Search Results (9,978)

This paper investigates the problem of identifying a time-dependent source term in distributed-order time-space fractional diffusion equations (FDEs) based on boundary observation data. Firstly, the existence, uniqueness, and regularity of the solution to the direct problem are proved. Using the regularity of the solution and a Gronwall inequality with a weakly singular kernel, the uniqueness and stability estimates of the solution to the inverse problem are obtained. Subsequently, the inverse source problem is transformed into a minimization problem of a functional using the Tikhonov regularization method, and an approximate solution is obtained by the conjugate gradient method. Numerical experiments confirm that the method provides both accurate and robust results. Full article

The spread of COVID-19 is influenced by multiple factors, including control policies, virus characteristics, individual behaviors, and environmental conditions, exhibiting highly complex nonlinear dynamic features. The time series of new confirmed cases shows significant nonlinearity and non-stationarity. Traditional prediction methods that rely solely on one-dimensional case data struggle to capture the multi-dimensional features of the data and are limited in handling nonlinear and non-stationary characteristics. Their prediction accuracy and generalization capabilities remain insufficient, and most existing studies focus on single-step forecasting, with limited attention to multi-step prediction. To address these challenges, this paper proposes a multi-module fusion prediction model—TSMixer-BiKSA network—that integrates multi-feature inputs, Variational Mode Decomposition (VMD), and a dual-branch parallel architecture for 1- to 3-day-ahead multi-step forecasting of new COVID-19 cases. First, variables highly correlated with the target sequence are selected through correlation analysis to construct a feature matrix, which serves as one input branch. Simultaneously, the case sequence is decomposed using VMD to extract low-complexity, highly regular multi-scale modal components as the other input branch, enhancing the model’s ability to perceive and represent multi-source information. The two input branches are then processed in parallel by the TSMixer-BiKSA network model. Specifically, the TSMixer module employs a multilayer perceptron (MLP) structure to alternately model along the temporal and feature dimensions, capturing cross-time and cross-variable dependencies. The BiGRU module extracts bidirectional dynamic features of the sequence, improving long-term dependency modeling. The KAN module introduces hierarchical nonlinear transformations to enhance high-order feature interactions. Finally, the SA attention mechanism enables the adaptive weighted fusion of multi-source information, reinforcing inter-module synergy and enhancing the overall feature extraction and representation capability. Experimental results based on COVID-19 case data from Italy and the United States demonstrate that the proposed model significantly outperforms existing mainstream methods across various error metrics, achieving higher prediction accuracy and robustness. Full article

Rapid global urbanization poses complex challenges that demand advanced data-driven forecasting solutions for smart cities. Traditional statistical and standalone Long Short-Term Memory (LSTM) models often struggle to capture non-linear dynamics and long-term dependencies in urban time-series data. This review critically examines hybrid LSTM models that integrate LSTM with complementary algorithms, including CNN, GRU, ARIMA, and SVM. These hybrid architectures aim to enhance prediction accuracy, integrate diverse data sources, and improve computational efficiency. This study systematically reviews principles, trends, and real-world applications, quantitatively evaluating hybrid LSTM models using performance metrics such as mean absolute error (MAE), root mean square error (RMSE), and the coefficient of determination (R²), while identifying key study limitations. The case studies considered include traffic management, environmental monitoring, energy forecasting, public health, infrastructure assessment, and urban waste management. For example, hybrid models have achieved substantial accuracy improvements in traffic congestion forecasting, reducing their mean absolute error by up to 29%. Despite the inherent challenges related to structural complexity, interpretability, and data requirements, ongoing research on attention mechanisms, model compression, and explainable AI has significantly mitigated these limitations. Thus, hybrid LSTM models have emerged as vital analytical tools capable of robust spatiotemporal prediction, effectively supporting sustainable urban development and data-driven decision-making in evolving smart city environments. Full article

With the exponential growth of engineering monitoring data, data-driven neural networks have gained widespread application in predicting retaining structure deformation in foundation pit engineering. However, existing models often overlook the spatial deflection correlations among monitoring points. Therefore, this study proposes a novel deep learning framework, CGCA (Convolutional Gated Recurrent Unit with Cross-Attention), which integrates ConvGRU and cross-attention mechanisms. The model achieves spatio-temporal feature extraction and deformation prediction via an encoder–decoder architecture. Specifically, the convolutional structure captures spatial dependencies between monitoring points, while the recurrent unit extracts time-series characteristics of deformation. A cross-attention mechanism is introduced to dynamically weight the interactions between spatial and temporal data. Additionally, the model incorporates multi-dimensional inputs, including full-depth inclinometer measurements, construction parameters, and tube burial depths. The optimization strategy combines AdamW and Lookahead to enhance training stability and generalization capability in geotechnical engineering scenarios. Case studies of foundation pit engineering demonstrate that the CGCA model exhibits superior performance and robust generalization capabilities. When validated against standard section (CX1) and complex working condition (CX2) datasets involving adjacent bridge structures, the model achieves determination coefficients (R²) of 0.996 and 0.994, respectively. The root mean square error (RMSE) remains below 0.44 mm, while the mean absolute error (MAE) is less than 0.36 mm. Comparative experiments confirm the effectiveness of the proposed model architecture and the optimization strategy. This framework offers an efficient and reliable technical solution for deformation early warning and dynamic decision-making in foundation pit engineering. Full article

This study conducted uniaxial creep tests on coal samples under both natural and water-saturated conditions for durations of about 180 days per sample to study the stability of coal pillar dams of the Daliuta Coal Mine underground reservoir. Combined with synchronized acoustic emission (AE) monitoring, the research systematically revealed the time-dependent deformation mechanisms and damage evolution laws of coal under prolonged water immersion and natural conditions. The results indicate that water-immersed coal exhibits a unique negative creep phenomenon at the initial stage, with the strain rate down to −0.00086%/d, attributed to non-uniform pore compaction and elastic rebound effects. During the steady-state creep phase, the creep rates under water-immersed and natural conditions were comparable. However, water immersion led to an 11.4% attenuation in elastic modulus, decreasing from 2300 MPa to 2037 MPa. Water immersion would also suppress AE activity, leading to the average daily AE events of 128, which is only 25% of that under natural conditions. In the accelerating creep stage, the AE event rate surged abruptly, validating its potential as an early warning indicator for coal pillar instability. Based on the identified long-term strength of the coal sample, it is recommended to maintain operational loads below the threshold of 9 MPa. This research provides crucial theoretical foundations and experimental data for optimizing the design and safety monitoring of coal pillar dams in CMURs. Full article

The real-time acquisition of an accurate underwater sound velocity profile (SSP) is crucial for tracking the propagation trajectory of underwater acoustic signals, making it play a key role in ocean communication positioning. SSPs can be directly measured by instruments or inverted leveraging sound field data. Although measurement techniques provide a good accuracy, they are constrained by limited spatial coverage and require a substantial time investment. The inversion method based on the real-time measurement of acoustic field data improves operational efficiency but loses the accuracy of SSP estimation and suffers from limited spatial applicability due to its stringent requirements for ocean observation infrastructures. To achieve accurate long-term ocean SSP estimation independent of real-time underwater data measurements, we propose a semi-transformer neural network (STNet) specifically designed for simulating sound velocity distribution patterns from the perspective of time series prediction. The proposed network architecture incorporates an optimized self-attention mechanism to effectively capture long-range temporal dependencies within historical sound velocity time-series data, facilitating an accurate estimation of current SSPs or prediction of future SSPs. Through the architectural optimization of the transformer framework and integration of a time encoding mechanism, STNet could effectively improve computational efficiency. For long-term forecasting (using the Pacific Ocean as a case study), STNet achieved an annual average RMSE of 0.5811 m/s, outperforming the best baseline model, H-LSTM, by 26%. In short-term forecasting for the South China Sea, STNet further reduced the RMSE to 0.1385 m/s, demonstrating a 51% improvement over H-LSTM. Comparative experimental results revealed that STNet outperformed state-of-the-art models in predictive accuracy and maintained good computational efficiency, demonstrating its potential for enabling accurate long-term full-depth ocean SSP forecasting. Full article

Air quality forecasts play a crucial role in informing the public about atmospheric pollutant levels that pose risks to human health and the environment. The accuracy of these forecasts strongly depends on the quality and resolution of the input data used in the modelling process. At HungaroMet, the Hungarian Meteorological Service, the CHIMERE chemical transport model is used to provide two-day air quality forecasts for the territory of Hungary. This study compares two configurations of the CHIMERE model: the current operational setup, which uses climatological averages from the LMDz-INCA database for boundary conditions, and a test configuration that incorporates real-time boundary conditions from the CAMS global forecast. The primary objective of this work was to assess how the use of real-time versus climatological boundary conditions affects modelled concentrations of key pollutants, including NO₂, O₃, PM₁₀, and PM_2.5. The model results were evaluated against observational data from the Hungarian Air Quality Monitoring Network using a range of statistical metrics. The results indicate that the use of real-time boundary conditions, particularly for aerosol-type pollutants, improves the accuracy of PM₁₀ forecasts. This improvement is most significant under meteorological conditions that favour the long-range transport of particulate matter, such as during Saharan dust or wildfire episodes. These findings highlight the importance of incorporating dynamic, up-to-date boundary data, especially for particulate matter forecasting—given the increasing frequency of transboundary dust events. Full article

Understanding the time-lagged response of land subsidence to groundwater level fluctuations and subsurface strain variations is crucial for uncovering its underlying mechanisms and enhancing disaster early warning capabilities. This study focuses on Dangshan County, Anhui Province, China, and systematically analyzes the spatio-temporal evolution of land subsidence from 2018 to 2024. A total of 207 Sentinel-1 SAR images were first processed using the Small Baseline Subset Interferometric Synthetic Aperture Radar (SBAS-InSAR) technique to generate high-resolution surface deformation time series. Subsequently, the seasonal-trend decomposition using the LOESS (STL) model was applied to extract annual cyclic deformation components from the InSAR-derived time series. To quantitatively assess the delayed response of land subsidence to groundwater level changes and subsurface strain evolution, time-lagged cross-correlation (TLCC) analysis was performed between surface deformation and both groundwater level data and distributed fiber-optic strain measurements within the 5–50 m depth interval. The strain data was collected using a borehole-based automated distributed fiber-optic sensing system. The results indicate that land subsidence is primarily concentrated in the urban core, with annual cyclic amplitudes ranging from 10 to 18 mm and peak values reaching 22 mm. The timing of surface rebound shows spatial variability, typically occurring in mid-February in residential areas and mid-May in agricultural zones. The analysis reveals that surface deformation lags behind groundwater fluctuations by approximately 2 to 3 months, depending on local hydrogeological conditions, while subsurface strain changes generally lead surface subsidence by about 3 months. These findings demonstrate the strong predictive potential of distributed fiber-optic sensing in capturing precursory deformation signals and underscore the importance of integrating InSAR, hydrological, and geotechnical data for advancing the understanding of subsidence mechanisms and improving monitoring and mitigation efforts. Full article

This paper presents an advanced time-series forecasting methodology that integrates multiple machine learning models to improve data prediction in enhanced long-range navigation (eLoran) systems. The analysis evaluates five forecasting approaches: multivariate linear regression, long short-term memory (LSTM) networks, random forest (RF), a fusion model combining LSTM and RF, and a dynamic weighting (DW) model. The results demonstrate that the DW model achieves the highest prediction accuracy while maintaining strong computational efficiency, making it particularly suitable for real-time applications with stringent performance requirements. Although the LSTM model effectively captures temporal dependencies, it demands considerable computational resources. The hybrid model utilises the strengths of LSTM and RF to enhance the accuracy but involves extended training times. By contrast, the DW model dynamically adjusts the relative contributions of LSTM and RF on the basis of the data characteristics, thereby enhancing the accuracy while significantly reducing the computational demands. Demonstrating exceptional performance on the ASF2 dataset, the DW model provides a well-balanced solution that combines precision with operational efficiency. This research offers valuable insights into optimising additional secondary phase factor (ASF) prediction in eLoran systems and highlights the broader applicability of real-time forecasting models. Full article

Climate change can severely impact plant-pollinator interactions and have serious effects on ecosystem services such as pollination. This study was carried out in 2023 and 2024, and it examined the effects of drought on flowering and nectar production in one cultivar of white mustard (Brassica alba cv. Palma), an important entomophilous crop of the temperate zone with several attributes that make it promising for sustainable agricultural practices. Drought-stressed plants delayed the flowering time, shortened the flowering duration, and developed significantly fewer flowers. Nectar production in white mustard depends on soil moisture levels and short-term changes in meteorological conditions (e.g., air humidity, air temperature). At reduced soil moisture, the total sugar yield per plant decreased by 60%, compared to control plants, resulting in lower availability of caloric food resources, which should be considered when developing strategies supporting pollinators. Changes in floral traits resulted in differences in the frequency of insect visits, which may exert a negative impact on white mustard pollination under drought stress and may have indirect consequences for seed yield resulting from increased drought intensity associated with climate change. The results provide important data for the management of the white mustard crop and indicate the need for broader evaluation of cultivars to promote drought-resistant B. alba cultivars. Full article

Objective: This study aimed to assess the effect of antibiotic de-escalation on 30-day mortality, duration of intravenous (IV) antibiotic therapy and length of hospital stay (LOS) in severe community-acquired pneumonia (sCAP). Methods: We performed a retrospective analysis of prospectively collected data from a cohort of adults diagnosed with sCAP and microbiologically confirmed etiology between 1995 to 2022. Two distinct time points of the de-escalation were analyzed: 3 and 6 days post-admission, corresponding, respectively, to the availability of microbiological results and the median time to clinical stability. Inverse propensity score-weighted binary logistic regression was used to adjust for potential confounders. Results: A total of 398 consecutive cases of sCAP were analyzed. No significant differences were observed between the de-escalation and non-de-escalation groups in terms of age, sex, comorbidities, or severity-related variables (such as impaired consciousness, shock, respiratory failure, or multilobar pneumonia). Patients in the de-escalation group had lower rates of leukopenia, bacteremia and empyema, and less need for mechanical ventilation, with variations depending on the timing of de-escalation. After adjusting for confounding factors in an inverse propensity score-weighted analysis, de-escalation within 3 or 6 days after admission was not associated with increased mortality risk (adjusted odds ratio [aOR] 1.48, 95% confidence interval [CI] 0.29–7.4; p = 0.63, and aOR 0.57, 95% CI 0.14–2.31, p = 0.43, respectively). Similar findings were observed for prolonged LOS. However, antibiotic de-escalation was related to a lower risk of prolonged IV antibiotic. Conclusions: Antibiotic de-escalation in microbiologically confirmed sCAP did not negatively impact clinical outcomes, supporting the safety of this strategy for optimizing antibiotic use in this serious infection. Full article

Fault detection and remaining useful life (RUL) prediction are critical tasks in self-healing network (SHN) environments and industrial cyber–physical systems. These domains demand intelligent systems capable of handling dynamic, high-dimensional sensor data. However, existing optimization-based approaches often struggle with imbalanced datasets, noisy signals, and delayed convergence, limiting their effectiveness in real-time applications. This study utilizes two benchmark datasets—EFCD and SFDD—which represent electrical and sensor fault scenarios, respectively. These datasets pose challenges due to class imbalance and complex temporal dependencies. To address this, we propose a novel hybrid framework combining Attention-Augmented Convolutional Neural Networks (AACNN) with transformer encoders, enhanced through Enhanced Ensemble-SMOTE for balancing the minority class. The model captures spatial features and long-range temporal patterns and learns effectively from imbalanced data streams. The novelty lies in the integration of attention mechanisms and adaptive oversampling in a unified fault-prediction architecture. Model evaluation is based on multiple performance metrics, including accuracy, F1-score, MCC, RMSE, and score*. The results show that the proposed model outperforms state-of-the-art approaches, achieving up to 97.14% accuracy and a score* of 0.419, with faster convergence and improved generalization across both datasets. Full article

This study analyzes the impact of the type of transported material (overburden, lignite, mixture) on the rate of core damage accumulation in Type St conveyor belts in open-pit mines. The research was conducted using the DiagBelt+ diagnostic system, which enables the assessment of belt core condition without dismantling the belt. Data were collected from over 100 conveyor belt loops, covering segments of varying lengths, ages, and operational histories. Damage density and area were assessed, and differences were analyzed depending on the material type. The results indicate that belt age and damage density vary significantly with material type, while the Resurs indicator (percentage of expected operating time) shows no clear dependence on the material type. A multiple regression analysis was also performed to predict failure density based on operational variables, such as Age, Resurs results, Loop Length, and Segment Length. The regression model explains approximately 46% of the variability in damage density, indicating the need for further research to improve predictive accuracy. The study emphasizes the importance of using non-destructive diagnostic systems to optimize maintenance planning and enhance conveyor belt reliability. Full article

Utilizing deep learning models to detect malicious anomalies within the traffic of application layer J1939 protocol networks, found on heavy-duty commercial vehicles, is becoming a critical area of research in platform protection. At the physical layer, the controller area network (CAN) bus is the backbone network for most vehicles. The CAN bus is highly efficient and dependable, which makes it a suitable networking solution for automobiles where reaction time and speed are of the essence due to safety considerations. Much recent research has been conducted on securing the CAN bus explicitly; however, the importance of protecting the J1939 protocol is becoming apparent. Our research utilizes long short-term memory models to predict the next binary data sequence of a J1939 packet. Our primary objective is to compare the performance of our J1939 detection system trained on data sub-fields against a published CAN system trained on the full data payload. We conducted a series of experiments to evaluate both detection systems by utilizing a simulated attack representation to generate anomalies. We show that both detection systems outperform one another on a case-by-case basis and determine that there is a clear requirement for a multifaceted security approach for vehicular networks. Full article

Accurate forecasting of river water flow helps to warn of floods and droughts in advance, provides a basis for the rational allocation of water resources, and at the same time, offers important support for the safe operation of hydropower stations and water conservancy projects. Water flow is characterized by time series, but the existing models focus on the positive series when LSTM is applied, without considering the different contributions of the water flow series to the model at different moments. In order to solve this problem, this study proposes a river water flow prediction model, named AT-BiLSTM, which mainly consists of a bidirectional layer and an attention layer. The bidirectional layer is able to better capture the long-distance dependencies in the sequential data by combining the forward and backward information processing capabilities. In addition, the attention layer focuses on key parts and ignores irrelevant information when processing water flow data series. The effectiveness of the proposed method was validated against an actual dataset from the Shizuishan monitoring station on the Yellow River in China. The results confirmed that compared with the RNN model, the proposed model significantly reduced the MAE, MSE, and RMSE on the dataset by 27.16%, 42.01%, and 23.85%, respectively, providing the best predictive performance among the six compared models. Moreover, this attention mechanism enables the model to show good performance in 72 h (3 days) forecast, keeping the average prediction error below 6%. This implies that the proposed hybrid model could provide a decision base for river flow flood control and resource allocation. Full article