10784 research outputs found
Sort by
Finite element method for minimizing geometric error in the bending of large sheets
No full textPublisher Copyright: © The Author(s), under exclusive licence to Springer-Verlag London Ltd., part of Springer Nature 2024.Minimizing geometric error in the bending of large sheets remains a challenging endeavor in the industrial environment. This specific industrial operation is characterized by protracted cycles and limited batch sizes. Coupled with extended cycle times, the process involves a diverse range of dimensions and materials. Given these operational complexities, conducting practical experimentation for data extraction and control of industrial process parameters proves to be unfeasible. To gain insights into the process, finite element models serve as invaluable tools for simulating industrial processes for reducing experimental cost. Consequently, the primary objective of this research endeavor is to develop an intelligent finite element model capable of providing operators with pertinent information regarding the optimal range of key parameters to mitigate geometric error in the bending of large sheets. This prediction model is based on response surface method to predict the bending diameter of the pipe taking into account three main process parameters: the plate thickness, the length, and the roll displacement. These results present promising prospects for the automation of the industrial process because the average geometric error in curvature is recorded at 0.97%, thereby meeting the stringent industrial requirement for achieving such bending with minimal equivalent plastic deformation.Peer reviewe
On the Improvement of Generalization and Stability of Forward-Only Learning via Neural Polarization
Get PDFPublisher Copyright: © 2024 The Authors.Forward-only learning algorithms have recently gained attention as alternatives to gradient backpropagation, replacing the backward step of this latter solver with an additional contrastive forward pass. Among these approaches, the so-called Forward-Forward Algorithm (FFA) has been shown to achieve competitive levels of performance in terms of generalization and complexity. Networks trained using FFA learn to contrastively maximize a layer-wise defined goodness score when presented with real data (denoted as positive samples) and to minimize it when processing synthetic data (corr. negative samples). However, this algorithm still faces weaknesses that negatively affect the model accuracy and training stability, primarily due to a gradient imbalance between positive and negative samples. To overcome this issue, in this work we propose a novel implementation of the FFA algorithm, denoted as Polar-FFA, which extends the original formulation by introducing a neural division (polarization) between positive and negative instances. Neurons in each of these groups aim to maximize their goodness when presented with their respective data type, thereby creating a symmetric gradient behavior. To empirically gauge the improved learning capabilities of our proposed Polar-FFA, we perform several systematic experiments using different activation and goodness functions over image classification datasets. Our results demonstrate that Polar-FFA outperforms FFA in terms of accuracy and convergence speed. Furthermore, its lower reliance on hyperparameters reduces the need for hyperparameter tuning to guarantee optimal generalization capabilities, thereby allowing for a broader range of neural network configurations.Peer reviewe
Methodology for the Geo-Referenced Urban-Scale Assessment of High Electrification Scenarios
No full textThis paper introduces a methodology for assessing high electrification scenarios and their impact on the electricity grid at an urban scale. The methodology uses a bottom-up approach with hourly resolution, starting with the georeferenced energy demand characterization of buildings. It then sets targets for decarbonisation technology deployment using urban-scale energy system analyses. These technologies are geospatially distributed using a weighted criteria allocation method and the AHP-TOPSIS method. The reference scenario's final configuration includes a baseline energy balance for each consumer and storage device management strategies. Power flow analysis is conducted, and solutions minimizing potential congestion in the electricity infrastructure are evaluated. Alternative scenarios include smart flexibility strategies and distributed solar generation. AI techniques are proposed for creating alternative scenarios, including PCA and clustering to identify representative cases. Technological solutions are then optimized for each case. A residential neighbourhood in Bilbao is used as a case study.Peer reviewe
Peer-to-Peer Energy Trading Approaches: Maximizing the Active Participation of the Prosumers in the Multi-carrier Energy Communities
No full textPublisher Copyright: © 2025 WILEY-VCH GmbH. Published 2025 by WILEY-VCH GmbH. All rights reserved.Peer-to-peer (P2P) energy trading represents a transformative shift in the multi-energy market, empowering prosumers, promoting the adoption of renewable energy, and fostering a more sustainable and resilient energy system. However, it also poses relevant technical, economic, and social challenges that must be addressed in order to encourage end-users to cooperate among themselves, implement a suitable decision-making process with minimal conflict of interests for all users, and deal with multiple actors in the multi-energy network requesting services with different objectives in mind. In this regard, various pilot projects and research have been carried out at different levels worldwide to overcome these issues. This chapter presents a comprehensive review of multiple aspects related to P2P energy trading in multi-energy markets, starting from the background and motivation to its architecture, market design, and technical restrictions.Peer reviewe
Gravity as a Key Constraint on Sensorimotor Coordination
No full textPublisher Copyright: © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.Sensing of the gravitation vertical is an essential component of the postural control system, and it is generally accepted that vestibular signals provide important, but not exclusive information about this important reference orientation. Less obvious is the need for vestibular information for tasks of visual orientation perception and sensorimotor coordination. In a recent study we tested patient suffering from vestibular disorders and compared their results with age-matched controls on a task of visual orientation matching. This comparison provides conclusive evidence that vestibular signal play a role in the multisensory perception of the vertical axis, a key constraint for the programming of arm movements and for arm-hand coordination.Peer reviewe
Automated MOLDAM Robotic System for 3D Printing: Manufacturing Aeronautical Mould Preforms
No full textPublisher Copyright: © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.Additive manufacturing by extrusion of thermoplastic materials is becoming a promising technology due to the flexibility for generating parts of various sizes with complicated geometries, reducing lead time. Specifically, the additive manufacturing of thermoplastics in pellet format allows material deposition rates of tens of kg/h, being able to generate thermoplastic pre-molds in a reasonable time that can help improve profitability compared to the manufacture of metal moulds, contributing to material saving. Under this approach, a new automated robotic cell has been developed to manufacture pre-moulds for the aeronautical sector integrating a Fanuc M900iB/700 robot and a high feed rate pellet extruder. Along the project, the printed material has been selected and characterized under certain conditions, taking into account the requirements of the most demanding applications. On the other hand, the behaviour of the designed parts under working con-ditions has been simulated. Finally, the implementation of the robotic system has allowed the preliminary characterization of the pre-mould geometrical performance under the customer’s requirements. This article presents the manufacturing results of one use case using a new automated robotic system for 3D printing including future challenges.Peer reviewe
Closed-Loop Platform for Human Movement Augmentation with Electrotactile Feedback
No full textPublisher Copyright: © 2024 IEEE.This work aims to fulfil and test Human Move-ment Augmentation with a closed-loop approach and a novel electrotactile feedback strategy. The user controls a Supernu-merary Robotic Arm by moving their foot, while receiving on their thigh electrotactile feedback conveying information on the robot's position. The feedback is provided through a matrix of 60 electrodes by stimulating different pads with different fre-quencies to represent 3D coordinates. A preliminary experiment was conducted with eight participants evenly distributed in Test and Control groups, who performed the task with and without electrotactile stimulation, respectively. After a learning phase, they were required to move the robot on targets displayed through augmented reality, without visual feedback of the robot. The Test group could rely on the electrotactile feedback and achieved fairly better performance than the Control group, with a success rate of 11±10% compared to the 8±5% of the Control group.Peer reviewe
Albumin Protein Impact on Early-Stage In Vitro Biodegradation of Magnesium Alloy (WE43)
No full textPublisher Copyright: © 2023 The Authors. Published by American Chemical Society.Mg and its alloys are promising biodegradable materials for orthopedic implants and cardiovascular stents. The first interactions of protein molecules with Mg alloy surfaces have a substantial impact on their biocompatibility and biodegradation. We investigate the early-stage electrochemical, chemical, morphological, and electrical surface potential changes of alloy WE43 in either 154 mM NaCl or Hanks’ simulated physiological solutions in the absence or presence of bovine serum albumin (BSA) protein. WE43 had the lowest electrochemical current noise (ECN) fluctuations, the highest noise resistance (Zn = 1774 Ω·cm2), and the highest total impedance (|Z| = 332 Ω·cm2) when immersed for 30 min in Hanks’ solution. The highest ECN, lowest Zn (1430 Ω·cm2), and |Z| (49 Ω·cm2) were observed in the NaCl solution. In the solutions containing BSA, a unique dual-mode biodegradation was observed. Adding BSA to a NaCl solution increased |Z| from 49 to 97 Ω·cm2 and decreased the ECN signal of the alloy, i.e., the BSA inhibited corrosion. On the other hand, the presence of BSA in Hanks’ solution increased the rate of biodegradation by decreasing both Zn and |Z| while increasing ECN. Finally, using scanning Kelvin probe force microscopy (SKPFM), we observed an adsorbed nanolayer of BSA with aggregated and fibrillar morphology only in Hanks’ solution, where the electrical surface potential was 52 mV lower than that of the Mg oxide layer.Peer reviewe
Using the Point Method to estimate failure loads in 3D printed graphene-reinforced PLA notched plates
No full textPublisher Copyright: © Published under licence by IOP Publishing Ltd.This work estimates failure loads in Fused Filament Fabrication (FFF) printed graphene-reinforced PLA (polylactic acid) plates containing different types of stress risers. With this aim, firstly, several notched plates are tested and conducted to fracture. Then, linear elastic Finite Element (FE) analyses are completed to define the corresponding stress profiles and, finally, the Point Method (PM) is applied to establish the failure criterion. This approach asserts that fracture conditions are achieved when the stress level equates the inherent strength (σ0) at a distance from the notch tip equal to L/2, so both parameters (related to each other through the material fracture toughness, Kmat) have been defined beforehand. The estimations of fracture loads obtained following this approach agree with the experimental results. Thus, the present work demonstrates the accuracy of the PM to estimate failure loads in this 3D printed material.Peer reviewe
Deep learning applications on cybersecurity: A practical approach
No full textPublisher Copyright: © 2023 Elsevier B.V.One of the most difficult challenges for computer systems has been security. On the other hand, new developments in machine learning are having an impact on almost every aspect of computer science, including cybersecurity. To analyze this impact, we have created three distinct cybersecurity-related problems to show the advantages of deep learning techniques. We examined their potential applications for SPAM filtering, detecting malicious software, and adult-content detection. We experimented with various techniques, such as Long Short-Term Memory (LSTMs) for spam filtering, Deep Neural Networks (DNNs) for malware detection, Convolutional Neural Networks (CNNs) combined with Transfer Learning for adult content detection and image augmentation methods. We are able to achieve an Area Under ROC Curve greater than 0.94 in every scenario, proving that excellent performance with a good relation between cost and effectiveness may be created without the need of complex designs.Peer reviewe