10784 research outputs found
Sort by
Supervised penalty-based aggregation applied to motor-imagery based brain-computer-interface
No full textPublisher Copyright: © 2023 The Author(s)In this paper we propose a new version of penalty-based aggregation functions, the Multi Cost Aggregation choosing functions (MCAs), in which the function to minimize is constructed using a convex combination of two relaxed versions of restricted equivalence and dissimilarity functions instead of a penalty function. We additionally suggest two different alternatives to train a MCA in a supervised classification task in order to adapt the aggregation to each vector of inputs. We apply the proposed MCA in a Motor Imagery-based Brain–Computer Interface (MI-BCI) system to improve its decision making phase. We also evaluate the classical aggregation with our new aggregation procedure in two publicly available datasets. We obtain an accuracy of 82.31% for a left vs. right hand in the Clinical BCI challenge (CBCIC) dataset, and a performance of 62.43% for the four-class case in the BCI Competition IV 2a dataset compared to a 82.15% and 60.56% using the arithmetic mean. Finally, we have also tested the goodness of our proposal against other MI-BCI systems, obtaining better results than those using other decision making schemes and Deep Learning on the same datasets.Javier Fumanal Idocin, Javier Fernandez, and Humberto Bustince’s research has been supported by the project PID2019-108392GB I00 ( AEI/10.13039/501100011033 ). Carmen Vidaurre research has been funded by the project RyC-2014-15671 .Peer reviewe
Hyperbolic mode resonance-based acetone optical sensors powered by ensemble learning
No full textPublisher Copyright: © 2024The current work describes and compares the performance of hyperbolic mode resonance (HMR)-based sensors for the detection of acetone at parts per billion (ppb) concentrations using ensemble machine learning (EML) techniques. A pair of HMR based-sensors with resonances located in the visible (VIS) and mid infrared (MIR) regions were obtained in order to train a set of ensemble machine learning models. The response of the detection system formed by both devices in the VIS and MIR regions, with the help of the EML system, allowed the limit of detection (LoD) of the sensors to be reduced by an order of magnitude. It is the first time that HMR-based sensors are shown in practical applications, at the same time that their performance is improved using EML techniques. This opens new avenues for the use of this type of HMR-based sensors for the detection of other substances, in addition to improving the performance of any optoelectronic sensor using EML techniques.Peer reviewe
Exploring the Effect of Microstructure and Surface Recombination on Hydrogen Effusion in Zn–Ni-Coated Martensitic Steels by Advanced Computational Modeling
No full textPublisher Copyright: © 2023 The Authors. Steel Research International published by Wiley-VCH GmbH.Ultrahigh-strength steel (UHSS) structures are plated with Zn–Ni coatings because of their excellent corrosion resistance properties, but the plating process is accompanied by the production of hydrogen. The presence of hydrogen in steel results in hydrogen embrittlement. Hence, during the production of UHSS parts, dedicated outgassing steps are employed to remove the diffusible hydrogen from the steel. In a production environment, the real effect of the outgassing process and the outgassing efficiency is unknown for parts coated with Zn–Ni. Hence, a finite element model is developed to capture the evolution of the hydrogen concentration profile in coated UHSS parts during outgassing to study the influence of coating morphology and microstructural features of steel. In order to develop the geometry of the model, scanning electron microscope images are analyzed to understand the microstructure and morphology of the coating. Numerical samples are generated by combining different coating morphologies with steel substrates of varying microstructural features to attain a series of samples with varying features. The results of the outgassing simulations clearly demonstrate the major role of the coating morphology on the hydrogen flux.Peer reviewe
Sustainable geopolymer concrete for thermoelectric energy harvesting
No full textPublisher Copyright: © 2023 The AuthorsConcrete is a widely used material that presents vast opportunities for energy harvesting applications. Among these, thermoelectric concrete shows promising potential for harvesting waste heat generated in urban and industrial environments. However, the development of thermoelectric concrete poses several challenges that need to be addressed, such as the need to accurately account for the intrinsic voltage. This factor is frequently disregarded, but it has a significant impact on the measurement of thermoelectric properties in improving the intrinsically low electrical conductivity and Seebeck coefficient. In this context, this study evaluates the performance of industrially scalable geopolymer-based concrete that incorporates recycled aggregates as low-cost and environmentally friendly additives for thermoelectric energy harvesting applications. The concrete exhibited an intrinsic voltage of 15 mV, and a new test protocol was proposed to exclude its impact on thermoelectric measurements. The geopolymer concrete demonstrated a significantly high Seebeck coefficient of 570 µV/k, surpassing all previously reported values for geopolymer-based materials. Furthermore, it was discovered that the thermoelectric behavior of geopolymer-based concrete is of ionic origin, indicating that further improvements can be made by adjusting the pore solution chemistry. This finding suggests that there is ample opportunity for innovation and optimization in the development of thermoelectric concrete."This work was born under the umbrella of the project PoroPCM ( PCI2019-103657 ) funded by MCIN/AEI/10.13039/ 501100011033 and co-founded by the European Union (Programación Conjunta Internacional 2019), the project NRG-STORAGE (GA 870114) funded by the European Commission. "The authors acknowledge the support provided by the PoroPCM project (PCI2019-103657), funded by MCIN/AEI/10.13039/501100011033 and co-funded by the European Union through the 2019 International Joint Programming Initiative. Additionally, the authors express gratitude for the funding received from the European Commission for the NRG-STORAGE project (GA 870114).".Peer reviewe
Improving the Performance of Electrotactile Brain–Computer Interface Using Machine Learning Methods on Multi-Channel Features of Somatosensory Event-Related Potentials
No full textPublisher Copyright: © 2024 by the authors.Traditional tactile brain–computer interfaces (BCIs), particularly those based on steady-state somatosensory–evoked potentials, face challenges such as lower accuracy, reduced bit rates, and the need for spatially distant stimulation points. In contrast, using transient electrical stimuli offers a promising alternative for generating tactile BCI control signals: somatosensory event-related potentials (sERPs). This study aimed to optimize the performance of a novel electrotactile BCI by employing advanced feature extraction and machine learning techniques on sERP signals for the classification of users’ selective tactile attention. The experimental protocol involved ten healthy subjects performing a tactile attention task, with EEG signals recorded from five EEG channels over the sensory–motor cortex. We employed sequential forward selection (SFS) of features from temporal sERP waveforms of all EEG channels. We systematically tested classification performance using machine learning algorithms, including logistic regression, k-nearest neighbors, support vector machines, random forests, and artificial neural networks. We explored the effects of the number of stimuli required to obtain sERP features for classification and their influence on accuracy and information transfer rate. Our approach indicated significant improvements in classification accuracy compared to previous studies. We demonstrated that the number of stimuli for sERP generation can be reduced while increasing the information transfer rate without a statistically significant decrease in classification accuracy. In the case of the support vector machine classifier, we achieved a mean accuracy over 90% for 10 electrical stimuli, while for 6 stimuli, the accuracy decreased by less than 7%, and the information transfer rate increased by 60%. This research advances methods for tactile BCI control based on event-related potentials. This work is significant since tactile stimulation is an understudied modality for BCI control, and electrically induced sERPs are the least studied control signals in reactive BCIs. Exploring and optimizing the parameters of sERP elicitation, as well as feature extraction and classification methods, is crucial for addressing the accuracy versus speed trade-off in various assistive BCI applications where the tactile modality may have added value.Peer reviewe
SEA: State-Exchange Attention for High-Fidelity Physics Based Transformers
No full textPublisher Copyright: © 2024 Neural information processing systems foundation. All rights reserved.Current approaches using sequential networks have shown promise in estimating field variables for dynamical systems, but they are often limited by high rollout errors. The unresolved issue of rollout error accumulation results in unreliable estimations as the network predicts further into the future, with each step's error compounding and leading to an increase in inaccuracy. Here, we introduce the State-Exchange Attention (SEA) module, a novel transformer-based module enabling information exchange between encoded fields through multi-head cross-attention. The cross-field multidirectional information exchange design enables all state variables in the system to exchange information with one another, capturing physical relationships and symmetries between fields. Additionally, we introduce an efficient ViT-like mesh autoencoder to generate spatially coherent mesh embeddings for a large number of meshing cells. The SEA integrated transformer demonstrates the state-of-the-art rollout error compared to other competitive baselines. Specifically, we outperform PbGMR-GMUS Transformer-RealNVP and GMR-GMUS Transformer, with a reduction in error of 88% and 91%, respectively. Furthermore, we demonstrate that the SEA module alone can reduce errors by 97% for state variables that are highly dependent on other states of the system. The repository for this work is available at: https://github.com/ParsaEsmati/SEA.Peer reviewe
Designing a low-cost wireless sensor network for particulate matter monitoring: Implementation, calibration, and field-test
No full textPublisher Copyright: © 2024 Turkish National Committee for Air Pollution Research and ControlPoor air quality can provoke severe impacts on health, necessitating environmental monitoring of atmospheric particulate matter (PM) to assess potential threats to human well-being. However, traditional continuous air quality monitoring systems are often costly and time-consuming in data treatment. Lately, there is a growing trend towards the use of low-cost wireless PM sensors, providing more detailed information than standard systems. This paper presents a system designed to measure air quality, specifically, a wireless sensor network composed of a distributed sensor network linked to a cloud system. The proposed system can efficiently measure air quality as it is cost-effective, small-sized, and consumes little power. Sensor nodes based on low-power long range (LoRa) motes transmit field measurement data to the cloud via a gateway, and a cloud computing system is implemented to store, monitor, process, and visualise the data. Advanced techniques were included in our cloud for data processing and analysis to optimise the detection of PM. Laboratory and field tests in the historic Riotinto mine validate the system's viability, offering real-time air quality information for nearby populations. Once calibrated, sensors demonstrate high accuracy, presenting mean error of −0.3% and low deviation (R2 = 0.96) when compared to regulatory systems for both low (<10 μgPM10/m3) and hazardous concentrations (300 μgPM10/m3), which makes them perfect as early warning systems for atmospheric pollution in mining.Peer reviewe
PROJECT FRACTESUS: FRACTURE TOUGHNESS ROUND ROBIN ON A HIGH COPPER WELD USING MINIATURE C(T) SPECIMENS – RESULTS AND DISCUSSION
No full textPublisher Copyright: Copyright © 2024 by a non-US government agency.The present FRACTESUS project aims to determine the effect of specimen size on the fracture toughness properties. Finite element models (FEM) are used to investigate the difference between large-size and miniature compact tension (MC(T)) specimens and quantitatively assess the resulting loss of constraint due to size reduction. The optimal range of usability of MC(T) specimens can therefore be determined and evidenced with experimental results. Large inter-laboratory testing is included in the FRACTESUS project to prove the repeatability and reproducibility of the small-scale testing of fracture toughness properties. Various materials relevant for most of the available reactor materials and irradiation conditions are investigated. The experimental round robin on unirradiated materials has been completed by 13 different laboratories. The reference temperatures T0 obtained from MC(T) specimens by the different laboratories are compared to the one obtained from larger (mostly 1T-C(T)) specimens for six different materials. The considered materials are relevant for the nuclear industry and consist of four base (15Kh2MFAA, A533B LUS, A533B JRQ, A508 Cl.3) and two weld (ANP-5 and 73W) materials. In this regard, the present publication will elaborate on the investigations and the results achieved during the round robin of the material ANP-5. The ANP-5 weld metal is a modified NiCrMo1 UP weld belonging to a 6-meter (236 inch) reactor pressure vessel (RPV) test weldment. This test weldment was manufactured and post weld heat treated by Klöckner Werke AG and is supposed to be representative for RPV weldments of 1st generation pressurized water reactors (PWR). The special feature of this material is the high copper content, which makes it particularly susceptible to irradiation-induced embrittlement. In the framework of the FRACTESUS project the material was investigated according to the ASTM E1921 standard by five partners: SCK CEN, University of Cantabria, Helmholtz-Zentrum Dresden-Rossendorf, United Kingdom Atomic Energy Authority and the Framatome GmbH. Nearly 100 miniature C(T) specimens were tested to determine the transition temperature T0 and to compare it with the results of standard sized specimen. Within this ANP-5 test program two test series performed different than expected. A sharp transition from brittle to ductile without a distinct transition range was observed. The results of the test program and reasons for the unexpected material behavior will be discussed hereinafter.Peer reviewe
MODELLING THE EFFECT OF RESIDUAL STRESSES ON DAMAGE ACCUMULATION USING A COUPLED CRYSTAL PLASTICITY PHASE FIELD FRACTURE APPROACH
No full textPublisher Copyright: Copyright © 2024 by ASME.Residual stresses are a crucial factor in assessing the integrity of welded joints. These stresses are known to influence the joint’s strength under additional loading, with the altered grain structure at and near the joint a complicating factor. Consequently, a mesoscale model is essential to understand the accumulation of damage in components subjected to external loading, as well as the impact of prior loads on failure. This study addresses the interplay between loading direction and grain morphology, explicitly investigating damage accumulation. The mesoscale model includes a coupled crystal plasticity and a phase field fracture model to estimate the deformation induced during a laser beam weld of 316H stainless steel. The displacement boundary condition was derived from a mechanical model of the weld, with the application of a Chaboche model. The temperature field required for the grain growth and mechanical models were obtained through a thermal fluid dynamics framework. Investigation of crack initiation and propagation was carried using a phase-field fracture model, which allowed the consideration of prior loading. This study indicated that the direction of loading plays an important role in damage susceptibility. The modified grain structure based on the welding simulation showed a different strain at failure compared to the 316H stainless steel parent material. The achieved strain at failure was found to be lower in normal loading compared to the transverse direction. Presently, the crystal plasticity model fails to estimate the macroscopic residual stresses, illustrated by damage propagation resulting in earlier than expected ductile failure upon reloading. The potential causes are addressed and discussed in detail.Peer reviewe