Search Results (9,827)

Accurate monthly runoff simulation is the foundation of watershed water resources management and flood–drought disaster prevention. To address the issues that traditional machine learning models struggle to capture long-term dependencies in runoff series and that existing deep learning methods ignore the differences in driving mechanisms across intra-annual hydrological periods, this paper proposes a Period-Adaptive Attention Informer (PA-Informer), in which a learnable period embedding is introduced into the attention mechanism to achieve adaptive representation of the memory length of each period. Based on autocorrelation analysis, K-means clustering and dynamic time warping, the hydrological year is divided into a dry period (November–February), a transition period (March–June and October) and a wet period (July–September). Using monthly precipitation, monthly mean temperature and antecedent runoff (lags 1–3) observed at the Changtaiguan Hydrological Station in the Chushandian Irrigation Area of the upper Huaihe River (1960–2020), a one-month-ahead monthly runoff simulation is performed with the meteorological variables of the target month taken as known; the record is split chronologically into a training set (1960–2002) and a test set (2003–2020) at a ratio of 7:3, and all preprocessing and tuning procedures are restricted to the training set. On the test set, PA-Informer achieves an R² = 0.957, NSE = 0.948, RMSE = 0.71 m³/s and MAE = 0.52 m³/s (mean over five random seeds), outperforming the best benchmark (Base-Informer) with a statistically significant difference (Wilcoxon signed-rank test, p = 0.0023). SHAP analysis reveals that dry-period runoff is driven mainly by temperature and antecedent runoff, transition-period runoff by precipitation and long-lag runoff, and wet-period runoff predominantly by precipitation with temperature acting as a suppressor. PA-Informer effectively integrates the physical laws of hydrological periodicity with a deep learning architecture, providing a new approach for high-accuracy monthly runoff simulation in irrigation areas. Full article

Background/Objectives: Workplace violence is a recognised psychosocial risk across many professions but remains insufficiently investigated in veterinary practice. Evidence on workplace violence among veterinary employees in Slovenia is limited. This study aimed to examine reported exposure to different forms of workplace violence involving animal owners, co-workers, and supervisors among employees of veterinary organisations in Slovenia and to assess associated emotional responses. Methods: A cross-sectional online survey was conducted between December 2024 and April 2025 among employees of all 182 registered veterinary practices and clinics in Slovenia. Of 1351 eligible employees, 178 completed the questionnaire (overall response rate: 13.17%), including 155 of 873 eligible veterinarians (veterinarian response rate: 17.75%). Respondents rated their agreement with statements describing verbal, physical, and psychological violence involving animal owners, co-workers, and supervisors, as well as associated negative emotional responses. Data were analysed using one-sample and paired-samples t-tests, principal component analysis, and multiple linear regression. Results: Respondents reported exposure to workplace violence involving animal owners, co-workers, and supervisors, with verbal and psychological forms particularly evident in interactions with animal owners. Fear, helplessness, and uncertainty were prominent emotional responses associated with violence involving animal owners, while violence involving co-workers and supervisors was also associated with negative emotional responses. Greater work experience was significantly associated with less intense negative emotional responses across all three perpetrator groups. Higher education and male gender were additionally associated with less intense emotional responses to violence involving animal owners. Conclusions: Workplace violence represents an important psychosocial concern in Slovenian veterinary organisations and may originate from both external and internal sources. Training, written protocols, and organisational and psychological support may warrant consideration as potential preventive measures; however, their effectiveness was not evaluated in this study and requires further investigation. The low response rate and potential non-response and self-selection bias limit the generalisability of the findings. Further research using validated event-based measures and stronger recruitment strategies is warranted. Full article

Fused deposition modeling (FDM) parts show condition-dependent variability in ultimate tensile strength (UTS), yet most machine learning (ML) studies report only deterministic point estimates, limiting use in safety-critical design. This work presents a probabilistic framework mapping FDM-PETG process parameters to orientation-specific conditional quantile estimates with pooled uncertainty, covering model selection, deployment, and process optimization. A controlled experiment varied 12 parameters across 111 printing conditions from a combined Taguchi and Latin Hypercube Sampling (LHS) design, yielding 333 ASTM D3039 tensile measurements. Six probabilistic models were trained on raw replicate data and ranked on point accuracy and calibration, weighted 75% toward uncertainty quantification (UQ). The mixture density network (MDN) showed the lowest composite rank. Orientation-specific (MDN-OR) and unified (MDN-UF) variants were compared; MDN-OR gave more conservative lower-quantile estimates for safety-critical design. Lower-quantile (q = 0.05) SHAP analysis identified filament consumption and wall loops as the strongest predictive associations with worst-case strength. For each orientation, the configuration with minimum filament consumption at maximum strength was selected, with its post hoc q = 0.05 lower bound confirmed by physical tensile testing. Within the studied single-material, single-printer setting, the framework shows how probabilistic process–property modeling can support uncertainty-aware parameter selection toward autonomous additive manufacturing. Full article

With the rapid advancement of industrial intelligence, equipment predictive maintenance (PdM) has emerged as a critical enabler for intelligent operation and maintenance (O&M) systems. However, conventional machine learning-based PdM methods are inherently constrained by scarce labeled fault samples, weak cross-condition generalization, and a lack of advanced causal reasoning capabilities in complex industrial process environments. Recent advances in large language models (LLMs) and generative artificial intelligence (GAI) have created unprecedented opportunities to fundamentally overcome these persistent PdM limitations. This paper presents a comprehensive review and future perspective on LLM applications for PdM in industrial process systems and proposes a hierarchical “Capability-Domain-Method (CDM)” taxonomy to establish a unified analytical framework for this fast-growing research field. Based on a systematic analysis of the 103 core research papers published between 2016 and 2026, this work comprehensively summarizes the latest research on LLMs for predictive maintenance from three core technical dimensions: data augmentation, model adaptation, and knowledge reasoning and decision-making. Data augmentation methods generate high-fidelity fault samples and degradation patterns to mitigate industrial data scarcity and class imbalance. Model adaptation techniques empower general-purpose pre-trained LLMs to adapt to specific equipment and complex industrial environments. Knowledge reasoning and decision-making approaches integrate external industrial domain knowledge, autonomous reasoning mechanisms, and maintenance decision processes to improve the performance of fault diagnosis, root cause analysis, and maintenance decision-making. Furthermore, this paper discusses key challenges related to operational reliability, deep domain adaptation, unified evaluation standards, and unified deployment efficiency. Finally, future research directions are advised, including physics-informed large models, industrial domain-specific models, autonomous maintenance agents, and standardized open benchmarking platforms, to accelerate the paradigm shift in PdM from data-driven pattern to knowledge-driven and autonomous decision-making. Full article

This study examined how elite women’s rugby union players perceived their off-season recovery experiences. An exploratory cross-sectional survey was completed by 140 elite women’s rugby union players (M age = 25.80, SD = 4.16) that captured global perceived readiness to train, recovery strategies, contextual demands, and preferred off-season structure. Quantitative data were analyzed descriptively, while responses to open-ended questions were explored using reflexive thematic analysis. Players reported a range of sport- and non-sport related demands during the off-season, including injury management, physical preparation, financial pressures, education and professional development, and external employment. Players reported a preferred off-season duration of 5.19 weeks on average (range 4–10 weeks), incorporating three broad components: a complete break from rugby, remote or personal training, and reintegration into the rugby environment. Overall, the findings indicate that elite women’s perceptions of off-season recovery reflect both rugby-related demands and wider occupational, educational and personal circumstances. The findings provide athlete-informed considerations for the design of off-season structures and identify areas for future prospective research. Full article

Physics-informed neural networks (PINNs) have developed from residual constrained neural approximations into a heterogeneous family of scientific computing formulations, combining governing equations, observations, numerical operators, and optimization. This review examines how formulation, training, and evaluation interact within an integrated computational methodology. A structured and iterative review assembled 853 records from Scopus, IEEE Xplore, ScienceDirect, MDPI, reference tracing, and topic-specific searches; 514 references are cited in the revised manuscript. The review analyzes representation, physics enforcement, differentiation, collocation, loss construction, optimization, architectural and theoretical developments, and multidimensional evaluation. The analysis shows that reliability depends on these coupled choices and that no universal architecture, optimizer, enforcement strategy, or sampling strategy exists. An evidence-linked co-design framework, qualitative design-performance dependency matrix, and failure-mode diagnostic map are therefore developed to connect design choices with performance and guide redesign. Low prediction or residual error alone does not establish conservation, stability, robustness, uncertainty calibration, or computational efficiency. The open PINN Review Atlas links the synthesis to evidence, metrics, mathematical formulations, and related data as an interactive resource to help researchers explore the field and support future studies. PINNs are likely to serve as problem-dependent trainable components of hybrid scientific computing workflows rather than universal replacements for traditional numerical solvers. Full article

Historical groundwater databases may contain systematic errors involving ground elevations, spatial coordinates, or hydraulic head measurements, which can affect hydrogeological interpretation and downstream analyses. Conventional outlier detection methods assume independent observations and may perform poorly in spatially structured groundwater data, where local hydrogeological features and spatial autocorrelation shape the observed field. Here, we present a graph-based framework for identifying persistent anomalous records in legacy groundwater datasets by combining a regional trend model with a graph attention network (GAT) to account for spatial dependence. Residuals from the hybrid model are compared across two contrasting hydrological seasons to assess temporal persistence. The approach was tested on a historical piezometric dataset of 161 wells from the Tavoliere di Foggia aquifer in southern Italy, with missing well-depth records imputed using missForest prior to spatial modelling. In this dataset, the proposed framework identified anomalous residual patterns that were not captured by standard univariate tests and showed seasonal stability in a subset of records, suggesting either potential database inconsistencies or localized hydrogeological conditions not resolved by the model. This persistence criterion showed no overlap with a classical, spatially explicit statistic (Local Indicators of Spatial Association) applied to the model’s final residual field but computed without reference to any of the trained model’s own parameters, indicating that the two diagnostics capture distinct properties of that residual field, and a core subset of the flagged wells proved reproducible across repeated model retraining under different random seeds. Tracing one flagged well back through the original field archive confirmed a concrete database error invisible to classical tests: a missing depth-to-water measurement had been silently treated as zero rather than propagated as missing, producing a spurious hydraulic head numerically identical to the well’s ground elevation. Cross-referencing declared well elevations against two independent national digital terrain models further showed that the largest elevation discrepancies were significantly enriched among the wells already flagged by both the proposed framework and classical tests, providing external corroboration from a source unconnected to the hydraulic head record itself. The framework therefore offers a spatially explicit screening tool to support quality control in legacy archives and to prioritize records for targeted physical verification. Full article

Background/Objectives: Aging is associated with declines in strength, balance, and mobility that contribute to falls, disability, and loss of independence. While exercise interventions improve physical function, limited data exist on community-dwelling adults aged 90 years and older. The purpose of this study was to evaluate the effect of age on physical function improvement following participation in an exercise program for older Veterans. Methods: We analyzed baseline and 3-month follow-up physical function data from 62 community-dwelling Veterans aged 60–99 years enrolled in the Veterans Affairs Gerofit, a supervised, individualized exercise program incorporating aerobic, resistance, balance, and flexibility training. Outcomes included lower extremity strength (30-s chair stands), mobility and balance (8-foot up-and-go), and cardiovascular endurance (6-min walk). Participants were grouped by age (60–69, 70–79, 80–89, 90–99). Paired samples t-tests assessed changes over time, and repeated-measures general linear models examined the effect of age on physical function improvement. Results: Across all participants, significant improvements were observed in chair stand and 8-foot up-and-go performance (both p < 0.001), whereas 6-min walk distance did not change significantly (p = 0.14). Chair stand performance improved in all age groups, including adults aged 90–99 (p < 0.01). Although baseline physical function declined with age, we did not detect differences in the rate of improvement in chair stand performance by age group (p = 0.69). Improvements in 8-foot up-and-go performance were observed in younger age groups and not in participants aged 90–99 years. Conclusions: Participation in a supervised exercise program was associated with significant improvements in lower extremity strength across all age groups, including adults aged 90–99 years. These findings support the inclusion of even the oldest adults in exercise programs to promote physical function. Full article

Background/Objectives: This survey examined referral decisions, the recognition and screening of warning features (red flags) of serious disease, regulatory interpretations, and barriers to direct-access physical therapy in Saudi Arabia. Methods: An English-language cross-sectional convenience survey (May–July 2026) used an investigator-developed questionnaire with 12 clinical vignettes (written patient cases) and 20 warning-feature items paired with screening questions. The primary outcome was choosing medical referral before physical therapy in all four urgent/critical vignettes: suspected gastrointestinal bleeding, cauda equina syndrome, abdominal aortic aneurysm, and spinal infection. This definition was finalised after data collection. Results: Of 131 Saudi physical therapists, 77 (58.8%; 95% confidence interval [CI] 50.2–66.8) chose referral before treatment in all four vignettes. Another 36 included referral in every vignette but would also start treatment while arranging referral in at least one; 18 chose treatment without referral at least once. In an exploratory six-predictor analysis, postgraduate academic qualifications were associated with higher odds of meeting the primary outcome (adjusted odds ratio 4.14, 95% CI 1.63–10.51; Holm-adjusted p = 0.017) and older age with lower odds (0.92/year, 95% CI 0.86–0.98; Holm-adjusted p = 0.047). Across 17 core warning features, 74.0% of responses identified a red flag; 49.1% of screening responses indicated checking often or always. Regulatory views differed: 45.8% believed direct access was permitted in limited settings. The highest-rated barriers concerned policy, law, reimbursement, and physician support. Conclusions: These findings identify priorities for evaluating training, clarifying regulations, and coordinating referral pathways. Vignettes and self-reported practices cannot establish safety in clinical practice, show that qualifications improve decisions, or determine workforce readiness. Full article

Background: Countermovement jump (CMJ) force–time variables are commonly used to monitor physical qualities linked to golf driving performance, but most evidence derives from single time-point, mean-based correlations. This study examined repeated-measures associations between CMJ variables and driving performance in sub-elite male golfers using Bayesian quantile regression. Methods: Twenty-eight male golfers (handicap 0.8 ± 1.9) completed two sessions, one week apart, of maximal CMJs and driver swings. Jump height, jump momentum and peak propulsive power were calculated from the CMJs, while clubhead speed (CHS), ball speed and smash factor were obtained from the driver swings. Results: Between-session absolute reliability was excellent for all variables (CV upper 95% credible interval [CrI] ≤ 3.8%), whereas relative reliability was excellent for CHS and all CMJ variables (ICC lower 95% CrI ≥ 0.96), moderate for ball speed (lower 95% CrI = 0.71) and poor for smash factor (lower 95% CrI = 0.00). All CMJ variables showed positive associations with CHS across percentiles in both sessions, with little evidence of session-related change for jump momentum or peak propulsive power. In contrast, the jump height–CHS association weakened in session two at the median and upper percentiles (from 1.3–1.4 to 1.0–1.2 mph per 0.05 m increase in jump height; 98–99% posterior probability of a decrease). Associations between CMJ variables and ball speed and smash factor were uncertain, with 95% CrI spanning zero in almost all cases. Conclusions: Repeated-measures designs may add value in applied golf settings. Jump momentum and peak propulsive power showed comparatively stable associations with CHS across sessions, although both scale with body mass, which may partly account for this stability. They may therefore be useful complementary monitoring variables, but their predictive value and responsiveness to training are yet to be established. Full article

The operational integration of solar energy into modern power grids, particularly for day-ahead markets and battery storage scheduling, requires highly accurate, multi-horizon irradiance forecasts to mitigate cloud-induced intermittency. This study evaluates a spatiotemporal, multi-horizon (24-h) Global Horizontal Irradiance (GHI) forecasting framework across a high-resolution 167-node microclimate in Tula, Tamaulipas, Mexico, using a robust 2020–2024 baseline. We benchmarked six predictive paradigms: diurnal persistence, linear regression, tree ensembles (Extra Trees and XGBoost), a zero-shot foundation model (TimesFM), and a deep CNN–LSTM network, using historical clear-sky irradiance sequences and strictly causal meteorological features. On the independent 2024 test set, XGBoost achieved the lowest macro-average RMSE ($120.87\mathrm{W}{\mathrm{m}}^{-2}$), but Diebold–Mariano testing confirmed that it was statistically equivalent to the CNN–LSTM architecture. XGBoost showed significant superiority at ultra-short horizons ($t+1$ and $t+2$) by efficiently capturing nonlinear radiative inertia, whereas CNN–LSTM minimized errors under sustained overcast regimes and produced the sharpest empirical 95% prediction intervals. At the 24-h mark, ordinary linear regression converged, whereas the complex models did not, highlighting a physical transition toward deterministic solar geometry. Remarkably, zero-shot TimesFM outperformed diurnal persistence without any domain-specific training. Conditional Tree SHAP interpretability demonstrated that the XGBoost model dynamically reallocated predictive weight, shifting from contemporaneous irradiance under clear skies to precipitable water during volatile, intermittent conditions. Full article

Background: Anecdotal evidence shows that small-sided games (SSGs) are the most preferred training method, as they integrate the reality of a match into training. However, it remains unclear which specific format/size best replicates the demands of competitive matches. Objective: This study aimed to quantify the physical demands associated with different SSG formats used by a South African Premier Soccer League team. Methods: Data were retrospectively collected from players within the same team using PlayerTek (10 Hz) GPS devices. Variables measured included total distance covered, high-intensity running distance, power plays, top-end speed, and distance covered per minute. Small-sided games were categorized by player number: small (5v5), medium (6v6 and 7v7), and large (8v8, 9v9, and 10v10). Results: Physical demands varied significantly across SSG formats (p = 0.001). Total distance covered was greater in 5v5 and 10v10 formats compared to 6v6 and 7v7. The 10v10 format showed higher high-intensity running distances than all other formats. Similar trends were observed for top-end speed and power plays, although 5v5 games recorded more power plays than 6v6, 7v7, and 9v9 formats. Additionally, 5v5 formats were associated with greater distance covered per minute than 7v7, 9v9, and 10v10 formats. Conclusions: Different SSG formats were characterized by distinct physical demands on players. Larger SSGs, particularly 10v10, were associated with greater high-intensity running and maximal-speed actions, whereas 5v5 formats were associated with greater relative work rates and repeated high-intensity efforts. These findings may assist coaches when selecting SSG formats according to specific training objectives. Full article

Learning-enabled control requires training to produce competent policies consistently when success requires every robot to finish. This study evaluates masked entity-attention multi-agent proximal policy optimization (MAPPO) for simulated landmark coverage by five differential-drive robots. Configurations were compared under a fixed reward, optimizer, stopping rule, and 100 held-out scenarios, using trained seeds as statistical units; the actor observes the robot–landmark scene but not the externally computed assignment. Across ten policies per principal configuration, masked entity attention achieved $94.6\pm 4.7\%$ deterministic full-team success, against $9.3\pm 12.4\%$ and $6.8\pm 5.1\%$ for wider (1.21 times its parameters) and ordinary fixed-index controls, every masked-attention seed exceeding every seed of both fixed-index controls. Near-parameter-matched Deep Sets and unmasked-attention controls reached $73.7\pm 14.0\%$ and $72.1\pm 13.1\%$, overlapping it. The same policies retained $91.4$–$96.3\%$ mean success from three to eight robots without retraining, and masked attention remained separated from both fixed-index controls at every rate in a 45-policy learning-rate design. Held-out scenes come from the training distribution; under shift, success fell to $70.9\%$ in an unseen larger workspace, while landmark, sensing, and actuation perturbations kept it above $93.3\%$. Masked entity-attention MAPPO gives the most consistent full-team completion among the tested designs; broader robustness and physical deployment remain unestablished. Full article

Physics-informed neural networks (PINNs) remain difficult to train for hyperbolic conservation laws. Artificial viscosity is a common remedy, and several methods reduce its coefficient during training. We examine adaptive moment estimation (Adam) followed by limited-memory Broyden–Fletcher–Goldfarb–Shanno (L-BFGS) optimisation. The schedule followed during Adam, the viscosity used during L-BFGS, and the checkpoint supplied between them can change together. We separate these factors for the one-dimensional Buckley–Leverett equation using a fixed multilayer perceptron with hyperbolic-tangent activations and 15 paired seeds per configuration. The central experiments impose a zero-diffusive-flux outlet condition. Lowering the L-BFGS viscosity from $3.5\times {10}^{-3}$ to $1.67\times {10}^{-4}$ increased full-grid relative $L^2$ error by $0.0413\pm 0.0069$. The schedule–viscosity interaction was $-0.0067\pm 0.0075$ and was not detected after multiplicity adjustment. A decay ending at $3.5\times {10}^{-3}$ improved on the tested low-endpoint decay before high-viscosity polishing, while its difference from constant high viscosity remained inconclusive. The higher of the two tested constant viscosities improved aggregate and pre-breakthrough accuracy but increased post-breakthrough error. Stage-matched curricula showed no detected accuracy advantage; their greater executed cost arose from intermediate diagnostic L-BFGS phases whose states were discarded. These results identify the terminal viscosity, checkpoint rule, outlet treatment, and optimiser exposure as controls needed to interpret schedule comparisons in this pipeline. Full article

Background: Composite materials are not easy to model using traditional experiments and physics-based approaches due to their diverse properties and multiscale nature. This review summarizes how artificial intelligence (AI) methodologies, including machine learning, deep learning, and generative AI, have been increasingly transforming and how they have been increasingly applied to composite materials research. It combines their possible use across the prediction of their properties, damage detection, structural health monitoring, and materials design. Methods: Searches were performed in Scopus and Google Scholar for articles published using combined keywords such as “(artificial intelligence OR machine learning OR deep learning OR Generative AI) AND (composite OR fiber reinforced) AND (prediction OR optimization OR characterization)”. The search strategy was designed to capture both fundamental developments in AI methodologies relevant to materials research and their specific applications to composite materials. The identified publications underwent a two-step selection process: (i) an initial review of titles and abstracts, and (ii) a detailed full-text evaluation. Studies were included if they provided explicit descriptions of AI models, clearly specified datasets, and applied quantitative performance assessments to composite materials-related applications such as material property prediction, damage detection, characterization, or manufacturing process optimization. For each selected article, relevant information was extracted in a structured and consistent manner, including the type of composite material, AI methodology, input parameters, dataset characteristics and size, validation methods, and application domain. These data were used to support both quantitative trend analysis and the qualitative identification of research gaps and future research opportunities in composite materials research. Results: Supervised learning methods, in particular artificial neural networks, support vector machines, random forests, and decision trees, were commonly used and exhibited durable potential accuracy in predicting properties of composite materials. Unsupervised methods such as principal component analysis (PCA) and clustering were comparatively underused. Generative AI, including generative adversarial networks (GANs) and variational autoencoders (VAEs), and physics-informed models emerged as growing approaches for synthetic data generation and improved interpretability, respectively. Limitations: The reviewed studies were frequently limited by small or heterogeneous datasets, weak generalizability testing beyond training conditions, high computational requirements and limited model interpretability (“black box” behavior); a risk-of-bias assessment across included studies was not formally conducted. Conclusions: AI is shifting composite materials research from trial-and-error toward data-driven, predictive design. Realizing its full potential will require open, well-annotated datasets, physics-aware and explainable models, and closed-loop, active-learning workflows linking prediction to experimental validation. Full article