Analog performance investigation of 10 nm Junctionless GAA FETs using Machine learning methods and deep learning analysis DOI Creative Commons

R. Ouchen,

Tarek Berghout, F. Djeffal

et al.

Research Square (Research Square), Journal Year: 2024, Volume and Issue: unknown

Published: Nov. 13, 2024

Abstract With the continuous downscaling of analog CMOS-based circuits, sensitivity nanoelectronic devices to design parameter variations has significantly increased. In this paper, we introduce a novel approach that combines numerical simulations with Machine Learning (ML) analysis explore key parameters ultra-low scale Junctionless Gate-All-Around (JL GAA) Field-Effect Transistors (FETs). Accurate 3D models incorporate quantum effects and ballistic transport are employed simulate I-V characteristics 10 nm JL GAA FET devices. The influence in device geometry doping concentration on Figures-of-Merit (FoMs), such as intrinsic gain (Av) cut-off frequency, is thoroughly analyzed. use high-k dielectric materials also explored for improving frequency response high-speed circuits. By leveraging ML techniques, study identifies optimal enhance performance metrics, enabling efficient prediction optimization behavior. Our results highlight importance channel radius enhancing Moreover, investigated FETs exhibit high performances, making them ideal candidates high-gain integration machine learning techniques further streamlines process, leading identification maximize device.

Language: Английский

Leveraging Digital Twins and AI for Enhanced Gearbox Condition Monitoring in Wind Turbines: A Review DOI Creative Commons
Houssem Habbouche, Yassine Amirat, Mohamed Benbouzid

et al.

Applied Sciences, Journal Year: 2025, Volume and Issue: 15(10), P. 5725 - 5725

Published: May 20, 2025

Wind power plays a significant role in sustainable energy production, but the reliability of wind turbines depends heavily on integrity their gearboxes. Gearbox failures can lead to downtime and operational disruption. In this context, paper provides an overview evolution gearbox monitoring techniques, culminating emergence digital twin (DT) technology. We explore application DT technology condition monitoring, focusing two critical components: bearings gears. This includes comprehensive review methodologies involving model-based approaches data-driven techniques using signal processing (SP) artificial intelligence (AI) algorithms. address challenges “learning with minimal knowledge” propose framework for effective Finally, we discuss future research directions potential contributions advancing field through continued development implementation DT-based solutions.

Language: Английский

Citations

0
Integrating Learning-Driven Model Behavior and Data Representation for Enhanced Remaining Useful Life Prediction in Rotating Machinery DOI Creative Commons
Tarek Berghout, Eric Bechhoefer, F. Djeffal

et al.

Machines, Journal Year: 2024, Volume and Issue: 12(10), P. 729 - 729

Published: Oct. 15, 2024

The increasing complexity of modern mechanical systems, especially rotating machinery, demands effective condition monitoring techniques, particularly deep learning, to predict potential failures in a timely manner and enable preventative maintenance strategies. Health data analysis, widely used approach, faces challenges due randomness interpretation difficulties, highlighting the importance robust quality analysis for reliable monitoring. This paper presents two-part approach address these challenges. first part focuses on comprehensive preprocessing using only feature scaling selection via random forest (RF) algorithm, streamlining process by minimizing human intervention while managing complexity. second introduces Recurrent Expansion Network (RexNet) composed multiple layers built recursive expansion theories from multi-model learning. Unlike traditional Rex architectures, this unified framework allows fine tuning RexNet hyperparameters, simplifying their application. By combining with RexNet, methodology explores behaviors deeper interactions between dependent (e.g., health indicators) independent variables Remaining Useful Life (RUL)), offering richer insights than conventional methods. Both RF undergo hyperparameter optimization Bayesian methods under variability reduction (i.e., standard deviation) residuals, allowing algorithms reach optimal solutions enabling fair comparisons state-of-the-art approaches. Applied high-speed bearings large wind turbine dataset, achieves coefficient determination 0.9504, enhancing RUL prediction. more precise scheduling imperfect predictions, reducing downtime operational costs improving system reliability varying conditions.

Language: Английский

Citations

2
Machine learning-assisted investigation of CIGS thin-film solar cell degradation using deep learning analysis DOI

A. Maoucha,

Tarek Berghout,

F. Djeffal

et al.

Journal of Physics and Chemistry of Solids, Journal Year: 2024, Volume and Issue: unknown, P. 112526 - 112526

Published: Dec. 1, 2024

Language: Английский

Citations

2
The Neural Frontier of Future Medical Imaging: A Review of Deep Learning for Brain Tumor Detection DOI Creative Commons
Tarek Berghout

Journal of Imaging, Journal Year: 2024, Volume and Issue: 11(1), P. 2 - 2

Published: Dec. 24, 2024

Brain tumor detection is crucial in medical research due to high mortality rates and treatment challenges. Early accurate diagnosis vital for improving patient outcomes, however, traditional methods, such as manual Magnetic Resonance Imaging (MRI) analysis, are often time-consuming error-prone. The rise of deep learning has led advanced models automated brain feature extraction, segmentation, classification. Despite these advancements, comprehensive reviews synthesizing recent findings remain scarce. By analyzing over 100 papers past half-decade (2019-2024), this review fills that gap, exploring the latest methods paradigms, summarizing key concepts, challenges, datasets, offering insights into future directions using learning. This also incorporates an analysis previous targets three main aspects: results revealed primarily focuses on Convolutional Neural Networks (CNNs) their variants, with a strong emphasis transfer pre-trained models. Other Generative Adversarial (GANs) Autoencoders, used while Recurrent (RNNs) employed time-sequence modeling. Some integrate Internet Things (IoT) frameworks or federated real-time diagnostics privacy, paired optimization algorithms. However, adoption eXplainable AI (XAI) remains limited, despite its importance building trust diagnostics. Finally, outlines opportunities, focusing image quality, underexplored techniques, expanding deeper representations model behavior recurrent expansion advance imaging

Language: Английский

Citations

2
Analog performance investigation of 10 nm Junctionless GAA FETs using Machine learning methods and deep learning analysis DOI Creative Commons

R. Ouchen,

Tarek Berghout, F. Djeffal

et al.

Research Square (Research Square), Journal Year: 2024, Volume and Issue: unknown

Published: Nov. 13, 2024

Abstract With the continuous downscaling of analog CMOS-based circuits, sensitivity nanoelectronic devices to design parameter variations has significantly increased. In this paper, we introduce a novel approach that combines numerical simulations with Machine Learning (ML) analysis explore key parameters ultra-low scale Junctionless Gate-All-Around (JL GAA) Field-Effect Transistors (FETs). Accurate 3D models incorporate quantum effects and ballistic transport are employed simulate I-V characteristics 10 nm JL GAA FET devices. The influence in device geometry doping concentration on Figures-of-Merit (FoMs), such as intrinsic gain (Av) cut-off frequency, is thoroughly analyzed. use high-k dielectric materials also explored for improving frequency response high-speed circuits. By leveraging ML techniques, study identifies optimal enhance performance metrics, enabling efficient prediction optimization behavior. Our results highlight importance channel radius enhancing Moreover, investigated FETs exhibit high performances, making them ideal candidates high-gain integration machine learning techniques further streamlines process, leading identification maximize device.

Language: Английский

Citations

0