Search Results (54,896)

Conventional small-signal security assessment requires repeated power flow calculations, system linearization, and eigenvalue analysis, resulting in considerable computational burden for online applications. Moreover, existing data-driven methods generally provide insufficient representation of branch power flows, line operating states, and incomplete PMU measurements. To address these limitations, this paper proposes an edge-enhanced graph attention network (EE-GAT) for real-time small-signal security assessment. The model takes bus voltage magnitude, phase angle, active and reactive power injections, and PMU availability masks as node features, while branch active and reactive power flows and line operating states are incorporated as edge features. An edge-enhanced multi-head attention mechanism is employed to jointly learn node state and branch coupling information, and a global attention readout mechanism is used to construct the system-level representation for secure/insecure classification. Tests on the New England 39-bus and NPCC 140-bus systems show that EE-GAT achieves accuracies of 97.75% and 96.47%, respectively, outperforming XGBoost, LSTM, CNN, GCN, and conventional GAT. The proposed model also maintains superior performance under incomplete PMU measurements and requires only 0.0034 s and 0.0051 s for the corresponding test batches, demonstrating its potential for online small-signal security screening. Full article

This study combines ligand- and structure-based in silico strategies to predict the inhibitory activity of natural flavonoids on the Polo-Like Kinase-1 (PLK-1) enzyme as candidate anticancer agents. This enzyme participates in mitosis and is overexpressed in cancer cells. Furthermore, it has been shown to have important implications for tumor metastasis, and its inhibitors are attractive starting points for drug development. First, classification models are developed using linear discriminant analysis and a multilayer perceptron neural network. Models with accuracy greater than 80%, validated using standard statistical performance metrics and applicability domain, are used for virtual screening identifying four compounds as potential antitumor drugs. Subsequently, the identified compounds are evaluated using a molecular docking methodology to verify their binding mode and interactions with the catalytic domain of PLK-1. Finally, the integration of molecular dynamics simulations, at 300 ns, with Molecular Mechanics/Generalized Born Surface Area (MM/GBSA) thermodynamic calculations demonstrates that the hydroxylation pattern of ring B in the flavonol scaffold is the fundamental chemical-structural determinant of electrostatic interactions and the architecture of water-mediated networks. Among the evaluated flavonoids, myricetin showed the most favorable overall computational profile, including the highest virtual-screening score and the most favorable mean MM/GBSA estimate, supporting its prioritization for experimental evaluation as a potential PLK-1 inhibitor. The integration of these approaches offers a robust methodological framework for proposing candidates with a higher probability of success, in subsequent stages of experimental validation, reducing time and costs in the early stages of drug development. Full article

Sales forecasting plays a critical role in business decision-making, including financial planning, resource allocation, and inventory management. Despite the growing adoption of machine learning in forecasting applications, controlled evaluations of how preprocessing decisions and contextual feature availability affect model performance remain limited. This study presents a controlled benchmarking evaluation that systematically examines preprocessing strategies and contextual feature availability across seven representative tree-based machine learning models, using two publicly available retail datasets: BigMart and Rossmann. Multiple preprocessing configurations are evaluated on both datasets, and ablation studies are conducted on the Rossmann dataset to quantify the independent contribution of seasonal and promotional features and the Customers variable to forecasting accuracy. Results show that GradientBoosting achieved the best performance on BigMart (R² = 0.611), while RandomForest and ExtraTrees led on Rossmann ${\mathrm{R}}^2=0.884$ under realistic conditions excluding the Customers variable, and ${\mathrm{R}}^2=0.965$ in an ablation scenario where Customers, typically unavailable at forecast time, is included. Ablation analysis confirmed that seasonal and promotional features are critical drivers of forecasting accuracy, and that the Customers variable is the single most influential predictor in the Rossmann dataset. These findings highlight the importance of preprocessing design and contextual feature availability in retail sales forecasting and provide practical insights for building effective forecasting pipelines. Full article

Continuous and noninvasive body fluid monitoring has attracted increasing attention in personalized healthcare, chronic disease management, and wearable point-of-care testing. Fiber-optic surface plasmon resonance (SPR) biosensors are particularly promising for this purpose because they combine label-free and real-time with miniaturization and low sample volume requirements. However, current body fluid sensing technologies and conventional bare metal SPR interfaces still face critical challenges, including insufficient analytical accuracy in complex biofluids, broad resonance linewidths, weak signal readability for trace biomarkers, and mechanical perturbations during wearable operation. These limitations highlight the need for nanomaterial-engineered fiber-optic SPR platforms that can convert interfacial molecular events into stable and sensitive signals. The review summarizes recent progress in nanomaterial-enabled fiber-optic SPR biosensors for continuous body fluid monitoring. Emphasis is first placed on nanomaterial mediated local electromagnetic field enhancement and plasmonic mode regulation. Subsequent discussion focuses on their functions in interfacial recognition, analyte enrichment, rapid mass transport, antifouling protection, and flexible integration for continuous operation. On this basis, representative sensing targets, material strategies, and device architectures for tears, urine, exhaled breath condensate, saliva and sweat are systematically analyzed. Finally, current challenges and future opportunities are discussed from the perspective of sensing reliability, wearable integration, and real sample validation. Full article

The rapid expansion of photovoltaic (PV) systems poses significant challenges to grid stability. Hybrid Energy Systems (HES) are intended to alleviate this volatility, yet their coordinated dispatch often remains suboptimal due to communication delays and ramp-rate constraints. Accurate ultra-short-term PV power forecasting is therefore essential, as it enables preemptive control and timely dispatch adjustments that unlock the full potential of HES. In this study, we propose a novel AI hybrid forecasting framework that integrates a rule-based model with a Decomposition Linear (DLinear) Long Short-Term Memory (LSTM) deep learning core, representing, to the best of our knowledge, a novel integration of a decomposition-based linear model (DLinear) with LSTM networks for ultra-short-term PV power forecasting. The DLinear component decomposes the time series into trend and remainder sequences, which are then independently modeled by separate LSTM networks to capture distinct dynamics. Using data from a 300 kWp PV power station, the framework achieves an average daily prediction accuracy exceeding 93% for both 5-min and 15-min horizons. The model reliably tracks power variations under sunny and rainy conditions, while under volatile cloudy weather its accuracy decreases but still captures essential fluctuation patterns. These results demonstrate the potential of the proposed framework for improving the dispatch and operational reliability of hybrid energy systems. However, further validation across additional seasons and sites is needed to establish broader generalizability. Full article

Ferroelectric field-effect transistors, which rely on ferroelectric polarization reversal to modulate the channel resistance, hold great promise for nonvolatile memory and neuromorphic computing. The polarization switching dynamics are critical for achieving high-speed, high-bit-density neuromorphic hardware. Here, we report a 1D-2D asymmetric heterojunction composed of a single-element tellurium (Te) nanowire and a magnetic metal, Fe₃GaTe₂. Piezoresponse force microscopy reveals reversible polarization switching at room temperature. Utilizing this ferroelectric heterojunction, we construct ferroelectric semiconductor field-effect transistors that exhibit tunable resistance states exceeding 7 bits, featuring an on/off ratio of 10³, a retention time exceeding 10³ s, and ultrafast switching down to 20 ns. Moreover, the transistor enables accurate recognition of six kinds of micro-defects with an accuracy of 97.1% on the workpiece surface by convolutional neural network. This work establishes the intrinsic relationship between ferroelectric polarization and resistance modulation, providing a device platform for next-generation multilevel storage and ultrafast neuromorphic computing networks. Full article

In busy airports, departure metering (DM) can reduce unnecessary runway-area queueing by holding aircraft at their stands, but its effectiveness is highly sensitive to the accurate estimation of taxi time. Existing studies generally pursue two separate approaches to improve the effectiveness of DM decisions: (1) improving the accuracy of taxi-time prediction, and (2) incorporating taxi-time uncertainty to make more robust decisions. However, prediction errors remain unavoidable, and uncertainty sets constructed directly from historical data of multiple flights fail to capture the heterogeneity between flights, so the previous approaches always result in suboptimal decisions. Thus, this research develops a learning-based two-stage stochastic DM that constructs the uncertainty sets from the outputs of machine-learning predictors and makes robust decisions under uncertainty. This framework can capture the uncertainty set for each flight more accurately and effectively reduce the influence of prediction errors on DM decisions. The DM model is formulated as a two-stage stochastic program integrating first-stage pushback decisions and second-stage runway sequencing, solved by the sample average approximation and scenario generation method. The proposed framework is evaluated using real-world data from Tokyo International Airport (RJTT). The results show that the framework can effectively reduce the overall queueing level and right-tail risk of runway-area queueing time compared with the certainty model, with reduced expected queueing time during peak hours. Additional equity and sensitivity analyses further show that these improvements are realized without being achieved at the expense of a small subset of flights and remain stable under schedule perturbations. This study demonstrates that learning-based uncertainty representations can enhance the robustness and operational quality of DM decisions. Full article

This paper presents a low-cost embedded monitoring system for real-time RMS voltage and RMS current acquisition in three-phase electrical networks. The proposed architecture is based on distributed Arduino Nano acquisition nodes equipped with ACS712 Hall-effect current sensors and isolated voltage transformers, while a Raspberry Pi 4 Model B is used as a centralized data acquisition and processing unit through the I²C communication protocol. The embedded acquisition nodes implement timer-controlled analog signal sampling using the internal 10-bit ADC of the ATmega328P microcontroller, allowing real-time acquisition of electrical waveforms for RMS computation. Unlike conventional low-cost IoT electrical monitoring systems focused mainly on basic parameter visualization and wireless communication, the proposed platform emphasizes synchronized three-phase RMS monitoring and experimental validation accuracy under real operating conditions. The proposed monitoring architecture is experimentally benchmarked against a FLUKE 435 professional power quality analyser used as a high-accuracy reference instrument. Experimental results demonstrate that the proposed low-cost embedded architecture can provide RMS voltage and RMS current measurements with acceptable accuracy for educational applications, experimental electrical platforms, and distributed IoT-based monitoring systems. The presented system does not aim to implement a fully IEC 61000-4-30-compliant power quality analyser but rather to validate the feasibility of low-cost embedded RMS monitoring architectures for real-time electrical applications. Full article

To address the recognition challenges caused by the high proportion, low resolution, and significant multi-scale variations of small objects in UAV small-object detection tasks, a UAV small-object detection and recognition algorithm based on CEF-YOLOv8n is proposed. The proposed algorithm uses YOLOv8n as the baseline network and introduces a Partial Convolution-based Cross Partial Feature (CPF) module into the backbone network to enhance the local feature extraction capability for low-resolution small objects. In the neck network, the concept of feature focusing and diffusion is adopted to construct a Focusing Generalized Feature Pyramid Network (FGFPN). A Feature Semantic Fusion Module (FSFM) based on a cross-attention mechanism is designed to complementarily fuse shallow detail features with deep semantic features, thereby enhancing information interaction among objects at different scales. In addition, a Lightweight Weight-Sharing Detection Head (LWSD) is proposed to improve the computational efficiency and real-time performance of the model while maintaining detection accuracy. Publicly available datasets are used for network training and detection evaluation, and comparative experiments are conducted with other algorithms. The results show that the proposed detection and recognition algorithm achieves 37.6% and 22.6% in terms of mAP50 and mAP50-95, respectively, representing improvements of 3.4 and 2.6 percentage points over the original YOLOv8n. Meanwhile, the number of parameters and FLOPs are reduced from 3.2 M and 8.7 G to 2.5 M and 6.9 G, respectively. Full article

Stock markets are complex adaptive systems whose feature dependencies evolve over time, and identifying which signals genuinely drive returns is fundamentally a problem of information. Traditional feature selection methods and linear models rely on statistical correlations and overlook causality, which weakens their predictive performance in non-linear, evolving markets. This paper develops an information-theoretic approach to causal feature selection based on Markov Blanket discovery. We apply the Iterative Parent–Child-based search of Markov Blanket (IPCMB), whose conditional-independence tests are conditional mutual information measures, to recover the Markov Blanket of next-month returns—the minimal feature set that carries all Shannon information about the target. We then pair it with a Classification and Regression Tree (CART), whose impurity-based splitting admits an information-gain interpretation, to predict the direction of Chinese A-share returns non-linearly. Using data on 2760 stocks, IPCMB selects 13 causal features from 72 candidates, and CART forecasts whether the next month’s return is positive. The empirical results show an accuracy of 58.0%, 7.7 percentage points above the all-features CART benchmark, and a 13-month cumulative return 35.88% higher than that of the CSI 300 index. The findings indicate that selecting features according to their causal information content and combining them with interpretable tree-based prediction can support more reliable investment decisions in emerging markets. Full article

Short-term wastewater flow forecasting is essential for pump-station scheduling, influent-load regulation, and overflow-risk warning in wastewater treatment systems. However, wastewater flow series often exhibit strong nonlinearity, large-amplitude fluctuations, and sensitivity to model hyperparameters, which makes accurate real-time prediction challenging. To address these issues, this study proposes an attention-enhanced hybrid deep learning framework optimized by an Improved Harris Hawks Optimisation algorithm. The prediction target is the wastewater flow value at the next target time step, and the model uses only historical wastewater flow observations as input. The framework integrates a Temporal Convolutional Network to extract local fluctuation and multi-scale temporal features, a Bidirectional Long Short-Term Memory network to capture contextual dependencies within the historical input window, and a Simple Attention Module to recalibrate high-dimensional features and emphasize key information. The improved optimisation algorithm further enhances hyperparameter search through hybrid initialization, early-stage differential evolution mutation, stagnation detection with diversity injection, and late-stage Lévy flight perturbation. Experimental results show that the unoptimized Temporal Convolutional Network (TCN)–Bidirectional Long Short-Term Memory (BiLSTM)–Simple Attention Module (SimAM) model achieved an R² of 0.9697, an RMSE of 231.1821, and an MAE of 177.5324 on the validation set. After Improved Harris Hawks Optimization (IHHO) optimization, these values improved to 0.9728, 218.8228, and 161.8777, respectively. Compared with the unoptimized model, Root Mean Square Error (RMSE) and Mean Absolute Error (MAE) were reduced by approximately 5.35% and 8.82%, respectively, while R² increased by 0.0031. These results indicate that the proposed framework improves prediction accuracy, stability, and robustness under the tested conditions. It provides a feasible technical solution for short-term wastewater flow forecasting and supports intelligent operation management of wastewater treatment systems. Full article

Underground distribution networks increasingly rely on aging cable assets whose parameter drift modifies propagation velocity, attenuation, and fault-initiated transient signatures, thereby reducing the reliability of conventional traveling-wave (TW) fault location and purely data-driven diagnosis. This paper proposes a physics-regularized hybrid learning framework for joint fault-type classification, feeder-area identification, and continuous fault localization in aging underground distribution feeders. The methodology integrates (i) an aging-aware simulation pipeline driven by a normalized aging-stress index $\alpha \in [0,1]$ that perturbs the per-unit-length cable matrices within a controlled domain; (ii) synchronized multi-sensor time–frequency representations of three-phase voltage and current transients; (iii) an area-aware multi-task architecture with fault-type and area-classification heads and area-specific local regression heads; and (iv) a propagation-consistency loss that depends explicitly on the model-predicted fault position and therefore contributes gradients during training. A branched underground feeder is evaluated using five synchronized sensing locations and a stratified dataset of $N_{\mathrm{tot}}=36,000$ simulated fault events covering 11 fault classes (SLG-A/B/C, LL-AB/BC/CA, DLG-ABG/BCG/CAG, LLL, and LLLG), six non-overlapping feeder areas, fault resistance $R_f\in [0.1,50]\Omega$, measurement noise $\mathrm{SNR}\in [20,40]\mathrm{dB}$, and a $20\mathrm{ms}$ transient window sampled at $200\mathrm{kHz}$. On a held-out test set of 7200 previously unseen event records drawn from the same simulation domain, the Hybrid model achieves a fault-type accuracy of $0.93$, an area-identification accuracy of $0.96$, a localization MAE of $0.011$ p.u., and a 95th-percentile absolute error of $0.027$ p.u. The proposed configuration outperforms the TW-TOA, purely data-driven Baseline, and physics-regularized PINN references across the reported diagnostic tasks within the prescribed simulator and parameter ranges. Time–frequency attribution is included only as a qualitative interpretability illustration and is not treated as quantitative evidence of explanation faithfulness. Accordingly, the results demonstrate comparative in-domain simulation performance rather than field or cross-simulator deployment readiness. Full article

Accurate thermospheric mass density is crucial for precision orbit determination, trajectory prediction, and collision warning in the increasingly congested Low Earth Orbit (LEO) region. Traditional high-precision instantaneous inversion methods rely on dedicated onboard accelerometers (e.g., CHAMP and GRACE satellites), which are costly, scarce, and unable to meet the expansive monitoring needs of modern mega-constellations. In recent years, orbit data inversion methods based on high-precision orbit determination data have been extensively studied. Among them, the interpolation–differentiation method, although capable of retrieving instantaneous atmospheric density, severely amplifies position errors and measurement noise, significantly limiting retrieval accuracy. To address this gap, this study proposes a novel atmospheric density retrieval scheme for accelerometer-free LEO satellites, introducing an orthogonal polynomial (Legendre and Chebyshev) sliding-window fitting approach. The proposed method is evaluated using high-precision simulated orbit data. Under instantaneous inversion conditions, the Legendre method yields an average relative error (MRE) of 16.55%, a normalized root mean square deviation (NRMSD) of 12.80%, and a correlation coefficient R of 0.9370. To further evaluate its potential in climatological modeling, this paper also proposes a local solar time (LST) bin-averaging method to assess atmospheric inversion accuracy. Under this LST averaging approach, the inversion performance achieves an MRE of 6.83% and a correlation coefficient R exceeding 0.98. This demonstrates the feasibility of the inversion method for analyzing diurnal effects on atmospheric field evolution. Full article

Breakthrough-curve prediction is central to fixed-bed biosorption column design, but data-driven modeling remains challenging with small, unevenly distributed datasets. This study compares a feedforward multilayer perceptron (MLP) with recurrent long short-term memory (LSTM) and bidirectional LSTM (Bi-LSTM) networks for simultaneous prediction of hexavalent chromium [Cr(VI)] and total chromium breakthrough curves in columns packed with Quercus crassipes acorn shell. A dataset of 14 experimental curves was used, covering variations in pH, influent concentration, flow rate, bed height, and biosorbent mass. Model performance was assessed using Mahalanobis distance (D_M)-based holdout selection, curve-shape evaluation, inter-model coefficient of variation (CoV), and SHAP attribution. All architectures achieved high internal cross-validation accuracy (R² = 0.9879–0.9979; RMSE = 0.0149–0.0309). However, complete-curve holdout performance depended on the excluded condition and chromium response. Under the central holdout condition (D_M = 0.317), R² remained high for both outputs (0.954–0.990). In non-central holdouts, errors also affected curve shape, including breakthrough timing and the approach to saturation. The largest deterioration occurred for Cr(VI) under the peripheral low-flow condition, where the MLP decreased to R² = 0.325, while LSTM and Bi-LSTM retained higher agreement. SHAP attribution confirmed process-consistent effects, with influent concentration and flow rate as dominant predictors and pH contributing more strongly to Cr(VI). These findings show that recurrent architectures can provide more robust representations of complex breakthrough profiles, while CoV–D_M analysis helps identify regions where additional experiments are needed. Full article

The BeiDou-3 Navigation Satellite System (BDS-3) broadcasts multi-frequency signals, including B1C, B2a, B1I, and B3I, offering a new technical approach for tilt monitoring of building structural safety. However, in building structural safety tilt monitoring based on the BDS-3 single-epoch algorithm, the engineering performance differences between the B1C/B2a new signal combination and the B1I/B3I traditional signal combination—in terms of monitoring accuracy, ambiguity fixing rate, computational efficiency, and tilt rate—have yet to be fully validated through comparative analysis. To address this issue, this paper proposes a building structural safety tilt monitoring method based on the BDS-3 single-epoch algorithm and conducts a field test on a multi-story building in Beijing. First, a BDS-3-based kinematic monitoring model is established, and an integer ambiguity error search band method based on the main and auxiliary frequencies is proposed. On this basis, three schemes are designed using medium Earth orbit (MEO), inclined geosynchronous orbit (IGSO), and geostationary Earth orbit (GEO) satellites, B1C/B2a (MEO/IGSO), B1I/B3I (MEO/IGSO), and B1I/B3I (MEO/IGSO/GEO), to comparatively analyze the accuracy, ambiguity fixing rate, computational efficiency, and measured tilt results of each scheme in building structural safety tilt monitoring. Experimental results show that all three schemes based on the BDS-3 single-epoch algorithm achieve millimeter-level monitoring accuracy and an ambiguity fixing rate exceeding 99%, with average computational times of 0.351 s, 0.338 s, and 4.572 s and corresponding building tilt rates of 0.22‰, 0.20‰, and 0.18‰, respectively, yielding an average tilt rate of 0.20‰. These results satisfy the 4‰ limit specified in the Code for Deformation Measurement of Building and Structure (JGJ 8-2016), thereby confirming the feasibility and effectiveness of the proposed method and offering a novel BDS-3 single-epoch algorithm for building tilt monitoring. Full article