Pub Date : 2025-11-04DOI: 10.1016/j.jappgeo.2025.106011
Qiang Liu , Liming Qiu , Yankun Ma , Dazhao Song , Miaomiao Yan , Wei Wang , Jie Liu , Limin Qie , Qi Jia , Peiwu Liao
The safety monitoring of high gas coal seam excavation is a critical measure for the prevention and control of coal and gas outburst accidents. In this paper, the geophysical methods such as direct current (DC) method and electromagnetic radiation (EMR) monitoring were used to evaluate the effectiveness of hydraulic flushing in the coal seam and to monitor and warn the safety of the coal seam roadway. The main conclusions are as follows: An outburst risk prediction method for coal roadway excavation process based on resistivity- electromagnetic radiation detection was proposed. The effective range of hydraulic flushing in the coal seam was found to be 8–12 m using the DC method. The reduction in gas content ranged from approximately 0.2 to 3 m3/t per unit, resulting in an overall decrease of around 40 %. EMR is effective in monitoring the dynamic events of the coal seam boring process. The signal was a fluctuation in the EMR signal after excavation began, reaching its maximum value during a coal burst. A method based on processing EMR-AE data to detect precursor signals is proposed. The Unified Precursor Index (UPI) of 0.75 is used as the early-warning threshold for coal burst events, indicating intense state changes in the coal mass. The UPI allows for the coal burst event to be detected 20 min in advance. The research provides a new perspective for the monitoring of coal rock dynamic disasters.
{"title":"Research on outburst risk prediction method for coal roadway excavation process based on resistivity- electromagnetic radiation detection","authors":"Qiang Liu , Liming Qiu , Yankun Ma , Dazhao Song , Miaomiao Yan , Wei Wang , Jie Liu , Limin Qie , Qi Jia , Peiwu Liao","doi":"10.1016/j.jappgeo.2025.106011","DOIUrl":"10.1016/j.jappgeo.2025.106011","url":null,"abstract":"<div><div>The safety monitoring of high gas coal seam excavation is a critical measure for the prevention and control of coal and gas outburst accidents. In this paper, the geophysical methods such as direct current (DC) method and electromagnetic radiation (EMR) monitoring were used to evaluate the effectiveness of hydraulic flushing in the coal seam and to monitor and warn the safety of the coal seam roadway. The main conclusions are as follows: An outburst risk prediction method for coal roadway excavation process based on resistivity- electromagnetic radiation detection was proposed. The effective range of hydraulic flushing in the coal seam was found to be 8–12 m using the DC method. The reduction in gas content ranged from approximately 0.2 to 3 m<sup>3</sup>/t per unit, resulting in an overall decrease of around 40 %. EMR is effective in monitoring the dynamic events of the coal seam boring process. The signal was a fluctuation in the EMR signal after excavation began, reaching its maximum value during a coal burst. A method based on processing EMR-AE data to detect precursor signals is proposed. The Unified Precursor Index (UPI) of 0.75 is used as the early-warning threshold for coal burst events, indicating intense state changes in the coal mass. The UPI allows for the coal burst event to be detected 20 min in advance. The research provides a new perspective for the monitoring of coal rock dynamic disasters.</div></div>","PeriodicalId":54882,"journal":{"name":"Journal of Applied Geophysics","volume":"244 ","pages":"Article 106011"},"PeriodicalIF":2.1,"publicationDate":"2025-11-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145467568","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-11-04DOI: 10.1016/j.jappgeo.2025.106020
Hao Hu , Shizhen Ke , Hongwei Shi , Yuhang Zhang , Hu Luo
The Maxwell-Garnett theory, a classical model for rock electrical properties, is widely used in interpreting formation dielectric characteristics and well-log responses. It enables inversion of key reservoir parameters such as water saturation and pore structure from dielectric logging data. However, its assumption of isolated inclusions may break down under conditions of high porosity or complex internal structures, leading to non-negligible prediction errors that could in turn affect the accuracy of inversion. This study employs Monte Carlo simulations to compute the effective permittivity of composite media consistent with the Maxwell-Garnett geometric assumptions and systematically analyzes the factors influencing model error and applicability. A Multiphase Iterative Maxwell-Garnett (MIMG) model is then proposed for media containing multiple types of inclusions. Results show that when inclusion permittivity exceeds that of the matrix, prediction errors increase markedly with volume fraction and permittivity contrast. In contrast, errors remain low when inclusion permittivity is lower. Regarding shape effects, errors increase with aspect ratio for oblate inclusions, while for prolate inclusions they either decrease then increase or decrease monotonically, depending on the permittivity contrast. Multiphase systems generally exhibit higher overall prediction errors than single-phase mixtures, indicating error accumulation. By iteratively introducing inclusions in a prescribed sequence, the MIMG model produces effective permittivity estimates more consistent with Maxwell-Garnett assumptions, thereby reducing prediction errors in multiphase systems and extending the theory's applicability in formation evaluation.
{"title":"Simulation of dielectric response in composite media based on Maxwell-Garnett theory and development of a multiphase dielectric model","authors":"Hao Hu , Shizhen Ke , Hongwei Shi , Yuhang Zhang , Hu Luo","doi":"10.1016/j.jappgeo.2025.106020","DOIUrl":"10.1016/j.jappgeo.2025.106020","url":null,"abstract":"<div><div>The Maxwell-Garnett theory, a classical model for rock electrical properties, is widely used in interpreting formation dielectric characteristics and well-log responses. It enables inversion of key reservoir parameters such as water saturation and pore structure from dielectric logging data. However, its assumption of isolated inclusions may break down under conditions of high porosity or complex internal structures, leading to non-negligible prediction errors that could in turn affect the accuracy of inversion. This study employs Monte Carlo simulations to compute the effective permittivity of composite media consistent with the Maxwell-Garnett geometric assumptions and systematically analyzes the factors influencing model error and applicability. A Multiphase Iterative Maxwell-Garnett (MIMG) model is then proposed for media containing multiple types of inclusions. Results show that when inclusion permittivity exceeds that of the matrix, prediction errors increase markedly with volume fraction and permittivity contrast. In contrast, errors remain low when inclusion permittivity is lower. Regarding shape effects, errors increase with aspect ratio for oblate inclusions, while for prolate inclusions they either decrease then increase or decrease monotonically, depending on the permittivity contrast. Multiphase systems generally exhibit higher overall prediction errors than single-phase mixtures, indicating error accumulation. By iteratively introducing inclusions in a prescribed sequence, the MIMG model produces effective permittivity estimates more consistent with Maxwell-Garnett assumptions, thereby reducing prediction errors in multiphase systems and extending the theory's applicability in formation evaluation.</div></div>","PeriodicalId":54882,"journal":{"name":"Journal of Applied Geophysics","volume":"244 ","pages":"Article 106020"},"PeriodicalIF":2.1,"publicationDate":"2025-11-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145467628","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-11-03DOI: 10.1016/j.jappgeo.2025.106014
Xiaotian Wang , Zhijiang Zheng , La Ta , Dongzhuo Xu , Haitao Zhou , Wenlong Liu , Yanqiang Wu , Guangqi Chen
Seismic fault interpretation is essential for understanding subsurface structures and has significant applications in resource exploration and earthquake assessment. Traditional methods rely on manual delineation or handcrafted seismic attributes, which are time-consuming and prone to subjective bias. Recent deep learning models, particularly CNNs, have improved fault segmentation but struggle with long-range dependencies and boundary continuity. To address these issues, we propose BDCNet, a novel Boundary Deformable Convolutional Network, which pioneers the Boundary Deformable Convolution and Mixed Boundary Loss as its key innovations. The Boundary Deformable Convolution dynamically adjusts convolutional sampling positions using a boundary-aware directional attention mechanism, improving the ability to capture long-range dependencies and refine fault boundaries. Mixed Boundary Loss integrates Binary Cross-Entropy loss, Dice loss, and a Boundary Aware loss, enhancing the sensitivity of the model to subtle fault structures and preserving boundary continuity. We validate BDCNet on publicly available seismic datasets and conduct extensive experiments. Results demonstrate that BDCNet outperforms widely used models such as U-Net, U-Net++, and DeepLabV3+, achieving superior performance in IoU, Dice, Precision, and Recall. By effectively capturing fault structures while preserving boundary continuity, BDCNet provides a robust and automated solution for seismic fault interpretation.
{"title":"Enhancing seismic fault segmentation for geological and engineering applications using the Boundary Deformable Convolutional Network","authors":"Xiaotian Wang , Zhijiang Zheng , La Ta , Dongzhuo Xu , Haitao Zhou , Wenlong Liu , Yanqiang Wu , Guangqi Chen","doi":"10.1016/j.jappgeo.2025.106014","DOIUrl":"10.1016/j.jappgeo.2025.106014","url":null,"abstract":"<div><div>Seismic fault interpretation is essential for understanding subsurface structures and has significant applications in resource exploration and earthquake assessment. Traditional methods rely on manual delineation or handcrafted seismic attributes, which are time-consuming and prone to subjective bias. Recent deep learning models, particularly CNNs, have improved fault segmentation but struggle with long-range dependencies and boundary continuity. To address these issues, we propose BDCNet, a novel Boundary Deformable Convolutional Network, which pioneers the Boundary Deformable Convolution and Mixed Boundary Loss as its key innovations. The Boundary Deformable Convolution dynamically adjusts convolutional sampling positions using a boundary-aware directional attention mechanism, improving the ability to capture long-range dependencies and refine fault boundaries. Mixed Boundary Loss integrates Binary Cross-Entropy loss, Dice loss, and a Boundary Aware loss, enhancing the sensitivity of the model to subtle fault structures and preserving boundary continuity. We validate BDCNet on publicly available seismic datasets and conduct extensive experiments. Results demonstrate that BDCNet outperforms widely used models such as U-Net, U-Net++, and DeepLabV3+, achieving superior performance in IoU, Dice, Precision, and Recall. By effectively capturing fault structures while preserving boundary continuity, BDCNet provides a robust and automated solution for seismic fault interpretation.</div></div>","PeriodicalId":54882,"journal":{"name":"Journal of Applied Geophysics","volume":"244 ","pages":"Article 106014"},"PeriodicalIF":2.1,"publicationDate":"2025-11-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145467566","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
We have performed Curie depth estimation from the aeromagnetic anomaly map of Bangladesh using the modified centroid method for scaling distribution of magnetic sources. The estimated Curie depth value ranges from 15 to 41 km and showcases the NNW-SSE increasing trend. It is observed that the Precambrian Stable Platform has shallower Curie depths than the Geosyncline Basin. Curie depth values are converted to geothermal gradient and heat flow anomalies ranging from 13–35 °C/km and 34‐96 mW/m2, respectively and found consistent with the geothermal gradient derived from the abandoned exploratory boreholes. Distinct anomalous regions are identified: 1) In the Stable Shelf region consisting of Rangpur Platform, shallow Curie depths are found ranging from 15 to 22 km; 2) moving further towards south, the region of Bogra shelf and SE sloping Hinge zone exhibits an intermediate Curie depths between 22 to 28 km; and 3) the Geosynclinal basin comprises various geological units have deeper Curie depth ranging from 28 to 38 km. The NW Stable Platform showcases a higher geothermal gradient (22–36 °C/km) and heat flow values (50–96 mW/m2) than the surroundings. Faults, interconnected fracture systems, and tectonics in the region are found as facilitators of basement heat transfer, making it a potential zone for future geothermal exploration. Apart from potential zones, an anomalously low geothermal gradient (<15 °C/km), corresponding to a lower heat flow (<40 mW/m2) and characterised by a deep Curie depth of 42 km, is observed near the Barisal-Chandpur High. This region exhibits deep-seated magnetic anomalies, complex tectonic settings, and lithospheric magnetic behaviour. Therefore, the observed results suggest complex geological processes, including continental-oceanic crustal transitions, crust-mantle interactions, compositional differences, surface heat distribution, and the structural characteristics of sedimentary layers strongly influence the Curie depth variation in Bangladesh.
{"title":"Thermal structure of Bangladesh using aeromagnetic data","authors":"Shubham Yadav , Abhey Ram Bansal , Mahak Singh Chauhan , Om Prakash","doi":"10.1016/j.jappgeo.2025.106018","DOIUrl":"10.1016/j.jappgeo.2025.106018","url":null,"abstract":"<div><div>We have performed Curie depth estimation from the aeromagnetic anomaly map of Bangladesh using the modified centroid method for scaling distribution of magnetic sources. The estimated Curie depth value ranges from 15 to 41 km and showcases the NNW-SSE increasing trend. It is observed that the Precambrian Stable Platform has shallower Curie depths than the Geosyncline Basin. Curie depth values are converted to geothermal gradient and heat flow anomalies ranging from 13–35 °C/km and 34‐96 mW/m<sup>2</sup>, respectively and found consistent with the geothermal gradient derived from the abandoned exploratory boreholes. Distinct anomalous regions are identified: 1) In the Stable Shelf region consisting of Rangpur Platform, shallow Curie depths are found ranging from 15 to 22 km; 2) moving further towards south, the region of Bogra shelf and SE sloping Hinge zone exhibits an intermediate Curie depths between 22 to 28 km; and 3) the Geosynclinal basin comprises various geological units have deeper Curie depth ranging from 28 to 38 km. The NW Stable Platform showcases a higher geothermal gradient (22–36 °C/km) and heat flow values (50–96 mW/m<sup>2</sup>) than the surroundings. Faults, interconnected fracture systems, and tectonics in the region are found as facilitators of basement heat transfer, making it a potential zone for future geothermal exploration. Apart from potential zones, an anomalously low geothermal gradient (<15 °C/km), corresponding to a lower heat flow (<40 mW/m<sup>2</sup>) and characterised by a deep Curie depth of 42 km, is observed near the Barisal-Chandpur High. This region exhibits deep-seated magnetic anomalies, complex tectonic settings, and lithospheric magnetic behaviour. Therefore, the observed results suggest complex geological processes, including continental-oceanic crustal transitions, crust-mantle interactions, compositional differences, surface heat distribution, and the structural characteristics of sedimentary layers strongly influence the Curie depth variation in Bangladesh.</div></div>","PeriodicalId":54882,"journal":{"name":"Journal of Applied Geophysics","volume":"244 ","pages":"Article 106018"},"PeriodicalIF":2.1,"publicationDate":"2025-11-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145467626","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Precise characterization of fracture systems in subsalt carbonate is a core challenge in oil and gas exploration in structurally complex areas. In the northeastern of the Amu Darya Basin in Central Asia, the combined effects of salt detachment and multiple stages of compressional deformation during the Himalayan period have led to the development of a multi-scale complex fracture network in the Callovian-Oxfordian carbonate, which are the main oil and gas producing layers in the region. This study introduces a refined methodology that integrates seismic frequency division with machine learning techniques to achieve hierarchical modeling of fracture systems. Applied in the eastern of the Amu Darya right bank, this approach enabled the construction of a high-resolution three-dimensional geological model of fracture-controlled reservoirs within subsalt carbonate formations. To resolve fractures of varying scales based on frequency sensitivity and required vertical resolution, the seismic data were decomposed into two dominant frequency domains: 25 Hz for large-scale faults with displacements exceeding 60 m, and 55 Hz for small-scale fractures with displacements less than 25 m. Hierarchical characterization of the fracture network was conducted using a lightweight and efficient three-dimensional UNet model, termed LightGEUnet, in conjunction with maximum likelihood attribute analysis. The LightGEUnet, through the grouped multi-axis Hadamard product attention (GHPA-M) and the multi-scale group aggregation fusion (GAF), demonstrates the advantages of low parameter count and high accuracy in both synthetic fault data and real seismic data, exhibiting excellent detection performance for large-scale faults. The likelihood attributes driven by high-frequency divided data, while effectively suppressing scattering noise from salt-gypsum rocks, simultaneously accomplish small-scale fracture detection, ultimately achieving a complete characterization of the fracture network distribution. This study provides a new pathway for fracture modeling in complex subsalt structural settings through the integration of “geology-machine learning,” offering novel ideas and experiences for the exploration and development of fracture-controlled carbonate gas reservoirs.
{"title":"Collaborative characterization of sub-salt carbonate fractures using LightGEUnet and seismic frequency division: A case study of the Eastern Amu Darya Right Bank","authors":"Yuzhe Tang , Hongjun Wang , Liangjie Zhang , Wenqi Zhang , Yunpeng Shan","doi":"10.1016/j.jappgeo.2025.106013","DOIUrl":"10.1016/j.jappgeo.2025.106013","url":null,"abstract":"<div><div>Precise characterization of fracture systems in subsalt carbonate is a core challenge in oil and gas exploration in structurally complex areas. In the northeastern of the Amu Darya Basin in Central Asia, the combined effects of salt detachment and multiple stages of compressional deformation during the Himalayan period have led to the development of a multi-scale complex fracture network in the Callovian-Oxfordian carbonate, which are the main oil and gas producing layers in the region. This study introduces a refined methodology that integrates seismic frequency division with machine learning techniques to achieve hierarchical modeling of fracture systems. Applied in the eastern of the Amu Darya right bank, this approach enabled the construction of a high-resolution three-dimensional geological model of fracture-controlled reservoirs within subsalt carbonate formations. To resolve fractures of varying scales based on frequency sensitivity and required vertical resolution, the seismic data were decomposed into two dominant frequency domains: 25 Hz for large-scale faults with displacements exceeding 60 m, and 55 Hz for small-scale fractures with displacements less than 25 m. Hierarchical characterization of the fracture network was conducted using a lightweight and efficient three-dimensional UNet model, termed LightGEUnet, in conjunction with maximum likelihood attribute analysis. The LightGEUnet, through the grouped multi-axis Hadamard product attention (GHPA-M) and the multi-scale group aggregation fusion (GAF), demonstrates the advantages of low parameter count and high accuracy in both synthetic fault data and real seismic data, exhibiting excellent detection performance for large-scale faults. The likelihood attributes driven by high-frequency divided data, while effectively suppressing scattering noise from salt-gypsum rocks, simultaneously accomplish small-scale fracture detection, ultimately achieving a complete characterization of the fracture network distribution. This study provides a new pathway for fracture modeling in complex subsalt structural settings through the integration of “geology-machine learning,” offering novel ideas and experiences for the exploration and development of fracture-controlled carbonate gas reservoirs.</div></div>","PeriodicalId":54882,"journal":{"name":"Journal of Applied Geophysics","volume":"244 ","pages":"Article 106013"},"PeriodicalIF":2.1,"publicationDate":"2025-11-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145467562","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Coalbed methane (CBM) represents a significant unconventional natural gas resource. Coal deformation exerts crucial influences on hydraulic fracturing effectiveness and fluid productivity in coal reservoirs, while logging-based identification accuracy of coal structure can directly guide zonal fracturing optimization in reservoir stimulation. To enhance the resolution of coal structure quantification using logging data, this study integrates principal component analysis (PCA) and multiple linear regression (MLR) methods with the geological strength index (GSI) chart specifically adapted for coal seams, establishing a logging coal structure index (LCSI). This index categorizes coal structures into four types: Type I (undeformed structure), Type II (brittle deformation structure), Type III (brittle-ductile transitional structure), and Type IV (ductile deformation structure). A well-specific quantitative index of coal deformation intensity (LCSIwell) was concurrently developed. High LCSIwell values predominantly cluster near regional reservoir-controlling structures, particularly at termination points of reservoir-controlling faults and within fold axial zones under compressional-shear stress. Furthermore, a hydraulic fracturing prediction (HFP) index (0–1 scale) was established through integration of entropy weight method (EWM) and grey relational analysis (GRA), with higher values indicating better hydraulic fracturing suitability. The study area was classified into four zones based on HFP Index. Analysis reveals that hydraulic fracturing effectiveness in coal seams is primarily controlled sequentially by coal structure, structural distance, and thickness of roof/floor strata. The northern and southwestern coal reservoirs exhibit brittle deformation, structural stability, and thicker roof-floor strata, thus constituting low difficulty zone (L zone) for hydraulic fracturing modification.
{"title":"Logging-based quantitative evaluation of coal deformation using PCA-MLR coupled with GSI: Implications for hydraulic fracturing zoning in structurally controlled CBM reservoirs","authors":"Quanliang Zou, Yingjin Wang, Guanqun Zhou, Xiaowei Hou","doi":"10.1016/j.jappgeo.2025.106015","DOIUrl":"10.1016/j.jappgeo.2025.106015","url":null,"abstract":"<div><div>Coalbed methane (CBM) represents a significant unconventional natural gas resource. Coal deformation exerts crucial influences on hydraulic fracturing effectiveness and fluid productivity in coal reservoirs, while logging-based identification accuracy of coal structure can directly guide zonal fracturing optimization in reservoir stimulation. To enhance the resolution of coal structure quantification using logging data, this study integrates principal component analysis (PCA) and multiple linear regression (MLR) methods with the geological strength index (GSI) chart specifically adapted for coal seams, establishing a logging coal structure index (<em>LCSI</em>). This index categorizes coal structures into four types: Type I (undeformed structure), Type II (brittle deformation structure), Type III (brittle-ductile transitional structure), and Type IV (ductile deformation structure). A well-specific quantitative index of coal deformation intensity (<em>LCSI</em><sub><em>well</em></sub>) was concurrently developed. High <em>LCSI</em><sub><em>well</em></sub> values predominantly cluster near regional reservoir-controlling structures, particularly at termination points of reservoir-controlling faults and within fold axial zones under compressional-shear stress. Furthermore, a hydraulic fracturing prediction (<em>HFP</em>) index (0–1 scale) was established through integration of entropy weight method (EWM) and grey relational analysis (GRA), with higher values indicating better hydraulic fracturing suitability. The study area was classified into four zones based on <em>HFP</em> Index. Analysis reveals that hydraulic fracturing effectiveness in coal seams is primarily controlled sequentially by coal structure, structural distance, and thickness of roof/floor strata. The northern and southwestern coal reservoirs exhibit brittle deformation, structural stability, and thicker roof-floor strata, thus constituting low difficulty zone (L zone) for hydraulic fracturing modification.</div></div>","PeriodicalId":54882,"journal":{"name":"Journal of Applied Geophysics","volume":"244 ","pages":"Article 106015"},"PeriodicalIF":2.1,"publicationDate":"2025-10-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145467563","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-10-31DOI: 10.1016/j.jappgeo.2025.106016
Feng Tan , Ping Yang , Jun-Xing Cao , Le Li , Zhihua Cui , Qian Ma
Dim-spot reservoirs, though theoretically as common as bright spots, remain challenging to detect because of seismic amplitude pitfalls caused by phase reversal and amplitude cancellation in P-wave to P-wave reflection waves. Two practical solutions are investigated to cope with the challenges in identifying dim-spot sandstones within the tight gas reservoirs of the Shaximiao Formation in China's Western Sichuan Depression. First, an optimized angle-domain stacking method for P-wave to P-wave reflection waves reduces polarity reversal effects by excluding incidence angles prone to destructive interference. Second, the P-wave to S-wave converted reflection waves directly resolve dim spots by eliminating phase reversal artifacts. Field validation in the Zhongjiang and Zitong areas confirmed that these approaches enhance reservoir detection. Previously obscured channel sandstones were successfully delineated, consistent with well-log data and seismic coherence attributes. The proposed methods show broad applicability to tight formations, offering a promising strategy to unlock overlooked hydrocarbon potential in similar basins globally.
{"title":"Challenges and solutions to seismic amplitude pitfalls in dim-spot sandstone reservoirs: A case study from the Shaximiao formation, the Western Sichuan Depression, China","authors":"Feng Tan , Ping Yang , Jun-Xing Cao , Le Li , Zhihua Cui , Qian Ma","doi":"10.1016/j.jappgeo.2025.106016","DOIUrl":"10.1016/j.jappgeo.2025.106016","url":null,"abstract":"<div><div>Dim-spot reservoirs, though theoretically as common as bright spots, remain challenging to detect because of seismic amplitude pitfalls caused by phase reversal and amplitude cancellation in P-wave to P-wave reflection waves. Two practical solutions are investigated to cope with the challenges in identifying dim-spot sandstones within the tight gas reservoirs of the Shaximiao Formation in China's Western Sichuan Depression. First, an optimized angle-domain stacking method for P-wave to P-wave reflection waves reduces polarity reversal effects by excluding incidence angles prone to destructive interference. Second, the P-wave to S-wave converted reflection waves directly resolve dim spots by eliminating phase reversal artifacts. Field validation in the Zhongjiang and Zitong areas confirmed that these approaches enhance reservoir detection. Previously obscured channel sandstones were successfully delineated, consistent with well-log data and seismic coherence attributes. The proposed methods show broad applicability to tight formations, offering a promising strategy to unlock overlooked hydrocarbon potential in similar basins globally.</div></div>","PeriodicalId":54882,"journal":{"name":"Journal of Applied Geophysics","volume":"244 ","pages":"Article 106016"},"PeriodicalIF":2.1,"publicationDate":"2025-10-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145467631","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Geophysical logging serves as a fundamental method in sandstone-type uranium exploration. Due to the large volume of logging data and the complexity of geological interpretation, artificial intelligence (AI) technologies provide efficient solutions for data processing and analysis. While lithological identification approaches in non-mineralized layers are similar to those used in oil, gas, and coal exploration, the mineralized layers present a unique challenge: anomalously high gamma-ray values are observed in some intervals without apparent lithological control, which can lead to misinterpretation. Therefore, accurate lithological classification of mineralized layers is therefore critical for the exploration and development of sandstone-type uranium deposits. This study focuses on the lithology of the mineralized lower Yaojia Formation in the Baolongshan uranium deposit, located in the southern Songliao Basin. A Random Forest regression model was employed to reconstruct anomalous intervals in the natural gamma-ray logging data, and an improved XGBoost model, optimized via Bayesian hyperparameter tuning, was used for lithological classification. Quantitative evaluations confirm the reliability of the reconstructed natural gamma-ray data, showing negligible differences in average values compared to non-anomalous data. Moreover, in the filled intervals where original gamma-ray values were excluded due to mineralization effects and subsequently reconstructed using the Random Forest model, the reconstructed gamma-ray curves not only exhibit morphological similarity to spontaneous potential curves but also show a mirror-image relationship with apparent resistivity and three-lateral resistivity curves, thereby further validating the reconstruction. Lithological classification accuracies for the mineralized layer across the three boreholes are 97.5 %, 96.5 %, and 98.6 %, with an average accuracy of 97.5 %. These results outperform those obtained from the original data, which yielded accuracies of 95.5 %, 96.0 %, and 97.3 %, with an average accuracy of 96.3 %. The AI-driven precise identification of lithology in mineralized layers significantly enhances the efficiency of sandstone-type uranium exploration and provides robust geological evidence for the evaluation and development of uranium resources.
{"title":"Reconstruction of geophysical logging and intelligent lithology identification of mineralized layer: A case study of the Baolongshan uranium deposit in the Southern Songliao Basin","authors":"Zhimo Zhang , Zhibing Feng , Li Jiang , Xiao Huang , Bocheng Zhang","doi":"10.1016/j.jappgeo.2025.106012","DOIUrl":"10.1016/j.jappgeo.2025.106012","url":null,"abstract":"<div><div>Geophysical logging serves as a fundamental method in sandstone-type uranium exploration. Due to the large volume of logging data and the complexity of geological interpretation, artificial intelligence (AI) technologies provide efficient solutions for data processing and analysis. While lithological identification approaches in non-mineralized layers are similar to those used in oil, gas, and coal exploration, the mineralized layers present a unique challenge: anomalously high gamma-ray values are observed in some intervals without apparent lithological control, which can lead to misinterpretation. Therefore, accurate lithological classification of mineralized layers is therefore critical for the exploration and development of sandstone-type uranium deposits. This study focuses on the lithology of the mineralized lower Yaojia Formation in the Baolongshan uranium deposit, located in the southern Songliao Basin. A Random Forest regression model was employed to reconstruct anomalous intervals in the natural gamma-ray logging data, and an improved XGBoost model, optimized via Bayesian hyperparameter tuning, was used for lithological classification. Quantitative evaluations confirm the reliability of the reconstructed natural gamma-ray data, showing negligible differences in average values compared to non-anomalous data. Moreover, in the filled intervals where original gamma-ray values were excluded due to mineralization effects and subsequently reconstructed using the Random Forest model, the reconstructed gamma-ray curves not only exhibit morphological similarity to spontaneous potential curves but also show a mirror-image relationship with apparent resistivity and three-lateral resistivity curves, thereby further validating the reconstruction. Lithological classification accuracies for the mineralized layer across the three boreholes are 97.5 %, 96.5 %, and 98.6 %, with an average accuracy of 97.5 %. These results outperform those obtained from the original data, which yielded accuracies of 95.5 %, 96.0 %, and 97.3 %, with an average accuracy of 96.3 %. The AI-driven precise identification of lithology in mineralized layers significantly enhances the efficiency of sandstone-type uranium exploration and provides robust geological evidence for the evaluation and development of uranium resources.</div></div>","PeriodicalId":54882,"journal":{"name":"Journal of Applied Geophysics","volume":"244 ","pages":"Article 106012"},"PeriodicalIF":2.1,"publicationDate":"2025-10-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145467629","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Accurate evaluation of the microwave heating sensitivity of hard rock in deep engineering, and its correlation with dielectric properties is crucial for predicting the microwave-mechanical rock breaking efficiency and assessing microwave-induced stress release effectiveness. Based on coaxial line transmission/reflection dielectric property test equipment, the dielectric properties of synthetic rocks made from different rock powders are tested, and the reliability of this method is verified. The dielectric properties and the microwave average heating rate of different rock minerals are tested, which provides the basis for the preparation of synthetic rocks. Taking engineering granite as an example, three main rock mineral powders are mixed thoroughly in different proportions, compressed into synthetic granite under 400 MPa to test their complex dielectric constant. A prediction model method for the dielectric properties of synthetic granite based on the mineral content is established. The research results indicate that the preparation pressure is positively correlated with the loss tangent of the sample and negatively correlated with the microwave penetration depth (Dp). The greater the Dp is, the smaller the loss tangent and the lower the microwave average heating rate of rocks. The classification standards for rock microwave heating sensitivity are discussed (microwave frequency of 2.45 GHz): strong sensitivity (Dp < 4 cm), moderate sensitivity (4 cm ≤ Dp ≤ 8 cm), and weak sensitivity (Dp > 8 cm). In engineering applications, the method for predicting rock microwave heating sensitivity provides a basis for selecting microwave parameters in different hard rock excavation areas or stress release areas.
{"title":"Method for predicting the microwave heating sensitivity of hard rocks: Application to microwave-mechanical excavation and microwave stress release in tunnels","authors":"Shiping Li, Xia-ting Feng, Chengxiang Yang, Feng Lin, Xiangxin Su, Tianyang Tong, Jiuyu Zhang","doi":"10.1016/j.jappgeo.2025.105995","DOIUrl":"10.1016/j.jappgeo.2025.105995","url":null,"abstract":"<div><div>Accurate evaluation of the microwave heating sensitivity of hard rock in deep engineering, and its correlation with dielectric properties is crucial for predicting the microwave-mechanical rock breaking efficiency and assessing microwave-induced stress release effectiveness. Based on coaxial line transmission/reflection dielectric property test equipment, the dielectric properties of synthetic rocks made from different rock powders are tested, and the reliability of this method is verified. The dielectric properties and the microwave average heating rate of different rock minerals are tested, which provides the basis for the preparation of synthetic rocks. Taking engineering granite as an example, three main rock mineral powders are mixed thoroughly in different proportions, compressed into synthetic granite under 400 MPa to test their complex dielectric constant. A prediction model method for the dielectric properties of synthetic granite based on the mineral content is established. The research results indicate that the preparation pressure is positively correlated with the loss tangent of the sample and negatively correlated with the microwave penetration depth (<em>D</em><sub><em>p</em></sub>). The greater the <em>D</em><sub><em>p</em></sub> is, the smaller the loss tangent and the lower the microwave average heating rate of rocks. The classification standards for rock microwave heating sensitivity are discussed (microwave frequency of 2.45 GHz): strong sensitivity (<em>D</em><sub><em>p</em></sub> < 4 cm), moderate sensitivity (4 cm ≤ <em>D</em><sub><em>p</em></sub> ≤ 8 cm), and weak sensitivity (<em>D</em><sub><em>p</em></sub> > 8 cm). In engineering applications, the method for predicting rock microwave heating sensitivity provides a basis for selecting microwave parameters in different hard rock excavation areas or stress release areas.</div></div>","PeriodicalId":54882,"journal":{"name":"Journal of Applied Geophysics","volume":"244 ","pages":"Article 105995"},"PeriodicalIF":2.1,"publicationDate":"2025-10-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145467564","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-10-30DOI: 10.1016/j.jappgeo.2025.106010
Maurizio Milano , Luigi Bianco , Mauro La Manna , Maurizio Fedi , Valentina Russo
This study shows that microgravity investigation can be a successful strategy to detect deep buried foundation in an urban context. Specifically, we focused on the challenging archaeological and engineering case of the “Basilica dello Spirito Santo” in Naples (Italy) where the foundation system was debated in last centuries due to its complex historical development. Here we show that microgravity data, processed through the Depth from Extreme Points (DEXP) transformation, inferred a quadrangular pattern consistent with the expected foundation reinforcements. Modelling indicates that the structure is located at ∼5 m depth, shallower than originally designed. Further geophysical investigations employing Ground Penetration Radar (GPR) reveal numerous shallow voids, interpreted as crypts and burial sites, although they did not yield conclusive evidence regarding the foundation structures. This would be likely due to weak permittivity contrasts with surrounding soils. Moreover, the data suggest the presence of a deeper elongated anomaly of uncertain origin, which could represent either a geological channel-like feature or an undocumented structure. The study demonstrates the effectiveness of multimethodological approaches in complex urban archaeological contexts, providing crucial information for both cultural heritage knowledge and restoration planning.
{"title":"Microgravimetric and GPR surveying for the detection of building foundations: the case of the “Basilica dello Spirito Santo” in Naples (Italy)","authors":"Maurizio Milano , Luigi Bianco , Mauro La Manna , Maurizio Fedi , Valentina Russo","doi":"10.1016/j.jappgeo.2025.106010","DOIUrl":"10.1016/j.jappgeo.2025.106010","url":null,"abstract":"<div><div>This study shows that microgravity investigation can be a successful strategy to detect deep buried foundation in an urban context. Specifically, we focused on the challenging archaeological and engineering case of the “Basilica dello Spirito Santo” in Naples (Italy) where the foundation system was debated in last centuries due to its complex historical development. Here we show that microgravity data, processed through the Depth from Extreme Points (DEXP) transformation, inferred a quadrangular pattern consistent with the expected foundation reinforcements. Modelling indicates that the structure is located at ∼5 m depth, shallower than originally designed. Further geophysical investigations employing Ground Penetration Radar (GPR) reveal numerous shallow voids, interpreted as crypts and burial sites, although they did not yield conclusive evidence regarding the foundation structures. This would be likely due to weak permittivity contrasts with surrounding soils. Moreover, the data suggest the presence of a deeper elongated anomaly of uncertain origin, which could represent either a geological channel-like feature or an undocumented structure. The study demonstrates the effectiveness of multimethodological approaches in complex urban archaeological contexts, providing crucial information for both cultural heritage knowledge and restoration planning.</div></div>","PeriodicalId":54882,"journal":{"name":"Journal of Applied Geophysics","volume":"244 ","pages":"Article 106010"},"PeriodicalIF":2.1,"publicationDate":"2025-10-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145419992","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号