Identification of reservoir types in deep carbonates has always been a great challenge due to complex logging responses caused by the heterogeneous scale and distribution of storage spaces.Traditional cross-plot analy...Identification of reservoir types in deep carbonates has always been a great challenge due to complex logging responses caused by the heterogeneous scale and distribution of storage spaces.Traditional cross-plot analysis and empirical formula methods for identifying reservoir types using geophysical logging data have high uncertainty and low efficiency,which cannot accurately reflect the nonlinear relationship between reservoir types and logging data.Recently,the kernel Fisher discriminant analysis(KFD),a kernel-based machine learning technique,attracts attention in many fields because of its strong nonlinear processing ability.However,the overall performance of KFD model may be limited as a single kernel function cannot simultaneously extrapolate and interpolate well,especially for highly complex data cases.To address this issue,in this study,a mixed kernel Fisher discriminant analysis(MKFD)model was established and applied to identify reservoir types of the deep Sinian carbonates in central Sichuan Basin,China.The MKFD model was trained and tested with 453 datasets from 7 coring wells,utilizing GR,CAL,DEN,AC,CNL and RT logs as input variables.The particle swarm optimization(PSO)was adopted for hyper-parameter optimization of MKFD model.To evaluate the model performance,prediction results of MKFD were compared with those of basic-kernel based KFD,RF and SVM models.Subsequently,the built MKFD model was applied in a blind well test,and a variable importance analysis was conducted.The comparison and blind test results demonstrated that MKFD outperformed traditional KFD,RF and SVM in the identification of reservoir types,which provided higher accuracy and stronger generalization.The MKFD can therefore be a reliable method for identifying reservoir types of deep carbonates.展开更多
The brittleness index plays a significant role in the hydraulic fracturing design and wellbore stability analysis of shale reservoirs.Various brittleness indices have been proposed to characterize the brittleness of s...The brittleness index plays a significant role in the hydraulic fracturing design and wellbore stability analysis of shale reservoirs.Various brittleness indices have been proposed to characterize the brittleness of shale rocks,but almost all of them ignored the anisotropy of the brittleness index.Therefore,uniaxial compression testing integrated with geophysical logging was used to provide insights into the anisotropy of the brittleness index for Longmaxi shale,the presented method was utilized to assess brittleness index of Longmaxi shale formation for the interval of 3155e3175 m in CW-1 well.The results indicated that the brittleness index of Longmaxi shale showed a distinct anisotropy,and it achieved the minimum value at β=45°-60°.As the bedding angle increased,the observed brittleness index(BI_(2_β))decreased firstly and increased then,it achieved the lowest value at β=40°-60°,and it is consistent with the uniaxial compression testing results.Compared to the isotropic brittleness index(β=0°),the deviation of the anisotropic brittleness index ranged from 10%to 66.7%,in other words,the anisotropy of brittleness index cannot be ignored for Longmaxi shale.Organic matter content is one of the main intrinsic causes of shale anisotropy,and the anisotropy degree of the brittleness index generally increases with the increase in organic matter content.The present work is valuable for the assessment of anisotropic brittleness for hydraulic fracturing design and wellbore stability analysis.展开更多
Since deep-sea gas hydrate-bearing sediments were drilled for the first time in the Blake Ridge in 1970,gas hydrates have been discovered at 53 drill sites in the continental margins of global oceans with internationa...Since deep-sea gas hydrate-bearing sediments were drilled for the first time in the Blake Ridge in 1970,gas hydrates have been discovered at 53 drill sites in the continental margins of global oceans with international scientific ocean drilling(DSDP/ODP/IODP Programs).As a result,massive amounts of geophysical well-logging data have been accumulated,which provide critical information for understanding the in-situ properties of gas hydrates and their host sediments.Gas hydrates have such physical and chemical properties as non-conductivity,low density,high acoustic velocity,and high hydrogen content,which form the basis of identifying gas hydrate reservoirs and predicting their distribution by well-logging data.A series of well-logging evaluation methods have been proposed to estimate gas hydrate saturation of sediments,including Archie equation,combined methods of density and nuclear magnetic resonance well logging,various forms of three-phase acoustic wave equations,and elastic wave velocity simulations based on different rock physical models.The distribution of gas hydrates is highly heterogeneous,which is mainly manifested in the selectivity of hydrate occurrence to the lithology of host sediments and to the nucleation sites within a host sediment of the same lithology.The scientific-ocean-drilling well logging data have also been preliminarily used for evaluating the heterogeneity of gas hydrate distribution and inferring the growth habit of gas hydrates in host sediments.Nevertheless,there still exist some problems.The formation models used in logging evaluation are in general oversimplified,in which only two or three stratal components are involved.The application of high-resolution logging while-drilling(LWD)data remains limited.Log interpretation is not closely integrated with core geology.Therefore,joint inversion of lithologic components,porosity and gas hydrate saturation based on more complex formation models,together with the applications of high-resolution LWD logging data and core calibration,may represent an important direction in future welllogging evaluation of gas hydrate reservoirs.展开更多
Most of the existing machine learning studies in logs interpretation do not consider the data distribution discrepancy issue,so the trained model cannot well generalize to the unseen data without calibrating the logs....Most of the existing machine learning studies in logs interpretation do not consider the data distribution discrepancy issue,so the trained model cannot well generalize to the unseen data without calibrating the logs.In this paper,we formulated the geophysical logs calibration problem and give its statistical explanation,and then exhibited an interpretable machine learning method,i.e.,Unilateral Alignment,which could align the logs from one well to another without losing the physical meanings.The involved UA method is an unsupervised feature domain adaptation method,so it does not rely on any labels from cores.The experiments in 3 wells and 6 tasks showed the effectiveness and interpretability from multiple views.展开更多
Coal bed methane is unconventional raw natural gas stored in coal seam with considerable reserves in China.In recent years,as the coal bed methane production,the safety and the use of resources have been paid more att...Coal bed methane is unconventional raw natural gas stored in coal seam with considerable reserves in China.In recent years,as the coal bed methane production,the safety and the use of resources have been paid more attentions.Evaluating coal bed methane content is an urgent problem.A BET adsorption isotherm equation is used to process the experimental data.The various parameters of BET equation under different temperatures are obtained;a theoretical gas content correction factor is proposed,and an evaluation method of actual coal bed methane is established.展开更多
Geophysical well logs are widely used in geological fields,however,there are considerable incompatibilities existing in solving geological issues using well log data.This review critically fills the gaps between geolo...Geophysical well logs are widely used in geological fields,however,there are considerable incompatibilities existing in solving geological issues using well log data.This review critically fills the gaps between geology and geophysical well logs,as assessed from peer reviewed papers and from the authors’personal experiences,in the particular goal of solving geological issues using geophysical well logs.The origin and history of geophysical logging are summarized.Next follows a review of the state of knowledge for geophysical well logs in terms of type of specifications,vertical resolution,depth of investigations and demonstrated applications.Then the current status and advances in applications of geophysical well logs in fields of structural geology,sedimentary geology and petroleum geology are discussed.Well logs are used in structural and sedimentary geology in terms of structure detection,in situ stress evaluation,sedimentary characterization,sequence stratigraphy division and fracture prediction.Well logs can also be applied in petroleum geology fields of optimizing sweet spots for hydraulic fracturing in unconventional oil and gas resource.Geophysical well logs are extending their application in other fields of geosciences,and geological issues will be efficiently solved via well logs with the improvements of advanced well log suits.Further work is required in order to improve accuracy and diminish uncertainties by introducing artificial intelligence.This review provides a systematic and clear descriptions of the applications of geophysical well log data along with examples of how the data is displayed and processed for solving geologic problems.展开更多
基金supported by the National Natural Science Foundation of China(No.U21B2062)the Natural Science Foundation of Hubei Province(No.2023AFB307)。
文摘Identification of reservoir types in deep carbonates has always been a great challenge due to complex logging responses caused by the heterogeneous scale and distribution of storage spaces.Traditional cross-plot analysis and empirical formula methods for identifying reservoir types using geophysical logging data have high uncertainty and low efficiency,which cannot accurately reflect the nonlinear relationship between reservoir types and logging data.Recently,the kernel Fisher discriminant analysis(KFD),a kernel-based machine learning technique,attracts attention in many fields because of its strong nonlinear processing ability.However,the overall performance of KFD model may be limited as a single kernel function cannot simultaneously extrapolate and interpolate well,especially for highly complex data cases.To address this issue,in this study,a mixed kernel Fisher discriminant analysis(MKFD)model was established and applied to identify reservoir types of the deep Sinian carbonates in central Sichuan Basin,China.The MKFD model was trained and tested with 453 datasets from 7 coring wells,utilizing GR,CAL,DEN,AC,CNL and RT logs as input variables.The particle swarm optimization(PSO)was adopted for hyper-parameter optimization of MKFD model.To evaluate the model performance,prediction results of MKFD were compared with those of basic-kernel based KFD,RF and SVM models.Subsequently,the built MKFD model was applied in a blind well test,and a variable importance analysis was conducted.The comparison and blind test results demonstrated that MKFD outperformed traditional KFD,RF and SVM in the identification of reservoir types,which provided higher accuracy and stronger generalization.The MKFD can therefore be a reliable method for identifying reservoir types of deep carbonates.
基金supported by the post-doctoral project of Petrochina Southwest Oil&Gas Field Company“Research on Deep Shale Geomechanics and Effective Fracturing Factors”(Grant No.20210302-31)the Program of Introducing Talents of Discipline to Chinese Universities(111 Plan)(Grant No.D18016)+2 种基金the Sichuan Science and Technology Program(Grant No.2020JDJQ0055)the Nanchong-SWPU Science and Technology Strategic Cooperation Foundation(Grant No.SXHZ033)the Youth Scientific and Technological Innovation Team Foundation of SWPU(Grant No.2019CXTD09).
文摘The brittleness index plays a significant role in the hydraulic fracturing design and wellbore stability analysis of shale reservoirs.Various brittleness indices have been proposed to characterize the brittleness of shale rocks,but almost all of them ignored the anisotropy of the brittleness index.Therefore,uniaxial compression testing integrated with geophysical logging was used to provide insights into the anisotropy of the brittleness index for Longmaxi shale,the presented method was utilized to assess brittleness index of Longmaxi shale formation for the interval of 3155e3175 m in CW-1 well.The results indicated that the brittleness index of Longmaxi shale showed a distinct anisotropy,and it achieved the minimum value at β=45°-60°.As the bedding angle increased,the observed brittleness index(BI_(2_β))decreased firstly and increased then,it achieved the lowest value at β=40°-60°,and it is consistent with the uniaxial compression testing results.Compared to the isotropic brittleness index(β=0°),the deviation of the anisotropic brittleness index ranged from 10%to 66.7%,in other words,the anisotropy of brittleness index cannot be ignored for Longmaxi shale.Organic matter content is one of the main intrinsic causes of shale anisotropy,and the anisotropy degree of the brittleness index generally increases with the increase in organic matter content.The present work is valuable for the assessment of anisotropic brittleness for hydraulic fracturing design and wellbore stability analysis.
基金Projects funded by the National Natural Science Foundation of China“Turbidites in the Abyssal Plain of the South China Sea and Their Tectonic Implications”(No.41676029)“Gravity Flow Deposits in the Ocean-Continent Transition Zone of the Northern South China Sea”(No.41876049).
文摘Since deep-sea gas hydrate-bearing sediments were drilled for the first time in the Blake Ridge in 1970,gas hydrates have been discovered at 53 drill sites in the continental margins of global oceans with international scientific ocean drilling(DSDP/ODP/IODP Programs).As a result,massive amounts of geophysical well-logging data have been accumulated,which provide critical information for understanding the in-situ properties of gas hydrates and their host sediments.Gas hydrates have such physical and chemical properties as non-conductivity,low density,high acoustic velocity,and high hydrogen content,which form the basis of identifying gas hydrate reservoirs and predicting their distribution by well-logging data.A series of well-logging evaluation methods have been proposed to estimate gas hydrate saturation of sediments,including Archie equation,combined methods of density and nuclear magnetic resonance well logging,various forms of three-phase acoustic wave equations,and elastic wave velocity simulations based on different rock physical models.The distribution of gas hydrates is highly heterogeneous,which is mainly manifested in the selectivity of hydrate occurrence to the lithology of host sediments and to the nucleation sites within a host sediment of the same lithology.The scientific-ocean-drilling well logging data have also been preliminarily used for evaluating the heterogeneity of gas hydrate distribution and inferring the growth habit of gas hydrates in host sediments.Nevertheless,there still exist some problems.The formation models used in logging evaluation are in general oversimplified,in which only two or three stratal components are involved.The application of high-resolution logging while-drilling(LWD)data remains limited.Log interpretation is not closely integrated with core geology.Therefore,joint inversion of lithologic components,porosity and gas hydrate saturation based on more complex formation models,together with the applications of high-resolution LWD logging data and core calibration,may represent an important direction in future welllogging evaluation of gas hydrate reservoirs.
基金Supported in part by the National Natural Science Foundation of China under Grant 61903353in part by the SINOPEC Programmes for Science and Technology Development under Grant PE19008-8.
文摘Most of the existing machine learning studies in logs interpretation do not consider the data distribution discrepancy issue,so the trained model cannot well generalize to the unseen data without calibrating the logs.In this paper,we formulated the geophysical logs calibration problem and give its statistical explanation,and then exhibited an interpretable machine learning method,i.e.,Unilateral Alignment,which could align the logs from one well to another without losing the physical meanings.The involved UA method is an unsupervised feature domain adaptation method,so it does not rely on any labels from cores.The experiments in 3 wells and 6 tasks showed the effectiveness and interpretability from multiple views.
文摘Coal bed methane is unconventional raw natural gas stored in coal seam with considerable reserves in China.In recent years,as the coal bed methane production,the safety and the use of resources have been paid more attentions.Evaluating coal bed methane content is an urgent problem.A BET adsorption isotherm equation is used to process the experimental data.The various parameters of BET equation under different temperatures are obtained;a theoretical gas content correction factor is proposed,and an evaluation method of actual coal bed methane is established.
基金supported by National Natural Science Foundation of China(Grant No.42002133)strategic cooperation project of PetroChina and CUPB(China University of Petroleum,Beijing)(ZLZX2020-01)Science Foundation of China University of Petroleum,Beijing(No.2462023QNXZ010).
文摘Geophysical well logs are widely used in geological fields,however,there are considerable incompatibilities existing in solving geological issues using well log data.This review critically fills the gaps between geology and geophysical well logs,as assessed from peer reviewed papers and from the authors’personal experiences,in the particular goal of solving geological issues using geophysical well logs.The origin and history of geophysical logging are summarized.Next follows a review of the state of knowledge for geophysical well logs in terms of type of specifications,vertical resolution,depth of investigations and demonstrated applications.Then the current status and advances in applications of geophysical well logs in fields of structural geology,sedimentary geology and petroleum geology are discussed.Well logs are used in structural and sedimentary geology in terms of structure detection,in situ stress evaluation,sedimentary characterization,sequence stratigraphy division and fracture prediction.Well logs can also be applied in petroleum geology fields of optimizing sweet spots for hydraulic fracturing in unconventional oil and gas resource.Geophysical well logs are extending their application in other fields of geosciences,and geological issues will be efficiently solved via well logs with the improvements of advanced well log suits.Further work is required in order to improve accuracy and diminish uncertainties by introducing artificial intelligence.This review provides a systematic and clear descriptions of the applications of geophysical well log data along with examples of how the data is displayed and processed for solving geologic problems.
