Search Results (178)

While gas-turbine combustors have received much research attention, the forced response of large atmospheric industrial flames is much less studied. To improve the understanding of thermoacoustic instabilities in industrial combustion systems, the forced response of a large natural-gas fired test furnace is computed using Scale-Adaptive Simulations (SASs) with a Flamelet Generated Manifold model. Two test burner configurations are compared. One produces a partially premixed flame (case P) and the other a non-premixed flame. Furthermore, the non-premixed configuration is simulated at both a slightly rich (case N) and a slightly lean set point (case NL). The flame is forced by perturbing the airflow using a superposition of sine waves at four discrete frequencies. That way, the gain and phase of the Flame Transfer Function (FTF) are determined in three simulations for a total of 12 discrete frequencies between 10 and 230 Hz. The results show very different behaviour of the partially premixed and non-premixed configurations. Case P is simulated to be a compact flame, with a maximum FTF gain of one around 70–80 Hz and a quasi-steady limit of 0.7. Case N and NL are characterised by slightly lifted flames acting as low-pass filters that quickly drop off towards higher frequencies. While the phase shift in case P is linearly dependent on frequency and can be related to its flame length, the non-premixed cases have a sharp initial phase shift that levels off with increasing frequency as the gain reduces to zero. Importantly, a non-zero phase shift at 0 Hz is observed for case NL. The nature of the combustion dynamics is further explored by a Proper Orthogonal Decomposition (POD) analysis. The FTFs are applied to predict the thermoacoustic stability using an Acoustic Network Model (ANM). This model is able to reproduce the stability of the cases observed in experiments. The results presented in this study provide insight on the effect of mixing and stoichiometry on the stability of large industrial furnaces. Full article

Implied volatility surfaces describe option-implied volatilities across strikes, and maturities and play a central role in derivative pricing and risk management. However, in practice, they are often incomplete due to illiquidity or sparse trading, requiring reliable reconstruction of missing regions. Existing approaches typically rely on parametric assumptions or latent space optimisation methods, which may be restrictive or computationally intensive. This study proposes a data-driven framework based on conditional generative adversarial networks (GANs) to map partially observed surfaces to completed ones in a single forward pass. The approach is evaluated in a controlled setting using synthetic data generated from the Heston stochastic volatility model, with varying levels of missingness (10–96%). The generator objective incorporates penalty terms enforcing the absence of call-spread, butterfly-spread, and calendar-spread arbitrage, together with a smoothness regulariser on the implied risk-neutral density. Compared with a conditional variational autoencoder (VAE), the Bates model, and the stochastic volatility-inspired (SVI) parameterisation, the proposed approach achieves lower reconstruction errors across all levels of missingness, including unseen cases, while preserving the no-arbitrage properties. An ablation study shows that the conditional GAN implicitly learns no-arbitrage behaviour, with density smoothness regularisation being the only constraint that meaningfully improves reconstruction quality. Full article

Digital Twins and Living Labs are increasingly used to support conservation, safety, accessibility, and visitor experience in cultural-heritage sites. Their practical value, however, depends on interfaces that can explain heterogeneous evidence, expose provenance, and operate under public-sector governance constraints. This paper presents a Sovereign Conversational Assistant (SCA) for the Libelium Heritage Living Lab, implemented as a small-language-model (SLM) and retrieval-augmented generation (RAG) stack that combines curated heritage and operational knowledge bases with provenance logging, refusal controls, and language enforcement. We first compare the Spanish public model BSC-LT/ALIA-40b-instruct-2601 with mistralai/Mistral-Small-3.2-24B-Instruct-2506 using 19 canonical test conditions executed over 155 repeated runs across five categories: historical queries, client experience, data analysis, hallucination resistance, and safety/ethics. Mistral passed all repeated runs, whereas ALIA passed 129/155 runs, showing strong factual and visitor-information behaviour but weaker numerical analysis, cross-lingual safety, and Spanish-language enforcement. To address external validity, we add a non-sovereign baseline comparison over the 13 canonical prompts against claude-opus-4-7, gemini-3.5-flash, and gpt-5.5 under the same RAG-conditioned harness. In this prompt-level comparison, mean final scores were ALIA 0.963, Claude Opus 4.7 0.938, Gemini 3.5 Flash 0.892, GPT-5.5 0.877, and Mistral 0.871; no pairwise difference was significant after Holm correction, and ALIA was non-inferior to the best external baseline at margins of 0.05 and 0.10, whereas Mistral was not. The contribution is therefore not a new RAG algorithm, but an operational method for deploying and evaluating a governance-aware, sovereign assistant for cultural-heritage Digital Twins, together with evidence that sovereign models can be competitive in controlled heritage RAG tasks while still requiring larger, human-calibrated benchmarks before stronger claims are made. Full article

Real-time energy management for large electric vehicle (EV) clusters requires both fast aggregate flexibility estimation and executable per-vehicle dispatch. Classical LP/MILP/MPC formulations provide strong feasibility and optimality guarantees when the model is fully specified, but their online solve time increases rapidly with cluster size; learning-based methods are fast but often rely on soft constraint penalties or external feasibility repair. We propose the Physics-Constrained Mamba-3 MIMO Aggregator (PC-M3), an amortized, constraint-aware sequence model that integrates a MIMO Mamba backbone, a history-dependent differentiable projection, a sparse routing layer, and an aggregation–disaggregation consistency loop, scaling AI-EMS from a single battery to ten-thousand-vehicle clusters in one forward pass. PC-M3 assigns every EV to one channel of a multi-input multi-output (MIMO) state-space recurrence and embeds the per-vehicle state-of-charge, power and energy constraints as a differentiable in-loop projection, jointly producing the cluster-level flexibility envelope and the per-vehicle charging trajectory. A sparse Routing-Mamba mixture-of-experts layer adaptively allocates capacity to behaviourally distinct sub-populations without supervised labels, and a consistency-trained aggregation–disaggregation loop binds the predicted envelope to the executed dispatch, forming a digital-twin-style predictive EMS pipeline that couples cluster dispatch with per-vehicle SoC evolution. On a single NVIDIA A100, PC-M3 sustains 0.34 s inference for 10,000 EVs over a 24-h horizon, about 18× faster than an Informer baseline and 2.4× faster than PowerMamba. Evaluated on the open ACN-Data and ElaadNL workplace and public charging corpora and on a 10,000-vehicle NREL dsgrid-TEMPO 2030 stress test, PC-M3 reduces the normalised envelope Hausdorff distance from 9.7% (PowerMamba) to 3.4%, cuts closed-loop cluster tracking RMSE from 1.45 MW (model predictive control) to 0.82 MW, and maintains zero observed feasibility violations with respect to the specified or imputed per-vehicle polytopes on every evaluated session. The framework provides a scalable, predictive, constraint-aware AI-EMS for V2G/G2V virtual-power-plant operation of large EV fleets. Full article

With society’s ever-growing concern for animal welfare, the equestrian industry has focused on passing and enforcing new rules to protect the main athlete, the horse. In jumping, courses go by quickly, with judges using the occurrence of conflict behaviours to assess any welfare breach. This study aimed to evaluate the overall frequency of such behaviours in Brazilian Sport Horses during elite show jumping competitions in Brazil. Conflict behaviour displays were analyzed in 120 different horse–rider pairs in jumping competitions in Brazil. All videos were observed at a slowed-down speed, with the frequency of behaviours being recorded. The observed behaviours included head shaking, tail swishing, neck hyperflexion, excessive pulling on the reins, kicking, bucking, rearing, and disobedience. All horses (100%) presented at least one type of conflict behaviour, with head shaking (100%, n = 120, IQR 3–7, range 1–18) being the most common one (p < 0.001) and tail swishing (45.83%, n = 55, IQR 1–10, range 1–29) being the second most common (p < 0.001). Most horses showed only one (41.67%, n = 50) or two (43.33%, n = 52) different types of behaviours, with few episodes of each throughout the course. Those findings were in line with studies in other disciplines, as the competition environment offers a series of challenging and stressful situations. The low levels of conflict observed in most horses indicate that the current horse welfare rules are working and must continue to be reinforced to consistently protect the horses. Full article

This study investigates the aerodynamic, aeroacoustic, and psychoacoustic behaviour of a side-by-side twin-propeller Unmanned Aerial Vehicle (UAV) system operating under both static and forward-flight conditions, with particular focus on the effects of asynchronous rotational speeds. Experiments were conducted using two identical five-bladed constant pitch propellers with a diameter of 9 in (228.6 mm) and a pitch to diameter ratio of 1. Rotational speed differences between 0 and 300 rpm were examined in 50 rpm increments at inflow velocities of 0 m/s, 14 m/s and 24 m/s. The results show that variations in rotational speed have a significant influence on both acoustic levels and perceived annoyance. Asynchronous operation causes the dominant tonal peak at the blade passing frequency to split into two components, reducing tonal reinforcement. This produces noise level reductions of approximately 2 dB in static and high advance ratio conditions, increasing to about 5 dB reduction at low advance ratios. Psychoacoustic metrics show greater sensitivity to tonal structure than to overall sound pressure level, with annoyance reductions of about 5% in static conditions and up to 15% at low advance ratios. A modest aerodynamic penalty of about 5% at $\Delta N=50$ rpm is observed, increasing with larger speed mismatches. Full article

Pedestrians are amongst the most vulnerable road user groups. Efforts to enhance pedestrian safety have mainly focused on intersections and midblock crossings. This study investigated the effect of bus stop environments on pedestrian safety in Kumasi, an area with a high incidence of pedestrian fatalities in Ghana. Crashes within a 50 m radius of bus stops were extracted using a spatial join. The Negative Binomial regression model was applied to model pedestrian crashes around bus stops as a function of three distinct non-collinear independent variable groups: road design features, bus stop characteristics, and pedestrian exposure measures. Formal bus stops were associated with higher crash rates than informal ones. The presence of medians and crosswalks was associated with lower crash rates, whereas wider carriageways were associated with higher crash rates. Higher crashes were linked to passing pedestrians and waiting pedestrians, while crossing pedestrians were associated with reduced crashes. These findings suggest that the combined effects of infrastructure and behavioural factors influence pedestrian safety at bus stops. Prioritising low-cost safety treatments, such as guard-railed waiting areas, marked crosswalks, medians, and raised crossings, around bus stops will yield substantial safety benefits for resource-constrained contexts and advance sustainable urban mobility. Full article

In this study, the commercial biobased cellulose–acetate-based blend ARBOBLEND^® 4655V was analysed with regard to its property changes after multiple mechanical recycling at three different processing temperatures (170 °C, 190 °C, and 210 °C). The results demonstrate that the material properties evolve in a distinct manner across the recycling cycles, depending on the processing temperature. While no relevant changes in the zero-shear viscosity, rheologically indicated average molecular weight, or thermal and mechanical properties were observed at 170 °C, moderate changes were observed at 190 °C, in particular an increase in the zero-shear viscosity and rheological indications of an increased average molecular weight, a broadening of the molecular weight distribution, and a change in crystallisation behaviour over the recycling passes. At 210 °C, a marked reduction in the zero-shear, rheological indications of a decreased average molecular weight, and a narrowing of the molecular weight distribution were observed. These rheology-based structural changes had an effect on the mechanical properties, such as the impact strength, the elongation at break, and the elongation at ultimate tensile stress, which were found to decrease with an increasing number of recycling passes. The study suggests that the processing temperature exerts a significant influence on the recycling behaviour of the cellulose–acetate-based blend ARBOBLEND^® 4655V. It is evident that even minor fluctuations in temperature can result in substantial changes to the physical, thermal, and mechanical properties of biopolymers. The findings underline the necessity of conducting recycling-related investigations with the processing temperature as a critical factor, particularly for temperature-sensitive materials such as cellulose–acetate-based compounds. Full article

In many non-tourism historical districts in China, property division has subdivided traditional dwellings into multi-household units. While such subdivision reshapes spatial sequences and connections, its consequences for everyday space use and circulation are rarely documented with continuous in situ evidence, partly because residential behaviour is temporally continuous and difficult to observe directly. This study examines two typical subdivision patterns in Subu Old Street: a longitudinal, single-axis serial dwelling (Case A) and a transversal, courtyard-centred dwelling (Case B). We formalize spatial units, connections, and operational nodes using a semantic ontology and map day-long Ultra-Wideband (UWB) trajectories to quantify occupancy and transition characteristics. Case A concentrates both staying and passing at the entrance-end kitchen, where activities overlap with through-movements and transition durations are short in most events but highly volatile with a long tail. Case B channels most transitions through the courtyard hub, keeping indoor rooms mainly for staying and producing longer but more stable transition durations. This study is positioned as a comparative exploratory case study of two representative subdivision patterns identified in Subu Old Street. Semantic ontology modelling, UWB-based behavioural tracking, and behavioural indicators are used together in a comparative analytical approach for examining how subdivision reorganises spatial structure and everyday residential behaviour. The results reveal pattern-specific differences in occupancy concentration, transition organisation, and movement duration. These findings are analytical observations derived from two representative cases. They provide a basis for spatial adjustment and micro-regeneration in still-inhabited subdivided traditional dwellings. Full article

Background/Objectives: Current clinical evidence suggests that cannabidiol (CBD) demonstrates therapeutic potential in the management of chronic pain, particularly in conditions involving inflammation. However, its therapeutic potential is severely limited by poor oral bioavailability, extensive first-pass metabolism, and the need for frequent high-dose administration, which compromises patient adherence and tolerability. Long-acting injectable (LAI) delivery systems offer a strategy to overcome these limitations by providing sustained plasma concentrations and reducing dosing frequency. This study aimed to develop and optimise CBD-loaded poly (lactic-co-glycolic acid) (PLGA) microspheres for LAI delivery and to evaluate poly(2-ethyl-2-oxazoline) (POx) as a functional and biocompatible alternative to the conventionally used poly (ethylene glycol) (PEG). Methods: CBD-loaded microspheres were prepared using emulsion–solvent evaporation technique. The formulations were optimised based on entrapment efficiency (EE), drug loading (DL), particle size distribution, surface morphology, thermal behaviour, in vitro release kinetics, and cytocompatibility using NIH 3T3 fibroblasts. Multiple in vitro release methodologies, including dialysis bag, shaking-flask, and USP Apparatus IV, were evaluated to identify the most discriminative and practical approach for long-term release assessment. Results: The optimised POx-based microspheres demonstrated superior control over particle size, yielding significantly smaller and more uniform particles compared with PEG-based microspheres (124 ± 1.47 µm vs. 218 ± 13.5 µm, respectively). Differential scanning calorimetry (DSC) confirmed molecular dispersion of CBD within the polymer matrix. In vitro release studies demonstrated sustained drug release over 20 days. Conclusions: POx represents a promising alternative to PEG for the formulation of CBD-loaded PLGA microspheres, offering enhanced physicochemical stability and biological compatibility. This platform supports the development of safe and effective long-acting injectable CBD therapies and consideration of POx as an alternative to PEG. Full article

Deep learning models have become central to automated Printed Circuit Board (PCB) defect detection. However, recent work has raised concerns about how reliably these models express confidence in their predictions, particularly when deployed in safety-critical inspection systems. This study conducts an empirical investigation of epistemic uncertainty across representative architectures used in PCB inspection: the two-stage Faster R-CNN detector, the one-stage YOLOv8 detector, and their corresponding classification counterparts, ResNet-50 and YOLOv8-Cls. Monte Carlo Dropout (MCD) was applied during inference to compute predictive entropy, mutual information, softmax variance, and bounding-box variability across multiple stochastic forward passes on both multiclass and binary inspection datasets. On the multiclass SolDef_AI dataset, Faster R-CNN achieved substantially stronger detection performance (mAP = 0.7607, F1 = 0.9304) and lower predictive entropy, with more stable localisation. In contrast, YOLOv8 produced markedly weaker performance (mAP = 0.2369, F1 = 0.3130) alongside higher entropy and greater bounding-box variability. On the binary Jiafuwen datasets, the YOLOv8-Cls model achieved higher overall performance (F1 = 0.6493) compared with the ResNet-50 classifier (F1 = 0.4904), reflecting its strength in simpler binary inspection tasks. Across uncertainty metrics, predictive entropy and mutual information were more sensitive to dataset size, showing higher and more variable values in the smaller multiclass dataset, whereas softmax variance and bounding-box variability appeared more architecture-dependent. These findings demonstrate that architectural choice, dataset structure, and task formulation jointly influence both performance and uncertainty behaviour. By integrating conventional metrics with uncertainty estimates, this study provides a transparent benchmark for assessing model confidence in automated optical inspection of PCBs. Full article

This study aimed to investigate the emotional changes in the brain activity of 34 special education teachers using electroencephalography (EEG) signals in response to common problem behaviours observed in students with Autism Spectrum Disorder (ASD), such as self-harm, aggression, tantrums, and stereotyped behaviours. Vignettes with Turkish narration and stimulus videos were used for each behaviour type to trigger emotions. EEG data were collected from the frontal, temporal, parietal, and occipital regions, and subjected to pre-processing steps such as band-pass filtering (0.5–40 Hz) and Independent Component Analysis (ICA), and various spectral and statistical features were extracted. To improve classification performance, feature selection was performed using the Least Absolute Shrinkage and Selection Operator (LASSO) method, and Support Vector Machine (SVM), Artificial Neural Network (ANN), Linear Discriminant Analysis (LDA), and Random Forest (RF) algorithms were used for classification. The machine learning techniques used achieved success rates of up to 97.66% F1 score in classifying teachers’ brain activity in response to different behavioural patterns. Teachers showed strong negative emotional responses to self-harm, aggression, and tantrums, while showing less response to the stereotypical behaviours. It is recommended that the study be replicated with different signals and teachers. Full article

Background/Objectives: Even with solid proof of its benefits for cardiovascular health and metabolism, adherence to the Mediterranean Diet (MD) in Spain has noticeably declined in recent years. The socioeconomic changes occurring in recent decades have prompted shifts in cooking habits and in how food is socially experienced, particularly among children and adolescents. The MD is more than just food: it is a cultural tradition and a lifestyle, rich in food and cooking skills, and food wisdom passed down over generations. When these practices fade, it affects both health and the environment, making them vital components in strengthening support for food knowledge, cooking abilities, and a healthier lifestyle. Considering these shifting dietary patterns and the growing need for targeted educational strategies, the present study aimed to investigate the association between cooking skills, MD-related health literacy, and adherence to the Mediterranean Diet across different developmental stages: childhood, adolescence, and adulthood in a sample of the Spanish population. Additionally, a secondary objective was to identify potential critical windows for intervention based on the strength of these associations. Methods: This cross-sectional study included 832 Spanish participants grouped by age: children and early adolescents (n = 408), older adolescents (n = 136), and adults (n = 288). Cooking skills were assessed using CooC11 for children and FCSk for older groups. Adults also completed Lit_MEDiet to assess MD-related health literacy. Adherence was measured with KIDMED (children/adolescents) and MEDAS (adults). Spearman correlations and standardized linear regressions were used. All statistical tests were two-sided, and statistical significance was set at p < 0.05. Results: In children, no significant association was found between cooking skills (CooC11) and KIDMED scores (β = 0.008; p = 0.875). Among adolescents, a strong positive association emerged between FCSk and KIDMED (β = 0.313; p < 0.001; ρ = 0.371), indicating a large, standardized effect and suggesting that this stage is particularly sensitive to food skills. In adults (18+), both food and cooking skills (FCSk) (β = 0.189; p = 0.001) and MD-related health literacy (Lit_MEDiet) (β = 0.187; p = 0.004) were moderately associated with MEDAS scores. Conclusions: These findings suggest that mid-adolescence could represent a favourable developmental window where food skills may hold potential to influence positive dietary behaviours. Regarding adults, the results indicate that combining practical and educational components appears to beneficial for dietary quality. Overall, this study supports the relevance of age-tailored public health strategies to potentially enhance long-term adherence to the Mediterranean Diet. Full article

Lateral passing distance (LPD) when motor vehicles overtake cyclists is a key safety metric, yet infrastructure-aware evidence remains limited. This study analyses 11,399 overtaking measurements from Austria’s OpenBikeSensor (OBS) project, spatially linked to the national road graph (GIP), with urban and rural networks examined separately. LPD was treated as a continuous dependent variable, and bivariate relationships were tested using nonparametric methods: Spearman’s rho/Kendall’s tau for metric predictors (speed limit, lane width, number of lanes) and Kruskal–Wallis tests with Dunn–Holm post hoc adjustments for categorical factors (Functional Road Class, Road Configuration, Infrastructure Type). Effect sizes and confidence intervals supported substantive interpretation. LPD was higher in rural than urban contexts, with compliance to Austria’s 2023 legal thresholds averaging 40% in cities (≥1.5 m) and 19% in rural areas (≥2.0 m). Positive correlations were found between LPD and lane width, speed limit, and functional class. The findings highlight infrastructure-sensitive patterns in sensor-generated LPD and emphasise the importance of clear cyclist allocation or physical separation, especially where high speeds or spatial constraints increase close-passing risk. Full article

This paper studies the volatility dynamics of the JSE Top40 Index by estimating a univariate GAS model with time-varying location, scale, and shape parameters (identity score scaling) and comparing its density and point-forecast performance against a stand-alone ARMA(3,2)–EGARCH(1,1) model and a hybrid ARMA(3,2)–EGARCH(1,1)–XGBoost framework. The GAS model is estimated on 3515 daily observations, and several conditional densities are examined. The Student-t GAS model (GAS–STD) obtains the lowest information criteria within the GAS family (AIC = 10,188.142; BIC = 10,243.626) and exhibits statistically significant persistence in location and scale dynamics. Statistical diagnostics provide evidence of correct density calibration (normalised log score = 1.1932; Uniform score = 0.4417), although residual skewness remains (IID-Test skewness $p=0.0134$). Out-of-sample analysis shows that GAS–STD performs strongly in density and risk forecasting, producing accurate 5% VaR and ES paths and passing coverage backtests (Kupiec LRuc $p=0.8414$; DQ $p=0.2281$). However, short-horizon point forecasts are most accurately produced by the Hybrid ARMA(3,2)–EGARCH(1,1)–XGBoost model (RMSE = 0.1386). The full Diebold-Mariano (DM) test confirms that all pairwise differences in predictive accuracy are statistically significant, and the model confidence set (MCS) procedure identifies the Hybrid model as the sole superior model at the 5% significance level, indicating that both ARMA(3,2)–EGARCH(1,1) and GAS–STD are statistically inferior. Simulation experiments illustrate that the tail behaviour of the Student-t distribution is sensitive to the degrees-of-freedom parameter $\nu$. For example, a Student-t distribution with $\nu =5$ exhibits total kurtosis of approximately $7.32$, indicating heavier tails compared to the Gaussian distribution. Overall, GAS–STD is a strong density and risk model for the JSE Top40, while the hybrid framework excels in short-term volatility forecasting. Full article