Search Results (22,338)

Gears, as critical power-transmission components in most power equipment, have a particularly urgent need for in situ inspection systems. Traditional gear inspection methods rely on contact inspection instruments, which are not only time-consuming, but also potentially damage the gear surface due to contact. This study delves into the detection requirements in the gear manufacturing process and establishes a rapid, non-contact detection mechanism and model using a CHCS. This model employs a CHCS to achieve high-speed, non-contact measurement on various surfaces with extremely high accuracy, enabling real-time monitoring of production process details, thereby improving production efficiency and ensuring product quality. Through actual inspection and comparison with a standard involute spur gear tooth profile model, this study implements a complete inspection system in a prototype. The results of gear inspection using a CHCS with an accuracy of 1 μm showed that the interquartile range of qualified gears under test (GUTs) was within 2.5 μm, and the beard line value was within 10 μm. The experiment demonstrated a layout equipped with a CHCS where the rotating axis represents the hobbing machine spindle. This method can be completed without moving the gear, enabling subsequent finishing processes. Full article

Accurate staging of bladder cancer is critical for guiding clinical management, particularly the distinction between non–muscle-invasive (T1) and muscle-invasive (T2–T4) disease. Although MRI offers superior soft-tissue contrast, image interpretation remains operator-dependent and subject to inter-observer variability. This study proposes an automated deep learning framework for MRI-based bladder cancer staging to support standardized radiological interpretation. A sequential AI-based pipeline was developed, integrating hybrid tumor segmentation using YOLOv11 for lesion detection and DeepLabV3 for boundary refinement, followed by three deep learning classifiers (VGG19, ResNet50, and Vision Transformer) for MRI-based stage prediction. A total of 416 T2-weighted MRI images with radiology-derived stage labels (T1–T4) were included, with data augmentation applied during training. Model performance was evaluated using accuracy, precision, recall, F1-score, and multi-class AUC. Performance uncertainty was characterized using patient-level bootstrap confidence intervals under a fixed training and evaluation pipeline. All evaluated models demonstrated high and broadly comparable discriminative performance for MRI-based bladder cancer staging within the present dataset, with high point estimates of accuracy and AUC, particularly for differentiating non–muscle-invasive from muscle-invasive disease. Calibration analysis characterized the probabilistic behavior of predicted stage probabilities under the current experimental setting. The proposed framework demonstrates the feasibility of automated MRI-based bladder cancer staging derived from radiological reference labels and supports the potential of deep learning for standardizing and reproducing MRI-based staging procedures. Rather than serving as an independent clinical decision-support system, the framework is intended as a methodological and workflow-oriented tool for automated staging consistency. Further validation using multi-center datasets, patient-level data splitting prior to augmentation, pathology-confirmed reference standards, and explainable AI techniques is required to establish generalizability and clinical relevance. Full article

The task of identifying high-value projects from vast investment portfolios presents a major challenge in the construction industry, particularly within the energy sector, where decision-making carries high financial and operational stakes. This complexity is driven by both the volume and heterogeneity of project documentation, as well as the multidimensional criteria used to assess project value. Despite this, research gaps remain: large language models (LLMs) as pretrained transformer encoder models are underutilized in construction project selection, especially in domains where investment precision is paramount. Existing methodologies have largely focused on multi-criteria decision-making (MCDM) frameworks, often neglecting the potential of LLMs to automate and enhance early-phase project evaluation. However, deploying LLMs for such tasks introduces high computational demands, particularly in privacy-sensitive, enterprise-level environments. This study investigates the application of the robustly optimized BERT model (RoBERTa) for identifying high-value energy infrastructure projects. Our dual objective is to (1) leverage RoBERTa’s pre-trained language architecture to extract key information from unstructured investment texts and (2) evaluate its effectiveness in enhancing project selection accuracy. We benchmark RoBERTa against several leading LLMs: BERT, DistilBERT (a distilled variant), ALBERT (a lightweight version), and XLNet (a generalized autoregressive model). All models achieved over 98% accuracy, validating their utility in this domain. RoBERTa outperformed its counterparts with an accuracy of 99.6%. DistilBERT was fastest (1025.17 s), while RoBERTa took 2060.29 s. XLNet was slowest at 4145.49 s. In conclusion, RoBERTa can be the preferred option when maximum accuracy is required, while DistilBERT can be a viable alternative under computational or resource constraints. Full article

Background/Objectives: Children and adolescents with intellectual disabilities (ID) have an elevated burden of obesity and cardiometabolic risk, yet factors associated with high blood pressure (BP) in this group remain insufficiently described. This study assessed the prevalence of hypertension (HTN) and isolated systolic hypertension (ISH) at a single visit and examined anthropometric and body composition correlates of elevated BP in children with ID. Methods: A cross-sectional study was conducted among 461 children and adolescents with ID aged 7–18 y attending special education schools in southeastern Poland. Anthropometric indicators (BMI, waist circumference [WC], hip circumference [HC], and waist-to-height ratio [WHtR]) and body composition parameters (BF%, MM%, FFM%, TBW%) were measured using standardized procedures. BP was assessed three times during one visit, and the average of the second and third readings was used. Receiver operating characteristic (ROC) analyses were used for exploratory assessment of discriminatory performance of anthropometric and body composition parameters, and multivariable logistic regression examined associations with elevated BP (HTN + ISH). Results: Overall, 13.9% of participants had HTN and 10.4% had ISH (combined prevalence: 24.3%). Abdominal obesity was present in 39.5% of participants, and elevated HC in 28.2%, both more common in girls. Higher BP categories were associated with greater WC, HC, BMI, and BF%, and lower MM%, FFM%, and TBW% (p < 0.0001). HC showed the highest discriminatory accuracy for HTN + ISH (AUC = 0.844), followed by MM%, BF%, and FFM%, whereas WHtR demonstrated limited discriminatory performance in ROC analyses. In multivariable models, WHtR ≥ 0.5 was associated with increased odds of elevated BP (OR = 4.25), whereas higher TBW% (≥55.38%) was inversely associated with elevated BP (OR = 0.17) in the total sample; similar patterns were observed in sex- and age-stratified analyses. Conclusions: Children with ID show a high prevalence of elevated BP at a single visit, including HTN-range and ISH-range values. Anthropometric indicators, particularly HC and WHtR, and BIA-derived body composition parameters reflecting higher fat mass and lower lean tissue proportion were associated with elevated BP. These exploratory findings suggest that simple anthropometric and body composition measures may help identify individuals who warrant further BP assessment, although longitudinal studies with repeated measurements are required before clinical application. Full article

Surface roughness is a critical quality attribute in laser cutting, directly influencing edge integrity, dimensional accuracy, and post-processing requirements. While most studies address surface roughness through forward modeling and optimization, practical manufacturing tasks often require solving inverse parameter selection problems, where process parameters must be chosen to satisfy prescribed surface quality requirements. In this study, surface roughness control in laser cutting is formulated within an inverse target-tracking framework based on response surface methodology (RSM). A quadratic response surface model is established using a Box–Behnken experimental design, with cutting speed, laser power, and assist-gas pressure as input factors. The fitted response surface provides an explicit forward mapping within a bounded operating window and serves as a local surrogate for methodological demonstration of target-oriented parameter estimation. Based on this surrogate model, a model-predicted feasible roughness range within the investigated design space is identified as R_a = 1.952–4.212 μm. For prescribed roughness targets within this interval, an inverse least-squares target-tracking formulation is employed to compute model-based parameter estimates. The inverse results are presented as continuous set-point maps and tabulated operating conditions, accompanied by a target-versus-predicted consistency check performed at the model level. Owing to the statistically significant lack-of-fit of the forward response surface, the inverse results presented in this study should be interpreted as theoretical, model-based estimates intended to illustrate the proposed framework rather than as experimentally validated process set-points. The proposed approach highlights both the potential and the limitations of inverse target-tracking strategies based on response surface models and underscores the need for statistically adequate models and independent experimental validation for industrial application. Full article

This study investigated heat-induced protein aggregation in skim camel milk by monitoring changes in the volume-weighted mean particle size (d_4,3) during isothermal heating (60–90 °C, up to 60 min, four temperature levels and 25 time–temperature conditions). Pronounced increases in d_4,3 with both time and temperature confirmed significant thermal aggregation. The reaction kinetics were described using a generalized exponential growth model, which fitted well at intermediate temperatures (e.g., coefficient of determination (R²) = 0.901 at 70 °C and 0.959 at 80 °C) but deviated at the lower (60 °C) and upper (90 °C) extremes, reflecting more complex behavior. Arrhenius analysis of the rate constant yielded an activation energy of 50.61 kJ mol⁻¹, lower than values typically reported for bovine milk systems, indicating that camel milk proteins require less thermal input to aggregate. In parallel, a machine learning model implemented as an artificial neural network (ANN) predicted d_4,3 from time-temperature inputs with high accuracy (R² > 0.97 across training, validation, and testing), capturing nonlinear patterns without mechanistic assumptions. Together, the kinetic and ANN approaches provide complementary insights into the heat sensitivity of camel milk proteins and offer predictive tools to support the optimization of thermal processing, formulation, and quality control in dairy applications. Full article

This paper addresses the core conflict between long-range communication and ultra-low power requirements in sensing nodes for Wireless Sensor Networks (WSNs) by proposing a wake-up receiver (WuRx) design featuring nanowatt-level power consumption and high sensitivity. Conventional architectures are plagued by low energy efficiency, poor demodulation reliability, and insufficient clock synchronization accuracy, which hinders their practical application in real-world scenarios like WSNs. The proposed design employs an event-triggered mechanism, where a continuously operating, low-power WuRx monitors the channel and activates the main system only after validating a legitimate command, thereby significantly reducing standby power. At the system design level, a key innovation is direct conjugate matching between the antenna and a multi-stage rectifier, replacing the traditional 50 Ohm interface, which substantially improves energy transmission efficiency. Furthermore, a mean-detection demodulation circuit is introduced to dynamically generate an adaptive reference level, effectively overcoming the challenge of discriminating shallow modulation caused by signal saturation in the near-field region. At the baseband processing level, a configurable fault-tolerant correlator logic and a data-edge-triggered clock synchronization circuit are designed, combined with oversampling techniques to suppress clock drift and enhance the reliability of long data packet reception. Fabricated in a TSMC 0.18 µm CMOS process, the receiver features an ultra-low power consumption of 305 nW at 0.5 V and a high sensitivity of −47 dBm, enabling a communication range of up to 400 m in the 920–925 MHz band. Through synergistic innovation at both the circuit and system levels, this research provides a high-efficiency, high-reliability wake-up solution for long-range WSN nodes, effectively promoting the large-scale application of WSN technology in practical deployments. Full article

Urban integrated energy systems (UIESs) are increasingly exposed to uncertainties arising from wind and photovoltaic variability, load fluctuations, and equipment failures, highlighting the need for refined reserve assessment and coordinated operation. This study develops a unified framework that jointly models renewable and load deviations together with a load-dependent failure probability model, using Monte Carlo sampling and K-means scenario reduction to obtain representative system states. A reserve-capacity-oriented optimisation model is formulated to minimise total operating cost—including thermal generation, energy-storage operation, and reserve cost—while satisfying power balance, reserve adequacy, unit operating limits, and state-of-charge constraints. Application to a UIES comprising a 1000 kW load, 800 kW photovoltaic unit, 100 kW wind turbine, five thermal power units (total capacity 1000 kW), and a 250 kW/370 kWh energy storage system shows that reserve requirements fluctuate between −100 kW (downward) and 500 kW (upward) across different scenarios, with uncertainty-driven reserves dominating and failure-related reserves remaining below 100 kW. The optimisation results indicate coordinated operation between thermal units and storage, with storage absorbing surplus renewable output, supporting peak shaving, and providing most upward and all downward reserves. The total operating costs under typical summer and winter scenarios are 2264.02 CNY and 3122.89 CNY, respectively, confirming the method’s ability to improve reserve estimation accuracy and support economical and reliable UIES operation under uncertainty. Full article

During the operation of above-ground main oil and gas pipelines, their elastic bends occur due to the properties of the soils in which the pipeline bases are installed, climatic factors, and the intersection of geodynamic zones. Exceeding the stress values in the pipeline wall above their permissible values leads to a rupture of the wall metal and major accidents. Most methods for estimating the values of bending stresses in the pipeline wall cannot be implemented during their operation, when the pipeline already has a bend, and the installation of any additional equipment on the pipeline requires additional investments. At the same time, the most widely used method for estimating bending stresses based on data from in-pipe diagnostics does not allow for evaluation in areas with varying internal diameters of the pipeline, as well as right-angle turns. The most promising method for estimating bending stresses is aerial laser scanning of pipelines, which consists of obtaining a cloud of points on the pipeline surface, estimating its spatial position, and calculating stress values. However, this method requires the development of more accurate algorithms for processing laser scanning data, and the method is associated with a number of difficulties that can be eliminated by developing the correct sequence of actions during scanning. Within the framework of this article, an algorithm has been developed for analyzing the coordinates of a cloud of points on the pipeline surface, which makes it possible to estimate the values of bending stresses in the pipeline wall. The influence of the unevenness of the thermal insulation surface on the stress assessment results was also studied, taking into account the minimum angle of the scanned pipeline sector, which ensures the accuracy of determining stress values up to 5% using the developed method. Full article

In remote sensing and low-altitude unmanned aerial vehicle(UAV) detection scenarios, small target detection is extremely challenging due to the low pixel proportion, sparse features, and complex backgrounds of targets. The reliability of low-altitude security, in particular, is directly dependent on the accuracy of small target detection. However, current methods still face three major limitations: insufficient detection accuracy for targets smaller than 20 pixels; artifacts and false textures introduced by Generative Adversarial Network-based enhancement, which lead to increased false detection rates; and the reliance of existing approaches on specialized architectures, resulting in weak generalization capability and difficulty in adapting to multi-scenario deployment requirements. To address these issues, this paper proposes a plug-and-play dual-mechanism collaborative enhancement framework named HD-BSNet. Firstly, a High-Frequency Differential Perception mechanism is designed to enhance the detailed feature representation of small targets. Secondly, a Background Semantic Modeling mechanism is introduced to learn key features that distinguish targets from the background. Additionally, a Parallel Multi-Scale Focus Module is constructed to further reinforce target features. Extensive experiments on three small target datasets demonstrate that the proposed method effectively improves the accuracy and generalization ability of small target detection. Full article

This paper addresses the opportunities and risks of integrating Large Language Models (LLMs) into structural engineering. Exclusive reliance on LLMs is inadequate in this field, because their probabilistic nature can lead to hallucinations and inaccuracies that are unacceptable in safety-critical domains which require rigorous calculations. To resolve this dilemma, we propose adopting Neuro-Symbolic Artificial Intelligence (NSAI), a hybrid approach that balances neural intuition with symbolic rigor. The NSAI architecture employs an intelligent query system to enrich user requests and delegate critical operations to deterministic external algorithms. This system is designed to enhance reliability and support regulatory compliance, as exemplified by the 3Muri chatbot case study, an NSAI (gemini-2.5-flash)-based intelligent assistant for structural analysis software. We developed 3Muri chatbot implementing AI processes. Our experimental results, based on over 200 questions submitted to the chatbot, show that this hybrid approach achieves 94% accuracy while keeping response times below 2 s. These results validate the feasibility of deploying AI systems in safety-critical engineering domains. Full article

Shrinkage of injection-molded parts is a major challenge for dimensional accuracy, especially for semi-crystalline polymers where crystallization induces pronounced volume change and heat release during cooling. Because packing pressure is effective only before gate or local solidification, multi-stage packing is commonly used to regulate the overall shrinkage behavior. In practice, however, the solidification/transition temperature taken from standard material tests does not necessarily represent the actual in-cavity state behavior under specific cooling rate and pressure history, which compromises the consistency of P–V–T-based shrinkage prediction. In this study, a modified P–V–T-based framework (Tait equation) is developed for polypropylene (PP) by introducing a Thermal Enthalpy Transformation Method (TETM) to determine a process-relevant solidification time and crystallization-completion temperature (including the corresponding target specific volume) directly from in-cavity melt temperature monitoring using an infrared temperature sensor. The novelty TETM utilizes the crystallization-induced enthalpy release to identify the temperature–time plateau, from which one can identify the effective solidification point. Because the Tait equation adopts a two-domain formulation (molten and solidified states), accurate identification of the domain-switching temperature is critical for reliable shrinkage prediction in practical molding conditions. In the experiment execution, the optimum filling time was defined using the minimum pressure required for melt-filling. Four target specific volumes, three melt temperatures, and two mold temperatures were examined, and a two-stage packing strategy was implemented to achieve comparable shrinkage performance under different target specific volumes. A conventional benchmark based on the solidification temperature reported in the Moldex3D material database was used for comparison only. The results show that the target specific volume determined by the TETM exhibits a more consistent and near-linear relationship with the measured shrinkage rate, demonstrating that the TETM improves the robustness of solidification-time identification and the practical usability of P–V–T information for shrinkage control. Full article

The ultrasonic pulse velocity (UPV) method is widely used for determining the dynamic modulus of elasticity of concrete. Traditionally, this approach requires assuming Poisson’s ratio (arbitrary values ranging from 0.1 to 0.25), regardless of the actual properties of the tested material. Such assumptions can lead to inaccurate estimations of the elastic modulus and limit the reliability of the method. In this study, an experimental methodology is proposed to enhance the accuracy of the estimation of the elastic modulus of concrete by combining digital image correlation (DIC) with UPV testing. The DIC technique is used during axial compression tests to directly measure the Poisson ratio of cubic concrete samples, while the dynamic modulus of elasticity is determined through UPV measurements. Subsequently, conversion models from the literature were applied to estimate the static modulus of elasticity from the dynamic modulus. The obtained values are compared with the experimental measurements of the static modulus, showing strong consistency and validating the proposed approach. The results highlight two key findings: (i) incorporating the actual Poisson ratio of the material significantly improves the precision of modulus predictions obtained via UPV, and (ii) DIC provides a reliable and adaptable tool for measuring Poisson’s ratio in concrete. Overall, the integration of DIC and UPV offers a robust and non-destructive framework for improving the assessment of mechanical properties of concrete. Full article

Soluble solid content (SSC) is a key indicator of strawberry quality. Conventional SSC measurement methods are destructive and impractical for large-scale applications. Therefore, this study developed a region-based hyperspectral imaging (HSI) and lightweight one-dimensional convolutional neural network (1D CNN) framework for nondestructive SSC prediction in strawberries. To evaluate spatial effects on predictive accuracy, the fruit surface was segmented into five groups (G1–G5). Three spectral preprocessing methods were applied with partial least squares regression and five convolutional neural network (CNN) architectures, including a simplified VGG-CNN. Larger regions generally improved prediction performance; however, the 50% region (G2) and 75% region (G3) achieved comparable performance to the full region, reducing data requirements. The simplified VGG-CNN model with SNV outperformed other models, exhibiting high accuracy with reduced computational cost, supporting its potential integration into portable and real-time sensing systems. The proposed approach can contribute to improved postharvest quality control and enhanced consumer confidence in strawberry products. Full article

Recently, deep learning methods have seen major advancements and are preferred for medical image analysis. Clinically, deep learning techniques for cancer image analysis are among the main applications for early diagnosis, detection, and treatment. Consequently, segmentation of breast histology images is a key step towards diagnosing breast cancer. However, the use of deep learning methods for image analysis is constrained by challenging features in the histology images. These challenges include poor image quality, complex microscopic tissue structures, topological intricacies, and boundary/edge inhomogeneity. Furthermore, this leads to a limited number of images required for analysis. The U-Net model was introduced and gained significant traction for its ability to produce high-accuracy results with very few input images. Many modifications of the U-Net architecture exist. Therefore, this study proposes the watershed encoder–decoder neural network (WEDN) to segment cancerous lesions in supervised breast histology images. Pre-processing of supervised breast histology images via augmentation is introduced to increase the dataset size. The augmented dataset is further enhanced and segmented into the region of interest. Data enhancement methods such as thresholding, opening, dilation, and distance transform are used to highlight foreground and background pixels while removing unwanted parts from the image. Consequently, further segmentation via the connected component analysis method is used to combine image pixel components with similar intensity values and assign them their respective labeled binary masks. The watershed filling method is then applied to these labeled binary mask components to separate and identify the edges/boundaries of the regions of interest (cancerous lesions). This resultant image information is sent to the WEDN model network for feature extraction and learning via training and testing. Residual convolutional block layers of the WEDN model are the learnable layers that extract the region of interest (ROI), which is the cancerous lesion. The method was evaluated on 3000 images–watershed masks, an augmented dataset. The model was trained on 2400 training set images and tested on 600 testing set images. This proposed method produced significant results of $98.53\%$ validation accuracy, $96.98\%$ validation dice coefficient, and $97.84\%$ validation intersection over unit (IoU) metric scores. Full article