it can learn.However,the introduction of angle-gather difference can force the deep learning network to focus on the lateral differences,thus improving the lateral accuracy of the prediction profile.The proposed deep learning network includes input and output blocks.First,angle gathers and the angle-gather difference are fed into two separate input blocks with Res Net architecture and Unet architecture,respectively.Then,three elastic parameters,including P-and S-wave velocities and density,are simultaneously predicted based on the idea of multi-task learning by using three separate output blocks with the same convolutional network layers.Experimental and field data tests demonstrate the effectiveness of the proposed method in improving the prediction accuracy of seismic elastic parameters.展开更多
Common-image gathers are extensively used in amplitude versus angle(AVA)and migration velocity analysis(MVA).The current state of methods for anisotropic angle gathers extraction use slant-stack,local Fourier transfor...Common-image gathers are extensively used in amplitude versus angle(AVA)and migration velocity analysis(MVA).The current state of methods for anisotropic angle gathers extraction use slant-stack,local Fourier transform or low-rank approximation,which requires much computation.Based on an anisotropic-Helmholtz P/S wave-mode decomposition method,we propose a novel and efficient approach to produce angle-domain common-image gathers(ADCIGs)in the elastic reverse time migration(ERTM)of VTI media.To start with,we derive an anisotropic-Helmholtz decomposition operator from the Christoffel equation in VTI media,and use this operator to derive the decomposed formulations for anisotropic P/S waves.Second,we employ the first-order Taylor expansion to calculate the normalized term of decomposed formulations and obtain the anisotropic-Helmholtz decomposition method,which generates the separated P/S wavefields with correct amplitudes and phases.Third,we develop a novel way that uses the anisotropic-Helmholtz decomposition operator to define the polari-zation angles for anisotropic P/S waves and substitute these angles to decomposing formulations.The polarization angles are then calculated directly from the separated vector P-and S-wavefields and converted to the phase angles.The ADCIGs are thusly produced by applying the phase angles to VTI ERTM.In addition,we develop a concise approximate expression of residual moveout(RMO)for PP-reflections of flat reflectors in VTI media,which avoids the complex transformations between the group angles and the phase angles.The approximate RMO curves show a good agreement with the exact solution and can be used as a tool to assess the migration velocity errors.As demonstrated by two selected examples,our ADCIGs not only produce the correct kinematic responses with regards to different velocity pertubatation,but also generate the reliable amplitude responses versus different angle.The final stacking images of ADCIGs data exhibit the identical imaging effect as that of VTI ERTM.展开更多
This is a case study of the application of pre-stack inverted elastic parameters to tight-sand reservoir prediction. With the development of oil and gas exploration, pre-stack data and inversion results are increasing...This is a case study of the application of pre-stack inverted elastic parameters to tight-sand reservoir prediction. With the development of oil and gas exploration, pre-stack data and inversion results are increasingly used for production objectives. The pre-stack seismic property studies include not only amplitude verse offset （AVO） but also the characteristics of other elastic property changes. In this paper, we analyze the elastic property parameters characteristics of gas- and wet-sands using data from four gas-sand core types. We found that some special elastic property parameters or combinations can be used to identify gas sands from water saturated sand. Thus, we can do reservoir interpretation and description using different elastic property data from the pre-stack seismic inversion processing. The pre- stack inversion method is based on the simplified Aki-Richard linear equation. The initial model can be generated from well log data and seismic and geologic interpreted horizons in the study area. The input seismic data is angle gathers generated from the common reflection gathers used in pre-stack time or depth migration. The inversion results are elastic property parameters or their combinations. We use a field data example to examine which elastic property parameters or combinations of parameters can most easily discriminate gas sands from background geology and which are most sensitive to pore-fluid content. Comparing the inversion results to well data, we found that it is useful to predict gas reservoirs using λ, λρ, λ/μ, and K/μ properties, which indicate the gas characteristics in the study reservoir.展开更多
Angle domain common imaging gathers(ADCIGs)serve as not only an ideal approach for tomographic velocity modeling but also as a crucial means of mitigating low-frequency noise.Thus,they play a significant role in seism...Angle domain common imaging gathers(ADCIGs)serve as not only an ideal approach for tomographic velocity modeling but also as a crucial means of mitigating low-frequency noise.Thus,they play a significant role in seismic data processing.Recently,the Poynting vector method,due to its lower computational requirements and higher resolution,has become a commonly used approach for obtaining ADCIGs.However,due to the viscoelastic properties of underground media,attenuation effects(phase dispersion and amplitude attenuation)have become a factor,which is important in seismic data processing.However,the primary applications of ADCIGs are currently confined to acoustic and elastic media.To assess the influence of attenuation and elastic effects on ADCIGs,we introduce an extraction method for ADCIGs based on fractional viscoelastic equations.This method enhances ADCIGs accuracy by simultaneously considering both the attenuation and elastic properties of underground media.Meanwhile,the S-wave quasi tensor is used to reduce the impact of P-wave energy on S-wave stress,thus further increasing the accuracy of PS-ADCIGs.In conclusion,our analysis examines the impact of the quality factor Q on ADCIGs and offers theoretical guidance for parameter inversion.展开更多
Conventional AVO inversion utilizes the trace amplitudes of CMP gathers. There are three main factors affecting the accuracy of the inversion. First, CMP gathers are based on the hypothesis of horizontal layers but mo...Conventional AVO inversion utilizes the trace amplitudes of CMP gathers. There are three main factors affecting the accuracy of the inversion. First, CMP gathers are based on the hypothesis of horizontal layers but most real layers are not horizontal. Greater layer dip results in a greater difference between the observed CMP gathers and their real location. Second, conventional processing flows such as NMO, DMO, and deconvolution will distort amplitudes. Third, the formulation of reflection coefficient is related to incidence angles and it is difficult to get the relationship between amplitude and incidence angle. Wave equation prestack depth migration has the ability of imaging complex media and steeply dipping layers. It can reduce the errors of conventional processing and move amplitudes back to their real location. With true amplitude migration, common angle gathers abstraction, and AVO inversion, we suggest a method of AVO inversion from common shot gathers in order to reduce the effect of the above factors and improve the accuracy of AVO inversion.展开更多
For the complicated reservoir description of the GD oilfield, P-wave and S-wave elastic impedance inversion was carried out using pre-stack seismic data to accurately identify the lithology of the reservoir. The joint...For the complicated reservoir description of the GD oilfield, P-wave and S-wave elastic impedance inversion was carried out using pre-stack seismic data to accurately identify the lithology of the reservoir. The joint inversion was performed using three or more partial stacks to overcome the singularity of post-stack seismic inversion that can not satisfy the requirements of complex reservoir description and to avoid the instability of the inversion result caused by low signal-noise ratio in the pre-stack gather. The basic theory of prestack elastic impedance inversion is briefly described in this paper and, using real data of the GD oilfield, the key steps of angle gather wavelet extraction, horizon calibration, S-wave velocity prediction, and elastic parameter extraction were analyzed and studied. The comprehensive interpretation of multiple elastic parameters determined from log analysis is a key to improving the effect ofprestack seismic inversion.展开更多
The extensive use of depth-imaged seismic data produced by pre-stack depth migration(PSDM)leads to the necessity to synthesize seismogram directly in depth domain.However,since seismic wavelet in depth domain is depen...The extensive use of depth-imaged seismic data produced by pre-stack depth migration(PSDM)leads to the necessity to synthesize seismogram directly in depth domain.However,since seismic wavelet in depth domain is dependent on media velocities.The time-domain convolution operation directly used in depth domain does not meet the linear time-invariant condition.In this paper,we present a new method for genuine depth-domain seismic synthesis.This method constructs the velocity-dependent seismic wavelets varying adaptively with the corresponding interval velocities in the depth direction and weights them by the reflectivities,then the synthetic seismic record is obtained by the superposition of these weighted depth-varying wavelets.We applied this method to synthesize the seismic record of both a multi-layered geological model and the field data.The results show that the method can accommodate the intrinsic velocity-dependent variation characteristics of seismic events in depth domain and avoid the redundant depth-to-time and time-to-depth transformations.展开更多
基金financially supported by the National Natural Science Foundation of China(Grant Nos.42130810,42204135,42174170,and 42074165)the Natural Science Foundation of Hunan Province(Grant No.2023JJ40716)。
文摘Pre-stack seismic inversion is an effective way to investigate the characteristics of hydrocarbon-bearing reservoirs.Multi-parameter application is the key to identifying reservoir lithology and fluid in pre-stack inversion.However,multi-parameter inversion may bring coupling effects on the parameters and destabilize the inversion.In addition,the lateral recognition accuracy of geological structures receives great attention.To address these challenges,a multi-task learning network considering the angle-gather difference is proposed in this work.The deep learning network is usually assumed as a black box and it is unclear what it can learn.However,the introduction of angle-gather difference can force the deep learning network to focus on the lateral differences,thus improving the lateral accuracy of the prediction profile.The proposed deep learning network includes input and output blocks.First,angle gathers and the angle-gather difference are fed into two separate input blocks with Res Net architecture and Unet architecture,respectively.Then,three elastic parameters,including P-and S-wave velocities and density,are simultaneously predicted based on the idea of multi-task learning by using three separate output blocks with the same convolutional network layers.Experimental and field data tests demonstrate the effectiveness of the proposed method in improving the prediction accuracy of seismic elastic parameters.
基金supported by the National Key R&D Program of China(2020YFA0710604 and 2017YFC1500303)the Science Foundation of the China University of Petroleum,Beijing(2462019YJRC007 and 2462020YXZZ047)the Strategic Cooperation Technology Projects of CNPC and CUPB(ZLZX2020-05).
文摘Common-image gathers are extensively used in amplitude versus angle(AVA)and migration velocity analysis(MVA).The current state of methods for anisotropic angle gathers extraction use slant-stack,local Fourier transform or low-rank approximation,which requires much computation.Based on an anisotropic-Helmholtz P/S wave-mode decomposition method,we propose a novel and efficient approach to produce angle-domain common-image gathers(ADCIGs)in the elastic reverse time migration(ERTM)of VTI media.To start with,we derive an anisotropic-Helmholtz decomposition operator from the Christoffel equation in VTI media,and use this operator to derive the decomposed formulations for anisotropic P/S waves.Second,we employ the first-order Taylor expansion to calculate the normalized term of decomposed formulations and obtain the anisotropic-Helmholtz decomposition method,which generates the separated P/S wavefields with correct amplitudes and phases.Third,we develop a novel way that uses the anisotropic-Helmholtz decomposition operator to define the polari-zation angles for anisotropic P/S waves and substitute these angles to decomposing formulations.The polarization angles are then calculated directly from the separated vector P-and S-wavefields and converted to the phase angles.The ADCIGs are thusly produced by applying the phase angles to VTI ERTM.In addition,we develop a concise approximate expression of residual moveout(RMO)for PP-reflections of flat reflectors in VTI media,which avoids the complex transformations between the group angles and the phase angles.The approximate RMO curves show a good agreement with the exact solution and can be used as a tool to assess the migration velocity errors.As demonstrated by two selected examples,our ADCIGs not only produce the correct kinematic responses with regards to different velocity pertubatation,but also generate the reliable amplitude responses versus different angle.The final stacking images of ADCIGs data exhibit the identical imaging effect as that of VTI ERTM.
基金supported by the National Basic Priorities Program "973" Project (Grant No.2007CB209600)China Postdoctoral Science Foundation Funded Project
文摘This is a case study of the application of pre-stack inverted elastic parameters to tight-sand reservoir prediction. With the development of oil and gas exploration, pre-stack data and inversion results are increasingly used for production objectives. The pre-stack seismic property studies include not only amplitude verse offset （AVO） but also the characteristics of other elastic property changes. In this paper, we analyze the elastic property parameters characteristics of gas- and wet-sands using data from four gas-sand core types. We found that some special elastic property parameters or combinations can be used to identify gas sands from water saturated sand. Thus, we can do reservoir interpretation and description using different elastic property data from the pre-stack seismic inversion processing. The pre- stack inversion method is based on the simplified Aki-Richard linear equation. The initial model can be generated from well log data and seismic and geologic interpreted horizons in the study area. The input seismic data is angle gathers generated from the common reflection gathers used in pre-stack time or depth migration. The inversion results are elastic property parameters or their combinations. We use a field data example to examine which elastic property parameters or combinations of parameters can most easily discriminate gas sands from background geology and which are most sensitive to pore-fluid content. Comparing the inversion results to well data, we found that it is useful to predict gas reservoirs using λ, λρ, λ/μ, and K/μ properties, which indicate the gas characteristics in the study reservoir.
基金supported by the National Natural Science Foundation of China(NSFC)under contract number 42274147 and 41874144。
文摘Angle domain common imaging gathers(ADCIGs)serve as not only an ideal approach for tomographic velocity modeling but also as a crucial means of mitigating low-frequency noise.Thus,they play a significant role in seismic data processing.Recently,the Poynting vector method,due to its lower computational requirements and higher resolution,has become a commonly used approach for obtaining ADCIGs.However,due to the viscoelastic properties of underground media,attenuation effects(phase dispersion and amplitude attenuation)have become a factor,which is important in seismic data processing.However,the primary applications of ADCIGs are currently confined to acoustic and elastic media.To assess the influence of attenuation and elastic effects on ADCIGs,we introduce an extraction method for ADCIGs based on fractional viscoelastic equations.This method enhances ADCIGs accuracy by simultaneously considering both the attenuation and elastic properties of underground media.Meanwhile,the S-wave quasi tensor is used to reduce the impact of P-wave energy on S-wave stress,thus further increasing the accuracy of PS-ADCIGs.In conclusion,our analysis examines the impact of the quality factor Q on ADCIGs and offers theoretical guidance for parameter inversion.
基金This project is sponsored by the "Pre-Cenozoic Marine Oil and Gas Resource Research around the Bohai Area" of the Knowledge Innovation Project of The Chinese Academy of Sciences (No. KZCX1-SW-18)
文摘Conventional AVO inversion utilizes the trace amplitudes of CMP gathers. There are three main factors affecting the accuracy of the inversion. First, CMP gathers are based on the hypothesis of horizontal layers but most real layers are not horizontal. Greater layer dip results in a greater difference between the observed CMP gathers and their real location. Second, conventional processing flows such as NMO, DMO, and deconvolution will distort amplitudes. Third, the formulation of reflection coefficient is related to incidence angles and it is difficult to get the relationship between amplitude and incidence angle. Wave equation prestack depth migration has the ability of imaging complex media and steeply dipping layers. It can reduce the errors of conventional processing and move amplitudes back to their real location. With true amplitude migration, common angle gathers abstraction, and AVO inversion, we suggest a method of AVO inversion from common shot gathers in order to reduce the effect of the above factors and improve the accuracy of AVO inversion.
文摘For the complicated reservoir description of the GD oilfield, P-wave and S-wave elastic impedance inversion was carried out using pre-stack seismic data to accurately identify the lithology of the reservoir. The joint inversion was performed using three or more partial stacks to overcome the singularity of post-stack seismic inversion that can not satisfy the requirements of complex reservoir description and to avoid the instability of the inversion result caused by low signal-noise ratio in the pre-stack gather. The basic theory of prestack elastic impedance inversion is briefly described in this paper and, using real data of the GD oilfield, the key steps of angle gather wavelet extraction, horizon calibration, S-wave velocity prediction, and elastic parameter extraction were analyzed and studied. The comprehensive interpretation of multiple elastic parameters determined from log analysis is a key to improving the effect ofprestack seismic inversion.
文摘The extensive use of depth-imaged seismic data produced by pre-stack depth migration(PSDM)leads to the necessity to synthesize seismogram directly in depth domain.However,since seismic wavelet in depth domain is dependent on media velocities.The time-domain convolution operation directly used in depth domain does not meet the linear time-invariant condition.In this paper,we present a new method for genuine depth-domain seismic synthesis.This method constructs the velocity-dependent seismic wavelets varying adaptively with the corresponding interval velocities in the depth direction and weights them by the reflectivities,then the synthetic seismic record is obtained by the superposition of these weighted depth-varying wavelets.We applied this method to synthesize the seismic record of both a multi-layered geological model and the field data.The results show that the method can accommodate the intrinsic velocity-dependent variation characteristics of seismic events in depth domain and avoid the redundant depth-to-time and time-to-depth transformations.
