Search Results (1,221)

Manufacturing workshops operate in dynamic and complex environments, where multiple orders are processed simultaneously through interdependent stages. This complexity makes it challenging to accurately predict the remaining completion time of ongoing orders. To address this issue, this paper proposes a data-driven prediction approach that analyzes key features extracted from multi-source manufacturing data. The method involves collecting heterogeneous production data, constructing a comprehensive feature dataset, and applying feature analysis to identify critical influencing factors. Furthermore, a deep learning optimization model based on a Convolutional Neural Network (CNN)–Bidirectional Long Short-Term Memory (BiLSTM)–Attention architecture is designed to handle the temporal and structural complexity of workshop data. The model integrates spatial feature extraction, temporal sequence modeling, and adaptive attention-based refinement to improve prediction accuracy. This unified framework enables the model to learn hierarchical representations, focus on salient temporal features, and deliver accurate and robust predictions. The proposed deep learning predictive model is validated on real production data collected from a discrete manufacturing workshop equipped with typical machines. Comparative experiments with other predictive models demonstrate that the CNN–BiLSTM–Attention model outperforms existing approaches in both accuracy and stability for predicting order remaining completion time, offering strong potential for deployment in intelligent production systems. Full article

This paper presents a hierarchical and step-by-step multi-objective genetic optimization method for the multi-port composite flux-modulated (MP-CFM) machine, aiming to propose a simpler and high-accuracy optimization strategy for such multi-port composite machines. As a specialized machine well-suited for hybrid power systems, the optimization design is innovatively conducted based on an analysis of the fundamental operating principles and working modes of the MP-CFM machines. Specifically, considering the complex structure of such composite machines, sensitivity analysis is employed, applying differentiated strategies based on parameter sensitiveness evaluation. Furthermore, to ensure the rationality of the optimization results, and also to reduce computational cost and improve convergence, the optimization is artfully developed hierarchically with multi-steps, in accordance with the multi-modes of the machine. Specific optimization objectives and variables are defined, respectively, for each mode to enhance the optimization efficiency. Finite element analysis results demonstrate the effectiveness of the proposed optimization strategy for such MF-CFM machines for hybrid power system applications. Full article

Complex human diseases, including cancer, are linked to genetic factors. Genome-wide association studies (GWASs) are powerful for identifying genetic variants associated with cancer but are limited by their reliance on case–control data. We propose approaches to expanding GWAS by using tumor and paired normal tissues to investigate somatic mutations. We apply penalized maximum likelihood estimation for single-marker analysis and develop a Bayesian hierarchical model to integrate multiple markers, identifying SNP sets grouped by genes or pathways, improving detection of moderate-effect SNPs. Applied to breast cancer data from The Cancer Genome Atlas (TCGA), both single- and multiple-marker analyses identify associated genes, with multiple-marker analysis providing more consistent results with external resources. The Bayesian model significantly increases the chance of new discoveries. Full article

Methane dry reforming (DRM) represents a promising route for the simultaneous valorization of CH₄ and CO₂ into syngas; however, conventional Ni-based catalysts suffer from rapid deactivation due to sintering and carbon deposition. In this work, we present a synergistically engineered Ni-based catalyst integrating hierarchical SiC confinement, Pd promotion via oleic-acid-assisted complexation, and MgO surface modification to overcome these challenges. Under optimized reaction conditions (CH₄/CO₂ = 1:1, 750 °C, GHSV = 36,000 mL g⁻¹ h⁻¹), the multifunctional NiPd/Si–xMg catalyst achieved steady-state conversions of 85% for CH₄ and 84% for CO₂, maintaining an H₂/CO ratio close to 1.0 over 100 h of continuous operation without noticeable deactivation. In contrast, the reference Ni/SiC and Ni/MgO catalysts exhibited initial conversions of 75–80% but declined by more than 50% within the same period, confirming the superior durability of the optimized system. Thermogravimetric analysis (TGA) revealed a drastic reduction in carbon deposition—from 119.0 mg C g⁻¹ for Ni/SiC to 81.4 mg C g⁻¹ for NiPd/Si-xMg—indicating enhanced coke resistance. Transmission electron microscopy (TEM) confirmed uniform Ni dispersion with an average particle size of 7.2 ± 1.8 nm, while H₂-TPR and CO₂-TPD analyses demonstrated improved reducibility and surface basicity. The combination of SiC confinement, Pd-induced hydrogen spillover, and MgO-mediated CO₂ activation effectively mitigated sintering and carbon accumulation, resulting in high activity, stability, and carbon tolerance. This integrated catalyst design provides a robust pathway toward industrially viable DRM systems for sustainable syngas production. Full article

The fracturing and driving oil technology used in shale oil provides a new approach for the development of offshore low-permeability reservoirs. However, the main control role of technical parameters is unclear, resulting in unsatisfactory accuracy and effectiveness of the enhanced oil recovery plan. For this reason, this study is based on the production and process data of five wells in the WZ oilfield. Fuzzy analytic hierarchical process analysis method (FAHP) was used to evaluate the parameter weights. Combined with numerical simulation technology, the evaluation results were verified by geological-engineering integration. The results show that in offshore low-permeability oilfields, the reservoir pressure coefficient has the greatest influence on the fracturing and oil repelling effect. The comprehensive weight reaches 0.450 compared to not adopting hydraulic fracturing oil displacement technology. This improves the recovery rate by 10.19% in 5 years. The surfactant concentration and the residual oil saturation of the reservoir rank are second, with a comprehensive weight of 0.219. Finally is the effective thickness of the reservoir, with a comprehensive weight of 0.113. In this study, the key parameters of fracturing and oil repelling in offshore low-permeability reservoirs are clarified. It provides theoretical basis and practical support for improving the success rate of well selection, layer selection and recovery capacity. Full article

The integration of distributed energy resources into distribution networks creates operational challenges, including voltage instability and power quality issues. While battery energy storage systems (BESSs) can address these challenges, research has focused primarily on transmission-level applications or single services. This paper bridges this gap through a comprehensive review of BESS technologies and control strategies for multi-service ancillary support in distribution networks. Real-world case studies demonstrate BESS effectiveness: Hydro-Québec’s $1.2$ MW system maintained voltage within 5% and responded to frequency events in under 10 ms; Germany’s hybrid 5 MW M5BAT project optimized multiple battery chemistries for different services; and South Africa’s Eskom deployment improved renewable hosting capacity by 15–70% using modular BESS units. The analysis reveals grid-forming inverters and hierarchical control architectures as critical enablers, with model predictive control optimizing performance and droop control ensuring robustness. However, challenges like battery degradation, regulatory barriers, and high costs persist. This paper identifies future research directions in degradation-aware dispatch, cyber-resilient control, and market-based valuation of BESS flexibility services. By combining theoretical analysis with empirical results from international deployments, this study provides utilities and policymakers with actionable insights for implementing BESS in modern distribution grids. Full article

Washing performance depends on the specific interactions between textiles, stains, detergents, mechanical action, temperature, and time. Its evaluation therefore requires a fundamental and practical understanding of the effects of the washing parameters, the type of soiling, and the tendency of the textiles to stain. Due to the complexity of these interactions, the evaluation of stain removal requires specific types of textiles, stains, and detergents. In this study, the temperature-specific behavior of detergents was studied in the washing process of cotton knitwear with a blank spot and 15 stains of different origin and composition at 60 °C and 90 °C. Despite the labeled composition of detergents, the detergent ingredients, surfactants, and bleaching agents were analyzed by titration methods. The evaluation of the total washing performance (TWP) and specific washing performance (SWP) was carried out by measuring reflectance as a spectral parameter. A hierarchical cluster analysis was carried out to compare the specific effects of detergents at both temperatures. The analysis of the detergents revealed fluctuations in the content of the surfactants and bleach. Some detergents with a higher surfactant content (SAS) showed poorer performance in washing at 60 °C compared to detergents with a lower SAS content. The dendrogram showed subtle similarities and dissimilarities between the detergents, which contributed to clarification of the total wash performance at both temperatures. The results proved that the quantitative indicators of the proportions of certain ingredients in a detergent are not the only criteria for evaluating the quality of a particular detergent. All detergents investigated showed a temperature-specific behavior, which was reflected in an increased TWP at 90 °C, while some detergents selectively improved the SWP at 90 °C. Full article

Sudden cardiac death (SCD) is a leading cause of mortality worldwide, with arrhythmia serving as a major precursor. Early and accurate prediction of SCD using non-invasive electrocardiogram (ECG) signals remains a critical clinical challenge, particularly due to the inherent asymmetric and non-stationary characteristics of ECG signals, which complicate feature extraction and model generalization. In this study, we propose a novel SCD prediction framework based on hierarchical feature fusion, designed to capture both non-stationary and asymmetrical patterns in ECG data across six distinct time intervals preceding the onset of ventricular fibrillation (VF). First, linear features are extracted from ECG signals using waveform detection methods; nonlinear features are derived from RR interval sequences via second-order detrended fluctuation analysis (DFA2); and multi-scale deep learning features are captured using a Temporal Convolutional Network-based sequence-to-vector (TCN-Seq2vec) model. These multi-scale deep learning features, along with linear and nonlinear features, are then hierarchically fused. Finally, two fully connected layers are employed as a classifier to estimate the probability of SCD occurrence. The proposed method is evaluated under an inter-patient paradigm using the Sudden Cardiac Death Holter (SCDH) Database and the Normal Sinus Rhythm (NSR) Database. This method achieves average prediction accuracies of 97.48% and 98.8% for the 60 and 30 min periods preceding SCD, respectively. The findings suggest that integrating traditional and deep learning features effectively enhances the discriminability of abnormal samples, thereby improving SCD prediction accuracy. Ablation studies confirm that multi-feature fusion significantly improves performance compared to single-modality models, and validation on the Creighton University Ventricular Tachyarrhythmia Database (CUDB) demonstrates strong generalization capability. This approach offers a reliable, long-horizon early warning tool for clinical SCD risk assessment. Full article

Deep-fat frying remains the predominant method of food preparation; however, increasing concerns regarding health and sustainability have prompted the search for safer alternatives. Palm olein is widely used as a frying medium but its consumption has been questioned due to the presence of contaminants (e.g., 3-monochloropropane-1,2-diol, 3-MCPD) and the challenges associated with its transportation from producing countries, creating a need for healthier and more sustainable alternatives. The present study aimed to assess the oxidative stability, physicochemical properties, and sensory characteristics of various oils used for deep-fat frying, with particular emphasis on identifying suitable replacements for palm olein. Five oils were evaluated: refined sunflower oil (RSO), RSO supplemented with tert-butylhydroquinone (RSO+TBHQ), RSO supplemented with rosemary extract (RSO+RE), high-oleic sunflower oil (HOSO), and palm olein (PO). Samples were evaluated before and after deep-frying of French fries, at 175 °C for 2.5 min, over a total of 12 consecutive frying cycles. The results demonstrated that palm olein and HOSO exhibited the highest oxidative stability (induction period determined by Rancimat method at 100 °C was 27 h and 26.2 h, respectively), whereas the addition of TBHQ (induction period 23.4 h) and rosemary extract (induction period 11.5 h) provided only a modest enhancement of RSO stability (induction period 9.6 h). Hierarchical cluster analysis grouped palm olein and HOSO together, confirming their similar stability, while RSOs formed a distinct cluster. These findings suggest that high-oleic sunflower oil represents the most promising, stable, and nutritionally advantageous alternative to palm olein, simultaneously supporting local production and improved dietary quality. Full article

The association between patterns of internet use for older adults’ well-being is unclear. We examined the association between the frequency and purpose of internet use and subjective well-being in older Japanese adults. We analyzed cross-sectional data from 2343 community-dwelling older adults (aged 65–84 years). Subjective well-being was measured using the World Health Organization Well-Being Index as a continuous score, and internet use was categorized by frequency and purpose. Hierarchical linear regression analysis was controlled for sociodemographic and health-related covariates. After full adjustment, only daily (B = 1.04, 95% CI [0.53, 1.56]) and dual-purpose use (i.e., for both practical and social communication purposes; B = 0.80, 95% CI [0.28, 1.31]) were independently associated with higher well-being. The analysis of the combined patterns further suggested that daily use was the primary factor. For older adults, regularity of internet use was more strongly associated with well-being than diversity of purpose. Daily integration appears to be a key factor for realizing benefits, suggesting that sustained practice is the foundational step in building the digital capital necessary for a flourishing later life. Longitudinal studies are needed to confirm these findings and untangle the causal relationship between sustained internet use and improved well-being among older adults. Full article

Few studies have explored the holistic public health risk assessment associated with toxic elements (TEs) and their bioaccessibility in integrated urban environmental media including soils, vegetables, atmospheric particles, dust, etc. Urban industrial complex areas like Qingshan-Chemical District (QCD) in the Chinese Wuhan city, located within the Yangtze River Economic Belt, face increasing environmental exposure risks due to industrial activities. This study innovatively assessed the hierarchical risks of toxic metals in 4 environmental media (air PM, dust, soil, vegetables) from the QCD based on field sampling and chemical analysis, and developed an improved fuzzy health risk assessment model based on toxic metals’ in vitro bioaccessibilities of different exposure pathways and triangular fuzzy numbers for handling parameter uncertainties. The study found that the highest health risks were associated with ingestion, particularly from consuming homegrown vegetables. Carcinogenic risks for arsenic (As), lead (Pb), and cadmium (Cd) via ingestion exceeded the admissible threshold of 1.00 × 10⁻⁶, with As showing the highest risk ([1.92 × 10⁻³, 2.37 × 10⁻³]), followed by Cd ([2.98 × 10⁻⁵, 3.67 × 10⁻⁵]) and Pb ([7.92 × 10⁻⁷, 1.48 × 10⁻⁶]). Inhalation risks from soil, dust, and air particulates were below the threshold, indicating lower respiratory concerns. Dermal exposure, especially from soil and dust, posed elevated carcinogenic risks for As ([7.47 × 10⁻⁶, 8.06 × 10⁻⁶]). With the screened priority risk control toxic metals and pathways, the targeted measures including relocating vegetable planting areas, promoting cultivation of low-enrichment crops, building vegetation buffer zones around the industrial park, etc., were proposed. Full article

This paper presents a novel hybrid and hierarchical architecture of a Convolutional Neural Network (CNN), based on MobileNetV2 and Support Vector Machines (SVM) for the classification of crop diseases (SVMobileNetV2). The system feeds from multispectral images captured by Unmanned Aerial Vehicles (UAVs) alongside data from IoT nodes. The primary objective is to improve classification performance in terms of both accuracy and precision. This is achieved by integrating contemporary Deep Learning techniques, specifically different CNN models, a prevalent type of artificial neural network composed of multiple interconnected layers, tailored for the analysis of agricultural imagery. The initial layers are responsible for identifying basic visual features such as edges and contours, while deeper layers progressively extract more abstract and complex patterns, enabling the recognition of intricate shapes. In this study, different datasets of tropical crop images, in this case banana crops, were constructed to evaluate the performance and accuracy of CNNs in detecting diseases in the crops, supported by transfer learning. For this, multispectral images are used to create false-color images to discriminate disease through spectra related to the blue, green and red colors in addition to red edge and near-infrared. Moreover, we used IoT nodes to include environmental data related to the temperature and humidity of the environment and the soil. Machine Learning models were evaluated and fine-tuned using standard evaluation metrics. For classification, we used fundamental metrics such as accuracy, precision, and the confusion matrix; in this study was obtained a performance of up to 86.5% using current deep learning models and up to 98.5% accuracy using the proposed hybrid and hierarchical architecture (SVMobileNetV2). This represents a new paradigm to significantly improve classification using the proposed hybrid CNN-SVM architecture and UAV-based multispectral images. Full article

Agriculture and its allied activities contribute to the primary sector in India and act as the basis for the country’s economy. Available agricultural landholdings are scattered as multiple plots across the country. Land fragmentation has led to problems achieving economies of scale and economies of scope; lower productivity, efficiency, and modernization; loss of biodiversity; and little scope for mechanization and technology. FPOs are small clusters of farmers who collaborate to enhance their bargaining strength through collective procurement, processing, and marketing efforts. To enhance the performance of FPOs at the grassroots level, the engagement of cluster-based business organizations (CBBOs) is vital. Millet FPOs are similar to voluntary farmer groups that are involved in the cultivation and promotion of millets. IIMR-promoted millet FPOs were selected purposively for the present study as they are involved in millet cultivation and farming. A total of 450 millet farmers from 15 FPOs and 3 states were randomly chosen for this action research study. The present research identified 10 key factors and collected farmers’ opinions toward member participation in millet FPOs using interpretive structural modeling. The ISM approach provided a clear understanding of how the selected factors interconnect hierarchically with each other as foundational drivers and dependent outcomes. The results from the MICMAC analysis demonstrated that foundational interventions, such as post-harvest technology availability (V2) and knowledge transfer by KVKs (V5), directly support higher-level objectives. Intermediate factors like economies of scale (V1) and market and credit linkages (V3) transform these services into operational advantages, while the outcome factors of business planning (V8), FPO branding (V7), and bargaining power (V9) emerge as dependent variables. The model demonstrates that V2 catalyzes improvements across the production, market, and institutional domains, cascading through intermediate enablers (V1, V4, V5, V6) to strengthen outcomes (V3, V7, V8, V9, V10). This hierarchy demonstrates that investing in post-harvest technology and complementary extension services is critical for building resilient millet FPOs and enhancing member participation. Full article

Student exam pass prediction (EPP) is a key task in educational assessment and can help teachers identify students’ learning obstacles in a timely manner and optimize teaching strategies. However, existing EPP models, although capable of providing quantitative analysis, suffer from issues such as complex algorithms, poor interpretability, and unstable accuracy. Moreover, the evaluation process is opaque, making it difficult for teachers to understand the basis for scoring. To address this, this paper proposes an approximate belief rule base (ABRB-a) student examination passing prediction method based on adaptive reference point selection using symmetry. Firstly, a random forest method based on cross-validation is adopted, introducing intelligent preprocessing and adaptive tuning to achieve precise screening of multi-attribute features. Secondly, reference points are automatically generated through hierarchical clustering algorithms, overcoming the limitations of traditional methods that rely on prior expert knowledge. By organically combining IF-THEN rules with evidential reasoning (ER), a traceable decision-making chain is constructed. Finally, a projection covariance matrix adaptive evolution strategy (P-CMA-ES-M) with Mahalanobis distance constraints is introduced, significantly improving the stability and accuracy of parameter optimization. Through experimental analysis, the ABRB-a model demonstrates significant advantages over existing models in terms of accuracy and interpretability. Full article

Multi-modal image segmentation is a key task in various fields such as urban planning, infrastructure monitoring, and environmental analysis. However, it remains challenging due to complex scenes, varying object scales, and the integration of heterogeneous data sources (such as RGB, depth maps, and infrared). To address these challenges, we proposed a novel multi-modal segmentation framework, DyFuseNet, which features dynamic adaptive windows and cross-scale feature fusion capabilities. This framework consists of three key components: (1) Dynamic Window Module (DWM), which uses dynamic partitioning and continuous position bias to adaptively adjust window sizes, thereby improving the representation of irregular and fine-grained objects; (2) Scale Context Attention (SCA), a hierarchical mechanism that associates local details with global semantics in a coarse-to-fine manner, enhancing segmentation accuracy in low-texture or occluded regions; and (3) Hierarchical Adaptive Fusion Architecture (HAFA), which aligns and fuses features from multiple modalities through shallow synchronization and deep channel attention, effectively balancing complementarity and redundancy. Evaluated on benchmark datasets (such as ISPRS Vaihingen and Potsdam), DyFuseNet achieved state-of-the-art performance, with mean Intersection over Union (mIoU) scores of 80.40% and 80.85%, surpassing MFTransNet by 1.91% and 1.77%, respectively. The model also demonstrated strong robustness in challenging scenes (such as building edges and shadowed objects), achieving an average F1 score of 85% while maintaining high efficiency (26.19 GFLOPs, 30.09 FPS), making it suitable for real-time deployment. This work presents a practical, versatile, and computationally efficient solution for multi-modal image analysis, with potential applications beyond remote sensing, including smart monitoring, industrial inspection, and multi-source data fusion tasks. Full article