Search Results (3,262)

The global atmospheric concentration of carbon dioxide (CO₂) has exceeded 420 ppm. Adsorption-based carbon capture technologies, offer energy-efficient, sustainable solutions. Relying on classical adsorption models like Langmuir to predict CO₂ uptake presents limitations due to the need for case-specific parameter fitting. To address this, the present study introduces a universal machine learning (ML) framework using multiple algorithms—Generalized Linear Model (GLM), Feed-forward Multilayer Perceptron (DL), Decision Tree (DT), Random Forest (RF), Support Vector Machine (SVM), and Gradient Boosted Trees (GBT)—to reliably predict CO₂ adsorption capacities across diverse zeolite structures and conditions. By compiling over 5700 experimentally measured adsorption data points from 71 independent studies, this approach systematically incorporates critical factors including pore size, Si/Al ratio, cation type, temperature, and pressure. Rigorous Cross-Validation confirmed superior performance of the GBT model (R² = 0.936, RMSE = 0.806 mmol/g), outperforming other ML models and providing comparable performance with classical Langmuir model predictions without separate parameter calibration. Feature importance analysis identified pressure, Si/Al ratio, and cation type as dominant influences on adsorption performance. Overall, this ML-driven methodology demonstrates substantial promise for accelerating material discovery, optimization, and practical deployment of zeolite-based CO₂ capture technologies. Full article

Detection of surface defects can significantly elongate mechanical service time and mitigate potential risks during safety management. Traditional defect detection methods predominantly rely on manual inspection, which suffers from low efficiency and high costs. Some machine learning algorithms and artificial intelligence models for defect detection, such as Convolutional Neural Networks (CNNs), present outstanding performance, but they are often data-dependent and cannot provide guarantees for new test samples. To this end, we construct a detection model by combining Mask R-CNN, selected for its strong baseline performance in pixel-level segmentation, with Conformal Risk Control. The former evaluates the distribution that discriminates defects from all samples based on probability. The detection model is improved by retraining with calibration data that is assumed to be independent and identically distributed (i.i.d) with the test data. The latter constructs a prediction set on which a given guarantee for detection will be obtained. First, we define a loss function for each calibration sample to quantify detection error rates. Subsequently, we derive a statistically rigorous threshold by optimization of error rates and a given guarantee significance as the risk level. With the threshold, defective pixels with high probability in test images are extracted to construct prediction sets. This methodology ensures that the expected error rate on the test set remains strictly bounded by the predefined risk level. Furthermore, our model shows robust and efficient control over the expected test set error rate when calibration-to-test partitioning ratios vary. Full article

In order to improve the poor prediction effect of current filling body strength design, a support vector machine (SVM) and Lib Toolbox were used to build an optimal match model or strength index of unclassified tailings filling body. Eight main factors were analyzed and screened as condition attributes, and backfill strength as a decision attribute. Next, we selected 72 groups of training samples and 6 groups of calibration samples. Our model adopts a radial basis function (RBF) as the kernel function and uses a grid search method to optimize parameters; it then tests the combination of optimal parameters by cross-validation. Results show that the mean error of regression prediction and verified predictions made by the SVM match model were 1.01%, which were more accurate than the BP neural network model’s predictions. On the premise that stope stability is ensured, the SVM match model may decrease cement consumption and the cost of backfill more effectively, and improve economic efficiency. Full article

Imaging technologies are widely used in surveillance, medical diagnostics, and other critical applications. However, under low-light conditions, captured images often suffer from insufficient brightness, blurred details, and excessive noise, degrading quality and hindering downstream tasks. Conventional low-light image enhancement (LLIE) methods not only require annotated data but also often involve heavy models with high computational costs, making them unsuitable for real-time processing. To tackle these challenges, a lightweight and unsupervised LLIE method utilizing a dual-stage frequency-domain calibration network (DFCNet) is proposed. In the first stage, the input image undergoes the preliminary feature modulation (PFM) module to guide the illumination estimation (IE) module in generating a more accurate illumination map. The final enhanced image is obtained by dividing the input by the estimated illumination map. The second stage is used only during training. It applies a frequency-domain residual calibration (FRC) module to the first-stage output, generating a calibration term that is added to the original input to darken dark regions and brighten bright areas. This updated input is then fed back to the PFM and IE modules for parameter optimization. Extensive experiments on benchmark datasets demonstrate that DFCNet achieves superior performance across multiple image quality metrics while delivering visually clearer and more natural results. Full article

Large-sized components with numerous small key local features are essential in advanced manufacturing. Achieving high-precision quality control necessitates accurate and highly efficient three-dimensional (3D) measurement techniques. A flexible measurement system integrating a fringe-projection-based 3D scanner with an industrial robot is developed to enable the rapid measurement of large object surfaces. To enhance overall measurement accuracy, we propose an enhanced calibration method utilizing a multidimensional ball-based calibrator to simultaneously calibrate for hand-eye transformation and robot kinematic parameters. Firstly, a preliminary hand-eye calibration method is introduced to compensate for measurement errors at observation points, leveraging geometric-constraint-based optimization and a virtual single point derived via the barycentric calculation method. Subsequently, a distance-constrained calibration method is proposed to jointly estimate the hand-eye transformation and robot kinematic parameters, wherein a distance error model is constructed to link parameter errors with the measured deviations of a virtual single point. Finally, calibration and validation experiments were carried out, and the results indicate that the maximum and average measurement errors were reduced from 1.053 mm and 0.814 mm to 0.421 mm and 0.373 mm, respectively, thereby confirming the effectiveness of the proposed method. Full article

Given the increases in renewable energy penetration, appropriately allocating transmission costs is important in generation and transmission investment decisions. This study examines how a generator-side transmission charge affects investment decisions by power generation companies (PC) and the transmission system operator (TSO) under two frameworks differing in who decides investment timing. We compare two frameworks: (1) TSO determines investment timing and the PC determines capacity (TL framework); (2) PC determines investment timing and capacity (GL framework). We examine how variations in generator-side charges and demand uncertainty affect the optimal investment timing, capacity, and social surplus. Regarding investment timing, increases in the generator-side charge led to earlier investment in the TL framework but delayed investment in the GL framework. Concerning investment capacity, the TL framework yielded greater capacity with low uncertainty, while the GL framework supported greater capacity with high uncertainty. The magnitude of the relative social surplus of the two frameworks was reversed according to the generator-side charge and uncertainty. Specifically, the GL framework became increasingly superior to the TL framework as uncertainty increased, and this advantage was amplified by a higher generator-side charge. Policymakers should consider uncertainty and calibrate the level of generator-side charge and the allocation of decision-making authority. Full article

Unbalanced accelerations occurring during tram travel have a significant impact on passenger comfort and safety, as well as on the rate of wear and tear on infrastructure and rolling stock. Ideally, these dynamic forces should be monitored continuously in real-time; however, traditional systems require high-precision accelerometers and proprietary software—investments often beyond the reach of municipally funded tram operators. To this end, as part of the research project “Accelerometer Measurements in Rail Passenger Transport Vehicles”, pilot measurement campaigns were conducted in Poland on tram lines in Gdańsk, Toruń, Bydgoszcz, and Olsztyn. Off-the-shelf smartphones equipped with MEMS accelerometers and GPS modules, running the Physics Toolbox Sensor Suite Pro app, were used. Although the research employs widely known methods, this paper addresses part of the gap in affordable real-time monitoring by demonstrating that, in the future, equipment equipped solely with consumer-grade MEMS accelerometers can deliver sufficiently accurate data in applications where high precision is not critical. This paper presents an analysis of a subset of results from the Gdańsk tram network. Lateral (x) and vertical (z) accelerations were recorded at three fixed points inside two tram models (Pesa 128NG Jazz Duo and Düwag N8C), while longitudinal accelerations were deliberately omitted at this stage due to their strong dependence on driver behavior. Raw data were exported as CSV files, processed and analyzed in R version 4.2.2, and then mapped spatially using ArcGIS cartograms. Vehicle speed was calculated both via the haversine formula—accounting for Earth’s curvature—and via a Cartesian approximation. Over the ~7 km route, both methods yielded virtually identical results, validating the simpler approach for short distances. Acceleration histograms approximated Gaussian distributions, with most values between 0.05 and 0.15 m/s², and extreme values approaching 1 m/s². The results demonstrate that low-cost mobile devices, after future calibration against certified accelerometers, can provide sufficiently rich data for ride-comfort assessment and show promise for cost-effective condition monitoring of both track and rolling stock. Future work will focus on optimizing the app’s data collection pipeline, refining standard-based analysis algorithms, and validating smartphone measurements against benchmark sensors. Full article

The rapid expansion of deep-sea sensing networks underscores the critical need for accurate underwater positioning of observation base stations. However, achieving precise acoustic localization, particularly at depths exceeding 4 km, remains a significant challenge due to systematic ranging errors, clock drift, and inaccuracies in sound speed modeling. This study proposes and validates a three-tier calibration framework consisting of a Dynamic Single-Difference (DSD) solver, a geometrically optimized reference buoy selection algorithm, and a Virtual Beacon (VB) depth inversion method based on sound speed profiles. Through simulations under varying noise conditions, the DSD method effectively mitigates common ranging and clock errors. The geometric reference optimization algorithm enhances the selection of optimal buoy layouts and reference points. At a depth of 4 km, the VB method improves vertical positioning accuracy by 15% compared to the DSD method alone, and nearly doubles vertical accuracy compared to traditional non-differential approaches. This research demonstrates that deep-sea underwater target calibration can be achieved without high-precision time synchronization and in the presence of fixed ranging errors. The proposed framework has the potential to lower technological barriers for large-scale deep-sea network deployments and provides a robust foundation for autonomous deep-sea exploration. Full article

This paper, based on the Johnson–Cook damage model, investigates the damage mechanism of high-strength steel tailor welded blanks (TWBs) (Usibor1500P and Ductibor500) during the forming process. Initially, specimens with varying notch sizes were designed and fabricated to perform uniaxial tensile tests to determine their mechanical properties. Then, the deformation process of the notched specimens was simulated using finite element software, revealing the distribution and variation of stress triaxiality at the fracture surface. By combining both experimental and simulation data, the parameters of the Johnson–Cook (J–C) damage model were calibrated, and the effects of temperature, strain rate, and stress triaxiality on material fracture behavior were further analyzed. Based on finite element analysis, the relevant coefficients for stress triaxiality, strain rate, and temperature were systematically calibrated, successfully establishing a J–C fracture criterion for TWB welds, Usibor1500P, and Ductibor500 high-strength steels. Finally, the calibrated damage model was further validated through the Nakajima-type bulge test, and the simulated Forming Limit Diagram (FLD) closely matched the experimental data. The results show that the analysis based on the J–C damage model can effectively predict the fracture behavior of tailor welded blanks (TWB) during the forming process. This study provides reliable numerical predictions for the damage behavior of high-strength steel laser-customized welded sheets and offers a theoretical basis for engineering design and material performance optimization. Full article

Aging railroad bridges present complex challenges due to advancing deterioration and outdated design assumptions. This study develops a comprehensive analytical approach for assessing an aging steel truss railroad bridge through finite element (FE) modeling, sensitivity analysis, and model updating, supported by field testing. An initial FE model of the bridge was created based on original drawings and field observations. Field testing using a laser Doppler vibrometer captured the bridge’s dynamic response (vibrations and deflections) under regular train traffic. Key structural parameters (material properties, section properties, support conditions) were identified and varied in a sensitivity analysis to determine their influence on model outputs. A hybrid sensitivity analysis combining log-normal sampling and a genetic algorithm (GA) was employed to explore the parameter space and calibrate the model. The GA optimization tuned the FE model parameters to minimize discrepancies between simulated results and field measurements, focusing on vertical deflections and natural frequencies. The updated FE model showed significantly improved agreement with observed behavior; for example, vertical deflections under a representative train were matched within a few percent, and natural frequencies were accurately reproduced. This validated model provides a more reliable tool for predicting structural performance and fatigue life under various loading scenarios. The results demonstrate that integrating field data, sensitivity analysis, and model updating can greatly enhance the accuracy of structural assessments for aging railroad bridges, supporting more informed maintenance and management decisions. Full article

Virtual reality (VR) can enrich neuropsychological testing, yet the ergonomic trade-offs of its input modes remain under-examined. Seventy-seven healthy volunteers—young (19–29 y) and middle-aged (35–56 y)—completed a VR Trail Making Test with three pointing methods: eye-tracking, head-gaze, and a six-degree-of-freedom hand controller. Completion time, spatial accuracy, and error counts for the simple (Trail A) and alternating (Trail B) sequences were analysed in 3 × 2 × 2 mixed-model ANOVAs; post-trial scales captured usability (SUS), user experience (UEQ-S), and acceptability. Age dominated behaviour: younger adults were reliably faster, more precise, and less error-prone. Against this backdrop, input modality mattered. Eye-tracking yielded the best spatial accuracy and shortened Trail A time relative to manual control; head-gaze matched eye-tracking on Trail A speed and became the quickest, least error-prone option on Trail B. Controllers lagged on every metric. Subjective ratings were high across the board, with only a small usability dip in middle-aged low-gamers. Overall, gaze-based ray-casting clearly outperformed manual pointing, but optimal choice depended on task demands: eye-tracking maximised spatial precision, whereas head-gaze offered calibration-free enhanced speed and error-avoidance under heavier cognitive load. TMT-VR appears to be accurate, engaging, and ergonomically adaptable assessment, yet it requires age-specific–stratified norms. Full article

In this study, an accurate attenuation measurement system with high attenuation capability (≥100 dB) is presented, covering a broad radio frequency range from 1 GHz to 25 GHz. The system employs a dual-channel intermediate frequency (IF) substitution method, utilizing a programmable inductive voltage divider (IVD) that provides precise voltage ratios at a 1 kHz operating IF, serving as the primary attenuation standard. To ensure optimal inter-channel isolation, essential for accurate high-attenuation measurements, an optical fiber assembly, consisting of a laser diode, a wideband external electro-optic modulator, and a photodetector, is integrated between the channels. A comprehensive performance evaluation is presented, with particular emphasis on the programmable IVD calibration technique, which achieves an accuracy better than 0.001 dB across all attenuation levels, and on the role of the optical fiber assembly in enhancing isolation, demonstrating levels exceeding 120 dB across the entire frequency range. The system demonstrates measurement capabilities with expanded uncertainties (k = 2) of 0.004 dB, 0.008 dB, and 0.010 dB at attenuation levels of 20 dB, 60 dB, and 100 dB, respectively. Full article

To mitigate the technical challenges associated with small-object detection, feature degradation, and spatial-contextual misalignment in UAV-acquired imagery, this paper proposes UAV-DETR, an enhanced Transformer-based object detection model designed for aerial scenarios. Specifically, UAV imagery often suffers from feature degradation due to low resolution and complex backgrounds and from semantic-spatial misalignment caused by dynamic shooting conditions. This work addresses these challenges by enhancing feature perception, semantic representation, and spatial alignment. Architecturally extending the RT-DETR framework, UAV-DETR incorporates three novel modules: the Channel-Aware Sensing Module (CAS), the Scale-Optimized Enhancement Pyramid Module (SOEP), and the newly designed Context-Spatial Alignment Module (CSAM), which integrates the functionalities of contextual and spatial calibration. These components collaboratively strengthen multi-scale feature extraction, semantic representation, and spatial-contextual alignment. The CAS module refines the backbone to improve multi-scale feature perception, while SOEP enhances semantic richness in shallow layers through lightweight channel-weighted fusion. CSAM further optimizes the hybrid encoder by simultaneously correcting contextual inconsistencies and spatial misalignments during feature fusion, enabling more precise cross-scale integration. Comprehensive comparisons with mainstream detectors, including Faster R-CNN and YOLOv5, demonstrate that UAV-DETR achieves superior small-object detection performance in complex aerial scenarios. The performance is thoroughly evaluated in terms of mAP@0.5, parameter count, and computational complexity (GFLOPs). Experiments on the VisDrone2019 dataset benchmark demonstrate that UAV-DETR achieves an mAP@0.5 of 51.6%, surpassing RT-DETR by 3.5% while reducing the number of model parameters from 19.8 million to 16.8 million. Full article

Climate concerns are driving the automotive industry to adopt advanced manufacturing technologies that aim to improve energy efficiency and reduce vehicle weight. In this context, lightweight structural materials such as aluminium alloys have gained significant attention due to their favorable strength-to-weight ratio. Laser welding plays a crucial role in assembling such materials, offering high flexibility and fast joining capabilities for thin aluminium sheets. However, welding these materials presents specific challenges, particularly in controlling heat input to minimize distortions and ensure consistent weld quality. As a result, numerical simulations based on the Finite Element Method (FEM) are essential for predicting weld-induced phenomena and optimizing process performance. This study investigates welding-induced distortions in laser butt welding of 1.5 mm-thick Al 6061 samples through FEM simulations performed in the SYSWELD 2024.0 environment. The methodology provided by the software is based on the Moving Heat Source (MHS) model, which simulates the physical movement of the heat source and typically requires extensive calibration through destructive metallographic testing. This transient approach enables the detailed prediction of thermal, metallurgical, and mechanical behavior, but it is computationally demanding. To improve efficiency, the Imposed Thermal Cycle (ITC) model is often used. In this technique, a thermal cycle, extracted from an MHS simulation or experimental data, is imposed on predefined subregions of the model, allowing only mechanical behavior to be simulated while reducing computation time. To avoid MHS-based calibration, this work proposes using thermal cycles acquired in-line during welding via infrared thermography as direct input for the ITC model. The method was validated experimentally and numerically, showing good agreement in the prediction of distortions and a significant reduction in workflow time. The distortion values from simulations differ from the real experiment by less than 0.3%. Our method exhibits a slight decrease in performance, resulting in an increase in estimation error of 0.03% compared to classic approaches, but more than 85% saving in computation time. The integration of real process data into the simulation enables a virtual representation of the process, supporting future developments toward Digital Twin applications. Full article

This study presents preliminary findings from an experimental campaign conducted on a pilot-scale green hydrogen production plant powered by a photovoltaic (PV) system. The integrated setup, implemented at the University “Mediterranea” of Reggio Calabria, includes renewable energy generation, hydrogen production via electrolysis, on-site storage, and reconversion through fuel cells. The investigation assessed system performance under different configurations (on-grid and selective stand-alone modes), focusing on key operational phases such as inerting, purging, pressurization, hydrogen generation, and depressurization. Results indicate a strong linear correlation between the electrolyser’s power setpoint and the pressure rise rate, with a maximum gradient of 0.236 bar/min observed at 75% power input. The system demonstrated robust and stable operation, efficient control of shutdown sequences, and effective integration with PV input. These outcomes support the technical feasibility of small-scale hydrogen systems driven by renewables and offer valuable reference data for calibration models and future optimization strategies. Full article