Search Results (410)

To solve the low-precision prediction problem of noisy non-stationary goaf subsidence sequences, this study aims to establish a high-accuracy hybrid prediction model for mining surface deformation monitoring. The Global Navigation Satellite System (GNSS) monitoring data of surface subsidence in goaf areas exhibits non-stationary and noisy characteristics, which limits the accuracy of traditional prediction models. In this paper, a hybrid prediction model, namely the Improved Radial Movement Optimization–Variational Mode Decomposition–Gated Recurrent Unit (IRMO-VMD-GRU) model, is proposed. The IRMO algorithm is employed to globally optimize the key parameters of VMD, achieving adaptive and stable decomposition of the settlement sequences. The obtained Intrinsic Mode Function (IMF) sub-sequences are input into the GRU network for independent training and prediction, followed by superposition and reconstruction. The model is validated using the GNSS monitoring data from three monitoring points at a coal mine in Shaanxi Province, China. The results show that the proposed model outperforms the comparison models in all four evaluation indicators, namely Mean Absolute Error (MAE), Root Mean Square Error (RMSE), Mean Absolute Percentage Error (MAPE), and Coefficient of Determination (R²), with all R² values exceeding 0.8. The model demonstrates superior fitting performance, correlation, and generalization ability, which provides important practical technical support for goaf subsidence early warning, geological disaster prevention and engineering safety management in mining areas. Full article

Mechanical Wear particle recognition is an important approach for equipment health monitoring and fault early warning. However, flow-field disturbances and high-speed particle motion in high-speed fluid environments can lead to image degradation, non-stationary electrostatic signals, and insufficient reliability of single-source recognition methods. Therefore, this study proposes a wear particle recognition method based on multi-source information fusion for high-speed fluid environments. The method establishes a multi-scale electrostatic sensing model to characterize the coupling relationship among particle material properties, motion states, and electrostatic response characteristics. Empirical mode decomposition and independent component analysis are combined for adaptive electrostatic signal denoising, and a Transformer network is used to extract multi-domain features. Meanwhile, an ECA-CNN model with an efficient channel attention mechanism is introduced to enhance the feature representation of degraded particle images. On this basis, a meta-learning-based sample-adaptive decision fusion framework is developed to achieve dynamic and complementary fusion of electrostatic and visual information. The experimental results demonstrate that the proposed method exhibits excellent recognition accuracy and robustness in the tested high-speed fluid environment of 10 m/s, achieving a fusion recognition accuracy of 96.0%, which is significantly superior to single-source recognition methods. Ablation experiments further show that removing the global scaling factor, guidance loss, interpolation loss, and category-specific weight generator decreases the average recognition accuracy by 0.7%, 1.2%, 0.4%, and 1.8%, respectively, confirming the contribution of each key module to fusion recognition performance. These findings provide a new technical approach for the online intelligent recognition of wear particles under high-speed fluid conditions and offer theoretical support and methodological guidance for condition monitoring, health assessment, and intelligent operation and maintenance of large-scale equipment. Full article

In low-voltage distribution systems, series arc faults caused by poor contact and loose connections are a leading cause of electrical fires. Due to the negative resistance characteristics of arcs, such faults are difficult to detect using conventional overcurrent or leakage protectors. Existing detection methods predominantly rely on wavelet-based feature extraction or threshold-based classifiers. Wavelet transforms require predefined basis functions and lack adaptability to non-stationary current signals from appliances such as induction cookers. Threshold-based classifiers produce excessive false alarms under varying load conditions, as normal non-stationary load waveforms share high-frequency characteristics with arc fault signatures. As a result, existing arc fault protectors exhibit high false alarm rates, limiting practical deployment. To address these limitations, this study proposes a method for diagnosing low-voltage series arc faults based on differential-sliding window higher-order cumulants (HoCs) and stacked autoencoders (SAEs). The method first employs a differential-sliding time window approach to extract HoC features from current signals across seven typical loads, establishing a feature vector database for arc fault patterns. A symmetric stacked autoencoder (SAE) is constructed, trained using layer-wise pretraining to optimize hyperparameters and select the model with the best generalization performance. Experimental results demonstrate that the proposed method achieves a detection accuracy of 96.4% with a false alarm rate of 0% across all tested loads. Full article

To address runoff forecasting inaccuracies caused by data gaps in reservoir operations, this paper proposes a collaborative modeling framework integrating deep learning, signal decomposition, uncertainty quantification, and transfer learning. Validated on the Wei River (source basin) and Yongding River (target basin) with similar hydrological characteristics, the framework first constructs a Pyraformer-BiLSTM-LSS point forecasting model to enhance characterization of non-stationary runoff sequences. Then, the BLSO-VMD optimization decomposition technique filters and reconstructs forecasting noise, improving model robustness. Subsequently, a probabilistic interval forecasting model is developed via multi-task learning to reliably quantify uncertainty. To tackle data scarcity in the target domain, a “decomposition–reconstruction–transfer” learning mechanism transfers model knowledge from the source domain to the target domain. Results show that the framework achieves excellent performance in the source domain and successfully transfers to the data-scarce target domain, significantly enhancing the accuracy and stability of both point and interval forecasts. By establishing a collaborative modeling framework combining transfer learning and multi-task learning, along with an adaptive signal decomposition method based on BLSO and a multi-scale deep learning model, this study effectively addresses the challenges of accuracy and reliability in runoff forecasting for data-scarce basins. It provides a transferable and scalable technical pathway for runoff simulation and reservoir operation in hydrologically underserved regions, supporting sustainable water resource management and ecological protection. Full article

Industrial sensor systems are increasingly vulnerable to both physical anomalies and cyberattacks, while their collected time series typically present complex periodic and non-stationary characteristics, along with dynamic spatial dependencies among sensors. To address these issues, this paper proposes a dual-branch graph learning framework with frequency gating for simultaneous industrial sensor anomaly and cyberattack detection. The model first divides the input time series into multiple patches and decomposes each patch into periodic and non-stationary components via frequency analysis. Two graph isomorphism network branches, namely periodic GIN (P-GIN) and non-stationary GIN (NS-GIN), are designed to model the spatial dependencies of the two components separately, where the graph structure is adaptively learned using a Gaussian kernel-based mechanism. Furthermore, a frequency gating module is introduced in the non-stationary branch to enhance the representation of abnormal and attack-related features. Hierarchical temporal encoding is performed via intra-patch attention and inter-patch attention to capture both local and long-range temporal dependencies. Extensive experimental results on real-world industrial sensor datasets demonstrate that the proposed method achieves superior performance compared with state-of-the-art methods in both anomaly detection and cyberattack detection tasks. Full article

This paper proposes a lightweight and adaptive anomaly detection framework for Internet of Things (IoT) time-series data that progressively combines statistical filtering with reinforcement learning (RL)-based decision mechanisms. Three classical statistical filters, Hampel, interquartile range (IQR), and Z-score, are initially evaluated under controlled IoT anomaly scenarios. While fixed-parameter configurations perform well under specific conditions, their performance degrades in non-stationary and heterogeneous environments. To address this limitation, a tabular Q-learning agent is introduced to dynamically select both filtering methods and their associated parameters according to scenario-specific signal characteristics. By extending the action space to include joint filter and parameter selection, the framework improves adaptability while reducing the need for manual tuning. A multi-agent reinforcement learning (MARL) formulation is further introduced to support distributed learning across heterogeneous IoT environments. The framework is additionally evaluated using real-world IoT temperature data augmented with controlled anomaly injection, enabling reproducible benchmarking under partially realistic sensing conditions. Experimental results show that both RL and MARL maintain stable detection performance across heterogeneous sensor streams. While MARL does not systematically outperform the single-agent approach in detection accuracy, it improves scalability and supports scenario-specific policy specialization, which is particularly relevant for distributed IoT deployments. Overall, the proposed approach provides a lightweight, interpretable, and computationally efficient solution for adaptive anomaly detection in resource-constrained IoT systems. Full article

Accurate prediction of traffic flow is the key to highways control. However, traditional time series forecasting methods cannot meet the accuracy requirements of long-term forecasting. This paper proposes a multi-channel univariate long-term highway inbound traffic flow forecasting framework with multidimensional signal feature decomposition of time series and a patch time series depth-separable convolutional neural network. Firstly, we propose a multidimensional decomposition block consisting of a principal feature decomposition block based on the Fourier transform, a backbone and noise decomposition block based on the Stationary Wavelet Transform, a cyclic signal enhancer based on threshold comparator, and a trend extraction block based on average pooling. Secondly, we propose to change the depth-separable convolution layer mode and stack multiple depth-separable convolution layers so as to capture the developmental characteristics of the time series signal. Furthermore, a feed-forward neural network layer is set up between the depth-separable convolution layers. Then, true time series decomposition is used in the training phase to compute the multidimensional feature loss, with the aim of improving the shortcoming of the tensor decomposition that does not allow for gradient propagation. Finally, weight aggregation is used to transform the multidimensional data into univariate time series data. Experimental results on real highway inbound traffic flow datasets show that the proposed method achieves better performance than the baseline model and effectively improves the prediction accuracy. Full article

This work presents a refined theoretical and numerical framework for shaping the axial intensity of finite-energy Bessel–Gaussian beams through programmable nonlinear phase modulation. Starting from the scalar Fresnel diffraction integral, we reformulate the propagation of a Gaussian-apodized axicon beam using a dimensionally consistent stationary-phase method. This analysis directly relates the radial phase gradient to the saddle-point trajectory, phase curvature, and on-axis intensity distribution. A Gaussian phase modulation (GPM) serves as a reference design to achieve a flattop axial profile while preserving the characteristic transverse Bessel ring structure. This work is validated against beam propagation simulations and previously reported spatial light modulator (SLM) experiments, confirming its accuracy within the paraxial regime. A parametric study then clarifies the scaling of wavelength, beam waist, axicon angle, and refractive index for extended focusing. Beyond standard GPM, several alternative nonlinear phase functions are systematically compared. High-performing profiles must replicate not only the amplitude scale but, more importantly, the radial phase-gradient structure of the Gaussian reference, which governs energy redistribution from annular regions to the axis. The results identify smooth, localized nonlinear functions as promising candidates for stable flattop Bessel beam generation. The proposed framework offers a flexible optical design for applications such as through-glass via (TGV) micromachining and light-sheet illumination, while prospective high-intensity laser plasma uses remain beyond the present linear model. Full article

The research investigates lubrication characteristics of a three-stage planetary transmission system under first and second gear conditions. A whole-system CFD model and a planetary carrier bearing CFD model are established. Oil distribution is simulated using a UDF dynamic mesh technique. A dedicated test bench is designed and built for a multi-stage planetary transmission system to measure oil flow data at the outlets of each planetary stage. By comparing the simulation and experimental results, the CFD model is confirmed. The oil distribution in the planetary transmission system is followed. In the first gear condition, the oil distribution within the second stage is significantly lower than that in the other two stages, and mainly converges onto the meshing surfaces of gears. In the second gear condition, the planetary carrier remained stationary, resulting in limited oil distribution in the first stage. Meanwhile, the third-stage planetary carrier bearings exhibit insufficient oil distribution across different gear conditions. To address this issue, several structural optimization structures for the numerical model of the third-stage planetary carrier bearings are compared in terms of theoretical oil supply rates and oil volume fraction distribution characteristics. Among these, constrained by the fixed positions between the oil inlet and oil holes, the structures with different numbers of oil holes in the planetary carrier lead to an oil flow rate reduction due to flow division and pressure loss induced by turbulence at high rotational speed, failing to meet the oil demand. Optimization of oil-hole diameter enlargement, the oil flow rate increases proportionally with the hole diameter. A diameter of 5 mm satisfies the theoretical oil flow rate demand, yet an asymmetric oil distribution is observed between the two inner bearings. Building upon the initial design with two oil holes, a 5 mm diameter design, a 1 mm axial leftward offset of the oil hole position, and a 20° oil-guiding inclination on the outer hub reduce the oil distribution asymmetry between the two inner bearings from 64.5% to 13%. The oil volume fraction increases from 0.005 to 0.069 in the inner bearing and from 0.001 to 0.013 in the outer bearing, resulting in a substantial improvement in overall bearing lubrication performance. Full article

Membrane distillation (MD) is a promising technology for high-salinity water treatment, but scaling still remains a critical limitation to stable operation. This study introduces a novel approach by exploring vibration signal design as a control variable for scaling mitigation in hollow-fiber DCMD, shifting from the conventional treatment of vibration as a fixed-frequency mechanical input. The influence of different vibration modes, including fixed, random, and patterned (music-derived structured non-stationary excitation) vibrations, on CaSO₄ scaling in hollow-fiber direct contact membrane distillation (DCMD) was systematically investigated. Bench-scale experiments were conducted using synthetic saline feed (35,000 mg/L NaCl and 2000 mg/L CaSO₄) under both outside-in and inside-out configurations. The results reveal that vibration modifies flux decline behavior by delaying the critical volume concentration factor (VCF_cr) and reducing post-critical scaling kinetics. In the outside-in mode, patterned vibration achieved the highest critical VCF (3.39) and lowest scale formation rate, indicating effective suppression of nucleation and crystal growth. In contrast, fixed-frequency vibration (100 Hz) was more effective in the inside-out mode, owing to resonance-induced amplification of vibration transmissibility (>140%), which enhanced local shear at the membrane surface. Spectral analysis shows that patterned vibration provides broadband and non-stationary excitation with multiple dominant frequencies, enabling continuous disruption of scaling processes, whereas random vibration lacks structured energy distribution. Furthermore, patterned vibration reduced energy consumption by 16–23% compared to fixed and random modes while maintaining comparable or superior fouling mitigation. These findings demonstrate that vibration pattern design, coupled with system resonance characteristics, is a key factor in optimizing MD performance and energy efficiency. Full article

Marathon heat stress is an increasing public health concern under climate change, particularly for mass participation endurance events. Using 967,878 runners from 18 U.S. marathon events between 2011 and 2019, this study examined the nonlinear association between race-day thermal exposure and marathon performance. A two-way fixed effects framework was used to account for race- and year-specific heterogeneity, demographic characteristics, race-day maximum air temperature, relative humidity, their interaction, and non-stationary exposure proxies. The results identified a humidity-dependent thermal optimal zone (TOZ). At the sample mean humidity level of 77.7%, the estimated the TOZ based on the race-day maximum air temperature was 13.0 °C, with a low-penalty range of 8.5–17.4 °C for predicted losses below 60 s. In the main specification, the temperature–humidity interaction was positive, suggesting that humidity-related penalties may increase under warmer conditions; however, race-year-level sensitivity analyses indicated that this interaction should be interpreted cautiously. Under 28.0 °C and 80% relative humidity, the model predicted a finish-time penalty of approximately 737.5 s. Stratified analyses showed that mass participation runners experienced larger high-temperature penalties than elite runners, and male runners aged 35–49 years showed the highest estimated thermal sensitivity at 28.0 °C. The UTCI modestly improved model calibration but produced unstable optimum estimates, supporting its use as a complementary biometeorological benchmark rather than as the primary basis for defining a marathon-specific TOZ. These findings suggest that a marathon heat-risk assessment should jointly consider air temperature, humidity, integrated biometeorological exposure, and subgroup-specific vulnerability. Full article

Biologically related time-series data characterize the dynamic evolution of biological systems, including genetic inheritance, disease diagnosis, and the biological microenvironment. However, accurate prediction of these data remains challenging due to their pronounced time-varying, non-stationary, and noisy characteristics. Existing approaches often fail to capture latent shifts of biologically related time series, limiting both predictive performance and time-varying pattern recognition capability. Thus, in this study, we first propose a time-varying neural network (TVNN) model that combines frequency-domain information with Koopman theory. TVNN-model Koopman transition matrices are used to model global dynamics and local time-varying behaviors for pattern extraction. Secondly, a time-varying pattern recognition large language model (TVPRLLM) is introduced to recognize and classify the extracted time-varying patterns, enabling the identification of potential pattern categories. Thirdly, we have developed a biology-related time-series predictive platform that can offer visualization, data analysis, and predictive services. Experimental results demonstrate that the TVNN model outperforms existing mainstream methods in predicting biology-related time-varying time series, and that it achieves competitive forecasting performance, though its behavior depends strongly on the design of the frequency-domain decomposition. Additional robustness analyses reveal that the choice of Fourier masking strategy can materially affect both RMSE and long-horizon stability. We further show that Koopman-derived time-varying representations are highly discriminative for dynamic state recognition. Full article

The use of biodiesel blends in heavy-duty diesel engines changes the relationship between engine operating conditions, fuel properties, and exhaust emissions, which may limit the reliability of data-driven emission models trained under a single fuel condition. This study investigates the cross-fuel transferability of virtual emission sensors for a heavy-duty diesel engine operating on B7 and B20 fuel blends. The analysis was carried out for three target signals: nitrogen oxides concentration, hydrocarbon concentration, and dry carbon dioxide concentration, using data from the World Harmonized Transient Cycle (WHTC) and World Harmonized Stationary Cycle (WHSC) tests. A structured modelling workflow was developed, including signal time alignment, construction of baseline, dynamic, and memory-based features, feature selection, and separate evaluation scenarios: within-domain, cross-cycle, and cross-fuel transfer. Three tree-based regression algorithms were compared: Random Forest (RF), Histogram-Based Gradient Boosting (HGB), and Extreme Gradient Boosting (XGBoost). XGBoost achieved the best predictive performance in the source domain and was selected as the reference model. The results showed that a change in cycle characteristics led to a significant decrease in predictive performance, whereas the transition from B7/WHTC to B20/WHTC resulted in a clearly smaller drop in the evaluation metrics. The relationship between engine operating signals and emission response remained partially transferable across fuels. The highest stability was observed for carbon dioxide, intermediate stability for nitrogen oxides, and the lowest stability for hydrocarbons. The findings support the development of robust data-driven virtual sensing methods for emission monitoring and calibration of heavy-duty diesel engines operating with biodiesel blends. Full article

Contactless transmission of mechanical power, which is characteristic of coaxial magnetic gears (CMGs), offers significant advantages over conventional mechanical gears, in particular, reduced maintenance frequency and inherent overload protection. At the same time, there is a lack of design methodologies for this type of gear based on the analysis and systematization of experience gained from already implemented designs. This paper presents a method for determining the maximum magnetic torques of CMGs on the basis of an equivalent magnetic-circuit model. The error associated with the proposed methodology does not exceed ±15%, which enables the influence of geometric parameters and the magnetic properties of materials on the key performance indicators of the gear to be assessed already at the preliminary design stage. A mathematical model of CMG dynamics was also developed, based on a quasi-stationary two-dimensional approximation of the magnetic field, accounting for the geometry of the magnetic circuit, the spatial distribution of the magnetic vector potential, and magnetic-circuit saturation. The proposed mathematical model was verified using the results of physical experiments. The discrepancy between the calculated and experimental values of the torque on the low-speed shaft in the steady state does not exceed 5.5%. Based on the optimization procedure, the dependence of the maximum linear torque density on the outer diameter of the CMG, the number of poles of the high-speed rotor, and the transmission ratio was determined. It was shown that, as the number of poles increases, the linear torque density also increases and, for example, for diameters of approximately 800 mm, it may exceed 100 N·m/m. Full article

Compressed sensing (CS) applied in the ambiguity domain offers an effective approach for recovering time–frequency distributions (TFDs) of non-stationary signals from sparse representations. Existing methods predominantly rely on the Wigner–Ville distribution (WVD) as the initial representation due to its simplicity and high auto-term concentration. However, the WVD performs poorly for signals with higher-order frequency-modulated (FM) components and is highly sensitive to interference and noise, which then propagate into the reconstruction. This paper introduces the systematic use of the L-Wigner distribution (LWD) as the initial representation for CS-based reconstruction, providing front-end adaptability to signal characteristics. By generating a controllable family of TFDs ranging from the spectrogram to cross-term-free polynomial WVDs, the LWD enables effective suppression of interference and noise while simultaneously enhancing auto-term localization for nonlinear FM components. The proposed LWD-based reconstruction framework is evaluated against the conventional WVD-based method using several state-of-the-art reconstruction algorithms, whose parameters are jointly optimized through a multi-objective meta-heuristic framework to ensure a fair comparison. Experiments on noisy synthetic signals and real-world gravitational-wave data demonstrate consistent performance gains. On synthetic signals, the proposed approach reduces the average reconstruction error index by up to 36.6%, improves the ${\ell }_1$-reconstruction error by up to 75.8%, and achieves complete elimination of cross-term energy. In addition, robustness analysis under additive white Gaussian noise shows up to a 75% improvement in ${\ell }_1$ performance. For real gravitational-wave data, the method reduces the mean reconstruction index by up to 5.8% while maintaining auto-term preservation and eliminating cross-term artifacts. These results establish the LWD-based initialization as an effective and general strategy for TFD reconstruction in complex signal environments. Full article