Search Results (15,664)

This paper investigates the optimization of maintenance strategies to improve the reliability of equipment at the Kapshagay Hydropower Plant (HPP), located in Kazakhstan. Operational data for the period 2020–2025 were analyzed to evaluate the effectiveness of existing maintenance systems. The analysis showed that the failure frequency of the main equipment averaged 3.8–4.2 events per year, while annual unplanned downtime reached 80–100 h, resulting in electricity generation losses of 2.5–3.2%. In addition, total maintenance costs were approximately 150 million KZT per year, with about 40% related to unplanned repairs. A reliability-centered maintenance model was developed using mathematical modeling and simulation tools such as Python 3.11 and SMath Solver 0.99.7920. The study integrates reliability theory, exponential failure modeling, and statistical performance analysis based on operational data from the Kapshagay HPP. Simulation-based validation was performed to compare baseline and optimized maintenance strategies under real operating conditions. After implementing the proposed model, equipment failure probability decreased by 15%, failure rate decreased by 28%, the mean time between failures increased from 120 days to 165 days, and repair duration decreased from 6 days to 4 days. Additionally, failure probability decreased from 0.10 to 0.07, while annual downtime decreased from 6.2 days to 4.1 days. Electricity generation losses decreased by approximately 18–22 GWh per year, while the annual economic benefit was estimated at 320–480 million KZTn. The results demonstrate that reliability-centered maintenance can increase equipment reliability by 20–30%, reduce maintenance costs by 10–12%, and improve electricity generation efficiency by 1.8–2.4%. The obtained results have practical significance for improving the technical and economic performance of hydropower plants. Full article

Reliable process monitoring and quality evaluation for resistance spot welding (RSW) have become more important now than ever. An ultrasonic probe embedded into welding electrodes has enabled the acquisition of data about molten pool formation throughout welding, but automation of high-performance ultrasonic data analyses is still necessary to fully realize a monitoring system. This work proposes a two-stage deep learning (DL) approach for automated ultrasonic data analysis for RSW processing monitoring. The first stage conducts semantic segmentation on ultrasonic M-scan welding process signatures, yielding masks for identified molten pool and stack regions from which weld penetration measurements can be directly extracted, as well as expulsion occurrences throughout welding. From input images and segmentation outputs, the second stage directly estimates resultant weld nugget diameters using an additional neural network. Both stages leveraged architectures based on TransUNet, mixing elements of both convolutional neural networks (CNN) and vision transformers, and the effect of cross-attention for stack-up sheet thickness data fusion was investigated via an ablation study. Additionally, in the diameter estimation stage, the ablation study included alternative feature extraction architectures in the network and investigated the provision of M-scans to the model alongside segmentation masks. In both cases, cross-attention was determined to improve performance, and in the case of diameter estimation, providing M-scans as input was found to be beneficial in general. With cross-attention, the segmentation approach yielded a mean intersection over union (IoU) of 0.942 on molten pool, stack, and expulsion regions in the M-scans with 13.4 ms inference time. With cross-attention, diameter estimates yielded a mean absolute error of 0.432 mm with 4.3 ms inference time, representing a significant improvement over algorithmic approaches based on ultrasonic time of flight. Additionally, the approach attained >90% probability of detection (POD) at 0.830 mm below the acceptable diameter threshold and <10% probability of false alarm (PFA) at 0.828 mm above the threshold. These results demonstrate a novel production-ready application of DL in ultrasonic nondestructive evaluation (NDE) and pave the way for zero-defect RSW manufacturing. Full article

By optimizing the carburizing heat treatment process, the grain size of the carburized layer of M50NiL steel was successfully refined to the sub-micron level. The mechanism for the generation of a large number of sub-micron crystal regions (SMCR) is that dislocations are entangled and linked due to the pinning effect of nanometer-sized carbides. In this study, a stacking fault energy (SFE) model for austenite in M50NiL steel was established. First-principles calculations were employed to investigate the effects of alloying elements, as well as the position and quantity of carbon (C) atoms, on the generalized stacking fault energy (GSFE). The variations in SFE were further analyzed in combination with differential charge density calculations. The simulation results revealed that the addition of alloying elements excluding nickel led to a reduction in the unstable stacking fault energy. Differential charge density analysis indicated that this decrease was associated with the weakening of Fe–Fe bonds in the L0 layer, where stacking faults occurred. When C atoms are interstitially dissolved near the L0 layer, the Fe–Fe bonds near the L0 layer are enhanced, and the unstable stacking fault energy is correspondingly increased. Compared with the pure iron system, the combined effect of alloying elements and C atoms in M50NiL steel maintained a relatively low level of both the unstable stacking fault energy and the stacking fault formation barrier, provided that C atoms were not dissolved in the L1 layer. This condition was favorable for dislocation slip. Meanwhile, the stable stacking fault energy significantly increased, enhancing the stability of austenite. Based on these simulation results, the relationship between the GSFE of austenite in M50NiL steel and the formation of subgrains and twins within the submicron crystalline regions of the carburized layer was discussed. Full article

The sale of raw milk via vending machines represents a well-established distribution model in many European countries, including Switzerland. As part of this study, data on the microbiological quality of raw milk sold via vending machines in Switzerland were collected. A total of 124 raw milk samples from 124 raw milk vending machines across Switzerland were analysed. In addition to standard hygiene parameters (TVC and E. coli), the scope of the investigation particularly included foodborne pathogens as well as methicillin-resistant Staphylococcus aureus (MRSA) and extended-spectrum β-lactamase (ESBL)-producing Enterobacterales. Isolates were further characterised by whole-genome sequencing. Shiga toxin-producing Escherichia coli (STEC) were detected in 3.2%, Staphylococcus aureus was detected in 12.1%, Listeria monocytogenes was detected in 2.4%, Campylobacter spp. were detected in 1.6%, Yersinia enterocolitica was detected in 29.8%, and Salmonella spp. were detected in 0% of the samples. MRSA and ESBL-producing Enterobacterales were each detected in 0.8% of samples. The results highlight the potential risk of foodborne infections associated with the consumption of untreated raw milk, as well as hygiene deficiencies linked to several raw milk vending machines. Based on the generated data, the importance of the requested heat treatment of raw milk in Switzerland is clearly underscored. Furthermore, more precise and binding guidelines for self-monitoring and the management of raw milk vending machines appear necessary. Full article

Triple-negative breast cancer (TNBC) remains one of the most aggressive breast cancer subtypes and is associated with limited therapeutic options, underscoring the urgent need for novel treatment strategies. In this study, a library of seventeen 1,3,5-triazine derivatives potentially targeting TNBC was developed using an activity-based approach. Compounds were synthesized via an ultrasound-assisted protocol, providing an efficient and environmentally friendly methodology. The synthesized library was evaluated in vitro against the human TNBC cell lines MDA-MB-468, MDA-MB-231, and Hs578T, as well as the non-tumorigenic epithelial cell line MCF10A. Compounds 9 and 17 exhibited the most promising antiproliferative activity against TNBC cell lines (MDA-MB-468: IC₅₀ = 36.62 µM for 9 and 38.29 µM for 17; MDA-MB-231: IC₅₀ = 37.32 µM for 9 and 32.86 µM for 17; Hs578T: IC₅₀ = 57.26 µM for 9 and 34.87 µM for 17), while maintaining acceptable selectivity toward non-cancerous cells. The lead compounds were further assessed in vivo using a Danio rerio model to evaluate general toxicity and cardiotoxicity. In addition, ADME parameters were predicted for all compounds using biomimetic chromatography. Overall, compounds 9 and 17 emerged as promising small-molecule candidates for TNBC treatment, requiring further toxicological evaluation in more human-relevant in vivo models. Full article

Dust pollution has emerged as a critical issue in a wide range of industrial applications, creating an urgent demand for effective strategies to mitigate particle deposition. Recent experimental studies have demonstrated that superhydrophobic coatings represent a promising class of self-cleaning materials, primarily attributed to their hierarchical rough structures and intrinsically low surface energy. Nevertheless, the underlying self-cleaning mechanisms of superhydrophobic surfaces have not yet been fully elucidated. This work examines particle deposition on superhydrophobic surfaces featuring stochastic roughness distributions through computational modeling. Surface topographies were generated using Fast Fourier Transform techniques. An integrated lattice Boltzmann–discrete element method (LBM–DEM) framework simulated particle transport in superhydrophobic-coated channels. Particle–fluid coupling was achieved via the immersed moving boundary approach, while particle–surface interactions employed a modified Johnson–Kendall–Roberts (JKR) adhesion model. Parametric studies quantified effects of particle size, interfacial energy, flow Reynolds number, and topographical statistics on deposition dynamics. Experimental validation demonstrates good agreement between numerical predictions and measurements. Smaller particles exhibit a lower tendency to deposit on superhydrophobic surfaces, whereas increasing surface energy significantly enhances particle deposition due to stronger adhesion forces and the suppression of particle resuspension. In addition, higher Reynolds numbers effectively reduce particle deposition. The revealed self-cleaning mechanisms provide theoretical guidance for the design of high-performance self-cleaning coatings, and the identified effects of particle and surface parameters offer practical insights for anti-pollution engineering applications. Full article

Foundation models, known as the large-scale, pretrained models capable of generalizing across diverse tasks, have significantly advanced the field of medical image analysis. While most early applications focused on 2D modalities, the unique challenges and opportunities associated with volumetric medical imaging have recently attracted growing interest. This study provides a comprehensive overview of the current landscape of foundation models tailored for volumetric medical image analysis, with a focus on CT, MRI, and PET imaging. We examine key components of these models, including 3D architectures, training strategies, and supported modalities. In addition, we highlight their contribution to major clinical tasks such as classification and prediction, segmentation, image registration, quality enhancement, and visual question answering. Critical challenges of these models, including high computational cost, limited and less diverse 3D datasets, and domain adaptation, are discussed alongside the promising solutions and future research directions. By synthesizing recent advances in volumetric foundation models and outlining key technical and clinical challenges, this review provides a thorough roadmap toward the development of scalable, generalizable, and clinically applicable AI systems for volumetric medical images. Full article

This study proposes a supervised machine learning framework for vibration-based fault diagnosis of rotating machinery using integrated data-driven and physics-informed feature sets. A dataset acquired under variable load and multiple operating conditions was used for model training. Parallel signal processing techniques were applied to capture fault-related information across multiple frequency bands including time-domain analysis, frequency-domain analysis, baseband analysis, and envelope analysis. From the corresponding signal representations, statistical, spectral, and physics-based features associated with characteristic fault frequencies were extracted and combined into integrated feature sets. The diagnostic performance of models trained using purely data-driven features was systematically compared with models incorporating integrated data-driven and physics-informed features. Support Vector Machine, Random Forests, Gradient Boosting, and an ensemble classifier were evaluated using accuracy, precision, recall, and F1-score metrics. The proposed framework employs a two-layer classification strategy, where the first layer performs multiclass fault identification, while the second layer evaluates the presence of imbalance as a coexisting fault. In addition, the influence of different feature groups as well as individual measurement axes and their combinations on diagnostic performance were analyzed. Validation using a new dataset measured in laboratory conditions confirmed the robustness and generalization capability of the proposed diagnostic framework. Full article

The nature of nonlinear dynamics of financial markets results in fractal geometry and chaotic behavior that can be viewed on a variety of scales in time. This paper conducts research on the fractal characteristics of the stock network and its contribution to the price of assets based on the Fractal Market Hypothesis (FMH). A multiscale network centrality measure is built based on high-frequency return dependencies to measure the self-similar, scale-invariant nature of inter-stock dependencies. The network factor and portfolio returns are then broken down with the wavelet packet decomposition (WPD) to obtain frequency-domain profiles, which characterize the variability of risk transmission in relation to investment horizons. The profiles are consistent with scaling properties of fractal, but the decomposition does not identify causal pathways on its own. Estimation of fractal dimension by use of the box-counting technique aided by the Hurst exponent analysis reveals that the A-share of China market exhibited long-range dependence and multifractal scaling. Network factor has the largest explanatory power in mid-frequency between the D5 and D6 bands of 32 to 128 days. This intermediary frequency concentration is consistent with the hypothesis of heterogeneous markets, in which the groups of investors with varying time horizons generate scale-related price dynamics. The addition of the network factor to a 6-factor specification lowers the GRS under the 5-factor specification by 31.45 to 17.82 on the same test-asset universe, indicating better cross-sectional coverage in the sample. The estimates of the Lyapunov exponents (0.039) as well as the correlation dimension ($D_2=4.7$) confirm the presence of low-dimensional chaotic processes of the network factor series, but these values are specific to the Chinese A-share market over the 2005–2023 sample period. These results provide a frequency-disaggregated use of network-based factor modeling and suggest that it can be applicable in multiscale portfolio risk management where the investor horizon is not uniform. Full article

Automated Short Answer Grading (ASAG) has garnered significant attention in the field of educational technology due to its potential to improve the efficiency, scalability, and consistency of student assessments. This study introduces a novel dataset of 651 student responses from a Database Transaction course exam at Beni-Suef University, referred to as the Beni-Suef Transaction Processing (BeSTraP) dataset. The BeSTraP is specifically designed to support ASAG evaluation. To assess ASAG performance, five approaches were employed: string-based similarity, semantic similarity, a hybrid of both, fine-tuning transformer-based models, and the application of Large Language Models (LLMs). The experimental results indicated that fine-tuned transformers, particularly GPT-2, achieved the highest Pearson correlation with human scores (0.8813) on the new dataset and maintained robust performance on the Mohler benchmark (0.7834). In addition to grading, the framework integrates automated feedback generation through LLMs, further enriching the assessment process. This research contributes (i) a novel, domain-specific dataset derived from an actual university examination, (ii) a comprehensive comparison of traditional and transformer-based approaches, and (iii) evidence of the efficacy of fine-tuned models in providing accurate and scalable grading solutions. The created dataset will be publicly available for the community. Full article

Underwater imaging is generally subject to complex degradation issues such as color distortion, contrast degradation, and detail blurring due to the selective absorption and scattering of light wavelengths by water. Existing deep learning methods have limitations in the collaborative optimization of local details and global color. To address this issue, this paper proposes a multi-scale enhancement network based on global and local feature fusion. By integrating the advantages of CNN and Transformer, it achieves joint optimization of global color correction and local detail enhancement. Specifically, MGLFNet extracts global and local features of the image through the global and local feature fusion block in the core component of the multi-scale convolution–Transformer block and performs dynamic fusion. Meanwhile, to extract features at different scales to enhance performance, we design a multi-scale convolution feed-forward network. Through the action of the fusion module and the feed-forward network, a color-rich and detail-clear enhanced image is obtained. A large number of experimental results show that MGLF-Net outperforms comparison methods in both qualitative and quantitative evaluations of visual quality, with PSNR and SSIM values of 25.37 and 0.918 on the UIEB dataset, respectively, as well as low memory usage and computational resource requirements. In addition, detailed ablation experiments prove the effectiveness of the core components of the model. Full article

Financial data analysis remains a significant challenge due to the inherent stochasticity, non-stationarity, and low signal-to-noise ratio of market data. Conventional methods often struggle to disentangle intrinsic trends from noise and frequently overlook the critical influence of investor sentiment on price dynamics. To address these issues, we propose an adaptive trading model named CBAM-BiLSTM-DDQN, which integrates signal decomposition, multi-source feature fusion, and deep reinforcement learning. First, we construct a comprehensive heterogeneous feature set by combining price signals decomposed via Variational Mode Decomposition (VMD) and investor sentiment indices extracted from financial texts. Subsequently, a Genetic Algorithm (GA) is employed to identify the most significant feature subset, effectively reducing dimensionality and redundancy. Finally, these optimized features are input into a Double Deep Q-Network (DDQN) agent equipped with a Convolutional Block Attention Module (CBAM) and a Bidirectional Long Short-Term Memory (BiLSTM) network to capture complex spatiotemporal dependencies. We evaluated this approach through simulated trading on three major Chinese stock indices—the Shanghai Stock Exchange Composite (SSEC), the Shenzhen Stock Exchange Component (SZSE), and the China Securities 300 (CSI 300). Experimental results demonstrate the superiority of our method over traditional strategies and standard baselines; specifically, the trading agent achieved robust cumulative returns across the SSEC and CSI 300 indices, confirming the model’s exceptional capability in balancing profitability and risk aversion in complex financial environments. Furthermore, additional experiments on individual stocks in the Chinese A-share market reinforce the robustness and generalization ability of our proposed model, validating its practical potential for diverse trading scenarios. Furthermore, additional experiments on individual stocks in the Chinese A-share market reinforce the robustness and generalization ability of our proposed model, validating its practical potential for diverse trading scenarios. Full article

Background: The generation of synthetic CT images from MRI scans represents a crucial step toward enabling MRI-only clinical workflows and supporting multi-modal integration in medical imaging, particularly in radiotherapy planning. Despite significant advancements in deep learning models, many current methods still struggle to reconstruct high-density structures, especially bone, and exhibit limited accuracy in density values. This shortcoming is largely attributed to the passage of excessive or noisy features through skip connections in the traditional U-Net architecture, which degrade the quality of information transmitted to the decoder, negatively impacting the clarity of anatomical boundaries and the pixel-wise accuracy of the resulting synthetic image. Methods: In this work, we propose an enhanced 3D U-Net architecture in which the Convolutional Block Attention Module (CBAM) is systematically integrated within each skip connection. The CBAM sequentially applies channel and spatial attention to adaptively reweight encoder feature maps before fusion with the decoder, thereby emphasizing anatomically relevant structures while suppressing irrelevant feature propagation. The model was trained and evaluated on the SynthRAD2023 (Task 1—Brain) MRI–CT dataset. To rigorously assess the contribution of the attention mechanism, a dedicated ablation study was conducted comparing three variants: 3D U-Net with Squeeze-and-Excitation (SE), Coordinate Attention (CA), and the proposed CBAM module. Performance was evaluated using Mean Absolute Error (MAE), Root Mean Square Error (RMSE), Peak Signal-to-Noise Ratio (PSNR), Structural Similarity Index (SSIM), and Normalized Cross-Correlation (NCC). Results: The ablation study demonstrated that the CBAM-enhanced model consistently outperformed both SE- and CA-based variants across all quantitative metrics. Specifically, the proposed method achieved an MAE of $38.2\pm 5.4$ HU and an RMSE of $51.0\pm 12.0$ HU, representing the lowest reconstruction errors among the evaluated models. In addition, it obtained a PSNR of $29.45\pm 2.10$ dB, SSIM of $0.940\pm 0.031$, and NCC of $0.967\pm 0.015$, indicating superior structural preservation and strong voxel-wise correspondence between synthesized and reference CT volumes. These results confirm that the sequential integration of channel and spatial attention provides a statistically and practically meaningful improvement for high-fidelity MRI-to-CT synthesis. Discussion and Conclusions: Generating high-resolution brain CT images from brain MRI scans using a 3D U-Net network enhanced with a CBAM module can contribute to supporting the clinical workflow by providing additional diagnostic data without the need for extra radiological examinations, thereby enhancing diagnostic efficiency and reducing radiation exposure. This technique helps reduce patient exposure to radiation and improves accessibility in resource-limited settings. Furthermore, this method is valuable for retrospective studies, surgical planning, and image-guided therapy, where complete multi-modal data may not always be available. Full article

In the Internet of Vehicles (IoV), connected vehicles generate high-load perception tasks with large-scale and multimodal sensitive data, imposing strict requirements on latency, computing, and privacy. Existing solutions still suffer from high task service latency and privacy risks. To address these issues, this paper proposes an integrated framework that jointly considers multi-flow task offloading, adaptive privacy preservation, and latency-aware resource incentive mechanism. Specifically, we propose a Location-Aware and Trust-based (LA-Trust) dual-node task offloading algorithm based on deep reinforcement learning (DRL), which treats pre-partitioned subtasks as multiple parallel flows and enables flow-level collaborative offloading optimization across neighboring nodes, allows subtask data uploading and processing to proceed concurrently, and incorporates node security into decision making. To further enhance privacy protection, a Distribution-Aware Local Differential Privacy (DA-LDP) algorithm is designed to adaptively inject artificial noise according to data heterogeneity, balancing privacy protection and task execution accuracy. In addition, a Delay-Cost Reverse Auction (DC-RA) algorithm is proposed to further reduce latency by introducing wireless channel modeling between idle vehicles and edge nodes into the incentive mechanism. Experimental results show that the proposed framework improves task execution accuracy by 38% and reduces offloading cost, delay, incentive cost, and auction communication latency by 64.41%, 64.64%, 19%, and 44%, respectively, while more than 60% of tasks are offloaded to high-trust nodes. Full article

Machine learning was employed to make fluid agnostic laminar heat transfer prediction in supercritical conditions, encompassing three fluids (sCO₂, sH₂O, sC₁₀H₂₂) representing a wide range of operating conditions. High-fidelity training data, consisting of both non-dimensional and dimensional (operating parameter) as inputs and Nu and T_wall as outputs, were generated from grid-converged, steady-state, computational fluid dynamic (CFD) simulations. The Random Forest (RF) algorithm outperformed the artificial neural networks (ANNs) across all scenarios on the small multi-fluid dataset (~1600 data points) employed during the training process. When using non-dimensional parameters as inputs, Nu prediction fidelities were better than T_wall predictions for both ML algorithms across both horizontal and vertical configurations. The RF model trained on data from a specific flow configuration (horizontal/vertical) could predict T_wall within an accuracy of +/−1% with dimensional, operational parameters as inputs while being agnostic to the working fluid. Furthermore, by including the gravity vector as an additional variable during the training process, the RF model could predict T_wall accurately in a mixed, multi-fluid dataset containing data from both horizontal and vertical configurations. Full article