最新刊期
卷 58 ， 期 2 ， 2026
-
Review on the Research and Development of
In-situ Utilization Technology of Deep Space Resources AI导读摘要：SignificanceA key aspect of enabling sustainable and affordable deep space programs is locating, extracting, and utilizing energy and resources available in deep space and converting them into the essential energy and products required for deep space exploration missions, with the aim of minimizing the dependence of exploration missions on Earth resources. This approach is known as In-Situ Resources Utilization (ISRU). This study reviews the current state of knowledge on in-situ resource utilization technologies for deep space resources in China and internationally, and examines the existing scientific challenges and future development trends of ISRU. It provides a reference for the research and development of in-situ resource utilization technologies in China.ProgressThe current research mainly focused on resource representation and mapping, in-situ consumables production, civil engineering construction, in-situ energy production and storage, environmental resource utilization, and ISRU analog testing. 1) The mechanical and mineralogical properties and the distribution characteristics of water and gas resources in geotechnical bodies on the Moon, Mars, and some asteroids were preliminarily demonstrated through integrated sample analysis and various remote sensing probes. The results indicated that deep-space environments contained many resources, such as water ice, 3He, metals, non-metals, and solar energy. These resources can provide essential support for human activities, including life support, power propulsion, food production, and manufacturing, as well as support sustainable and affordable deep-space exploration. 2) Numerous technologies and pieces of equipment for lunar soil excavation, preparation of anhydrous concrete, and production of oxygen and water were proposed and manufactured through indoor testing. 3) The construction of lunar and Mars bases represented an ambitious objective for human space colonization, and various structural configurations were proposed while considering the complex environmental and geological conditions of deep space. However, existing lunar and Mars bases remained at the conceptual design stage, and quantitative design from a civil engineering perspective has not yet been conducted. Systematic and comprehensive quantitative calculations for the bearing capacity of lunar soil, the selection of lunar base foundations, detailed loading design, structural stability, and disaster protection remained insufficient. 4) Energy is the cornerstone for the long-term operation and maintenance of future lunar bases. Abundant energy resources were identified in deep space, such as solar energy on the Moon and wind energy on Mars, and feasible in-situ energy production and storage technologies were proposed while considering the complex deep-space environment. However, these energy production and storage technologies remain at the conceptual stage and lack systematic theoretical foundations, technological development, equipment support, and indoor testing and verification under deep-space environmental conditions. 5) The high vacuum, low gravity, and weak stability zone of the Moon represent valuable environmental resources that can hardly be reproduced on Earth, and Mars is the closest external body in the solar system with environmental conditions similar to those of the Earth. However, due to limited exploration and development on the Moon and Mars, these special environmental resources cannot yet be effectively utilized, and no systematic technologies have been developed for the comprehensive utilization of the extreme environments of the Moon and Mars. 6) NASA possesses significant technological leadership in In-Situ Resources Utilization, and a series of relatively mature technologies and related equipment have been developed over recent decades. In addition, a series of simulated field experiments was conducted, completing tasks such as resource characterization and mapping, soil mining, and in-situ resource treatment and product preservation studies (anhydrous concrete, oxygen), which demonstrated the feasibility of NASA’s existing technologies and equipment.Conclusions and ProspectsOver the past half-century, NASA, by virtue of the advantages of Apollo’s abundant sample reserves and advanced deep space exploration technologies, has conducted many studies on the geology and engineering of the Moon and Mars. Therefore, it has established an advantage of "early knowledge and early leadership" in the field of deep space resource utilization. However, research on ISRU technology and equipment has only recently begun in China. Many technologies remain at the conceptual design stage and lack corresponding operable equipment or practical design programs, particularly in the areas of in situ resource extraction and transfer, in situ product production and storage, lunar base structure and protection design, and energy production and storage. Therefore, it is necessary to conduct targeted feasibility technology research and develop operational equipment that accounts for the complex environments of the Moon and Mars. Then, these technologies should be integrated into an implementable ISRU technology and equipment system to achieve sustainable and affordable development for China’s future deep space exploration.关键词：resource in deep space;in-situ resource utilization;lunar base;lunar soil707|1170|0更新时间：2026-03-26
SCIENTIFIC FRONTIERS
- 摘要：ObjectiveOn the transport belt used for normal coal flow, large coal gangue, anchor rods, and other foreign objects can be present. When large coal gangue or other foreign objects accumulate at the coal drop port, issues such as coal stacking and coal blockage occur. Anchor rods and other foreign objects can become entangled with transport belt components, causing surface scratches or even severe belt tearing, which seriously affects the normal coal flow transport. Deep learning methods previously applied demonstrate inferior baseline network performance compared to the YOLOv8 (You Only Look Once) model and fail to incorporate targeted lightweight optimization for edge deployment scenarios. Currently, computer vision-based detection methods do not achieve performance improvements over the YOLOv8 model in coal mine target detection tasks. Therefore, this study proposes a foreign object detection method for coal mine conveyor belts, YOLOv8‒SPCD, which is developed based on an improved YOLOv8 framework.MethodsThe YOLOv8‒SPCD model introduced several key improvements to enhance the detection performance of the original YOLOv8 model. First, the coal belt foreign body dataset was constructed based on existing mine images. The labelme tool was utilized to annotate the image data, and the images were divided into the training set (train), validation set (val), and test set (test) based on a ratio of 8:1:1. Then, SPD‒Conv was utilized to replace the convolutional component in the Backbone, and the spatial blocks of the input feature map were rearranged into the channel dimension to increase the number of channels, reduce the spatial resolution, and retain richer information during the feature extraction stage. Next, partial convolution was introduced to improve the C2f structure in the original network. The computation of redundant feature maps was reduced, while the spatial features of the input images were still effectively extracted by applying convolution only to part of the input channels. Then, a lightweight cross-scale feature fusion module (CCFM), was utilized to improve the Neck component and enhance the detection capability of the model for objects at different scales. Finally, to eliminate the adverse effect of the penalty term in the original loss function on convergence speed and to obtain faster and more effective regression results, the improved Inner‒DIoU function was introduced to optimize the bounding box regression loss of the network, enabling faster convergence and more accurate localization of belt foreign bodies during training.Results and DiscussionsGroups 1 to 4 experiments were independent experiments in which the improved modules were modified separately on the baseline network, allowing the impact of each individual module on the baseline network to be clearly observed. In the third group of experiments, the CF‒Neck structure was utilized to replace the original Neck component, and the mAP value remained unchanged even though the number of model parameters was reduced by 37%, indicating that CF‒Neck enhanced the detection capability of the model for objects at different scales. In the fourth group of experiments, Inner‒DIoU was utilized to replace the CIoU loss function, and the experimental indicators, such as mAP@0.5 and FPS, were improved, indicating that Inner‒DIoU effectively enhanced the fitting performance of the model. The ninth group of experiments corresponded to the YOLOv8‒SPCD model proposed in this study. The model weight was reduced to 43% of the baseline network, GFLOPs was reduced to 59% of the original value, mAP@0.5 was increased by 4.3 percentage points, mAP@0.5:0.95 was increased by 4.1 percentage points, and FPS was slightly improved. The effectiveness of the proposed method for detecting foreign objects on coal mine belts was thus verified. The training loss curves of the YOLOv8‒SPCD model with Inner‒DIoU and without Inner‒DIoU were compared in this study, and the results showed that the convergence speed of the YOLOv8‒SPCD model with Inner‒DIoU was significantly faster than that of the model without Inner‒DIoU. The Box Loss, which measured the discrepancy between the actual boundary box and the predicted boundary box of the target object, and the Classification Loss, which measured the accuracy of the model in predicting each target category, were both significantly reduced. The distribution focal loss (DFL), which was utilized to correct errors in predicting object boundary frames, remained similar to that before modification during training, indicating that the fitting performance of the proposed model on the mine image dataset was superior to that of the original model. The proposed model was also compared to mainstream target detection models such as YOLOv3-tiny, YOLOv5n, YOLOv6n, SSD, and Faster R‒CNN. The comparison results showed that the proposed model exhibited clear advantages.ConclusionsThe YOLOv8 model provides a feasible technical solution for detecting the presence of coal gangue, bolts, and other foreign matter during the coal conveying process on conveyor belts. The improved model integrates a series of enhancement strategies, including SPD‒Conv, PConv, the CCFM, and the Inner idea, demonstrating the broad application potential of the YOLOv8 model in coal mine target detection. This work provides a prerequisite for deployment at the mine edge. Then, the research objective is to deploy the improved model on embedded equipment at the mine edge end, realize practical algorithm application, and further optimize the model during the deployment process.关键词：YOLOv8;foreign body identification;SPD‒Conv;Partial Convolution;cross-scale feature fusion module317|508|0更新时间：2026-03-26
- 摘要：ObjectiveAccurate prediction of deflection variation holds significant importance for bridge operation and maintenance. The complex and nonlinear dynamic characteristics of bridge deflection consistently challenge traditional prediction models, as hysteresis in the deflection response and interference from irregular waveforms in historical monitoring data reduce prediction accuracy. This study proposes a deflection separation-prediction model for bridges by integrating wavelet optimization and long short-term memory networks to capture multi-scale features of deflection signals and account for external influences.MethodsFirstly, an Internet of Things monitoring system was employed to investigate the deflection behavior of in-service bridges. Focusing on the Xiongshang High-speed Railway Bridge over the Daguang Expressway, sensors installed on the structure were utilized to record variations in deflection, dynamic load, and temperature. Secondly, given the decoupling of different deflection components across multiple time scales, a wavelet-based optimization approach was applied to decompose historical monitoring data into trend deflection generated by prestress loss and noise deflection induced by external influences such as temperature and dynamic loads. Thirdly, based on the decomposed deflection components and the associated external factors, two LSTM-based time series prediction models were developed, including a multi-factor model for noise deflection and a single-factor model for trend deflection. Vehicle load, temperature, and noise deflection served as the inputs for the noise model, while trend deflection was used as the sole input for the trend model. Separate predictions were conducted, and the final cumulative bridge deflection was obtained by summing both predicted components based on the principle of time series superposition. Traditional models were adopted for comparison across short-term, medium-term, and long-term periods to evaluate prediction accuracy. Prediction performance was assessed using three metrics: correlation coefficient (R), root mean square error (ERMS), and mean absolute error (EMA). Fourthly, a comparative analysis was performed between the proposed model and the single LSTM prediction model to demonstrate the necessity of the combined model for forecasting bridge deflection. Fifthly, in order to verify the necessity of incorporating external factors, the proposed model was compared to a time series model including a single external factor and another excluding external influences, emphasizing differences in prediction accuracy. Sixthly, the maximal information coefficient was introduced to identify the dominant factors affecting noise deflection by analyzing its correlation with temperature and dynamic load.Results and Discussions1) Comparison of the prediction results for short-term, medium-term, and long-term periods with the BP neural network and LSSVM models showed that the prediction accuracy for all three models remained similar in the short and medium periods. However, in the long-term deflection prediction, the DWT‒LSTM-based bridge deflection separation model achieved the highest accuracy and demonstrated stronger generalization ability, with correlation coefficients of 0.86 and 0.77, ERMS of 2.18 and 2.20 mm, and mean absolute errors (EMA) of 2.05 and 1.91 mm. In contrast, the LSSVM model produced ERMS values of 2.82 and 3.52 mm, with EMA values of 2.45 and 3.13 mm. The BP neural network produced ERMS values of 3.06 and 3.53 mm, with EMA values of 2.89 and 3.24 mm. Compared to the LSSVM model, the DWT‒LSTM deflection separation model reduced ERMS by 22.70% and 37.50% and reduced EMA by 39.26% and 38.98%. Compared to the BP neural network, the DWT‒LSTM deflection separation model reduced ERMS by 28.76% and 37.68% and reduced EMA by 29.07% and 41.05%. 2) Compared to the DWT‒LSTM deflection separation model, the prediction accuracy decreased when the single LSTM model was used. The ERMS reached 3.74 mm, and the EMA reached 3.45 mm. These relatively large deviations indicated that this model had limited suitability for bridge deflection prediction. 3) Compared to the time series models that considered only temperature, only vehicle load, or excluded external factors, the model that excluded external factors exhibited the lowest prediction accuracy, with ERMS of 3.91 mm and EMA of 3.38 mm. Among the models that considered a single external factor, the time series model that considered load showed higher prediction accuracy, with ERMS of 2.81 mm and EMA of 2.65 mm, outperforming the temperature-only model, which had ERMS of 2.97 mm and EMA of 2.83 mm. In contrast, the DWT‒LSTM deflection separation model achieved the highest accuracy, with ERMS of only 2.18 mm and EMA of only 2.05 mm. 4) Analysis of the dominant factors that influenced noise deflection using the Maximal Information Coefficient (MIC) showed correlation coefficients of 0.35 for temperature and 0.51 for load, indicating that vehicle load has a greater impact on noise deflection than temperature.ConclusionsThis study presents a DWT‒LSTM-based bridge deflection separation prediction model that is suitable for predicting long-term deflection variation patterns. Compared to traditional prediction models, the proposed model shows higher accuracy, reduced errors, and improved capability in addressing time-lag effects, providing a new approach and method for long-term bridge deflection prediction.关键词：long short‒term memory;wavelet optimization;bridge deflection;multi-factor prediction;vehicle impact parameters216|331|0更新时间：2026-03-26
- 摘要：ObjectiveVisual inspection is a critical technology for roadside perception in vehicle-road cooperative systems. However, in practical applications, achieving both high detection accuracy and computational efficiency simultaneously remains challenging due to limited computing resources. This study proposes a novel method based on an improved YOLOv5 combined with CombineSORT for image recognition and target tracking, which achieves strong detection performance while maintaining low computational time cost, as demonstrated through experimental results.MethodsFirstly, Multi-scale Feature Enhancement (MFE) was applied to the FPN of YOLOv5 to extract shallow target details. This module was mainly composed of Scale Fusion, CombineFPN, and Pixel‒Region Attention. A super-efficient IOU (SEIOU) loss function and network pruning were applied to improve convergence and reduce model complexity. In this process, the loss was calculated based on differences in length, width, and diagonal between the detection boxes and the ground-truth boxes, while batch normalization (BN) layer sparsification was applied for convolutional channel filtering. Secondly, by combining DeepSORT, StrongSORT, and Bot-SORT, a new multi-target tracking method named CombineSORT was presented. In this approach, the basic framework of DeepSORT was adopted, and the BotNet with ResNet50 as the backbone was utilized to extract appearance features. Kalman filtering was replaced by polynomial fitting to improve trajectory smoothness, while the joint similarity matrix from StrongSORT was utilized to match targets with trajectories. Based on the operational procedure of the proposed algorithm, a series of experiments was designed to validate its effectiveness. Using images from real intersections, ablation tests verified the effectiveness and data volume contribution of each improved module. The algorithm was then compared to classical methods using intersection video streams with varying traffic volumes, all of which were executed on a mobile edge computer (MEC) with limited computing resources.Results and DiscussionsThrough ablation tests, the original YOLOv5 achieved an mAP@90 of 0.894 with a parameter quantity of 21.2 M. Scale Fusion, CombineFPN, and Pixel‒Region Attention increased the mAP@90 of the original model to 0.91, 0.923, and 0.916, respectively, while the parameter quantity increased to 24.4, 25.3, and 24.1 M, respectively. The YOLOv5 model integrating all three modules achieved an mAP@90 of 0.939 with a parameter quantity of 31.0 M, after which network pruning reduced the parameter quantity to 6.6 M while maintaining an mAP@90 of 0.937. Through three groups of real intersection experiments, the average recall rates for Group 1 to 3 were 97.68%, 95.83%, and 96.76%, while the multiple object tracking accuracy (MOTA) values were 0.944, 0.890, and 0.910, respectively. Among all target categories, pedestrians and non-motorized vehicles exhibited relatively poor detection performance. Especially in Group 2, the recall rate and MOTA for pedestrians were 89.98% and 0.75, respectively, while those for non‒motorized vehicles were as low as 84.5% and 0.675. This behavior occurred because these two target types had relatively small sizes and did not strictly follow traffic rules, which caused frequent occlusion and increased trajectory prediction difficulty. In addition, the recall rates of buses and trucks were nearly 3 percentage points lower than those of cars, especially in the group, where the recall rates were only 94.81% and 94.92%, respectively. This issue occurred because box trucks and buses exhibited similar appearance features, which increased the likelihood of misidentification from rear perspectives. When comparing the overall processing performance of different algorithms at low-volume intersections, the worst test result achieved a recall rate of 96.54% with a MOTA value of 0.938, while the best result achieved a recall rate of 97.69% with a MOTA value of 0.946. These results indicated that most algorithms achieved good detection performance under sparse target conditions, and lightweight models demonstrated advantages when considering computational resource constraints. However, for high-volume intersections, although the lightweight algorithm based on EfficientNet and ByteTrack exhibited the shortest computation delay, its recall rate and MOTA value were only 91.75% and 0.817, respectively. In contrast, algorithms based on YOLOv5, YOLOX, YOLOv7, and the improved YOLOv5 proposed in this study achieved recall rates ranging from 95.26% to 96.28%, while algorithms combined with DeepSORT, StrongSORT, Bot-SORT, and CombineSORT achieved MOTA values ranging from 0.887 to 0.901. However, most of these methods exhibited computation times exceeding 80 ms, which prevented real-time operation. Among algorithms with computation times below 80 ms, the proposed method based on improved YOLOv5 and CombineSORT achieved the best overall performance, with a recall rate of 96.27% and an MOTA value of 0.900, which confirmed its ability to balance detection accuracy and computational efficiency.ConclusionsThis study focuses on traffic target perception from a fixed roadside perspective, and the results demonstrate the effectiveness and accuracy of the proposed algorithm. Compared to other commonly used algorithms, the proposed approach simultaneously achieves higher detection accuracy and lower time cost at high-volume intersections, indicating strong application potential in vehicle road collaboration scenarios. For improved engineering practices, further research can be conducted to enhance recognition and tracking performance based on continuous image sequences under adverse weather conditions.关键词：vehicle-road cooperative;roadside perception;Image recognition;YOLOv5;CombineSORT239|334|0更新时间：2026-03-26
-
Depression Detection Based on Dual‒Path DCGAN Data Generation and Classification‒Regression Network AI导读
摘要：ObjectiveAccurate evaluation of depression scores in patients with depression provides effective support for clinical auxiliary diagnosis and enables the development of personalized diagnosis and treatment plans, improving the overall accuracy of clinical diagnosis and intervention and contributing significantly to patient health outcomes. Existing research on voice-based depression detection exhibits several limitations, including complex feature extraction processes, single-mode data augmentation, and uncontrollable prediction bias in regression-based estimation. This study proposes a dual‒path DCGAN for data generation and introduces a classification-regression network model for depression score prediction, enabling effective auxiliary diagnosis of depression severity.MethodsFirstly, based on the audio characteristics of depressed patients, six types of emotional features were selected from existing speech features, and corresponding two-dimensional feature maps were constructed for each audio signal sample. For MFCC features, the Teager energy operator was fused with MFCC to form MFCC‒TEO features, which further highlighted differences in energy distribution. In addition, the dual‒path deep convolutional generative adversarial network proposed in this study was utilized to enhance the two-dimensional feature maps of each depression level to expand the dataset, increase feature map diversity, and improve model robustness and generalization. Simultaneously, an evaluation index based on spatial and frequency domain characteristics was proposed to screen generated feature maps and retain high-quality samples. Finally, a classification regression network was introduced into the prediction framework to reduce prediction bias by narrowing the prediction confidence interval. For residual networks within the classification framework, multi-scale convolution was introduced to enhance information interaction among features, which enabled the residual network to fully perceive multi-level information contained in the feature maps.Results and DiscussionsFeature validity tests were conducted for the six selected emotional features, in which MFCC, MFCC‒TEO, LPCC, and Jitter features were sequentially added based on short-term energy, zero-crossing rate, and sound intensity, and accuracy (Acc), root mean square error (RMSE), and mean absolute error (MAE) under different input configurations were calculated. Experimental results showed that Acc, RMSE, and MAE were 89.76%, 6.17, and 2.08, respectively, when MFCC was added. When MFCC‒TEO was added, Acc, RMSE, and MAE reached 92.07%, 5.49, and 1.58, respectively. When MFCC‒TEO and LPCC were added, Acc, RMSE, and MAE further improved to 93.41%, 5.09, and 1.39, respectively. When MFCC‒TEO, LPCC, and Jitter were added, Acc, RMSE, and MAE reached 94.73%, 4.55, and 1.11, respectively. These results demonstrated that when MFCC‒TEO was used as an input feature, Acc increased by 2.31 percentage points, while RMSE and MAE decreased by 0.68 and 0.50, respectively, compared to using MFCC alone, which indicated that combining MFCC with TEO enhanced the representation of energy distribution differences. The MFCC‒TEO coefficient exhibited stronger depression characterization capability than the MFCC coefficient. Subsequent incorporation of LPCC and Jitter features further improved prediction accuracy to a certain extent. In the data enhancement experiments, when the original dataset was utilized to predict depression scores, Acc, RMSE, and MAE were 80.51%, 8.47, and 3.94, respectively. After data enhancement using the dual deep convolutional generative adversarial network, Acc, RMSE, and MAE improved to 94.73%, 4.55, and 1.11, respectively. Compared to the original dataset, prediction accuracy significantly improved, with Acc increasing by 14.22 percentage points, and RMSE and MAE decreasing by 3.92 and 2.83, respectively, which demonstrated that DP‒DCGAN-based data enhancement effectively expanded the dataset. In the prediction network, the classification accuracy of the original ResNet was 93.28%, while the MSC‒ResNet achieved a classification accuracy of 94.73%, representing an improvement of 1.45 percentage points. These results confirmed that the multi-scale convolution strategy extracted richer global and contextual information, after which the residual network captured detailed information, enabling the network to fully perceive multi-scale characteristics within the input feature maps and ultimately improve overall model performance.ConclusionsThis study proposes a depression diagnosis model based on a deep generation network and a classification regression framework. The MFCC‒TEO feature is obtained by introducing TEO into MFCC, and six features, including TEO, are extracted to construct a two-dimensional feature map incorporating time‒frequency, linear, and nonlinear properties. Feature maps corresponding to each depression score in the original dataset are enhanced to increase feature diversity, and evaluation indicators are proposed to screen high-quality feature maps from both spatial and frequency domain perspectives by constructing a DP‒DCGAN network. High-quality and diversified feature maps significantly improve the overall performance of the model. Finally, the proposed MRVN classification regression network is applied to predict depression scores. A multi-scale convolution module is added to the ResNet classification network to address the limitation of single-scale receptive fields in feature extraction by integrating the unique characteristics of the feature maps proposed in this study. In addition, the input data can be predicted on a more uniform scale by combining classification and regression strategies, reducing large prediction deviations commonly observed in regression tasks.关键词：audio;depression detection;generative adversarial network;category‒regression;Residual Network249|380|0更新时间：2026-03-26 - 摘要：ObjectiveSix-degree-of-freedom (6-DoF) unmanned aerial vehicle (UAV) air combat scenarios present substantial challenges for strategy learning when reinforcement learning methods are applied. These challenges stem from high-dimensional state spaces, continuously coupled action domains, and strongly nonlinear flight dynamics. Conventional end-to-end deep reinforcement learning (DRL) approaches struggle to achieve rapid convergence, to identify effective maneuver strategies, and to generalize learned policies beyond narrowly constrained conditions. In addition, reward functions often rely on handcrafted rules derived from human expertise, which do not ensure that higher reward values correspond to genuinely effective combat strategies. This study addresses these limitations by introducing a hierarchical framework based on time scale separation theory. The proposed framework employs a two-stage training procedure that accounts for differences in how flight parameters influence state variables across multiple time scales, improving learning efficiency, enhancing strategy quality, and increasing generalization capability in complex and diverse combat environments.MethodsA novel algorithm, termed TTS‒PPO, was developed. TTS‒PPO stood for a Two-Stage Training framework leveraging Time-Scale separation within Proximal Policy Optimization. The method focused on partitioning the 6 DoF UAV air combat decision-making process into short-cycle and long-cycle segments, which reflected differences in how control inputs influenced state variables across distinct time scales. A time-division framework was established. The short‒cycle component addressed rapid rotational and attitude adjustments. Instead of allowing the DRL procedure to directly manage these fine-grained actions, a Proportional-Integral-Derivative (PID) controller was employed to output real-time joystick commands. This configuration allowed classical low-level stability and attitude control to be handled independently, which reduced the complexity encountered by the DRL policy at the higher strategic level. With low-level stability assured, the DRL agent focused on tactical and strategic decision-making. The long-cycle component used Proximal Policy Optimization to manage trajectory planning and tactical maneuvers. The long-cycle PPO agent effectively decoupled strategic decision-making from low-level actuation tasks by issuing high-level commands to guide the PID-driven short-cycle layer. This hierarchical decomposition allowed learning to proceed more efficiently. The long-cycle agent encountered a reduced problem space and concentrated on discovering effective combat strategies without being burdened by the complexities of rapid stabilization maneuvers. Time scale separation was further implemented within the state space. Environmental states were divided into long-cycle and short-cycle groups. The long-cycle states captured slowly evolving features such as relative positions, energy conditions, and global situational parameters, whereas the short-cycle states encompassed rapidly changing variables such as angular rates and orientation deviations. Aligning state variables with their corresponding time scales accelerated learning and improved policy robustness. A relative situation transformation module was introduced to refine and compress the state representation, which ensured that the agent received relevant information at appropriate decision intervals and minimized computational complexity at each step. A two-stage training strategy was employed. In the first stage, single-step rewards designed for specific subtasks, such as pursuit or strike, were introduced with a lower decision frequency to assist the agent during the initial “cold start” period. This incremental guidance supported the stabilization of fundamental behavioral patterns and facilitated the acquisition of essential tactical principles. During this phase, the agent overcame early-stage instability, which resulted in more reliable initial policies. In the second stage of training, single-step rewards were removed, and only sparse terminal rewards were retained. In the second stage of training, single-step rewards were removed, and only sparse terminal rewards were retained, while the decision frequency was increased. In the absence of frequent intermediate rewards, the policy emphasized long-term outcomes rather than short-term objectives. The higher decision frequency enabled more refined tactical adjustments and encouraged the emergence of maneuvers that improved overall performance. The gradual transition from a guided, intermediate-reward scenario to a sparse-reward, high-frequency regime allowed the policy to progress from basic stability toward advanced strategic competence. A simulation environment was constructed using an open-source F-16 UAV model coupled with the JSBSim flight dynamics engine to evaluate the effectiveness of the proposed hierarchical DRL algorithm founded on time scale separation theory. This configuration provided realistic 6 DoF conditions and supported one-on-one close-range air combat simulations. Ablation experiments were conducted to assess the contribution of individual components within the TTS‒PPO framework. One configuration trained the agent against a non-maneuvering linear opponent, which served as a controlled baseline for determining whether the learned policy can scale from simple engagements to more complex combat scenarios.Results and DiscussionsThe results demonstrated that the TTS‒PPO approach, which incorporated hierarchical decomposition and time scale separation, achieved faster convergence and improved final performance metrics compared to baseline end-to-end DRL methods that lacked time scale separation or a two-stage training procedure. Assigning state variables to short-cycle and long-cycle categories, together with hierarchical action decomposition, significantly reduced overall problem complexity. Training convergence speed improved by approximately 67%, which reduced computational costs and enabled more frequent iterative policy refinements. With enhanced efficiency, the DRL agent discovered more stable and effective combat strategies within fewer training episodes. Generalization performance was evaluated by testing agents trained under different variants of the approach across various initial conditions, velocities, and adversary tactics. Comparisons were conducted among three agent types: an agent trained with PPO on full-state inputs without time scale division (FS‒PPO), an agent using time scale-separated states with a single-stage training approach (TS‒PPO), and the two-stage time scale-separated TTS‒PPO. The agent trained with TTS‒PPO outperformed both FS‒PPO and TS‒PPO agents in pairwise confrontations, which indicated that combining time scale separation with two-stage training not only enhanced learning speed but also enabled the agent to acquire more generalizable combat principles rather than narrowly optimizing for a specific scenario. Further validation involved testing the TTS‒PPO-trained agent against rule-based expert opponents. The policy derived from TTS‒PPO successfully defeated these expert systems. Even when training was conducted exclusively against a simple linear adversary, the learned policy surpassed expert-level strategies, which confirmed that hierarchical time scale separation and the two-stage training design facilitated the development of adaptable policies with robust tactical proficiency. The ability to transfer from minimal training complexity to outperforming expert opponents highlighted the scalability and versatility of the learned strategies.ConclusionsAccordingly, the hierarchical DRL algorithm, grounded in time scale separation theory and employing a two-stage training strategy, addressed significant challenges associated with applying DRL to 6-DoF UAV air combat tasks. The method substantially improved training efficiency and enhanced both the robustness and generalization capability of the resulting policies by decomposing decision-making into short-cycle and long-cycle phases, introducing a PID-controlled low-level stabilization layer, and separating state variables based on their respective time scales. The hierarchical framework enabled the agent to focus on strategic maneuvers at the long-cycle level, while the short-cycle PID layer managed rapid stabilization tasks. Time scale-aware state representations and a staged training procedure guided the policy from basic stability to advanced tactical competence. The observed increases in convergence speed and the ability to manage a range of adversarial conditions highlight the value of applying time scale separation principles in challenging reinforcement learning domains. The TTS‒PPO framework can serve as a reference for addressing other complex reinforcement learning problems characterized by distinct time scale dynamics, fostering more efficient, generalizable, and strategically effective decision-making in advanced autonomous systems.关键词：time-scale separation;PID;PPO;two-stage training;TTS‒PPO333|492|0更新时间：2026-03-26
-
Research on Automatic Identification Method of Loess Earthquake Landslide Based on VGG‒Unet Model AI导读
摘要：ObjectiveLoess areas are highly prone to frequent earthquake-induced landslides. Rapid and accurate identification in loess regions is essential to obtain a detailed spatial distribution of loess earthquake landslides. The identification results of landslides reflect their development status and spatial distribution, serving as a theoretical foundation for research on landslide safety, vulnerability, and management. Field investigations and remote sensing technology are two primary approaches for landslide identification. However, relying solely on field surveys consumes substantial human, material, and financial resources. Similarly, depending exclusively on remote sensing technology for landslide identification does not ensure the authenticity and reliability of the database. This research applies the VGG‒Unet network for the automatic identification of loess earthquake landslides based on landslide data obtained from field surveys.MethodsFirstly, based on loess earthquake landslides investigated by the earthquake landslide research team of the College of Disaster Prevention Science and Technology during field surveys in Gansu Province and Ningxia Hui Autonomous Region, the quantitative and qualitative parameters for extracting loess earthquake landslides on the satellite imagery platform through visual interpretation were systematically summarized. The quantitative parameters included stratigraphy, topography, slope, slide length, and slide width characteristics, and the qualitative parameters included planar morphology, profile, and color tone characteristics. These features supported the visual interpretation of remote sensing satellite imagery and facilitated the extraction of a database of loess landslides triggered by seismic events. Ultimately, a total of 494 original images were extracted, covering 1 052 loess landslides induced by seismic events, which served as input images to improve the model's generalization ability. Secondly, due to the exceptional feature extraction capability of the VGG16 network, its simple yet effective structural design, and its extensive pre-trained weights, it facilitated the construction of a Unet network model using it as the primary backbone feature extraction network, enhancing performance. Therefore, this study applied the VGG‒Unet network model using the Python programming language based on the PyTorch framework under the Windows 10 environment, with a GPU of NVIDIA Quadro P2000 20.9 GB, to train and validate loess earthquake landslides on the satellite imagery platform and to achieve automatic segmentation of loess earthquake landslides in the validation area, improving recognition accuracy. Three databases were established to optimize the segmentation prediction performance of the model. Database 1 contained 494 original loess earthquake landslide maps, Database 2 contained 920 loess earthquake landslide maps obtained after expansion through rotation, and Database 3 contained 613 maps obtained after expansion through cropping. The database 1 was input into the original Unet model and the VGG‒Unet model, corresponding to Experiment 1 and Experiment 2, respectively. Then, Databases 2 and 3 were input into the model that produced better results in the first two tests, namely the VGG‒Unet model, corresponding to Experiment 3 and 4, respectively. The four sets of tests were compared and analyzed to determine the optimal model performance metrics. Landslides within the study area were extracted from satellite imagery with coverage areas of 0.5～1.0 km2, ensuring that each image contained an appropriate number of landslide patches and effective training data, providing the model with sufficient landslide pixels for feature learning. A total of 494 raw images were obtained, covering 1 052 landslides, designated as Database 1. A portion of the images was expanded through rotation, yielding 920 image panels, designated as Database 2, while another portion was expanded through cropping, yielding 613 image panels, designated as Database 3. Database 1 was input into models utilizing ResNet and VGG16 as backbone features, namely the original Unet model and the VGG‒Unet model, corresponding to Experiments 1 and 2, respectively. Then, Databases 2 and 3 were input into the model that yielded superior results in the preceding two experiments, namely the VGG‒Unet model, corresponding to Experiment 3 and 4, respectively. Finally, the VGG‒Unet model, which demonstrated the best performance among the four tests, was utilized to generate prediction results for typical loess earthquake landslide areas within the validation area. The researchers compared the predicted identification results of the model with the field-investigated landslide locations one by one to ensure experimental accuracy.Results and DiscussionsThe results showed that the VGG‒Unet model demonstrated the best training and validation performance on the dataset that was cropped and expanded to 613 frames, and performed particularly well on the unknown dataset. The accuracy was 89.57%, the MIoU was 70.13%, the F1 value was 81.11%, and the Recall reached 80.53%, and all model performance indices were high. In contrast, on the original 494-frame dataset, the performance of the original Unet model was the lowest, followed by the VGG‒Unet model, and both were lower than the results of Test 3. However, on the dataset expanded through rotation to 920 frames, although the training performance metrics of the VGG‒Unet model reached the highest values among the three tests, the validation performance metrics decreased to the lowest. This method identified loess landslides triggered by earthquakes. The model showed higher accuracy in identifying loess earthquake landslides with well-developed topographic and profile features, whereas it tended to miss landslides with ambiguous tonal and profile features, as well as small-scale landslides, and tended to misinterpret complex terrains such as uneven strata. Overall, the VGG‒Unet model effectively predicted areas prone to loess landslides. The clearer and more complete the characteristic markers of loess landslides were, the more effectively the VGG‒Unet model performed in segmentation tasks. Its high precision and accuracy enabled computers to rapidly and accurately identify target areas and determine their locations. For landslides with blurred or incomplete characteristic markers, although the model produced a limited number of misclassifications and omissions during segmentation, most areas were still accurately segmented and identified.ConclusionsTherefore, the method proposed in this study can efficiently segment and recognize the same type of landslide areas and can automatically segment and recognize the locations of landslides before field investigations. This approach assists field teams in locating loess earthquake-induced landslides and provides technical support for large-scale landslide disaster investigations.关键词：VGG‒Unet model;loess earthquake landslide;landslide identification;satellite images;field investigation211|556|0更新时间：2026-03-26
INTELLIGENCE INTERDISCIPLINARY SCIENCE AND ENGINEERING
- 摘要：ObjectiveAccurately assessing the probability of slope failure induced by the anisotropic spatial variability of soil is crucial for slope reinforcement and risk evaluation. A current limitation in existing research is the insufficient discussion of the reinforcement effects of anti-slide piles and limit analysis under anisotropic random fields. This study introduces the discrete mechanism of limit analysis into general rotational anisotropic random fields and employs the upper-bound analysis method to evaluate slope reliability.MethodsFirstly, the random field was generated using the Cholesky decomposition method. The autocorrelation function was transformed through a coordinate system conversion before the decomposition of the autocorrelation coefficient matrix to ensure that the random field demonstrated general rotational anisotropy. This procedure generated representative random fields for the soil strength parameters c and φ, which exhibited transverse isotropy and general rotational anisotropy, respectively. Then, the discrete mechanism upper-bound analysis theory was employed to generate velocity discontinuity surfaces in spatially variable soils, which rigorously satisfied the associated flow rule in plasticity mechanics. Specifically, at any spatial location, the angle between the tangential velocity along the velocity discontinuity surface and the actual velocity equaled the soil's internal friction angle. Next, to evaluate the reinforcement effect of anti-slide piles, the elastic-plastic analysis theory was applied to determine the ultimate lateral soil pressure acting on the piles. Assuming a linear distribution of lateral soil pressure along the piles, an energy balance equation was established for the slope reinforced with anti-slide piles. Finally, the strength reduction method was employed to calculate the safety factor at the critical state of slope failure. The influences of fluctuation scales, coefficients of variation, and correlation coefficients on the reliability of slopes reinforced with anti-slide piles were compared for the two types of anisotropic spatial variability random fields.Results and DiscussionsThe verification of the calculation for the homogeneous pile-reinforced slope was conducted by setting all unit grid parameters of the random field to equal values, simulating a homogeneous condition. The safety factors for Examples 1 and 2 were calculated as 1.63 and 1.46, respectively, with an average error of 1.8% in the verification results. For the anisotropic spatially variable slope, the error in the failure probability was on the order of 10-3. In addition, the variation trends of the failure probability for the anisotropic random field based on coordinate transformation remained consistent regardless of the differences between the minimum and maximum fluctuation angles. When analyzing the influence of fluctuation scales, the probability density function (PDF) curves of the safety factor became higher and narrower as the minimum fluctuation scale decreased. The influence of the minimum fluctuation scale on the PDF curve was more significant than that of the maximum fluctuation scale. As the minimum and maximum fluctuation scales increased, the PDF curves for general rotational anisotropy exhibited a greater tendency to become shorter and wider compared to those for transverse anisotropy. Regarding the influence of the coefficient of variation, as Vc and Vφ increased, the corresponding PDF curves became shorter and wider while shifting to the left, resulting in a significant increase in failure probability. As the coefficient of variation of the internal friction angle (Vφ) increased, the distribution of the critical sliding surface of the slope became more dispersed. This result indicated that when the soil's Vφ value was high, ensuring slope stability could require increasing the burial depth of the anti-slide piles. Regarding the influence of the cross-correlation coefficient, the failure probability (Pf) increased as ρc,φ increased. Under identical cross-correlation coefficients, transverse anisotropic slopes exhibited greater reliability than general rotational anisotropic slopes. Regarding the installation position of the anti-slide piles, for transverse anisotropic random fields, Pf initially decreased as the height of the pile arrangement increased, before slightly increasing. For general rotational anisotropic random fields, Pf generally decreased as the height of the pile arrangement increased. The failure probability of the general rotational anisotropic spatially variable pile-reinforced slope was higher than that of the transverse anisotropic pile-reinforced slope. The optimal pile position Xf/Lx (the ratio of the distance from the pile installation position to the toe of the slope to the distance from the top of the slope to the toe of the slope) for the general rotational anisotropic spatially variable pile-reinforced slope approached 0.8. In contrast, for the transverse anisotropic pile-reinforced slope, it ranged between 0.6 and 0.7. When designing spatially variable pile-reinforced slopes, neglecting the influence of anisotropy on design parameters can lead to risks in determining the anti-slide pile installation position.ConclusionsThe results demonstrate that the optimal pile position is relatively insensitive to spatial variability parameters. However, a significant difference in the optimal pile position is observed between the two typical anisotropic random fields. In particular, under general rotational anisotropy conditions, the use of the Xf/Lx=0.6 pile position in practical engineering can pose potential risks. In addition, the internal friction angle of soil shows more significant fluctuations in spatial distribution. When the slope becomes unstable, the slip surface distribution tends to become more discrete and deeper. Therefore, increasing the embedding depth of anti-slide piles in engineering projects is essential to improve the reliability of slope stability.关键词：soil slope;anisotropic spatial variability;Monte Carlo simulation;upper-bound analysis;reinforcement with anti-slide piles;discrete mechanisms254|393|0更新时间：2026-03-26
- 摘要：ObjectiveThe widespread occurrence of expansive soil landslides poses a significant threat to the long-term stability and safe operation of the Danjiangkou Reservoir. Expansive soil is rich in hydrophilic minerals such as montmorillonite and illite, which are highly sensitive to water. The structural characteristics, expansion and contraction behavior, and strength attenuation properties of expansive soil constitute the main factors leading to the instability and failure of expansive soil landslides and exhibit typical features of shallow depth, traction movement, gentle slope conditions, seasonality, and repeatability. At present, most related studies focus on the effects of rainfall on the shallow failure of expansive soil landslides. However, for large expansive soil landslides affected by reservoir water, the combined effects of rainfall and reservoir water on landslide deformation and failure require further clarification. Therefore, investigating the mechanisms of instability and failure of expansive soil landslides under the combined influence of rainfall and reservoir water level fluctuations is essential for implementing meteorological early warning, risk assessment, and engineering mitigation of landslides at the Danjiangkou Reservoir.MethodsUsing the Jingang Village landslide in Yunyang District, Shiyan City, a typical cross-section along the main sliding direction was selected to establish a two-dimensional slope experimental model based on a comprehensive analysis of geological surveys, GNSS monitoring, and data on rainfall and reservoir water levels. The model was vertically divided into the sliding mass, sliding zone, and sliding bed, while three additional horizontal surfaces were arranged from the front edge to the rear edge. Each monitoring profile was equipped with moisture content sensors and pore water pressure sensors, and displacement sensors were installed on the surface of the landslide model. Based on the step-like deformation characteristics of the landslide corresponding to fluctuations in reservoir water levels and rainfall conditions, and incorporating meteorological and hydrological data from the area surrounding the Jingang Village landslide, it was observed that the reservoir area experienced frequent heavy rainfall during periods of high water level operation and during phases of rising water levels. Three physical simulation experiments were designed. Condition one simulated the reservoir water level rising from low to high, remaining stable for a period, and then dropping back to a low level, with the cycle repeated several times. Condition two simulated continuous heavy rainfall during periods of high reservoir water level. Condition three was developed based on condition one and simulated continuous heavy rainfall while the water level rose from low to high. Real-time monitoring of the landslide failure process, including pore water pressure, moisture content data, and slope deformation, facilitated the investigation of the instability mechanisms of expansive soil landslides under the combined influence of rainfall and reservoir water level fluctuations.Results and DiscussionsThe main research content and results were summarized as follows: 1) Model experiments showed that the failure process of the Jingang Village landslide manifested as tractional failure. The process began with localized deformation at the toe of the slope and then progressively developed upward, eventually leading to overall failure in the mid-front section. This process formed an open face and a steep step at the landslide fracture, followed by stepwise collapse; 2) Reservoir water fluctuations and rainfall produced different response patterns in the internal pore water pressure and moisture content of the landslide. Only during fluctuations in the reservoir water level did the pore water pressure and volumetric water content at the front edge of the slope exhibit cyclic changes, while no response occurred at the middle and rear edges. After rainfall was applied, pore water pressure and volumetric water content outside the influence range of the reservoir water began to respond, increased with continued rainfall, and eventually stabilized. The increase in moisture content within the sliding zone was significantly higher than that at the slope surface, indicating that water ultimately accumulated near the sliding zone; 3) In situ monitoring and physical model experiments demonstrated that a drop in reservoir water level produced sustained deformation at the front edge of the landslide. Rainfall caused overall deformation of the landslide, which was significantly greater than the deformation induced by the drop in reservoir water level. Landslides that experienced repeated cycles of rising and falling reservoir water levels were more susceptible to significant deformation under heavy rainfall conditions and could potentially become unstable; 4) The continuous deformation of expansive soil landslides affected by the combined action of rainfall and reservoir water resulted from the dual effects of external added stress and the reduction of internal strength parameters. Persistent heavy rainfall rapidly infiltrated the slope body along fractures and raised the groundwater level, while the strength of the sliding zone soil continued to decrease due to prolonged soaking by groundwater. This process served as a key factor driving the continued deformation of large expansive soil landslides.ConclusionsBased on the analysis of the deformation mechanism of expansive soil landslides under the combined effects of rainfall and reservoir water, together with geological surveys and in-situ monitoring data such as rainfall, reservoir water level, and GNSS monitoring, and based on the analysis of landslide deformation characteristics, a physical model was designed and constructed. The changes in relevant landslide parameters under different heavy rainfall and reservoir water conditions were monitored in real time, reflecting the deformation and failure mechanisms of expansive soil landslides under the combined effects of rainfall and reservoir water by installing sensors such as displacement, pore water pressure, and water content sensors. The findings provide reference significance for the study of expansive soil landslide failure theory and for practical disaster prevention and mitigation efforts.关键词：Expansive soil landslide;Deformation failure mechanism;physical model test;Reservoir water level fluctuation;rainfall275|321|0更新时间：2026-03-26
- 摘要：ObjectiveIn situations involving mountain floods, intense rainfall-induced erosion, and the breaching of clusters of landslide dams that lead to upstream soil and water loss, the inflow to landslide dams often becomes a sand-laden flow. Compared to clear water inflow, sand-laden flows produce more complex infiltration, erosion, and deposition processes, substantially influencing both the dam breaching process and the peak flow rates. This study investigates the mechanisms through which the sediment concentration and particle size of sand-laden flows influence dam breaching processes and examines the erosion-deposition dynamics and seepage-clogging phenomena induced by sand-laden flows during landslide dam breaching.MethodsFive sets of model experiments were conducted using a flume measuring 5 m in length and 0.4 m in width. Clear water inflow served as the control group, while the effects of different volumetric concentrations of sand-laden flow (0.01 and 0.03) and various maximum sediment sizes (0.006 5, 0.125 0, and 0.500 0 mm) on the dam breaching process were investigated. The primary method involved pre-experimental testing to adjust the composition of the sand-laden water flow, supplemented by thorough mixing in the water supply tank and the flow stabilization tank to achieve a stable and target sand concentration in the dam body area. The design of the landslide dam model was based on the range of dimensionless parameters related to dam height, dam volume, and reservoir volume derived from a database of landslide dam cases, and the material selection referenced empirical data from the 2008 Tangjiashan landslide dam. The experiments primarily used water level gauges to record the breach outflow process, while high-resolution cameras and grid systems monitored changes in the erosion rates of the landslide dams. Timed quantitative sampling methods tracked variations in sediment concentration. In addition, sampling, drying, and sieving techniques collected data on the moisture content and grain size distribution of the residual dam. The main steps of the experiment included material preparation, placement of the movable bed, layered construction of the dam body, execution of model tests, recording of experimental data, and subsequent sample analysis.Results and DiscussionsThe failure mode of the landslide dams in this series of experiments was consistently classified as overtopping failure, and the breaching process was primarily characterized by layered erosion accompanied by slope instability. The breach evolution process was divided into two stages: the initiation stage and the development stage. During the breach initiation stage, headward erosion predominated, and the dam height showed little noticeable reduction. The breach development stage was characterized by a rapid decrease in dam height, and peak flow rates were reached early in this stage. The water level in front of the dam reached its peak at the end of the breach initiation stage, whereas the peak flow rate occurred early in the breach development stage, and a secondary peak developed due to upstream slope instability that partially blocked the breach. The concentration of sand-laden flow in the dam area was inversely correlated with the breach flow rate, whereas the concentration measured at the end of the flume area showed a positive correlation. The mean particle size of the residual dam decreased longitudinally within the dam area and remained relatively stable in the downstream movable bed area. Comparative tests of sand-laden flows with different volumetric concentrations (Tests 1~3) showed that as the initial volumetric sand concentration increased from 0 to 0.03, the peak breach flow rate increased from 3.2 to 3.6 L/s, the time to peak increased from 76 to 90 s, the maximum water level in front of the dam increased from 24.6 to 26.9 cm, the residual dam height increased from 6.5 to 8.2 cm, and the final average breach width increased from 18.8 to 20.6 cm. The analysis indicated that higher sediment concentrations in the flow enhanced the erosion capacity during the breach development stage; however, the depositional layer formed during the process influenced the downward erosion rate during both the breach initiation stage and the late breach development stage. Comparative tests of sand-laden flows with different maximum particle sizes showed that as the maximum particle size increased from 0.006 5 to 0.500 0 mm, the peak breach flow rate slightly increased from 3.3 to 3.4 L/s, the time to peak increased from 79 to 90 s, and the maximum water level in front of the dam increased from 26.0 to 26.5 cm. The residual dam height remained relatively constant at approximately 7.5 cm, and the final average breach width also showed minimal variation at approximately 19.5 cm. The analysis indicated that the particle size of the sand-laden flow had minimal influence on the erosion rate of the landslide dam, mainly because the effect of particle size on the viscosity of the flow became negligible when particle sizes exceeded 0.1 mm. The comparative analysis of clear water inflow (Test 1) and sand-laden flows (Tests 2~5) showed that during the breach initiation stage and the late stage of breach development, low-velocity and high-concentration sand-laden flows generated strong depositional effects. These effects caused a slight reduction in erosion rates, an increase in the maximum water level in front of the dam, and a greater residual dam height. In contrast, during the early to middle stages of breach development, high-velocity sand-laden flows demonstrated greater erosive capacity and reduced deposition, which was reflected in increased erosion rates and higher peak outflow. In addition, sand-laden flows during the dam breaching process produced a pronounced clogging effect, which resulted in a multilayered structure within the dam body composed of depositional layers, retention layers, and original layers. The dense structure of the retention layer reduced the permeability of the sand-laden flow, lowered the position of the saturation zone on the breach slopes, and increased the scale of slope failures.ConclusionsThe experimental results indicate that, compared to clear water, the viscosity coefficient of sand-laden flow is higher, which enhances its erosive capability and also intensifies deposition. As the concentration increases, the viscosity of the sand-laden flow increases markedly, resulting in higher erosion rates, larger breach sizes, and greater peak outflow rates. As the particle size increases, the viscosity coefficient of the sand-laden flow exhibits negligible variation, leading to only minor changes in erosion rates and peak outflow rates. The seepage-clogging effect refers to the accumulation of sand particles within the surface layer of the dam body during the infiltration of sand-laden water flow, accompanied by the retention of these particles inside the dam. This retention layer decreases the infiltration rate, which lowers the saturated zone along the side slope of the breach. This change results in an increased scale of collapses on the side slope.关键词：landslide dam;Sand-Laden Flow;flume test;Overtopping;peak outflow discharge413|710|0更新时间：2026-03-26
MECHANISM OF LANDSLIDE\-DAMMED LAKE AND ITS CONTROL
- 摘要：ObjectiveThe determination of indicator weights in multi-criteria group decision-making (MCGDM) represents a critical issue in architectural programming and post-occupancy evaluation (POE). Traditional approaches often model expert weights and indicator weights independently, which leads to potential inconsistencies and suboptimal decision outcomes. This study aims to develop an iterative group decision-making model that simultaneously optimizes expert weights and indicator weights, enhancing the accuracy, reliability, and practical applicability of weight determination in complex engineering contexts.MethodsAn iterative optimization model was proposed to integrate the determination of expert weights and indicator weights. First, an expert weight optimization model was established based on the consistency between individual expert evaluations and aggregated group evaluations. This model incorporated the relative importance of different indicators through a weighted distance metric derived from a specially defined A-norm, which generalized the standard Euclidean distance. Second, indicator weights were calculated by aggregating individual expert evaluations weighted by their corresponding credibility levels. Finally, an iterative framework was designed to enable mutual calibration between expert weights and indicator weights until convergence was achieved. The model was implemented algorithmically and validated using two real-world case studies, including green performance evaluation during architectural programming and building post-occupancy evaluation.Results and DiscussionsNumerical experiments demonstrated that the proposed iterative model converged reliably under various initial conditions. In the green performance evaluation case, which involved nine experts and 96 indicators, the model converged within eight iterations, with differences between successive iterations falling below 10‒10 in the infinity norm. The resulting expert weights exhibited significant variation, reflecting differences in expertise across building types, such as office buildings and tourism buildings. Similarly, in the POE case involving five experts and 69 indicators, the model converged within seven to nine iterations, and the resulting weights effectively captured functional differences between museum buildings and stadium buildings. Compared to baseline methods that assigned equal weights to experts, the proposed model generated more discriminative and contextually appropriate indicator weights. For example, in the office building case, the highest weighted indicators included room temperature, sunshade facilities, and noise control, whereas in the tourism building case, rest spaces and accessibility planning received higher weights, which were consistent with functional priorities. The model also demonstrated robustness to different initial conditions and maintained convergence regardless of the order of weight updates.ConclusionsThis study presents a novel iterative group decision-making model that simultaneously determines expert and indicator weights by leveraging their mutual influence. The proposed model improves upon existing approaches by incorporating indicator importance into the calibration of expert weights, achieving more consistent and context-aware weight assignments. The convergence and practical applicability of the model are empirically validated through real-world case studies in architectural programming and POE. The method provides a mathematically rigorous and engineering-oriented approach to MCGDM, with potential applicability in other domains that involve multi-expert and multi-criteria evaluation processes.关键词：architectural programming and POE;multi-criteria group decision-making;indicator weights;expert weights;iterative model247|310|0更新时间：2026-03-26
- 摘要：ObjectiveDue to the concave appearance, spatial layout, and mirror reflection characteristics of building curtain walls, photothermal focusing problems frequently occur. Reflected sunlight and excessive solar radiation highly concentrated by concave curtain walls induce severe glare and deteriorate the thermal environment around concave buildings, and can even cause burns to people or objects in the focused area or trigger fire hazards. Therefore, it is critical for architects to rapidly observe and analyze photothermal focusing so that such problems can be effectively avoided through design adjustments at the design stage. However, existing static calculation and simulation analysis methods exhibit limitations in three-dimensional building model establishment, visualization of calculation results, and user friendliness for the primary decision-makers, architects, and thus fail to effectively support the optimization of design schemes for buildings with concave curtain walls. Therefore, this study aims to develop a new method that enables architects to dynamically simulate and calculate photothermal focusing of curtain walls during the design stage and to rapidly perform analytical evaluations.MethodsBased on the principle of mirror reflection and the corresponding ray tracing method, the causes of photothermal focusing induced by concave building curtain walls were identified. The logical framework for dynamic simulation calculation and analysis of photothermal focusing in curtain walls was clarified. Through theoretical derivation, calculation models for Ray Tracing and Sunny Day Solar Radiation were developed, and the corresponding battery modules were prewritten in Grasshopper. Then, a new method was established by adopting Rhino, a drawing software widely used by architects, as the modeling and visualization platform, while the modified Grasshopper served as the simulation and calculation engine. Since the Sunny Day solar radiation model constituted the most important theoretical foundation of the proposed method, this model, along with its four calculation modules, namely the calculation module of dynamic geometric characteristics of the sun, the calculation module of solar radiation, the calculation module of reflected radiation intensity of curtain walls, and the calculation module of solar radiation intensity and ground temperature in the focus area, as well as the associated formulas, was systematically and logically introduced in detail. Finally, the application of the proposed method in the simulation, calculation, and analysis of photothermal focusing problems in curtain walls was demonstrated step by step, and the results obtained for a case study building using the proposed method were compared to field measurement results reported in previous studies and references.Results and DiscussionsThe newly developed method exhibited several advantages. First, by adopting Rhino as the modeling and visualization platform, the method was highly accessible to architects, as the three-dimensional building models created in Rhino did not require reconstruction and could be directly used for simulation. In addition, simulation results could be promptly transmitted back to Rhino, enabling the visualization of numerical results as graphical outputs, which allowed architects to observe potential scenarios associated with different design schemes and to efficiently analyze photothermal focusing issues. Second, through the calculation module of dynamic geometric characteristics of the sun, the modified Grasshopper was able to trace both direct and reflected rays from curtain walls, enabling the identification of whether photothermal focusing problems would occur, as well as their timing and location, and the specific glass units responsible for their formation. Through the calculation module of solar radiation, the modified Grasshopper calculated the intensity of direct solar radiation at any time throughout the year. Through the calculation module of reflected radiation intensity of curtain walls, the intensity of reflected solar radiation from curtain walls was determined. Through the calculation module of solar radiation intensity and ground temperature in the focus area, the combined intensities of direct and reflected solar radiation on the ground within focus areas or hotspots, as well as the corresponding ground temperature, were calculated. Overall, the proposed method enabled architects to observe the dynamic evolution of photothermal focusing phenomena associated with different curtain wall design schemes over an entire year during the design stage and to rapidly determine which design schemes would induce photothermal focusing, when and where such focusing would occur, how the focusing intensity would evolve, and which curtain wall units would contribute to the focusing effect. Finally, comparisons between the simulation results of the case study building at the Vernal Equinox and field measurement results further demonstrated that the proposed method allowed architects to obtain accurate and comprehensive results efficiently, supporting improved decision-making aimed at avoiding or mitigating photothermal focusing problems induced by buildings with concave curtain walls.ConclusionsThe newly developed method utilizes Rhino as the modeling and visualization platform and a modified Grasshopper environment as the simulation and calculation engine. This approach enables architects to adjust curtain wall design schemes in a timely manner to avoid photothermal focusing effects or to reduce focusing intensity during the design stage.关键词：Sun rays;Concave curtain wall;solar radiation;Focus area;visualization227|538|0更新时间：2026-03-26
GREEN BUILDING AND SMART CONSTRUCTION
- 摘要：ObjectiveDue to the rapid development of infrastructure construction, the consumption of concrete has increased significantly. However, natural aggregate remains in shortage in many regions because of environmental constraints. Therefore, the use of locally available desert sand in concrete production in structural engineering receives increasing attention in northwest China's Xinjiang Province. The Xinjiang region is characterized by limited fine natural aggregate resources, and nearly a quarter of the area is covered with desert sand. Therefore, using desert sand as a fine aggregate resource addresses the imbalance between the supply and demand of river sand, and it reduces the cost of concrete production and transport. The column is one of the essential components of a structure, so numerous studies appear in previous literature. However, relatively limited research examines the axial compression behavior of reinforced desert sand concrete short columns. Therefore, the primary objective of this study is to investigate the influence of desert sand on the axial compression behavior of reinforced concrete short columns. The failure mode, load axial displacement curve, load concrete strain curve, load longitudinal steel strain curve, load stirrup strain curve, and ductility behavior of reinforced desert sand concrete short columns are analyzed.MethodsTaking the desert sand replacement ratio, longitudinal steel ratio, and stirrup ratio as the main parameters, the study designed eighteen reinforced concrete short columns made with desert sand in accordance with "Code for design of concrete structures" (GB 50010—2010). All columns were constructed with dimensions of 550 mm in length, 150 mm in width, and 150 mm in height. The test was conducted on an electro-hydraulic servo pressure testing machine with an axial loading capacity of 5 000 kN, and the loading velocity was set to 0.01 mm/min. Four displacement sensors with a measuring range of 100 mm were arranged along the central axis of the four surfaces of each specimen to record the axial displacement. Strain gauges were attached to the longitudinal steel bar, stirrup, and concrete surface to measure their respective strains.Results and DiscussionThe test results showed that 1) the reinforced desert sand concrete short columns failed by forming vertical through cracks, and the failure process and pattern of reinforced desert sand concrete short columns were similar to those of reinforced ordinary concrete short columns. 2) When the desert sand replacement ratio increased from 0 to 80%, the peak load changed by -9.04%, 4.62%, 2.00%, and -5.52%, and the ductility coefficient changed by -16.20%, 5.40%, 1.80%, and -9.46%, respectively; the concrete's longitudinal and transverse strains of the specimens increased at first, then decreased, and then increased again; the longitudinal steel strain and stirrup strain of specimens with a desert sand replacement ratio of 20% and 80% were larger, while the longitudinal steel strain and stirrup strain of specimens with a desert sand replacement ratio of 40% and 60% were smaller. The main reason for this behavior was that the concrete contained more pores than ordinary concrete when the desert sand replacement ratio was low (20%), while the internal pores were filled by desert sand when the replacement ratio ranged from 40% to 60%. The axial compression properties decreased again because the strength of desert sand particles was lower than that of river sand when the replacement ratio reached 80%. 3) When the longitudinal steel ratio increased from 0.89% to 3.57%, the peak load increased by about 8.53% and 15.80%, and the ductility coefficient increased by about 27.44% and 39.98%, respectively; the concrete's longitudinal and transverse strains decreased gradually; the longitudinal steel strain and stirrup strain of the specimens also decreased gradually; the results were consistent with those of ordinary reinforced concrete columns. 4) When the stirrup ratio increased from 1.22% to 1.52%, the peak load increased by about 11.44% and the ductility coefficient increased by about 21.58%; the concrete's longitudinal and transverse strains decreased gradually; the longitudinal steel strain and stirrup strain of the specimens decreased gradually; the results were also consistent with those of ordinary reinforced concrete columns. 5) The predicted axial compression strength using "Code for the Design of Concrete Structures" of China Code GB 50010—2010 and "Structural Concrete Building Code" of US Code ACI318 was higher than the experimental values by 21% and 38%, respectively; after introducing the influence of desert sand on the strength of concrete and the matching effect coefficient of desert sand concrete and steel bar, the study established a formula for the axial compression strength of reinforced desert sand concrete short columns. The calculated results agreed well with the experimental results.Conclusions1) The reinforced desert sand concrete short columns with desert sand replacement ratios of 20% and 80% demonstrated poor performance, while those with replacement ratios of 40% and 60% exhibited good performance. Therefore, it was appropriate to use desert sand concrete specimens with replacement ratios of 40% to 60%, and a 60% replacement ratio was recommended to make full use of desert sand resources. 2) When the longitudinal steel ratio increased from 0.89% to 3.57% and the stirrup ratio increased from 1.22% to 1.52%, the axial compression properties of the reinforced desert sand concrete short column improved gradually, and the behavioral characteristics were comparable to those of the ordinary reinforced concrete short column. 3) The formula for calculating the axial bearing capacity of reinforced desert sand concrete short columns was proposed by incorporating the influence of desert sand on concrete strength and the matching effect coefficient of desert sand concrete and steel bars. It can provide a theoretical basis for the engineering application of the reinforced desert sand concrete short column.关键词：desert sand concrete;short column;axial compressive property;ductility property;carrying capacity210|515|0更新时间：2026-03-26
- 摘要：ObjectiveStudying the shear characteristics of the interface between granite residual soil and concrete is crucial for ensuring the safety of geotechnical engineering applications, including pile foundations, anchor rods, and underground pipelines. At present, the shear characteristics of the interface are primarily influenced by the water content of the soil and the roughness of the concrete surface. This study investigates the effects of these two factors on the shear characteristics through experimental and numerical analysis.MethodsBased on the standard structural surface roughness curve, a three-dimensional model representing surfaces with different roughness levels was established. TPU molds with varying roughness values were produced using 3D printing technology, and concrete test blocks exhibiting different surface roughness characteristics were cast. Indoor large-scale direct shear tests were conducted under four water content levels (14%, 19%, 24%, and 29%) shear stress displacement curve, shear strength, and vertical displacement were analyzed. Finally, discrete element software was employed to investigate meso-mechanical characteristics, including porosity, force chain distribution, and fabric anisotropy, of granite residual soil during the shearing process.Results and DiscussionsThe test results indicated that the interfacial shear strength initially increased and decreased with increasing water content. At a water content of 19%, the maximum interfacial shear strength and interfacial cohesion were achieved, with values of 82.4 kPa and 30.17 kPa, respectively, under a normal stress of 100 kPa. Water content exerted a pronounced influence on the stress-strain behavior of the interface. At lower water content levels ranging from 14% to 19%, shear softening behavior was observed, whereas higher water content levels from 24% to 29% exhibited shear hardening behavior. At a water content of 14%, the strains corresponding to peak interface shear stress were 3.26%, 3.84%, and 3.96 % under normal stresses of 100 kPa, 200 kPa, and 300 kPa, respectively, indicating enhanced shear ductility with increasing normal stress. As the concrete roughness coefficient increased, the interface shear strength increased gradually; however, once the critical roughness value was reached, further increases in roughness did not result in additional shear strength enhancement. When the concrete surface was relatively smooth, the interface shear strength mainly originated from frictional resistance between the soil and concrete, and the shear failure surface was located at the interface. When the concrete surface became rougher, grooves on the surface interlocked with soil particles, generating a biting effect that increased the interface shear strength and caused the shear failure surface to shift from the interface into the soil, forming a localized shear band. With further increases in roughness beyond the critical value, additional soil particles filled the surface grooves, which reduced the effective roughness of the concrete surface and led to a marginal occlusion effect. Therefore, the shear failure surface no longer propagated deeper into the soil and even exhibited slight regression, resulting in no further increase in interface shear strength. Numerical simulation results further demonstrated that the particle porosity within the upper 5 cm range of the sample increased during shearing, whereas the porosity in the remaining regions showed an opposite trend. The influence of porosity on particle behavior gradually decreased with increasing distance from the interface. In addition, the principal anisotropy directions of both normal and tangential contact forces deflected consistently with the shear direction.ConclusionsA critical roughness value was identified at the soil‒concrete interface. Before reaching this critical roughness, the shear strength, interfacial cohesion, and internal friction angle of the interface increased with increasing roughness. However, once this critical value (JRC equals to 12.8) was exceeded, further increases in these parameters ceased, and a stable state was reached. Simulation results indicated pronounced anisotropy in both normal and tangential contact forces, with the principal anisotropic direction aligned with the direction of maximum compressive stress, enhancing resistance to shear deformation. In practical foundation engineering applications, the water content of foundation soil can be appropriately controlled based on experimental findings. In addition, the surface morphology of concrete can be designed with reference to the critical roughness value, improving interface shear resistance.关键词：soil‒structure interface;moisture content;roughness;shear characteristics;numerical simulation294|641|0更新时间：2026-03-26
- 摘要：ObjectiveThe alluvial plain in the lower reaches of the Yellow River exhibits typical sedimentary stratigraphic characteristics and is mainly composed of interbedded layers of clay, fine sand, and silty sand. Its bearing capacity is relatively low. In major infrastructure projects in this region, rotating drilling variable-diameter piles (expanded foundation piles) demonstrate advantages in construction efficiency and settlement control. However, the vertical compressive performance and failure mechanisms of such piles under the specific geological conditions of the Yellow River alluvial layer remain insufficiently investigated. This study aims to comprehensively examine the vertical bearing capacity, load transfer characteristics, and ultimate bearing capacity of rotating drilling variable-diameter piles within this typical interbedded silty soil-clay layer, providing a theoretical and practical basis for optimizing pile design in similar geological environments.MethodsA series of investigations were conducted, including field experimentation, numerical simulation, and theoretical analysis. Full-scale field tests were conducted at an airport expansion site located on the Yellow River alluvial plain. Six instrumented test piles were installed using rotary drilling: three piles with a shaft diameter of 1 000 mm and three piles with a diameter of 800 mm, each featuring an enlarged base with a diameter twice that of the shaft and constructed at a precise distance above the pile tip. The piles were extensively equipped with strain gauges at multiple depths along the shaft and earth pressure cells at the base to continuously monitor internal forces and end-bearing pressures. The load tests were performed as slow maintained tests, and the piles were incrementally loaded to 1.2 times their code-calculated ultimate load to fully characterize their pre-yield and post-yield behavior and capture the complete load-settlement trajectory. Complementary three-dimensional finite element models were developed using PLAXIS 3D software to obtain insights beyond the limits of physical measurements. The complex stratified soil profile was modeled, and the constitutive model Hardening Soil was selected to simulate the mechanical behavior of the soil. The parameters used in the constitutive model were carefully calibrated and validated against the field test data to ensure model reliability. For further comparison, identical models of conventional equal-diameter piles with the same dimensions were established and subjected to the same loading and boundary conditions. In addition, a refined theoretical analytical method was applied. This method incorporated specific modifications for calculating the settlement of the enlarged base by considering its geometry and for estimating the in-situ soil deformation modulus from laboratory data. The proposed theoretical analytical method predicted the complete nonlinear load-settlement response and produced a more accurate back-calculation of the ultimate bearing capacity.Results and DiscussionsThe on-site tests demonstrated that the load-settlement (Q-S) curve of the variable-diameter pile exhibited a gradual and progressive failure mode without a distinct yield point, which was consistent with the characteristics of friction-end-bearing piles. The release of the frictional resistance and end-bearing force of the pile body did not occur simultaneously. The resistance of the upper portion of the pile body reached its limit earlier than that of the lower portion, and the complete release of the pile body's frictional resistance generally occurred before the complete release of the end-bearing force. The widened bottom section played a crucial role and contributed more than 79% to 85% of the ultimate bearing capacity of the test pile. This load distribution not only reduced the total settlement of the pile body but also effectively delayed the elastic and plastic deformation of the pile body itself. Therefore, it provided more favorable conditions for the gradual mobilization of lateral resistance of the pile body in deeper and stiffer strata compared to straight piles. In cases of premature failure, these lateral resistances will not have been fully mobilized. In addition, compared to straight piles, the ratio of elastic compression of the pile body to the total settlement at the pile top of the variable-diameter pile was significantly lower, which indicated that its structural load transfer more closely followed general mechanical behavior. The analysis of the relative displacement between the pile and the soil showed that sliding initiated in the upper soil layer and gradually expanded downward as the load increased until it extended to the expansion base area under the ultimate load. The presence of the enlarged base prevented sudden failure of the pile body. The numerical simulation results confirmed the experimental observations. The ultimate bearing capacity of the variable-diameter pile obtained from the calibrated model was approximately 10% higher than that of the geometrically equivalent straight pile. The evolution of the plastic zone in the surrounding soil differed significantly: for the variable-diameter pile, the plastic yielding phenomenon initiated simultaneously near the pile head and in the expansion base area and then expanded downward from the top. During the failure process, the lateral extent remained within approximately 2 meters around the pile body. In contrast, for straight piles, the plastic zone expanded more extensively in both vertical and horizontal directions and frequently extended to the model boundary under the ultimate load, indicating a broader range of soil failure. The analysis of settlement influence showed that the surface settlement caused by the variable-diameter pile was significantly smaller than that produced by the straight pile and remained more localized. Through theoretical calculations, a predicted load-settlement curve that was highly consistent with the measured data was obtained, with errors typically ranging from 6% to 15%. Using this method, the ultimate bearing capacity obtained from the back-calculation agreed well with the numerical simulation results and the maximum test load. This agreement indicated that, within this stratum, the actual ultimate bearing capacity of the variable-diameter pile was approximately 10% to 20% higher than the value calculated using the empirical formula in the current Chinese standard (JGJ 94—2008).ConclusionsIn the alluvial layer of the lower reaches of the Yellow River, the rotating drilling variable-diameter pile behaves as a typical friction end-bearing pile. The release of pile body resistance does not occur simultaneously but instead proceeds segment by segment along the pile depth direction. Specifically, the upper part first releases its resistance completely, after which the lower part begins to release its resistance. The enlarged pile bottom significantly enhances the end-bearing effect, reduces and delays settlement, and raises a more complete release of the pile body frictional resistance, improving the overall vertical bearing capacity. Compared to the straight column-type pile, the pile body compression of the variable-diameter pile is relatively smaller, and the total settlement is also reduced. Its performance more closely resembles that of a non-rigid pile and produces a more localized influence on the surrounding soil. Its failure mechanism is characterized by the simultaneous formation of plastic zones at the pile head and the enlarged pile bottom. During loading, the plastic zone gradually expands from the top toward the bottom. In the ultimate state, the failure process is progressive, and the significant plastic deformation of the surrounding soil is limited to the area within approximately 2 meters from the pile body. This behavior differs from the extensive failure zone typically observed in straight piles. Under these specific geological conditions, the actual ultimate bearing capacity of this variable diameter pile is approximately 10% to 20% higher than the value calculated using current standard normative methods. The combination of on-site testing, numerical simulation, and revised theoretical analysis provides a reference for evaluating and predicting the performance of such piles.关键词：The Yellow River alluvial formation;variable diameter pile;static load test;stress performance;ultimate bearing capacity;numerical simulation233|562|0更新时间：2026-03-26
- 摘要：ObjectiveThis study selects grey brick masonry from ancient structures in Longshengzhuang, Inner Mongolia, as the research object to address the issues associated with long-term freeze‒thaw cycles and impact-induced damage to sand particles in ancient buildings in Inner Mongolia. Ball pressure impact tests conducted on grey bricks before and after freeze‒thaw cycles are utilized to investigate the impact damage behavior of grey bricks in ancient buildings.MethodsAn engineering material durability damage testing system simulating natural environmental conditions and an automatic ball drop impact apparatus were utilized to conduct ball pressure impact tests on ancient building grey bricks before and after freeze‒thaw cycling. The digital image correlation (DIC) technique was applied to analyze strain field distributions within the indentation zones of the bricks. Dynamic hardness and coefficients of restitution were calculated to evaluate elastic‒plastic behavior. In addition, ultra-depth three-dimensional microscopy and scanning electron microscopy were employed to observe damage morphology and to investigate underlying damage mechanisms.Results and DiscussionsAs the number of freeze‒thaw cycles increased, the apparent hardness, compressive strength, elastic modulus, and Poisson's ratio of the grey brick samples showed a declining trend. The impact process of ancient building grey bricks was divided into two stages. In the first stage, the brick surface was compressed under ball impact, and the impact load rapidly increased to its peak value. In the second stage, elastic deformation of the brick generated an impulse force acting on the ball, which appeared as a secondary rise in the impact load time curve. With increasing impact height, both the peak load and secondary rebound force initially increased gradually and then increased rapidly, while the time interval associated with secondary impact was extended. Under identical impact heights, freeze‒thaw cycling caused an increase in peak impact load, whereas the secondary rebound time and secondary impact peak load continuously decreased. Without freeze‒thaw cycling, the peak load and secondary rebound peak load were 1 240.1 N and 380.4 N, respectively. After 10 freeze‒thaw cycles, the peak impact load increased by approximately 22.3%, while the secondary rebound peak load decreased by approximately 26.4%. These effects became more pronounced after 20 cycles. Following 50 freeze‒thaw cycles, the surface impact peak load increased by approximately 92.2%, and the secondary impact peak load decreased by approximately 63.9%, accompanied by a significant reduction in the secondary impact time interval. After impact loading, strain within the damaged zone exhibited a ring-shaped diffusion pattern centered on the impact location, and radial strain decreased with increasing distance from the impact center. At impact heights of 23, 27, 31, 35, and 39 cm, the maximum radial strains were 3 531, 4 972, 5 780, 7 861, and 9 231, respectively. At identical impact heights, increasing numbers of freeze‒thaw cycles caused the radial strain in the central indentation region to increase by 729 to 5 490. Before 20 freeze‒thaw cycles, strain increases at the indentation edges were minimal, whereas after 20 cycles, edge strain increased significantly relative to earlier stages. The dynamic hardness and coefficient of restitution of ancient building grey bricks decreased as impact height increased. At an impact height of 23 cm, these values were 53.29 MPa and 0.41, respectively, while at 39 cm, they decreased to 31.65% and 24.39% of their initial values, respectively. Under identical impact heights, increasing freeze‒thaw cycles reduced the bricks' resistance to plastic deformation, and both dynamic hardness and coefficient of restitution progressively declined. Without freeze‒thaw cycling, dynamic hardness and coefficient of restitution were 53.29 MPa and 0.41, respectively. After 50 freeze‒thaw cycles, these values decreased to 54.91% and 34.15% of their original values. Under impact loading, the surface of the ancient building's grey bricks formed spherical crown-shaped indentations, with particle shedding observed around the indentation. Freeze‒thaw cycling caused surface pulverization, spalling, and cracking, while internal pore structures gradually expanded and became larger and more rounded, reducing resistance to plastic deformation. After impact loading, indentation depth, diameter, surface area, and volume increased progressively; the indentation surfaces became rougher, and particle shedding occurred within the impact pits. Unfrozen bricks exhibited the smallest indentation damage zones in terms of depth, area, and volume, indicating the highest resistance to plastic deformation. After 50 freeze‒thaw cycles, indentation depth and diameter increased by 51.26% and 15.66%, respectively, compared to the unfrozen condition. The experimental results provided valuable guidance for durability assessment, protection strategies, and restoration practices for ancient buildings in Inner Mongolia.ConclusionsFreeze‒thaw cycles significantly deteriorate the mechanical properties and impact resistance of grey bricks used in ancient buildings, leading to reduced dynamic hardness, decreased coefficient of restitution, and increased surface damage under impact loading. The experimental results provide valuable insights for durability assessment and for the development of protection and repair strategies for ancient buildings in Inner Mongolia.关键词：ancient building grey brick;impact load;freeze‒thaw cycle;digital image correlation(DIC);strain cloud diagram;damage mec-hanism208|766|0更新时间：2026-03-26
- 摘要：ObjectiveSteel tube confined concrete (STCC) columns have an increasingly broad application prospect in high-cold and high-intensity areas. In steel-concrete composite structures, the cooperative behavior of steel and concrete is an essential prerequisite for the normal stress performance of such structures. However, due to the influence of ambient temperature on the mechanical properties of steel and concrete and the interaction at their interface, the seismic performance of the structure cannot be ignored. Therefore, to investigate the influence of ambient temperature on the seismic performance of steel tube confined concrete columns, quasi-static tests of steel tube confined concrete columns under different ambient temperatures are conducted. Based on the test results, the influence of ambient temperature on the seismic indices of the specimens is analyzed, and the underlying mechanism by which ambient temperature affects the seismic performance of steel tube confined concrete columns is revealed. In addition, a modified formula for calculating the horizontal bearing capacity of steel tube confined concrete columns considering temperature effects is proposed to provide a necessary basis for the seismic design of steel tube confined concrete structures in high-cold and high-intensity areas.MethodsA set of quasi-static test devices suitable for different ambient temperatures was designed and developed. Quasi-static tests on nine steel tube confined concrete columns were conducted, with ambient temperature and concrete strength selected as the primary research parameters. Comparative tests were performed on three groups of steel tube confined concrete columns under different environmental temperature conditions (-40 °C, 20 °C, and 60 °C) and three groups with different concrete strengths (C40, C50, and C60). The influence mechanism of ambient temperature on the seismic performance of steel tube confined concrete columns was systematically revealed by comparing and analyzing seismic performance indices, including failure mode, horizontal bearing capacity, ductility performance, stiffness degradation, and energy dissipation capacity of each specimen. Considering the limitation that existing calculation methods for the horizontal bearing capacity of steel tube confined concrete columns did not consider the influence of ambient temperature, the temperature influence coefficient of horizontal bearing capacity was regressed based on the test results in this study. The existing calculation formula for horizontal bearing capacity was multiplied by the temperature influence coefficient k0, and the accuracy and reliability of the proposed formula were verified through comparisons with the experimental results of this study and other reported test results in the literature.Results and DiscussionsThe failure modes of specimens under identical temperature conditions were similar and were characterized by typical compression-bending plastic hinge failure. Under all temperature conditions, buckling occurred 10～30 mm above the stiffener at the base of the specimens, with initial local buckling gradually propagating toward both sides. Higher concrete strength resulted in more pronounced local buckling above the stiffener. In addition, steel tube tearing was observed under all temperature conditions. Under normal and high temperature conditions, tearing occurred in the middle of the buckled region, whereas under low temperature conditions, cracks developed along the lower side of the buckle, which was attributed to the significant increase in concrete strength in cold environments. Compared to normal temperature conditions, under low temperature conditions (-40 °C), the initial stiffness and horizontal bearing capacity of the specimens increased by 16.23% and 20.88%, respectively. However, the displacement ductility coefficient decreased by 37.52%, indicating a substantial reduction in ductility and energy dissipation capacity. Under high temperature conditions (60 °C), the initial stiffness and horizontal bearing capacity decreased by 12.52% and 4.25%, respectively, while the displacement ductility coefficient decreased by 5.27%, accompanied by a reduction in energy dissipation capacity. Although increasing concrete strength enhanced horizontal bearing capacity, its overall influence on the seismic performance of steel tube confined concrete columns remained relatively limited. The calculated values obtained using existing formulas showed relative errors of no more than 9% when compared to experimental results under normal temperature conditions, satisfying engineering accuracy requirements. However, under high and low temperature conditions, the maximum relative error between calculated and experimental values reached 21.73%, demonstrating the necessity of considering ambient temperature effects and modifying the existing horizontal bearing capacity calculation formulas accordingly.ConclusionsAmbient temperature has a significant influence on the seismic performance of steel tube confined concrete columns. Ambient temperature and concrete strength exhibit different degrees of influence on the initial stiffness, horizontal bearing capacity, ductility, and energy dissipation capacity of the specimens. These effects are mainly attributed to the additional hoop constraint induced by ambient temperature variations and the corresponding changes in material properties. The ductility of steel tube confined concrete columns under different ambient temperatures is lower than that observed at room temperature; however, the underlying mechanisms of degradation differ. In low-temperature environments, the increased brittleness of the core concrete and the accelerated accumulation of damage lead to a significant reduction in ductility. In high-temperature environments, the reduction in the viscous performance of the steel concrete interface weakens the confinement effect of the outer steel tube on the core concrete, resulting in decreased ductility. A modified formula for calculating the horizontal bearing capacity of steel tube confined concrete columns that considers the influence of ambient temperature is proposed. The relative error between the calculated and experimental values is less than 10%, indicating good agreement and providing a theoretical basis for the design of this type of structure in alpine regions. The design of steel tube confined concrete structures in alpine regions should comprehensively consider the effects of temperature-induced additional confinement and temperature-induced material property changes on the mechanical behavior of the structure.关键词：steel tube confined concrete column;ambient temperature;quasi-static test;seismic performance;horizontal bearing capacity;temperature-induced additional constraint effects224|365|0更新时间：2026-03-26
-
Experimental and Theoretical Research of Built-in Steel Tube Reinforced Concrete Beams with Openings 增强出版 AI导读
摘要：ObjectiveTo reduce storey height or increase clear height, openings are usually made in the beam web to improve space utilisation and reduce construction costs. However, this reduces the stiffness and load-bearing capacity of the beam, so it is necessary to provide additional reinforcements. The design theory and construction method of conventional inclined reinforcement for openings is relatively complex, particularly when there are a large number of openings. Therefore, a new method of reinforcing openings and calculating their load-bearing capacity is necessary for application in engineering.MethodsTo simplify construction and optimise the force transfer mechanism, a new built-in steel tube was proposed as an alternative to the conventional diagonal reinforcement bars of the beam with openings. A mechanical test was conducted on the beam with openings to investigate the influence of the reinforcement method on cracking development, failure modes, material strains, and load-deflection curves of the beams. The tested beams were one solid web beam (L1), and two beams with openings (L2, which adopted the traditional inclined reinforcement method, and L3, which adopted the built-in steel tube reinforcement method). Additionally, numerical models of the beams were established using the ABAQUS to analyse the internal force development. Finally, an additional compressive stress field (strut action) was introduced to create a strut-and-tie model and analyse the shear capacity of the beam with openings.Results and DiscussionsThe results showed that the openings and reinforcement methods had a significant impact on crack distribution, the location of the maximum crack and crack width. For the beam without openings, the maximum crack was concentrated in the mid-span region and the crack widths in the two side regions were relatively small; bending damage was the main failure mode. For the beam with openings, the maximum crack width occurred near the beam's opening when using either traditional inclined reinforcement or built-in steel tube reinforcement. The crack width in the mid-span region was relatively small and tended to result in shear damage. The ultimate load-bearing capacity of the beam with built-in steel tube reinforcement was similar to that of the beam with traditional inclined reinforcement. Using the built-in steel tube reinforcement improved stress concentration and displacement. Steel strain analysis showed that the steel tube was under compression in the area near the opening. Concrete strain distribution and development was smoother in L3 than in L2. The reinforcement method significantly affected the local load-bearing mechanism of the opening. Unlike with the traditional inclined reinforcement method, an additional compression action was generated in the opening area due to the restraining effect of the built-in steel tube. The displacement trend and the location of the maximum displacement of the tested beams differ slightly due to the openings. Comparisons showed that using built-in steel tube reinforcement effectively improved stress concentration and displacement development. Compared to the bending moment distribution of L1, the bending moment in the opening area experienced local increases or decreases due to the effect of sudden shear changes, particularly near the supports. The reinforcement method of the openings had a significant impact on the local moment distribution. For L2, the bending moment trends of the concrete and steel bars were opposite, particularly at mid-span. For L3, the bending moment trends of the concrete and steel bars were more similar, with the steel tube carrying part of the bending moments. The comparison showed that the opening region in L2 was under tension and the opening region in L3 was under compression, and that the built-in steel tube reinforcement changed the force mechanism in these regions. This behaviour strengthened the compressive properties of the concrete and improved the mechanical properties of the beams with openings. For L2 and L3, the maximum tensile (compressive) strains were concentrated on either side of the opening. Compared to L2, the strain distribution in L3 was more uniform without any sharp, sudden changes. The principal stress trajectory in the shear zone of the beams with steel tube reinforcement was extracted. The built-in steel tube-reinforced beam at the beam end opening had an additional compression transfer path. This path was approximated by connecting the corner near the top of the beam end opening and the closest stirrup at the beam end with a straight line, forming an additional strut-tie model (STM). The new strength method led to the formation of the concrete compressive stress field near to the opening, and the additional compressive action shortened the force transfer path. To establish the calculation method of STM for beams with openings, taking into account the additional compression action. The modelling process of STM included firstly determining the tension ties, and then determining the position of compression struts and tension ties in three paths. For the beams with openings, there were two failure modes, i.e. bending failure and shear failure. This paper focused on the shear failure mode, assuming the destruction of tension tie or compression strut. The proposed strut-and-tie model can be used to predict the shear capacity of the beam strengthened with the built-in steel tube.ConclusionsCompared to conventional inclined reinforcement methods, the built-in steel tube reinforcement offers greater construction convenience. Furthermore, this method increases the beams' shear resistance and reduces crack width at the bottom of the beam. The built-in steel tube reinforcement method creates an additional compressive effect around the opening region, thereby shortening the load transfer path. The strut-tie model can be used to analyse the shear capacity of concrete beams with openings reinforced by built-in steel tube.关键词：beams with openings;strengthen method;built-in steel tube;ultimate load-bearing capacity;crack;principal stress vectors312|787|0更新时间：2026-03-26 - 摘要：ObjectiveGeosynthetic clay liner (GCL), as a typical anti-seepage material, is extensively employed in hydraulic engineering. In river regulation and water storage projects, laying the GCL on the riverbed cushion significantly enhances the anti-seepage performance of river courses. For rivers deficient in fine-grained materials, adopting pebble soil as the upper and lower cushion layers of the GCL serves as an effective measure to reduce project investment. However, during the construction compaction process, pebbles induce jacking, puncturing, and extrusion damage to the GCL, which can even lead to the loss of anti-seepage functionality in severe cases. In practical engineering, monitoring the motion behavior of pebbles during compaction remains challenging, resulting in the inability to quantitatively evaluate the impact of pebbles on the GCL. Therefore, this study adopts numerical analysis methods to investigate the migration law of pebble soil and its effects on the GCL. Based on the numerical results, a set of safety evaluation indices for the GCL is proposed, providing a theoretical basis for guiding field construction.MethodsFirstly, undisturbed soil samples collected from the construction site were remolded in the laboratory. Particle size analysis and compaction tests were conducted to determine the particle gradation and compaction curve of the remolded pebble soil. Triaxial compression tests were further performed to obtain the macroscopic mechanical parameters, such as peak strength and strain, of the pebble soil, which provided an experimental basis for calibrating the meso-scale parameters of the numerical model. Secondly, based on the discrete element method (DEM), the geometric morphology of irregular pebbles, the operational process of vibrating rollers, and the structural characteristics of the GCL were accurately simulated. Using the laboratory test results and the maximum elongation rate index of the GCL specified in relevant standards, the numerical model was employed to replicate the laboratory testing procedures to calibrate the meso-scale parameters of the pebble soil and the GCL and to verify their effectiveness. Then, an integrated DEM model consisting of a vibrating roller, a pebble soil protective layer, a GCL, and a lime-mixed pebble soil support layer was established to simulate the dynamic construction compaction process. The effects of key factors, including pebble distribution position, morphology, pebble content, and compaction thickness, on the compaction characteristics and the overall particle distribution pattern of the backfill soil layer were systematically investigated. Based on the spatial distribution characteristics of pebbles, such as motion frequency, trajectory, and morphology, the migration mechanism of pebbles under vibrating compaction loads was revealed. Finally, the effects of different compaction passes, pebble content, and soil layer thickness on the extrusion, jacking, and embedding behaviors of the GCL were analyzed. The deformation behavior of the GCL during construction compaction was clarified, and deformation-based evaluation indices were proposed to assess the safety performance of the GCL throughout the compaction process.Results and DiscussionsAfter the soil was spread, leveled, and stabilized under its self-weight, the initial porosity of the backfill soil was 13.5%. With an increase in the number of compaction passes, the porosity of the backfill soil exhibited a pattern of rapid reduction, slight fluctuation, and eventual stabilization. At the completion of 15 compaction passes, the porosity decreased to 12.6%, while the minimum porosity of 12.58% was recorded at the 9th pass. Based on the observed porosity variation pattern, the porosity essentially reached a stable condition after 7 compaction passes. Although additional compaction can further reduce porosity, a rebound phenomenon was observed after the 9th pass. The cumulative settlement of the backfill soil demonstrated a trend of gradual increase, minor fluctuation, and final stabilization. Regarding the growth rate of cumulative settlement, negative growth occurred after the 7th compaction pass, followed by a slight increase and subsequent stabilization with minor fluctuations. A combined evaluation of porosity and cumulative settlement indicates the existence of an optimal number of compaction passes during construction. Under the extrusion action of vibrating rollers, irregular pebbles experience rotational and translational movements during compaction. The interaction between pebbles and GCL is strongly influenced by pebble morphology, and different contact modes result in extrusion, jacking, and embedding effects on the GCL. Two primary contact modes occur between triangular-like pebbles and GCL: 1) point contact between pebble corners and the GCL, which generates significant jacking forces and presents a high risk of GCL puncture; and 2) surface contact, which provides a larger contact area and lower puncture risk, but can induce overall pebble embedding into the GCL during compaction, leading to lateral migration of adjacent bentonite particles and attenuation of anti-seepage performance at contact locations. Elliptical-like pebbles exhibit two contact states with the GCL, including contact through the long axis, which corresponds to surface contact behavior similar to that of triangular-like pebbles, and contact through the short axis, which resembles point contact behavior. Circular-like pebbles predominantly form surface contact with the GCL, resulting in relatively uniform extrusion. When elliptical-like and triangular-like pebbles are distributed on both the upper and lower sides of the GCL, point contact occurs, causing staggered extrusion effects. Point contact tends to raise jacking and embedding of the GCL by pebbles, leading to a high puncture risk, particularly when pebbles are present on both sides of the GCL. Three safety evaluation indices for GCL were proposed, namely the local elongation rate, pebble embedding depth, and bentonite coordination number. The local elongation rate of the GCL increases with the number of compaction passes, and the maximum local elongation rate observed at representative positions during the entire compaction process was 3.79%, which is significantly lower than the standard limit of 14.00%. The pebble embedding depth is closely related to pebble motion frequency, with a maximum recorded value of 2.1 mm. The distribution characteristics of the bentonite coordination number are influenced by both the contact mode between pebbles and the GCL and the number of compaction passes. Under point contact conditions, bentonite particles exhibit high motion frequency and low coordination numbers, indicating a reduction in the local thickness of the GCL, whereas under surface contact conditions, bentonite particles display lower motion frequency and a more uniform distribution. The coordination number remained above the theoretical limit during the first 7 compaction passes; however, localized coordination numbers fell below the theoretical limit as the number of compaction passes increased.ConclusionsBased on a comprehensive evaluation of the three indices, effective control of compaction passes during construction is identified as a critical measure for ensuring the structural integrity and anti-seepage performance of the GCL.关键词：Lime-mixed gravel soil;Vibration rolling discrete element model;Geosynthetic clay liner elongation;Gravel embedding value;bentonite coordination number206|532|0更新时间：2026-03-26
- 摘要：ObjectiveCurrently, most studies on decoupled charges that use water as a coupling medium focus on rock damage and crack propagation. The impact of shock wave propagation and the water medium on the blasting effect is generally reflected through the damage observed in rock specimens. This study investigates the propagation of shock waves and blast products when a water-decoupled charge structure detonates in an air field. In addition, it analyzes the pressure and equivalent stress around the borehole to examine the blasting behavior of axial water-coupled charge structures in an air field.MethodsFirst, various axial water decoupling charge structures were designed, and the propagation images of shock waves and blast products, as well as the shock wave pressures at fixed positions, were obtained when the charge structures were detonated in an air field by using a high-speed schlieren system and an overpressure testing system. The propagation speed time curves of the shock waves and blast products in the schlieren field were obtained by marking the shock wave front and the blast product boundaries in the images. The study summarized the influence patterns of the water coupling position on the blasting effect by analyzing the charge structures under different conditions. Numerical simulations were conducted to obtain the shock wave pressure values along the vertical and horizontal directions and at various positions along the borehole axis. Computational models for different charge structures were established using the ANSYS finite element numerical simulation software. The detonation process was simulated numerically, and the simulation results were integrated with the schlieren experimental results to determine the evolution of the propagation speed and pressure of the explosive shock waves in the air flow field and the pressure distribution inside the charge.Results and DiscussionsThe results obtained from the schlieren system and the overpressure testing system for different water coupled charge structures in field experiments were as follows: For Structures 1#～3#, the maximum shock wave speeds in the horizontal direction were 419.5, 411.1, and 438.6 m/s, respectively. Structure 4# generated two shock waves with maximum propagation speeds of 403.5 and 398.3 m/s. The decay trends of the shock wave speeds in the vertical direction for Structures 1#～3# were similar. The shock wave speed in the vertical direction rapidly decayed from about 600 m/s to about 480 m/s within 60 μs after initiation, before separating from the blast products. After separation, the shock wave speed decayed more slowly and gradually decreased to about 430 m/s before stabilizing. In contrast, the maximum speeds of the two shock waves in Structure 4# were lower than those recorded for the other three structures. The overpressure peaks in the horizontal direction for Structures 1#～4# were 567.80, 464.76, 709.12, and 387.73 Pa, respectively. Structure 2# displayed the lowest overpressure peak among Structures 1#～3#, while the overpressure peak of Structure 3# was 34.46% higher than that of Structure 4#. The overpressure peaks in the vertical direction for Structures 1#～4# were 592.8, 655.8, 606.9, and 520.5 Pa, respectively. In the numerical simulations, pressure values for different water coupled charge structures were obtained from measurement points arranged in the vertical direction, horizontal direction, and within the borehole wall. For Structure 1#, the peak pressure at measurement point 2 was 207.62 kPa, and at measurement point 5 it was 137.97 kPa, representing a decrease of 33.55%. For Structure 2#, the peak pressure at measurement point 2 was 300.40 kPa, and at measurement point 5 it was 142.80 kPa, representing a decrease of 52.46%. For Structure 3#, the pressure decrease was 32.84%, and for Structure 4#, it was 50.11%. Along the vertical borehole direction, Structures 1#～3# all reached pressure peaks at corresponding measurement points at approximately 0.033, 0.057, 0.087, 0.123, and 0.170 ms. The borehole wall pressure peak curve for Structure 1# displayed a bell-shaped distribution, while Structures 2# and 3# exhibited an M-shaped distribution, and Structure 4# showed a valley-shaped distribution. The analysis indicated that the water medium affected the shock wave pressure decay process. On the side with water coupling, the shock wave pressure decayed more slowly, which potentially increased the range of the explosion impact.ConclusionsThe explosion energy of the axial water non-coupled structure is more concentrated along the vertical axis of the borehole compared to the borehole axis. After the detonation of the explosive, the shock wave propagates in a flattened spherical form into the air flow, with the blast products moving outward with the shock wave and gradually separating from it, which results in greater displacement in the vertical direction. The shock wave pressure generated by the explosion in the axial water non-coupled charge structure is influenced by the medium along its propagation path. The side with water coupling exhibits lower peak shock wave pressure, but the attenuation rate decreases more slowly over the same distance, and the shock wave duration remains longer, extending the blast effect range on that side. The position of the charge segment in the axial water non-coupled structure affects the blasting performance. The borehole wall pressure in the charge segment is substantially higher than that in the water medium segment, and the peak pressure decreases with increasing distance from the charge segment. Therefore, the damage primarily occurs within the charge segment. The damage range can be effectively controlled by modifying the number, length, and position of the explosives in the charge structure.关键词：high-speed schlieren;axial charging;water decoupling;pressure of bore wall;ANSYS AUTODYN333|657|0更新时间：2026-03-26
- 摘要：ObjectiveThis study accurately assesses the seismic safety of near-fault subway station structures by developing an advanced methodology and procedure. Currently, this assessment faces two major challenges, namely the scarcity of near-fault seismic records and the lack of practical modeling methods. This study proposes a deterministic physics-based simulation approach to construct strong ground motion fields that couple source, path, and site effects, and then uses the simulated motions as inputs to assess the seismic performance of critical station structures located at near-fault sites.MethodsFirstly, this study introduced a comprehensive fault-to-station response simulation approach that comprised two main components: the FK-SE-FE hybrid method for simulating seismic wave propagation and the GP14.3 source model for synthesizing broadband ground motions. Specifically, the first component employed the FK-SEM hybrid method to simulate wave propagation from the source to the site, while the SEM-FEM hybrid method was utilized to simulate wave propagation from the site to the structure. The second component integrated high-frequency stochastic perturbations into the low-wavenumber deterministic asperity source model to ensure the bandwidth effectiveness of the synthesized ground motion, constructing the GP14.3 kinematic hybrid source model. This integrated simulation approach enabled a comprehensive analysis from fault rupture to structural response and effectively considered several key physical processes within the seismic scenario chain, including: 1) fault rupture mechanics; 2) regional crustal wave propagation; 3) local site effects; and 4) soil-structure interaction. Secondly, the underground stations of the Tianjin Metro were taken as an example, and six important stations were selected as research subjects based on the existing rail transit network. A refined finite element model of the subway station structure was established while explicitly considering soil-structure interaction. In addition, against the backdrop of the Tianjin hidden fault, and based on active fault detection, seismic hazard assessment, and geotechnical exploration drilling data for the target area, a kinematic hybrid source model of the target fault and the regional underground velocity structure was constructed. Finally, the synthesized broadband seismic waves were employed as input motions to investigate the dynamic responses and damage risks of various subway station structures under the prescribed earthquake scenarios.Results and DiscussionsThe synthesized ground motions revealed substantial variations in acceleration waveforms and response spectra across the selected stations. Peak ground acceleration (PGA) values at specific locations reached up to 2.32 m/s2, which significantly exceeded the standard design basis seismic acceleration value of 0.15g for general construction projects in Tianjin. This finding highlighted the necessity for site-specific seismic assessments. Horizontal seismic motions were generally more pronounced than vertical motions, with horizontal PGAs reaching values up to twice those of the vertical components in some cases, indicating the importance of incorporating multi-directional seismic inputs in structural dynamic analyses. The structural deformation patterns exhibited considerable variability among the stations. Tianjin South station experienced the most severe deformation and the highest damage risk due to its proximity to the fault trace, with maximum interstory drift ratios approaching the plastic deformation threshold, which indicated a high potential for significant structural damage. In contrast, stations such as Yingkou Road, Tianjin West, and Tianjin stations exhibited moderate deformation levels, whereas Haihe Education Park and Binhai Airport stations experienced minimal deformation, indicating relatively lower risk profiles. Lateral wall relative displacements demonstrated nonlinear variations along the structural height. Displacements in the y-direction were larger than those in the x-direction, reflecting the complex interactions between multi-directional seismic inputs and the soil-structure system. The severity of potential structural damage varied among the stations, with Tianjin South station exhibiting the highest displacement values, followed by moderate risk levels at Yingkou Road, Tianjin West, and Tianjin stations, and minimal risk levels at Haihe Education Park and Binhai Airport stations. These results demonstrated the effectiveness of the proposed FK‒SE‒FE hybrid method combined with the kinematic hybrid source model in capturing the spatial variability of seismic motions and the complex interactions within the soil-structure system. The method provided a precise and practical tool for the seismic risk assessment of near-fault subway stations, emphasized varying vulnerability levels, and highlighted the necessity for site-specific seismic design considerations.ConclusionsThis study presents an advanced methodology and systematic procedure for evaluating the seismic response of near-fault subway stations by employing the FK‒SE‒FE hybrid method in conjunction with the GP14.3 kinematic hybrid source model. The results demonstrate the effectiveness of this integrated approach in accurately capturing complex interactions and the spatial variability of seismic motions. The analysis indicates that stations located in proximity to the fault line, such as Tianjin South, are subjected to significantly higher risks of severe deformation and structural damage, highlighting the necessity for enhanced seismic design measures in these locations. In contrast, stations situated farther from the fault, including Haihe Education Park and Binhai Airport, exhibit comparatively lower seismic risk levels. The proposed procedure provides meaningful insights for engineers and researchers seeking to enhance the resilience and safety of subway infrastructure in earthquake-prone regions. Overall, the results highlight the critical importance of site-specific seismic design considerations and the implementation of advanced simulation techniques for mitigating seismic hazards.关键词：Scenario earthquake;seismic wave propagation;hybrid source model;subway station;full-process response319|824|0更新时间：2026-03-26
- 摘要：ObjectiveThis study addresses the deterioration behavior and service life of ancient architectural blue bricks in saline environments in central and western Inner Mongolia. It simulates damage development in ancient wall bricks subjected to soluble salt corrosion, analyzes corrosion mechanisms and remaining service life prediction for blue bricks under saline conditions, and provides theoretical guidance for preserving ancient blue bricks and predicting their remaining service life in saline environments of central and western Inner Mongolia.MethodsBased on the characteristics of saline-alkali soil environments in central and western Inner Mongolia, the primary types and concentrations of soluble salts in the soil were identified, and three groups of solutions were prepared at 10 to 20 times the soluble salt concentration, including a 5% Na₂SO₄ solution, a 5% NaCl solution, and a composite salt solution consisting of 2.5% Na₂SO₄ + 2.5% NaCl. A distilled water solution was used as the control group for comparative analysis. Blue bricks were subjected to accelerated corrosion tests for 5, 10, 15, 20, 25, 30, 35, and 40 cycles under alternating wet and dry conditions. Each corrosion cycle consisted of 12 h of immersion followed by 12 h of drying at 60 °C. Macroscopic changes in brick mass, compressive strength, and relative dynamic modulus were measured. Scanning electron microscopy (SEM) was utilized to examine changes in internal micropore structures before and after testing, while X-ray diffraction (XRD) was employed to analyze phase transformations. The service life of the bricks was predicted using the Wiener process theoretical model. Considering the service life requirements of ancient structures and the initial compressive strength of the bricks, the failure time threshold was determined. The remaining service life of ancient bricks in the saline environments of central and western Inner Mongolia was predicted by quantifying the relationship between damage rate and damage severity under saline conditions.Results and DiscussionsAs the number of salt wet and dry corrosion cycles applied to the blue bricks gradually increased, changes in visual morphology indicated that the bricks exhibited no significant damage after wet and dry cycling in the distilled water solution. Under Na₂SO₄ solution cycles, damage and material loss at the edges and corners of the specimens progressively intensified, accompanied by crack initiation and propagation. After NaCl solution cycles, the edges and corners of the specimens became increasingly loose and rough, and small cracks gradually developed into penetrating cracks. Under composite salt solution wet and dry cycling, the edges and corners of the blue bricks gradually became loose and rough, with damage and material loss continuously worsening. Overall, as the number of accelerated salt wet and dry corrosion cycles increased, the degree of damage to the blue bricks consistently intensified. Analysis of macroscopic performance indicated that the mass, compressive strength, and relative dynamic modulus of the blue bricks exhibited differentiated degradation trends. The severity of degradation followed the order of 5% Na₂SO₄ solution > 2.5% Na₂SO₄ + 2.5% NaCl composite salt solution > 5% NaCl solution. Microscopic morphology analysis using SEM and XRD demonstrated that corrosion degradation of the blue bricks resulted primarily from salt crystallization pressure within internal pores, mineral dissolution and hydrolysis, mudding, and interionic replacement reactions. Under salt wet and dry cycling, both the macroscopic appearance and microscopic structure of the blue bricks exhibited varying degrees of damage, and deterioration of the microscopic structure led to a continuous decline in macroscopic mechanical properties.ConclusionsUnder accelerated salt wet‒dry cycle corrosion testing, the predicted service life of green bricks based on mass loss was calculated as 1 086, 1 802, and 1 111 hours in 5% Na₂SO₄ solution, 5% NaCl solution, and 2.5% Na₂SO₄ + 2.5% NaCl composite salt solution, respectively. The predicted service life based on compressive strength was calculated as 1 109, 1 490, and 1 196 hours, respectively. The average predicted service lives were 1 097, 1 646, and 1 153 hours. Failure thresholds were reached after 30, 45, and 35 wet‒dry cycles, respectively, corresponding to 1.64, 2.01, and 1.84 years under actual saline environmental conditions. The remaining service life of blue bricks in Na₂SO₄, NaCl, and composite salt solutions in central western Inner Mongolia was determined to be 120, 139, and 124 years, respectively by combining the service life of ancient buildings in Inner Mongolia with the compressive strength of original green bricks, and through quantitative analysis of the relationship between salt wet‒dry cycle damage rates and green brick damage severity in saline environments. Under the regional saline conditions, the remaining service life of blue bricks in ancient structures is at least 120 years. The predicted remaining service life of blue bricks is consistent with, or even exceeds, that of the ancient structures in which they are used. Through evaluation of the remaining service life of ancient buildings, the predicted lifespan of their blue bricks falls within a reliable range. These experimental findings provide guiding principles for evaluating the durability of ancient architectural blue bricks and for their conservation and restoration.关键词：salt‒dry‒wet cycle;Ancient building green brick;salt crystallization pressure;microstructure;phase transition;residual life prediction505|2056|0更新时间：2026-03-26
- 摘要：ObjectiveA spur dike is an in-stream structure that is widely employed for the protection of riverbank erosion. In river engineering, traditional regulating structures exhibit certain limitations, such as occupying large areas, requiring significant quantities of stone, being difficult to maintain, and involving high construction costs, which make them unsuitable for use in some regions. Therefore, scholars propose the application of floating dike structures in river regulation engineering. These structures exhibit advantages such as minimal impact on sediment transport, improved adaptability to upstream water and sediment discharge, and reduced impact on the riverbed and ecological environment. However, research on the effects of floating dikes on flow structure and the riverbed remains insufficient, and the understanding of their practical application is limited. Considering practical application, ships are used as substitutes for floating dikes; therefore, this structure is also referred to as a ship-shaped floating dike. This study focuses on the effects of the submergence ratio (D is submergence depth/water depth), length ratio (K is ship length/river width), arrangement angle (θ > 0 indicates the dike oriented downstream), and transverse bed slope at the location of the ship-shaped floating dike on the surrounding flow field and turbulence characteristics. The findings provide a hydrodynamic basis for further understanding the influence of ship-shaped floating dikes on sediment transport and riverbed evolution, as well as technical support and reference for their application in practical river engineering projects.MethodsThis study investigated the influence of geometric characteristics, including submergence depth and ship length, as well as arrangement parameters, including ship angle and transverse bed slope, of ship-shaped floating dikes on the surrounding flow field and turbulence structure, with the objective of identifying the optimal arrangement configuration. The same flume dimensions (length is 1.4 m, width is 0.3 m, and height is 0.05 m) and simulation conditions as those used in the physical model experiments were adopted to enhance the reliability of the numerical model. The RNG k‒ε turbulence model was employed to simulate the flow structure around the ship-shaped floating dike. Considering the negligible variation in the water surface, the rigid-lid assumption was applied, with a constant inlet velocity and free outflow conditions at the outlet. The velocity-pressure coupling equations were solved using the semi-implicit SIMPLE algorithm.Results and DiscussionsThe numerical simulation results for velocity distribution and Reynolds shear stress across different longitudinal sections were consistent with those obtained from physical experiments, confirming the reliability of the numerical model. The three-dimensional flow structure and turbulence kinetic energy distribution were analyzed by considering variations in the submergence ratio (D), length ratio (K), arrangement angle (θ), and transverse bed slope (SL) of the ship-shaped floating dike. The numerical results were verified against experimental observations. The results demonstrated that: 1) As D increased from 0.2 to 0.8, the maximum relative streamwise velocity below the ship-shaped floating-dike (SSFD), defined as the ratio of local maximum flow velocity/inlet average flow velocity (U/U0max), gradually increased, from 1.52 to 1.69. The length of the back-flow area and U/U0max increased from 1.0L to 3.5L (L is the length of SSFD) and from ‒0.35 to ‒0.51. Increasing D also intensified near-bed flow velocity and enhanced turbulence intensity downstream of the SSFD. 2) As K increased, the effective flow-passed cross-section decreased, resulting in increased flow velocity below and outside the SSFD. The area of the back-flow and low-velocity zones expanded with increasing K, whereas the length and intensity of the back-flow region remained nearly unchanged. Turbulence kinetic energy gradually increased and formed a concentrated core region near the bottom and downstream of the SSFD. 3) When the SSFD was tilted upward with θ = ‒30º, a strong blocking effect was observed, with U/U0max upstream of the SSFD reduced to 0.4, while near the shore and bottom of the SSFD, U/U0max = 1.53, which was less than θ = 30º with U/U0max = 1.63, but both were less than θ=0 with U/U0max = 1.72. The distribution of the slow-flow zone behind the SSFD was larger when θ = ‒30º. When θ = 30º, the flow remained relatively stable, characterized by weak turbulence kinetic energy, and localized acceleration zones and circulation structures developed downstream of the SSFD. When the SSFD was positioned on the deeper side of the channel, the flow concentrated beneath the structure, resulting in an intensified back-flow region. The transverse slope of the riverbed enhanced downstream circulation. The results indicated that the SSFD achieved optimal slow-flow and sedimentation promotion effects when it was arranged on the deeper side, tilted downward, with a submergence ratio
- 摘要：ObjectiveSoft clay is widely distributed in coastal areas of China. Due to poor engineering properties such as high water content, high porosity, high compressibility, low permeability, and low strength, infrastructure constructed in these areas faces significant challenges. Cement-based stabilization of soft clay is a commonly adopted approach for reinforcing soft foundations. However, existing research primarily focuses on the behavior of stabilized clay at a single scale, such as the micro-, meso-, or macro-scale, while cross-scale correlations remain insufficiently understood. In response, this study proposes a multi-scale linkage framework to bridge the mechanical properties of stabilized clay across different scales.MethodsTwo artificial clays with distinct clay mineral compositions, including kaolinite-rich kaolin and montmorillonite-rich bentonite, as well as the primary strength-producing constituents of cement clinker, C3S and C3A, were selected to prepare stabilized samples. In addition to conventional macroscopic unconfined compressive strength tests, nanoindentation and small-strain bending element testing techniques were applied to characterize the micro- and meso-mechanical properties of the stabilized clays. Then, big data deconvolution statistical analysis and a homogenized self-consistent model were employed to quantitatively establish cross-scale correlations within the stabilized clays.Results and DiscussionsA large volume of indentation hardness and elastic modulus data was obtained from the nanoindentation tests. The elastic modulus distribution cloud maps of the stabilized samples visually reflected the micro- and nano-scale mechanical characteristics resulting from interactions between hydration products and clay minerals. Statistical analysis indicated that the average elastic moduli of KS, KA, BS, and BA samples were 23.2, 13.4, 18.6, and 16.4 GPa, respectively, demonstrating that cement clinker and clay mineral composition significantly influenced the micro- and nano-mechanical properties of stabilized soils. During the hydration process, the alkaline environment progressively eroded clay particles and their aggregates, raising the dissolution of active SiO2 and Al2O3 colloids within the clay minerals. These precipitated amorphous oxide colloids reacted with Ca(OH)2 generated during hydration to form additional cementitious products such as C‒S‒H and C‒A‒H. As a result, volcanic ash reactions continued over prolonged curing and service periods, influencing material strength and stiffness. Based on PDF and CDF deconvolution results, the volume fractions of distinct material phases were quantitatively determined, and the volume contents obtained from both methods were generally consistent. The primary hydration product in C3S-stabilized clay was C‒S‒H, and the average C‒S‒H content in BS was 36.8%, which was slightly higher than that in KS at 34.7%. Similarly, C‒A‒H was the dominant hydration product in C3A-stabilized clay, and the C‒A‒H content in BA reached 29.1%, exceeding that in KA at 21.4%. These results indicated differences in reaction mechanisms and reaction intensities among various clay minerals and cement clinker compositions. The volcanic ash reaction intensity of bentonite was significantly greater than that of kaolin, leading to increased formation of secondary C‒S‒H and C‒A‒H. The elastic modulus E of each sample was obtained using the small-strain bending element test, and the meso-scale elastic moduli of KS, KA, BS, and BA were 1 416.76, 911.26, 1 480.26, and 995.81 MPa, respectively. Following the bending element tests, the samples were subjected to unconfined compressive strength testing, which indicated macro-scale elastic moduli of 256.72, 102.21, 268.68, and 132.79 MPa for KS, KA, BS, and BA, respectively. Relationships among micro, nano, meso, and macro elastic moduli were established by integrating elastic modulus results obtained from nanoindentation, cross-scale modeling, small-strain bending element testing, and unconfined compressive strength testing. The micro-scale elastic moduli of KS, KA, BS, and BA were determined to be 10.3 to 14.4 times and 56.5 to 122.7 times greater than their corresponding meso- and macro-scale elastic moduli, respectively.ConclusionsBoth cement clinker and clay minerals have significant influences on the micro-, meso-, and macro-scale mechanical properties of stabilized clay. A multi-scale linkage framework was developed by investigating the mechanical properties of stabilized clay at micro, meso, and macro scales, which deepens the understanding of the reaction mechanisms between clay minerals and hydration products in stabilized clay and provides guidelines for predicting the macroscopic mechanical properties of stabilized clay.关键词：hydration products;clay minerals;multiscale mechanical characterization;nano indentation;big data statistical analysis385|660|0更新时间：2026-03-26
-
Spatial Matrix Displacement Method of Reciprocal Structures Based on Condensed Nodal Coding System AI导读
摘要：ObjectiveThe rod articulation in the spatial matrix displacement method is simulated by directly eliminating the rows and columns associated with the hinge point in the stiffness matrix, which is only applicable when two rods are directly articulated on the axial line. For reciprocal structures, the hinge points are located at the rod edges, and the presence of eccentricity allows the bending moments to be transmitted between rods, so the direct modification of stiffness matrices is not applicable to this type of reciprocal system. This study proposes a static calculation method applicable to such systems to accurately and efficiently simulate the reciprocal hinge point and more conveniently calculate the internal forces and deformations of reciprocal structures.MethodsFirst, the generalized nodal displacement consisting of nine components is defined at the hinge point, including the three translations at the hinge point and the three rotations at the two beam end nodes. The displacement relationship of the beam end nodes at the reciprocal connection is simplified by introducing the generalized displacements at the hinge point to describe the displacements of the nodes at the beam ends. Based on the traditional spatial matrix displacement method, the elements and nodes of the spatial frame structure require coding. Because the nodal displacements at the beam end nodes are described by the generalized displacements at the hinge point, the entire structural nodal coding is condensed by replacing the two codes at the beam end nodes with one code at the hinge point. For a structure with N beam end nodes and Nm hinge points, the number of nodal codes decreases from N to N‒Nm after condensation, and the dimension of the nodal displacement vector is also reduced. Due to the adoption of two sets of nodal coding, integrating the whole stiffness matrix by using the traditional positional localization approach becomes difficult. Therefore, the equation of internal virtual work done by the internal force of the beam element on the virtual displacement of the beam end nodes and the equation of external virtual work done by the equivalent node load on the virtual displacement of the beam end nodes are established separately. Based on the principle of virtual work, in which the virtual work of internal force equals the virtual work of external force, the virtual work equation of the structure is obtained as the whole stiffness equation of the structure under the condensed nodal coding system. At the same time, the whole stiffness matrix of the structure under the condensed coding system is obtained and, similar to the traditional matrix displacement method, remains symmetric. After introducing boundary conditions, the whole stiffness matrix is modified, and the nodal displacements of the structure under the condensed nodal coding system are obtained. In addition, the displacements and internal forces of the structure under the original coding system are calculated.Results and DiscussionsIn the numerical examples, the proposed method is first applied to calculate and analyze the nodal displacements and internal forces of the space lap structure with two rods, and ANSYS software is employed to perform finite element analysis for verification. In the finite element model, Beam 189 and Solid 45 are utilized to simulate the rods, and the results show that regardless of the element type, the finite element results agree well with the theoretical results. Further, the reciprocal structure transformed from the 63 type of Archimedes paving is determined by the proposed method to obtain the nodal displacements and internal forces, and comparative validation is again conducted by using ANSYS finite element software. The results show that the error between the proposed method and ANSYS finite element results remains within 3%. The two sets of results are generally consistent but do not perfectly match because short stiff arms are introduced in the ANSYS model to simulate the articulation of the beam end nodes. The stiff arm is not absolutely rigid, although it has much greater stiffness than the rods, so the ANSYS finite element model does not fully satisfy the flat cross-section assumption.ConclusionsCompared to finite element analysis, the spatial rigid matrix displacement method under the condensed nodal coding system presents clear advantages for analyzing reciprocal structures. The finite element method requires additional short stiff arms to simulate the reciprocal connection, which generates 15 nodal degrees of freedom. In contrast, the proposed method uses only 9 degrees of freedom for the generalized nodal displacements at the actual hinge points, effectively reducing the computational degrees of freedom of the whole structure and providing a more concise and efficient analysis approach. In addition, in the finite element method, enforcing the flat cross-section assumption requires the stiffness of the two short rigid arms to be extremely large, which can cause convergence difficulties or even non-convergence in nonlinear calculations when inappropriate stiffness values are used. Because the proposed method does not include short rigid arm elements, it exhibits better numerical stability.关键词：Reciprocal structure;Condensed nodal coding;stiffness matrix;Spatial matrix displacement method;finite element method387|479|0更新时间：2026-03-26 - 摘要：ObjectiveThis study accurately calculates the stress and deformation fields of deep-buried pressure pipelines subjected to high internal water pressure, with a specific focus on the influence of fractures in the concrete layer. The mechanisms of load transfer and distribution within the pipeline bearing system are elucidated by considering various influencing factors, including internal water pressure levels, the interlayer gap between structural components, and the geological categories of the surrounding rock.MethodsThe bearing behavior of pressure pipes was simplified as a plane strain problem, and the stress and deformation fields were solved using the power series method of complex function theory. The core of this approach lies in determining the potential functions for the steel lining and surrounding rock layers, as well as the contact stresses between each structural layer. 1) Considering the presence of an interlayer gap, the bearing process was divided into two stages: the free bearing stage and the joint bearing stage. During the free bearing stage, only the steel lining underwent radial expansion under internal water pressure until its outer boundary came into contact with the inner boundary of the concrete layer. At this stage, the corresponding internal water pressure and the form of the steel lining potential function were determined based on the prescribed deformation value at the outer boundary of the steel lining. 2) In the joint bearing stage, when the internal water pressure exceeded a critical threshold, the steel lining and concrete layer achieved complete contact and continued to expand simultaneously. At this point, the concrete layer and surrounding rock acted as an external constraint system and jointly supported a portion of the internal water pressure. 3) Under high internal water pressure, the concrete layer experienced significant circumferential tensile stress, which inevitably led to the formation of radial cracks and the loss of circumferential load-bearing capacity. In this case, the concrete layer transmitted only radial loads and was therefore modeled as an equivalent spring layer with a defined stiffness coefficient. 4) Based on the deformation compatibility conditions among the steel lining, concrete layer, and surrounding rock during the joint bearing stage, the contact stresses and the forms of the potential functions between the structural layers were determined. 5) The final stress distribution and deformation characteristics of the pressure pipe structure were obtained, which enabled the calculation of the load-sharing ratio of the surrounding rock by applying the superposition principle of elasticity theory.Results and DiscussionsThis study investigated the influence of internal water pressure, the gap size between structural layers, and the classification of surrounding rock on the joint load-bearing behavior of pressure pipelines. Comparative analyses were conducted with and without consideration of concrete cracking effects. The key findings were summarized as follows: 1) Internal water pressure induced radial expansion of the pressure pipeline, and its effects, manifested as stress increments and deformations, diminished progressively outward from the inner boundary. When the interlayer gap was small, the external concrete layer and surrounding rock bore a significant portion of the internal water pressure. Therefore, in engineering design, if the surrounding rock is structurally sound and stable and the quality of concrete filling is ensured, the thickness of the steel lining can be appropriately reduced to optimize construction costs. 2) When cracks developed in the concrete layer, leading to the loss of circumferential load‒bearing capacity, the corresponding internal water pressure was redistributed between the steel lining and surrounding rock based on their respective stiffness characteristics. This redistribution resulted in increased radial stress, deformation, and circumferential stress in both components. Due to its high stiffness and direct exposure to internal water pressure, the steel lining assumed a larger share of the load, while the contribution of the surrounding rock decreased. The extent of this redistribution depended on the overall structural parameters and the magnitude of the applied internal water pressure. 3) Given that the deformation of the steel lining under internal water pressure typically occurred at the millimeter scale, even minor interlayer gaps significantly affected load transfer and distribution within the composite system formed by the steel lining, concrete layer, and surrounding rock. For example, based on the parameters adopted in this study, the load-sharing ratio of the surrounding rock decreased from 71.31% under zero-gap conditions (t = 0 mm) to 44.35% when the gap increased to 1 mm. 4) The classification of the surrounding rock had a significant influence on the load-sharing ratio of the surrounding rock. Without considering concrete cracking, the difference in circumferential tensile stress in the steel lining between Class Ⅱ and Class Ⅴ surrounding rock conditions reached 40.93 MPa, with a corresponding difference of 20.56% in the load-sharing ratio of the surrounding rock. When concrete cracking was considered, these differences became more pronounced, as the maximum difference in circumferential stress increased to 118.03 MPa, radial deformation differed by up to 1.6 mm, and the load-sharing ratio of the surrounding rock varied by as much as 59.31%. These results indicated that Class Ⅴ surrounding rock contributed minimally to resistance against internal water pressure under such conditions, implying that the steel lining became the primary load-bearing component. Reinforcement measures such as grouting were required to enhance the contribution of the surrounding rock.ConclusionsA power series solution method based on complex variable functions was proposed to address the calculation problem of load transfer and distribution in deep-buried pressure pipelines under high internal water pressure conditions. The stress state, deformation at any position of the bearing structure, and the surrounding rock load-sharing ratio can be accurately determined by solving the potential functions of the steel liner and surrounding rock, as well as the contact stresses between to address the calculation challenges associated with load transfer and distribution in deep-buried pressure pipelines subjected to high internal water pressure, the bearing components. The proposed method accounted for the reduction in circumferential bearing capacity caused by cracking of the concrete layer by representing it as a spring layer characterized by a specific stiffness coefficient. Based on this modeling approach, the variation patterns of the joint bearing sharing ratio under different internal water pressures, gap sizes, and surrounding rock types were systematically investigated. The proposed analytical calculation method adopts a framework-based approach, allowing for high adaptability. In addition, supplementary conditions can be incorporated based on specific engineering characteristics, enabling targeted and refined analysis.关键词：deep‒buried pressure pipeline;steel lining‒concrete‒surrounding rock;sharing rate of surrounding rock;high internal hydraulic pressure;gap;surrounding rock type202|280|0更新时间：2026-03-26
- 摘要：ObjectiveGeotechnical engineering in cold regions typically involves rock masses containing joints and fractures. During seasonal freeze-thaw cycles, water trapped within these internal structural planes undergoes freezing and thawing, resulting in freeze-thaw damage to the rock mass. This damage leads to the deterioration of mechanical properties and overall structural stability, significantly increasing the risk of engineering accidents. As engineering activities in cold regions continue to expand, the problem of freeze-thaw degradation in fractured rock masses becomes increasingly significant. Therefore, this study investigates the degradation effects and damage mechanisms of saturated fractured rock masses during cyclic freeze-thaw processes at the micromechanical level. The objective is to accurately assess how primary fractures influence the development of secondary fractures during freeze-thaw cycles and to determine whether micromechanical damage alters the loading-fracture characteristics of rock masses. This research provides new insights into the mechanisms of freeze-thaw damage and failure prediction methods for rock slopes in cold regions and advances the interdisciplinary field of rock mechanics and numerical simulation.MethodsFractures within rock masses were connected through rock bridge zones. When external loads caused rock bridge failure, step-like sliding occurred on slopes with discontinuous joints of varying steepness. Therefore, in laboratory experiments, three carefully selected rock specimens with fractures (high dip angle of 75°, low dip angle of 15°, rock bridge length of 15 mm, dip angle of 90°) were subjected to freeze-thaw cycles and uniaxial compression tests after saturation. CT scans were performed on the specimens after they reached predetermined freeze-thaw cycles (0, 15, 30 cycles) or after loading. Using Avizo software, three-dimensional reconstructed images enabled analysis of micro-parameter evolution and quantitative fractal geometric characterization. In numerical simulations, a pore-water-saturated rock model was established in Particle Flow Code (PFC) based on the discrete element method to represent the phase-change expansion of water and ice particles. Parallel-bonding contact models were selected to represent particle cohesion, and the micro-parameters were calibrated using laboratory mechanical test results and trial-and-error methods. The entire freeze-thaw cycle was controlled using Fish language to simulate the coupled expansion of pore water and ice particles during phase transitions. Microcrack evolution diagrams, crack rose diagrams, displacement plots, and force chain diagrams were recorded and processed for different freeze-thaw cycle counts (0, 10, 15, 20, 25, 30 cycles). The extent to which internal structural degradation influenced local strain development during loading deformation and determined failure characteristics remained unclear. Therefore, subsequent uniaxial compression simulations were conducted on freeze-thaw-cracked saturated sandstone. These simulations employed deformation-controlled loading at a rate of 0.05 mm/min until specimen failure occurred. Microcrack evolution diagrams, displacement plots, and force chain diagrams were recorded and processed for different loading stages.Results and DiscussionsIn indoor experiments, the processed CT images underwent three-dimensional reconstruction to evaluate the effects of freeze-thaw damage on distinct fracture zones independently. The reconstructed 3D model was segmented into Fracture Zone I, the rock bridge region, and Fracture Zone Ⅱ. As the number of freeze-thaw cycles increased, both the increment in Euler number and the increment in fracture volume of the specimens gradually increased, indicating an accelerated progression of freeze-thaw damage during the later stages. Second, the study calculated the surface porosity of each layer based on the CT scan images. The increment in surface porosity during the later freeze-thaw stage exceeded that observed in the earlier stage, which further validated the accelerated progression of freeze-thaw damage in the later phase. In addition, the study revealed that damage was concentrated in Fracture Zone I during both the early and late freeze-thaw stages, while the rock bridge region exhibited minimal damage. In numerical simulations, microcrack development during the early freeze-thaw phase primarily occurred on specimen surfaces and around joints. As the number of freeze-thaw cycles increased, microcracks progressively expanded inward from the specimen periphery and joint margins. Then, crack clustering emerged around the joints and exhibited localized cumulative effects. The intensity of crack clustering was significantly higher around low-angle joints than around high-angle joints. After 20 freeze-thaw cycles, secondary cracks induced by freeze-thaw damage became observable around the low-angle joints, which caused severe specimen degradation. This finding further confirmed that low-angle joints sustained greater damage than high-angle joints during freeze-thaw cycles. At any freeze-thaw cycle count, tensile microcracks significantly outnumbered shear cracks and dominated the damage pattern. Further analysis of the displacement and force chain diagrams revealed that particle displacement and freeze-induced swelling forces were primarily concentrated around the fractures. After 25 freeze-thaw cycles, a distinct low-force chain line appeared from the left tip of the low-angle fractures to the specimen boundary. This occurred because secondary fractures generated by freeze-thaw damage propagated to the boundary and caused stress release. Regarding the evolution of microcracks and fracture mechanisms during the simulated uniaxial compression of freeze-thaw fractured sandstone, the study identified four distinct loading stages: (Ⅰ) crack-free stage, (Ⅱ) slow crack development stage, (Ⅲ) rapid crack propagation stage, and (Ⅳ) through-crack failure stage. A comprehensive evaluation of fracture evolution under loading for both unfrozen specimens and specimens subjected to 30 freeze-thaw cycles revealed several similarities. Fractures initiated on the right side of the low-angle fractures, propagated along the rock bridge regions, and connected with the high-angle fractures. Both conditions exhibited block detachment on the left side of the low-angle fractures. However, clear differences also emerged. Multiple inner wing cracks were observed around the low-angle fractures in the specimens subjected to 30 freeze-thaw cycles during failure. Crack propagation was more intense, and failure was more complete in these specimens, which indicated a greater destructive potential in freeze-thaw-cycled specimens.ConclusionsThis study employs the discrete element method to simulate freeze-thaw cycles and uniaxial compression in fractured, water-saturated sandstone by modeling the phase change expansion of water-ice particles. The results are validated through laboratory mechanical tests and computed tomography (CT) analysis. At the microscale, the evolution of freeze-thaw damage and the characteristics of microfractures in fractured, water-saturated sandstone are investigated. The results indicate that freeze-thaw damage primarily concentrates near low-angle fractures, which trigger secondary crack propagation during the later stages of the cycles. Tensile microcracks dominate throughout the freeze-thaw process, while stress concentration occurs around fractures due to freeze expansion caused by water-ice phase transitions. During loading-induced failure, microcracks exhibit a "slow-moderate-rapid" evolution trend. Tensile microcracks significantly outnumber shear microcracks, indicating a pronounced tensile failure mode. Secondary cracks induced by freeze-thaw cycles become interconnected under loading. As the number of freeze-thaw cycles increases, crack propagation intensifies, ultimately leading to more complete fragmentation. This study provides a theoretical foundation for improving operational safety in geotechnical engineering in cold regions.关键词：freeze‒thaw cycle;fractured rock mass;discrete element;Mesoscopic damage;CT scan;Rupture characteristics242|324|0更新时间：2026-03-26
- 摘要：ObjectiveDeep underground engineering projects, such as transportation tunnels and mine roadways, increasingly encounter high ground temperature conditions, under which the rheological properties of cement grout and the mechanical strength of hardened grout deteriorate significantly, weakening the surrounding rock reinforcement effect. This study develops a novel modified graphene oxide (MGO) using free-radical copolymerization and investigates the effects of MGO on the strength evolution of cement grout and its underlying mechanisms to overcome the challenges of reduced grout fluidity caused by traditional nano-material incorporation and strength reduction resulting from water reducer addition.MethodsMGO was synthesized via free-radical copolymerization using industrial-grade graphene oxide (GO), silane coupling agent (VTMS), Methallyl Polyoxyethylene Ether (MAPE), acrylic acid (AA), ammonium persulfate (APS), and mercaptopropionic acid (MPA) as raw materials. The synthesis ratio of the reactants was strictly controlled, where GO and VTMS were reacted at a mass ratio of 4:1, followed by copolymerization with SGO, MAPE, MPA, APS, and AA at a ratio of 1.0:2.5:0.1:0.2:1.0. The chemical structure and interplanar spacing of MGO were analyzed using Fourier-transform infrared spectroscopy (FTIR) and X-ray diffraction (XRD). Scanning electron microscopy coupled with energy-dispersive X-ray spectroscopy (SEM-EDS) was employed to visualize the morphology of MGO and the hardened grout. The dispersibility of MGO in alkaline cement paste was evaluated by ultraviolet-visible spectroscopy (UV‒Vis), using a saturated CaOH₂ solution as the simulated medium. The cement grouts were prepared using P·O 42.5 Portland cement as the base material and were supplemented with triethanolamine (TEA) and sodium chloride (NaCl). Three water-to-cement (W/C) ratios of 0.50, 0.60, and 0.75 were designed, and MGO, GO, or polycarboxylate superplasticizer (PCE) was added at a fixed dosage of 0.05% for performance comparison, while a blank group without any additives served as the control. At 25 °C, 35 °C, and 45 °C, the flowability of the fresh grout was measured using a Marsh funnel viscometer and a truncated cone, and the flow spread was determined by averaging two perpendicular diameters. The bleeding rate was quantified using a graduated cylinder method by calculating the percentage ratio of bleeding height to the total grout height after 2 hours of static placement. Unconfined compressive strength tests were conducted on standard cylindrical samples (ϕ50 mm × h100 mm) that were cured at 25 °C, 35 °C, or 45 °C under 95% relative humidity for 7, 14, and 28 days. The microstructure of the hardened grout was observed by SEM, and porosity was quantified using gray-level threshold analysis of the obtained images. Thermogravimetric analysis (TGA) was employed to quantify the bound-water content and calcium hydroxide (CH) content in the hydration products. The CH crystallite size was calculated from the XRD data of the hardened cement grout using the Scherrer equation.Results and DiscussionsFTIR analysis showed the appearance of new CH₂/CH₃ and Si—OH peaks in MGO, while the original C=O peak was intensified, which confirmed successful polymer grafting on GO. XRD results showed that the characteristic MGO peak shifted by 2.7° toward a lower 2θ value, indicating an increase in interlayer spacing after polymer grafting. SEM-EDS analysis indicated that polymer grafting increased the carbon content and blunted the wrinkle edges of MGO. UV‒Vis analysis revealed that MGO exhibited a higher initial absorbance and a smaller decrease within 5 min compared to GO, indicating improved dispersion stability. Regarding fluidity, the incorporation of MGO reduced the Marsh funnel time of the grout with a W∶C ratio of 0.6 from 44.3 s to 40.2 s and increased the flow spread range from 218 mm to 221 mm at 25 ℃. Similar trends were observed at 35 ℃ and 45 ℃. The addition of MGO slightly increased the bleeding rate of the grout; however, the initial fluidity exhibited only a weak correlation with bleeding behavior. In terms of compressive strength, the 28 d strength of the hardened grout with a W/C ratio of 0.6 and containing MGO increased to 34.33 MPa at 25 ℃, 33.53 MPa at 35 ℃, and 28.73 MPa at 45 ℃, which exceeded the corresponding strengths of the grouts without MGO by 11.4%, 27%, and 26.7%, respectively. SEM-based porosity analysis revealed that the incorporation of MGO resulted in a denser hardened grout structure, with a porosity of approximately 1.03%. SEM observations showed increased amounts of C‒S‒H and CH in the samples containing MGO, which was attributed to the additional nucleation sites induced by MGO during early hydration. TGA results indicated that, compared to the control sample, the grout containing MGO exhibited 1.39% more bound water and 1.98% more CH, demonstrating that the incorporation of MGO enhanced the cement hydration rate. The CH crystallite sizes calculated using the Scherrer method were 37.74 nm for the {001} plane and 35.73 nm for the {101} plane in the grout containing MGO, which were 21.6% and 28.5% smaller than those of the control group, respectively. These results indicated that MGO effectively reduced the CH crystallite size, refining the microstructure of the grout. The formation of finer CH crystallites enhanced the gel bonding within the grout containing MGO and contributed to the observed strength improvement.ConclusionsThe results demonstrate that within the temperature range of 25 ℃ to 45 ℃, the incorporation of MGO not only significantly improves the 28 d strength of hardened cement grout by approximately 11% to 27%, but also enhances the fluidity of fresh cement grout, improving both grout fluidity and mechanical performance. Compared to graphene oxide, the developed MGO exhibits a more pronounced effect on enhancing the strength of hardened cement grout while avoiding an obvious reduction in fluidity. Uniformly dispersed MGO in cement grout provides additional nucleation sites for the formation of C‒S‒H gels and CH crystals during cement hydration, and plays a critical role in refining the size of CH crystals, resulting in a denser microstructure of the hardened grout. A denser microstructure leads to higher strength. These findings provide in-depth insights into the improvement of grouting quality in rock masses, particularly for deep formations with high ground temperatures.关键词：rock grouting;cement grout materiats;graphene oxide;high ground temperature;strength of hardened cement grout193|315|0更新时间：2026-03-26
- 摘要：ObjectiveThe construction of pile foundation projects in cold regions requires quantitative analyses of displacement and stress in the surrounding soil, which can be achieved using cavity expansion methods. Classical cavity expansion methods fail to account for uneven temperature distribution, temperature variation, and the unique mechanical behavior of frozen soil. Therefore, this study proposes a thermo‒elasto‒plastic model based on a strength criterion with a parabolic curve, also referred to as the parabolic strength criterion, which is applicable to frozen soil and provides analytical expressions for displacement and stress.MethodsFirstly, the geometry and boundary conditions were determined, and several fundamental assumptions of the proposed model were introduced. A Dirichlet boundary condition, defined as a fixed surface temperature, was considered when addressing the heat diffusion problem. Secondly, the governing equations of the thermo‒elastic cavity expansion model were introduced and systematically combined, through which the explicit expressions for displacement and stress in the elastic state were derived. With the increase in internal radial pressure, the state of the surrounding soil gradually transitioned from an elastic state to an elasto‒plastic state, whereby the surrounding soil was divided into elastic and plastic zones governed by different equations. Thirdly, to account for the nonlinear strength behavior of frozen soil, the parabolic strength criterion was introduced, indicating that the strength initially increased and decreased as the mean stress increased. The stress distribution within the plastic zone was derived by combining the equilibrium equations of continuum mechanics with the parabolic strength criterion of frozen soil. It was clearly impossible to derive an explicit expression for displacement using conventional plasticity theory because the parabolic strength criterion was based on triaxial experimental results. Therefore, the proposed model assumed that the compressive equations remained unchanged after the transition from the elastic state to the plastic state. The expression for the displacement of the internal wall surface was derived by substituting the compressive equations into the kinematic relations of continuum mechanics. Finally, an example involving pile foundation construction was analyzed using the proposed model, and the calculation results were compared to those obtained from a precedented model and a numerical method.Results and DiscussionsComparing the modeling procedure with that of the classical cavity expansion model indicated that both stress and displacement were directly interconnected with the temperature distribution. A logarithmic temperature distribution was first obtained by solving the heat diffusion equation, and this distribution was used in the derivation of elastic stress and displacement. For stress and displacement in the elastic state, the influence of temperature was clearly reflected in the term containing the thermal expansion coefficient, which was generally small for solids such as ice, rock, and soil. A significant finding was that when compressive stress was defined as positive, the radial stress consistently represented the maximum principal stress, while the circumferential stress represented the minimum principal stress. This finding facilitated the determination of the plastic zone and enabled the substitution of principal stresses into the parabolic strength criterion. Within the plastic zone, the influence of temperature became considerably more complex because temperature variations not only altered the parameters of the parabolic strength criterion but also affected the subsequent calculations of stress and displacement. The application of the parabolic strength criterion made it challenging to obtain explicit expressions for stress and displacement in the plastic zone, whereas quantitative results were achievable through numerical analysis. The quantitative results of the example demonstrated several distinctive characteristics of the proposed model. For instance, with increasing internal radial pressure, the plastic boundary gradually moved outward, reaching 1 m when the pressure was 1 690 kPa and 2 m when the pressure was 6 306 kPa. The quadratic approximation of the plastic boundary coordinate exhibited good agreement with the results. The circumferential stress at the boundary surface of the plastic zone was found to be discontinuous, which was attributed to stress redistribution in the frozen soil. The displacement of the internal wall surface was calculated and was 83.5 mm when the pressure was 1 690 kPa and 170.0 mm when the pressure was 5 000 kPa. The characteristic of compressive hardening was clearly observed from the stress‒displacement curve. When the results were compared to those of the precedented model based on the Tresca strength criterion and the numerical method based on PFEM, the stress‒displacement relationship was considered acceptable from a geotechnical perspective. The sensitivity of model parameters was further analyzed and was primarily associated with the effects of temperature and the coefficient of plasticity-growing. Internal temperature groups of ‒5 ℃, ‒10 ℃, and ‒15 ℃ were respectively applied, resulting in smaller plastic zones, indicating higher frozen soil strength, and lower circumferential stress drops at the plastic boundary. In addition, the coefficient of plasticity-growing was shown to be effective in further adjusting the calculated results.ConclusionsThe proposed model is developed based on a strength criterion characterized by a parabolic curve applicable to frozen soil, with temperature variation explicitly considered. A thermoelastic constitutive equation is employed in which temperature is treated as a variable to derive the stress and displacement expressions of the soil. The model can incorporate more complex construction conditions and is verified to be effective and rational from a geotechnical perspective. The results provide valuable guidance for in situ soil testing, as well as for the design and construction of pile foundations in frozen soil regions.关键词：cavity expansion;frozen soil;temperature field;strength;elasto‒plastic analysis86|203|0更新时间：2026-03-26
HYDRAULIC & CIVIL ENGINEERING
- 摘要：ObjectiveThe fault detection results of the selected parameters require effective fusion to obtain comprehensive system detection results to accurately monitor faults occurring during the operation of the air-conditioning and refrigeration system. Considering the nonlinear and non-Gaussian variation characteristics of temperature and pressure in the selected parameter variables, extracting the abnormal fluctuation characteristics of the air-conditioning and refrigeration system from individual variables becomes the primary problem addressed. Because single-variable monitoring data are susceptible to wild values, noise, and other disturbances, relying only on a single variable to detect faults in the air-conditioning system can produce inaccurate results. Therefore, integrating the fluctuation characteristics of multiple related variables of the air-conditioning and refrigeration system based on the abnormal fluctuation characteristics of each variable using appropriate methods to obtain the overall fault detection results of the system, becomes another problem addressed in this study. Therefore, this research proposes a multivariate fluctuation fusion-oriented fault detection method for air-conditioning and refrigeration systems based on the research concept of "single-variable first, and then overall".MethodsThis study proposed a fault detection method for air conditioning and refrigeration systems based on the fusion of multivariate fluctuation features. First, the abnormal fluctuation characteristics of a single variable were extracted through prediction and smoothing techniques to achieve the Gaussianization of non-Gaussian variables and obtain fault detection evidence that reflected the fluctuation characteristics of the variable. Second, the nonlinear correlation among multiple variables in the air conditioning system was captured using the distance correlation coefficient (dCor). Finally, the evidential reasoning rule with dependent evidence (ERr‒DE) method was utilized to fuse the fluctuation characteristic evidence of multiple related variables in the air conditioning system to obtain the overall fault detection results at the system level. The main innovations of the study included 1) the proposed fluctuation feature extraction method for non-Gaussian univariate variables based on prediction and smoothing techniques, which achieved the Gaussianization transformation of monitoring data for non-Gaussian variables in the air conditioning system, extracted the fluctuation features of univariate variables, and provided the basis for fault detection at the univariate level, and 2) the adoption of a correlation evidence inference method to effectively fuse fluctuation features of multiple relevant variables in the air conditioning system to obtain overall fault detection results at the system level, which provided a new solution for fault detection in air conditioning and refrigeration systems.Results and DiscussionsThis study verified the effectiveness of the proposed method through experiments and analyzed its advantages in detecting system faults at different levels. It also provided a basis for optimizing the operation and maintenance management of air conditioning and refrigeration systems. In this study, fuzzy inference theory, deep learning, and confidence rule base methods were selected as comparative approaches for fault monitoring. Compared to fuzzy inference theory and neural network theory, the diagnostic accuracy improved by 15.75% and 13.75%, respectively. Compared to the confidence rule base method, the accuracy improved by 30.75%. Therefore, the comparison experiments indicated that the performance of the diagnostic model based on the ERr-DE method was superior to other methods. In addition, the method proposed in this study not only obtained fault detection results for individual system variables but also fused them to generate overall fault detection results for the system. This capability allowed system managers to comprehensively understand the fault status of the system at different levels. Specifically, the single-variable fault detection results were utilized to detect early signs of system abnormalities, providing alarms for front-line duty personnel. In addition, the fused overall fault detection results were utilized to assess the overall operational status of the system, which provided a basis for decision-making when optimizing system operation and maintenance strategies.ConclusionsThis study proposes a new fault detection method for air-conditioning and refrigeration systems. The method is based on prediction and smoothing techniques to extract the fluctuation characteristics of the system's single-variable time series. On this basis, the ERr‒DE model is employed to achieve effective fusion of evidence generated by the refrigeration system's multi-correlation variables and to obtain the overall fault detection results of the system. The proposed method is applied to a fault detection experiment of a certain type of air-conditioning refrigeration system to verify its effectiveness. The obtained results enable system operation and maintenance managers to understand the fault status of the system at different levels, providing a reliable decision-making basis for early abnormal alarms and the overall operation and maintenance of the system.关键词：air conditioning refrigeration system;fault detection;fluctuation feature extraction;evidence reasoning rule with dependent evidence;feature fusion228|454|0更新时间：2026-03-26
-
Incremental Attribute Reduction Algorithm for Dominance-based Neighborhood Relative Decision Entropy AI导读
摘要：ObjectiveIn a big data environment, data is continuously and dynamically updated, which imposes significant limitations and challenges on traditional machine learning algorithms. Incremental learning is a learning paradigm that focuses exclusively on newly changed data based on the learning outcomes of existing models, substantially improving learning efficiency during data update processes. Incremental attribute reduction is a widely adopted method and strategy for achieving efficient attribute reduction in dynamic dataset environments. However, in ordered information systems with dynamically updated and evolving objects, existing incremental attribute reduction methods neglect the potential classification information provided by rough approximation sets, which limits their effectiveness in supporting attribute reduction.MethodsIn existing research results, scholars jointly proposed a relative decision entropy model by utilizing the roughness and dependency of upper and lower approximation sets to combine information entropy. This entropy model considered both the classification information of the upper and lower approximation sets of attributes and the classification information from the perspective of attribute information entropy, and it demonstrated better performance in attribute uncertainty measurement and attribute reduction. At the same time, the basic calculation unit of relative decision entropy was the upper and lower approximation sets of attributes. When the object changed, there was no need to calculate the information granules of the updated object. Therefore, this study extended relative decision entropy to mixed ordered information system environments. First, the relative decision entropy model of dominance-based neighborhoods was proposed. Then, the relative decision entropy of dominance-based neighborhoods was constructed based on the dominance-based neighborhood relation. The relative decision entropy of dominance-based neighborhoods was reconstructed in the form of a matrix, and a non-incremental attribute reduction algorithm for hybrid ordered information systems was designed. Finally, for the two scenarios of increasing and decreasing objects in a hybrid ordered information system, the matrix-form incremental update of the relative decision entropy of the dominance-based neighborhood was analyzed and studied, and incremental attribute reduction algorithms were constructed using this update mechanism.Results and DiscussionsIn the experimental stage, 8 public datasets were selected for simulation experiments to compare the incremental algorithm with non-incremental algorithms. In the case of increasing the number of objects in the dataset, the number of reduced attributes for the non-incremental and incremental algorithms was basically similar across the 8 datasets, with averages of 14.75 and 14.62 attributes, respectively. The classification accuracy of the non-incremental algorithm and the incremental algorithm was also basically similar across the 8 datasets, with average SVM classification accuracies of 86.12% and 86.65%, and average NB classification accuracies of 87.22% and 87.24%, respectively. The processing time of the incremental algorithms was significantly shorter than that of the non-incremental algorithms. The processing time of the non-incremental algorithms on the 8 datasets was 4 215.13 seconds, whereas the processing time of the incremental algorithms on the 8 datasets was only 55.38 seconds. The performance of the incremental algorithms was significantly superior. In the case of reducing dataset objects, the number of reduced attributes for the non-incremental and incremental algorithms was basically similar across the 8 datasets, with averages of 16.00 and 15.62, respectively. The classification accuracy of the non-incremental algorithm and the incremental algorithm was basically similar across the 8 datasets, with average SVM classification accuracies of 86.25% and 86.74%, and average NB classification accuracies of 87.03% and 87.48%, respectively. The processing time of the incremental algorithms was significantly shorter than that of the non-incremental algorithms, with the non-incremental algorithms consuming 1 053.56 seconds and the incremental algorithms consuming only 36.92 seconds. The effectiveness of the incremental algorithm was verified through these experimental results. An experimental comparison was conducted between the incremental algorithm proposed in this study and four superior comparative algorithms to verify the superiority of the algorithm. In the case of increasing the number of objects in the dataset, the number of reduced attributes for the four comparative algorithms and the algorithm proposed in this study were 14.76, 16.75, 14.87, 14.62, and 14.62, respectively. The SVM classification accuracies of the four comparative algorithms and the algorithm proposed in this study were 87.53%, 83.84%, 89.20%, 86.84%, and 89.69%, respectively. The NB classification accuracies of the four comparative algorithms and the algorithm proposed in this study were 86.10%, 84.51%, 88.00%, 87.44%, and 89.81%, respectively. In terms of processing time, the proposed algorithm was 35%, 32%, 40%, and 7% faster than the four compared algorithms, respectively. In the case of reducing dataset objects, the number of reduced attributes for the four comparative algorithms and the algorithm proposed in this study were 16.00, 18.37, 15.90, 15.82, and 15.62, respectively. The SVM classification accuracies of the four comparative algorithms and the algorithm proposed in this study were 85.92%, 84.06%, 87.67%, 86.74%, and 89.50%, respectively. The NB classification accuracies of the four comparative algorithms and the algorithm proposed in this study were 85.82%, 83.85%, 88.75%, 87.43%, and 90.15%, respectively. In terms of processing time, the proposed algorithm was 55%, 46%, 63%, and 12% faster than the four compared algorithms, respectively. Comparing these results demonstrated that the incremental algorithm in this study selected fewer attributes, achieved higher classification accuracy, and significantly outperformed the comparative algorithms in terms of efficiency.ConclusionsThe experimental results demonstrate that the proposed incremental algorithm exhibits superior attribute reduction performance on dynamic datasets, significantly enhancing the efficiency of dynamic attribute reduction while maintaining the number of attribute selections and classification accuracy. At the same time, the proposed incremental algorithm selects fewer attributes and achieves higher classification accuracy compared to similar algorithms. Most importantly, the algorithm demonstrates higher computational performance.关键词：attribute reduction;object changes;incremental;ordered information system;dominance-based neighborhood relative decision entropy205|314|0更新时间：2026-03-26 - 摘要：ObjectiveIn recent years, applications of unmanned aerial vehicles (UAVs) have experienced rapid growth in both military and civilian domains, including aerial surveillance, remote sensing, and search and rescue. Beyond individual use, UAV swarms are widely employed as micro-UAV array radars. High-precision detection is achieved through beamforming by treating a swarm as a large movable antenna aperture. Therefore, high-precision flight control becomes essential. However, outdoor formations often rely on relative position-based strategies to avoid GPS inaccuracies. Although this approach remains theoretically feasible, micro-UAVs can deviate because of turbulence, and positional errors can be inherited by subsequent UAVs, leading to significant degradation in detection performance. Therefore, this study aims to evaluate the impact of such inherent positional perturbations on the direction of arrival (DoA) estimation performance.MethodsThis study focused on the lower bound of estimation accuracy to assess the efficiency and precision of estimators in the presence of calibration errors. Because the analytical calculation of the Cramér‒Rao bound (CRB) with random nuisance parameters was challenging, the hybrid Cramér‒Rao bound (HCRB) was adopted, as it represented an effective balance between computability and tightness. Based on the assumption of the inheritability of UAV array perturbations, the HCRB was derived for the DoA of a single unknown source. A swarm of N UAVs was considered as a uniform linear array, where each UAV carried an omnidirectional antenna. The model jointly analyzed perturbations in the parallel and perpendicular directions to the aperture under a far-field narrowband signal scenario. In addition, numerical simulations and Monte Carlo experiments based on the matched filter (MF) estimator were conducted to validate the theoretical derivation.Results and DiscussionsNumerical simulations analyzed the HCRB performance under various influencing factors. The results demonstrated that as the positional perturbation variance increases, the HCRB for both low and high SNR cases increases from the CRB level without interference, confirming that the error-free CRB represents the lower bound of the HCRB. Monte Carlo experiments exhibited trends consistent with these theoretical results. In addition, as the SNR increases, the HCRB of DoA becomes more sensitive to positional errors. Although the HCRB decreases with higher SNR, smaller positional variance consistently produces lower HCRB values. Regarding the number of UAVs, although the HCRB monotonically decreases as N increases from 1 to 128, it did not converge to the CRB without perturbations. Finally, the impact of different DoA angles indicated that the HCRB is non-uniform and axially symmetric, achieving its lowest value at 0° and increasing as the absolute value of the DoA increases.ConclusionsThe inherent nature of positional errors arising from relative position-based formations, which has been overlooked in previous literature, significantly affects UAV array performance. The derivation and quantitative analysis of the HCRB in this study confirm that the inheritability of perturbations leads to a performance floor that cannot be eliminated simply by increasing the SNR or the number of UAVs. These findings provide an important theoretical basis for assessing efficiency and predicting the performance of mobile array radars in practical flight environments, particularly for applications such as interferometric synthetic aperture radar (InSAR).关键词：UAV;array;positional errors;DoA;Cramer-Rao bound342|705|0更新时间：2026-03-26
INFORMATION ENGINEERING
- Technical support is provided by Beijing Founder electronics co., LTD 京ICP备09064830号-19
京公网安备11010802024621 - It is recommended to read the content of this site in Chrome&IE9+. Please switch to extreme mode in browser 360.
- Cookies We use cookies to help provide and enhance our service and tailor content. By continuing, you agree to the use of cookies.
0