Producing complex fracture networks in a safe way plays a critical role in the hot dry rock (HDR) geothermal energy exploitation. However, conventional hydraulic fracturing (HF) generally produces high breakdown press...Producing complex fracture networks in a safe way plays a critical role in the hot dry rock (HDR) geothermal energy exploitation. However, conventional hydraulic fracturing (HF) generally produces high breakdown pressure and results only in single main fracture morphology. Furthermore, HF has also other problems such as the increased risk of seismic events and consuption of large amount of water. In this work, a new stimulation method based on cyclic soft stimulation (CSS) and liquid nitrogen (LN2) fracturing, known as cyclic LN2 fracturing is explored, which we believe has the potential to solve the above issues related to HF. The fracturing performances including breakdown pressure and fracture morphology on granites under true-triaxial stresses are investigated and compared with cyclic water fracturing. Cryo-scanning electron microscopy (Cryo-SEM) tests and X-ray computed tomography (CT) scanning tests were used for quantitative characterization of fracture parameters and to evaluate the cyclic LN2 fracturing performances. The results demonstrate that the cyclic LN2 fracturing results in reduced breakdown pressure, with between 21% and 67% lower pressure compared with using cyclic water fracturing. Cyclic LN2 fracturing tends to produce more complex and branched fractures, whereas cyclic water fracturing usually produces a single main fracture under a low number of cycles and pressure levels. Thermally-induced fractures mostly occur around the interfaces of different particles. This study shows the potential benefits of cyclic LN2 fracturing on HDR. It is expected to provide theoretical guidance for the cyclic LN2 fracturing application in HDR reservoirs.展开更多
In modern physics and fabrication technology,simulation of projectile and target collision is vital to improve design in some critical applications,like;bulletproofing and medical applications.Graphene,the most promin...In modern physics and fabrication technology,simulation of projectile and target collision is vital to improve design in some critical applications,like;bulletproofing and medical applications.Graphene,the most prominent member of two dimensional materials presents ultrahigh tensile strength and stiffness.Moreover,polydimethylsiloxane(PDMS)is one of the most important elastomeric materials with a high extensive application area,ranging from medical,fabric,and interface material.In this work we considered graphene/PDMS structures to explore the bullet resistance of resulting nanocomposites.To this aim,extensive molecular dynamic simulations were carried out to identify the penetration of bullet through the graphene and PDMS composite structures.In this paper,we simulate the impact of a diamond bullet with different velocities on the composites made of single-or bi-layer graphene placed in different positions of PDMS polymers.The underlying mechanism concerning how the PDMS improves the resistance of graphene against impact loading is discussed.We discuss that with the same content of graphene,placing the graphene in between the PDMS result in enhanced bullet resistance.This work comparatively examines the enhancement in design of polymer nanocomposites to improve their bulletproofing response and the obtained results may serve as valuable guide for future experimental and theoretical studies.展开更多
Hydraulic fracturing stimulation technology is essential in the oil and gas industry.However,current techniques for predicting rock fracture pressure in hydraulic fracturing face significant challenges in precision an...Hydraulic fracturing stimulation technology is essential in the oil and gas industry.However,current techniques for predicting rock fracture pressure in hydraulic fracturing face significant challenges in precision and reliability.Traditional approaches often result in inadequate accuracy due to the complex and diverse nature of underground formations.However,recent advances in computational power and optimization techniques have enabled the application of machine learning in mining operations,resulting in improved prediction and feedback.In this study,various machine learning techniques are employed to predict hydraulic fracturing pressure based on the concept of mechanical specific energy.Additionally,the study interprets the models through feature importance analysis.Thefindings suggest that most machine learning models deliver highly accurate predictions.Feature importance analysis indicates that for an approximate assessment of fracture pressure,the characteristics of well depth and torque are sufficient.For more precise predictions,incorporating additional characteristics from the mechanical specific energy framework into the machine learning model is essential.The study emphasizes the feasibility of employing machine learning methods to predict fracture pressure and their usefulness in determining optimal engineering sites.展开更多
Accurate prediction of fatigue life under multiaxial loading conditions remains challenging due to complex stress–strain interactions.In this study,we integrate machine-learning(ML)regression with variance-based sens...Accurate prediction of fatigue life under multiaxial loading conditions remains challenging due to complex stress–strain interactions.In this study,we integrate machine-learning(ML)regression with variance-based sensitivity analysis(SA)to predict multiaxial fatigue life in CuZn37 brass and to identify the dominant mechanical factors influencing fatigue damage.Several surrogate models were evaluated,with the Gaussian Process model achieving the highest accuracy(R^(2)=0.991)while maintaining robust generalization across loading paths.Gradient Boosting,Random Forest,and Penalized Spline Regression models also demonstrated strong predictive capabilities.Importantly,the SA explicitly accounted for statistical dependencies among input parameters,revealing that normal strain–stress interactions account for over 40%of the total variance in fatigue life.In contrast,shear-related parameters exhibited secondary,compensatory effects.These results highlight the importance of capturing parameter dependencies in fatigue modeling and demonstrate that ML-based surrogates can help provide both high-fidelity predictions and physical insights under complex multiaxial loading conditions.展开更多
Soil liquefaction assessment remains a crucial and complex challenge in seismic geotechnical engineering due to various liquefaction records and limited information,which entails a more generalized off-the-shelf model...Soil liquefaction assessment remains a crucial and complex challenge in seismic geotechnical engineering due to various liquefaction records and limited information,which entails a more generalized off-the-shelf model that can achieve favourable performance on different data sources.In this work,a deep learning model is built and investigated on the soil liquefaction prediction and a modified transfer learning scheme between different data sources is presented.Various datasets,including shear wave velocity-based,CPT-based,SPT-based,and real cases,are collected and utilized to verify the effectiveness and accuracy of the proposed model.Because different data sources in soil liquefaction generally share several geotechnical and mechanical parameters,we work to combine model prior information,feature mapping and data reconstruction in transfer learning models to tackle multi-source domain adaption,which can be further applied to other predictive analysis and facilitate online learning models in geotechnical engineering.Also,the deep learning model is compared with several classical machine learning and ensemble learning models and the modified transfer learning model is formulated by comparing different feature transformation techniques integrated with various feature-based and instance-based transfer learning methods.The accuracy and effectiveness of the deep learning and modified transfer learning models have been validated in the numerical study.展开更多
For soil liquefaction prediction from multiple data sources,this study designs a hierarchical machine learning model based on deep feature extraction and Gaussian Process with integrated domain adaption techniques.The...For soil liquefaction prediction from multiple data sources,this study designs a hierarchical machine learning model based on deep feature extraction and Gaussian Process with integrated domain adaption techniques.The proposed model first combines deep fisher discriminant analysis(DDA)and Gaussian Process(GP)in a unified framework,so as to extract deep discriminant features and enhance the model performance for classification.To deliver fair evalu-ation,the classifier is validated in the approach of repeated stratified K-fold cross validation.Then,five different data resources are presented to further verify the model’s robustness and generality.To reuse the gained knowledge from the existing data sources and enhance the generality of the predictive model,a domain adaption approach is formu-lated by combing a deep Autoencoder with TrAdaboost,to achieve good performance over different data records from both the in-situ and laboratory observations.After comparing the proposed model with classical machine learn-ing models,such as supported vector machine,as well as with the state-of-art ensemble learning models,it is found that,regarding seismic-induced liquefaction prediction,the predicted results of this model show high accuracy on all datasets both in the repeated cross validation and Wilcoxon signed rank test.Finally,a sensitivity analysis is made on the DDA-GP model to reveal the features that may significantly affect the liquefaction.展开更多
This paper proposes an accurate,efficient and explainable method for the classification of the surrounding rock based on a convolutional neural network(CNN).The state-of-the-art robust CNN model(EfficientNet)is applie...This paper proposes an accurate,efficient and explainable method for the classification of the surrounding rock based on a convolutional neural network(CNN).The state-of-the-art robust CNN model(EfficientNet)is applied to tunnel wall image recognition.Gaussian filtering,data augmentation and other data pre-processing techniques are used to improve the data quality and quantity.Combined with transfer learning,the generality,accuracy and efficiency of the deep learning(DL)model are further improved,and finally we achieve 89.96%accuracy.Compared with other state-of-the-art CNN architectures,such as ResNet and Inception-ResNet-V2(IRV2),the presented deep transfer learning model is more stable,accurate and efficient.To reveal the rock classification mechanism of the proposed model,Gradient-weight Class Activation Map(Grad-CAM)visualizations are integrated into the model to enable its explainability and accountability.The developed deep transfer learning model has been applied to support the tunneling of the Xingyi City Bypass in the high mountain area of Guizhou,China,with great results.展开更多
基金supported by the Youth Program of the National Natural Science Foundation of China(52004299)Major Project of the National Natural Science Foundation of China(52192621)+2 种基金the National Science Foundation for National R&D Program for Major Research Instruments of China(51827804)Beijing Outstanding Young Scientist Program(BJJWZYJH01201911414038)the National Science Foundation for Distinguished Young Scholars of China(51725404).
文摘Producing complex fracture networks in a safe way plays a critical role in the hot dry rock (HDR) geothermal energy exploitation. However, conventional hydraulic fracturing (HF) generally produces high breakdown pressure and results only in single main fracture morphology. Furthermore, HF has also other problems such as the increased risk of seismic events and consuption of large amount of water. In this work, a new stimulation method based on cyclic soft stimulation (CSS) and liquid nitrogen (LN2) fracturing, known as cyclic LN2 fracturing is explored, which we believe has the potential to solve the above issues related to HF. The fracturing performances including breakdown pressure and fracture morphology on granites under true-triaxial stresses are investigated and compared with cyclic water fracturing. Cryo-scanning electron microscopy (Cryo-SEM) tests and X-ray computed tomography (CT) scanning tests were used for quantitative characterization of fracture parameters and to evaluate the cyclic LN2 fracturing performances. The results demonstrate that the cyclic LN2 fracturing results in reduced breakdown pressure, with between 21% and 67% lower pressure compared with using cyclic water fracturing. Cyclic LN2 fracturing tends to produce more complex and branched fractures, whereas cyclic water fracturing usually produces a single main fracture under a low number of cycles and pressure levels. Thermally-induced fractures mostly occur around the interfaces of different particles. This study shows the potential benefits of cyclic LN2 fracturing on HDR. It is expected to provide theoretical guidance for the cyclic LN2 fracturing application in HDR reservoirs.
基金B.M.and X.Z.appreciate the funding by the Deutsche Forschungsgemeinschaft(DFG,German Research Foundation)under Germany’s Excellence Strategy within the Cluster of Excellence PhoenixD(EXC 2122,Project ID 390833453).
文摘In modern physics and fabrication technology,simulation of projectile and target collision is vital to improve design in some critical applications,like;bulletproofing and medical applications.Graphene,the most prominent member of two dimensional materials presents ultrahigh tensile strength and stiffness.Moreover,polydimethylsiloxane(PDMS)is one of the most important elastomeric materials with a high extensive application area,ranging from medical,fabric,and interface material.In this work we considered graphene/PDMS structures to explore the bullet resistance of resulting nanocomposites.To this aim,extensive molecular dynamic simulations were carried out to identify the penetration of bullet through the graphene and PDMS composite structures.In this paper,we simulate the impact of a diamond bullet with different velocities on the composites made of single-or bi-layer graphene placed in different positions of PDMS polymers.The underlying mechanism concerning how the PDMS improves the resistance of graphene against impact loading is discussed.We discuss that with the same content of graphene,placing the graphene in between the PDMS result in enhanced bullet resistance.This work comparatively examines the enhancement in design of polymer nanocomposites to improve their bulletproofing response and the obtained results may serve as valuable guide for future experimental and theoretical studies.
文摘Hydraulic fracturing stimulation technology is essential in the oil and gas industry.However,current techniques for predicting rock fracture pressure in hydraulic fracturing face significant challenges in precision and reliability.Traditional approaches often result in inadequate accuracy due to the complex and diverse nature of underground formations.However,recent advances in computational power and optimization techniques have enabled the application of machine learning in mining operations,resulting in improved prediction and feedback.In this study,various machine learning techniques are employed to predict hydraulic fracturing pressure based on the concept of mechanical specific energy.Additionally,the study interprets the models through feature importance analysis.Thefindings suggest that most machine learning models deliver highly accurate predictions.Feature importance analysis indicates that for an approximate assessment of fracture pressure,the characteristics of well depth and torque are sufficient.For more precise predictions,incorporating additional characteristics from the mechanical specific energy framework into the machine learning model is essential.The study emphasizes the feasibility of employing machine learning methods to predict fracture pressure and their usefulness in determining optimal engineering sites.
文摘Accurate prediction of fatigue life under multiaxial loading conditions remains challenging due to complex stress–strain interactions.In this study,we integrate machine-learning(ML)regression with variance-based sensitivity analysis(SA)to predict multiaxial fatigue life in CuZn37 brass and to identify the dominant mechanical factors influencing fatigue damage.Several surrogate models were evaluated,with the Gaussian Process model achieving the highest accuracy(R^(2)=0.991)while maintaining robust generalization across loading paths.Gradient Boosting,Random Forest,and Penalized Spline Regression models also demonstrated strong predictive capabilities.Importantly,the SA explicitly accounted for statistical dependencies among input parameters,revealing that normal strain–stress interactions account for over 40%of the total variance in fatigue life.In contrast,shear-related parameters exhibited secondary,compensatory effects.These results highlight the importance of capturing parameter dependencies in fatigue modeling and demonstrate that ML-based surrogates can help provide both high-fidelity predictions and physical insights under complex multiaxial loading conditions.
文摘Soil liquefaction assessment remains a crucial and complex challenge in seismic geotechnical engineering due to various liquefaction records and limited information,which entails a more generalized off-the-shelf model that can achieve favourable performance on different data sources.In this work,a deep learning model is built and investigated on the soil liquefaction prediction and a modified transfer learning scheme between different data sources is presented.Various datasets,including shear wave velocity-based,CPT-based,SPT-based,and real cases,are collected and utilized to verify the effectiveness and accuracy of the proposed model.Because different data sources in soil liquefaction generally share several geotechnical and mechanical parameters,we work to combine model prior information,feature mapping and data reconstruction in transfer learning models to tackle multi-source domain adaption,which can be further applied to other predictive analysis and facilitate online learning models in geotechnical engineering.Also,the deep learning model is compared with several classical machine learning and ensemble learning models and the modified transfer learning model is formulated by comparing different feature transformation techniques integrated with various feature-based and instance-based transfer learning methods.The accuracy and effectiveness of the deep learning and modified transfer learning models have been validated in the numerical study.
文摘For soil liquefaction prediction from multiple data sources,this study designs a hierarchical machine learning model based on deep feature extraction and Gaussian Process with integrated domain adaption techniques.The proposed model first combines deep fisher discriminant analysis(DDA)and Gaussian Process(GP)in a unified framework,so as to extract deep discriminant features and enhance the model performance for classification.To deliver fair evalu-ation,the classifier is validated in the approach of repeated stratified K-fold cross validation.Then,five different data resources are presented to further verify the model’s robustness and generality.To reuse the gained knowledge from the existing data sources and enhance the generality of the predictive model,a domain adaption approach is formu-lated by combing a deep Autoencoder with TrAdaboost,to achieve good performance over different data records from both the in-situ and laboratory observations.After comparing the proposed model with classical machine learn-ing models,such as supported vector machine,as well as with the state-of-art ensemble learning models,it is found that,regarding seismic-induced liquefaction prediction,the predicted results of this model show high accuracy on all datasets both in the repeated cross validation and Wilcoxon signed rank test.Finally,a sensitivity analysis is made on the DDA-GP model to reveal the features that may significantly affect the liquefaction.
文摘This paper proposes an accurate,efficient and explainable method for the classification of the surrounding rock based on a convolutional neural network(CNN).The state-of-the-art robust CNN model(EfficientNet)is applied to tunnel wall image recognition.Gaussian filtering,data augmentation and other data pre-processing techniques are used to improve the data quality and quantity.Combined with transfer learning,the generality,accuracy and efficiency of the deep learning(DL)model are further improved,and finally we achieve 89.96%accuracy.Compared with other state-of-the-art CNN architectures,such as ResNet and Inception-ResNet-V2(IRV2),the presented deep transfer learning model is more stable,accurate and efficient.To reveal the rock classification mechanism of the proposed model,Gradient-weight Class Activation Map(Grad-CAM)visualizations are integrated into the model to enable its explainability and accountability.The developed deep transfer learning model has been applied to support the tunneling of the Xingyi City Bypass in the high mountain area of Guizhou,China,with great results.
