Skip to main content

Tecnologías Inteligentes para la Fabricación, el diseño y las Operaciones en entornos iNdustriales

PLEC2023-010251

•
Funding agency name Agencia Estatal de Investigación
Funding agency acronym AEI
Program Programa Estatal para Impulsar la Investigación Científico-Técnica y su Transferencia
Subprogram Subprograma Estatal de Generación de Conocimiento
Call Proyectos de I+D+i en Líneas Estratégicas
Call year 2023
State Plan 2021-2023 Plan Estatal de Investigación Científica y Técnica y de Innovación
Beneficiary centre UNIVERSIDAD POLITECNICA DE MADRID

Resultados relacionados

Resultados totales (Incluyendo duplicados): 17
Encontrada(s) 1 página(s)

DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Transfer learning-enhanced deep reinforcement learning for aerodynamic airfoil optimization subject to structural constraints

  • Ramos, David
  • Lacasa, Lucas
  • Valero, Eusebio
  • Rubio, Gonzalo
© 2025 Author(s). Published under an exclusive license by AIP Publishing., The main objective of this paper is to introduce a transfer learning-enhanced deep reinforcement learning (DRL) methodology that is able to optimize the geometry of any airfoil based on concomitant aerodynamic and structural integrity criteria. To showcase the method, we aim to maximize the lift-to-drag ratio C L / C D while preserving the structural integrity of the airfoil—as modeled by its maximum thickness—and train the DRL agent using a list of different transfer learning (TL) strategies. The performance of the DRL agent is compared with Particle Swarm Optimization (PSO), a traditional gradient-free optimization method. Results indicate that DRL agents are able to perform purely aerodynamic and hybrid aerodynamic/structural shape optimization that the DRL approach outperforms PSO in terms of computational efficiency and aerodynamic improvement and that the TL-enhanced DRL agent achieves performance comparable to the DRL one, while further saving substantial computational resources., The authors thank useful discussions within the NUMATH group. D.R., E.V., and G.R. acknowledge funding from Project TIFON (No. PLEC2023-010251) funded by MCIN/AEI/10.13039/501100011033, Spain, funding from the European Union (Project HERFUSE) under GA No. 101140567. Views and opinions expressed are those of the authors only and do not necessarily reflect those of the European Union or Clean Aviation Joint Undertaking. Neither the European Union nor Clean Aviation JU can be held responsible for them. L.L. acknowledges partial financial support from via Project MISLAND (No. PID2020-114324GB-C22), CSxAI (PID2024-157526NB-I00) and the Maria de Maeztu Program for units of Excellence (Grant No. CEX2021-001164-M), the three of them funded by MICIU/AEI/10.13039/501100011033, and funding from the European Commission Chips Joint Undertaking Project No. 101194363 (NEHIL). G.R. acknowledges partial financial support received by the Grant DeepCFD (Project No. PID2022-137899OB-I00) funded by MICIU/AEI/10.13039/501100011033 and by ERDF, EU. Finally, all authors gratefully acknowledge the Universidad Politécnica de Madrid (www.upm.es) for providing computing resources on Magerit Supercomputer., With funding from the Spanish government through the "María de Maeztu Unit of Excellence" accreditation (CEX2021-001164-M), Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Towards certification: A complete statistical validation pipeline for supervised learning in industry

  • Lacasa, Lucas
  • Pardo, Abel
  • Arbelo, Pablo
  • Sánchez-Domínguez, Miguel
  • Bascones, Noelia
  • Yeste, Pablo
  • Martínez-Cava, Alejandro
  • Rubio, Gonzalo
  • Gómez, Ignacio
  • Valero, Eusebio
  • Vicente, Javier de
Machine and Deep Learning methods are gradually being integrated into industrial operations, although at different speeds for different types of industries. The aerospace and aeronautical industries have been recently pushing to develop a roadmap for concepts of design assurance and integration of neural network-related technologies in the aeronautical sector. Within such a roadmap, there is currently no clear definition of what would constitute a certifiable validation protocol for an industrial surrogate model. This paper aims to bridge this gap and contribute to the paradigm of AI-based certification in the context of supervised learning, by proposing a practical and comprehensible validation pipeline of a surrogate model that integrates concepts and methods from deep learning, mathematical optimization and statistical data science. This pipeline is represented as a directed graph of ten logical steps/nodes, incorporating both linear-sequential procedures and feedback loops to iteratively refine the surrogate model. Each of these logical steps solves specific machine learning problems by merging previously scattered key statistical methods with some novel algorithmic solutions. We provide details and best practices for each of the steps and we illustrate the application of this pipeline in a realistic supervised problem arising in aerostructural design: predicting the likelihood of different stress-related failure modes during different flight maneuvers based on a (large) set of features characterizing the aircraft internal loads and geometric parameters., UPM and Airbus have collaborated in the context of a working group established by Airbus as part of the project ONEiRE(P220021406) funded by the Spanish Research Agency Centro para el Desarrollo Tecnológico y la Innovación (CDTI), Spain. The authors also acknowledge funding from project TIFON (PLEC2023-010251) funded by MCIN/AEI/10.13039/501100011033/, Spain, and from projects DYNDEEP, Spain (EUR2021-122007), MISLAND, Spain (PID2020-114324GB-C22), and the María de Maeztu project CEX2021-001164-M (LL) also funded by MICIU/AEI/10.13039/501100011033, Spain ., With funding from the Spanish government through the "María de Maeztu Unit of Excellence" accreditation (CEX2021-001164-M), Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

A Certifiable Machine Learning-Based Pipeline to Predict Fatigue Life of Aircraft Structures

  • Ladrón, Ángel
  • Sánchez-Domínguez, Miguel
  • Rozalén, Javier
  • Sánchez, Fernando R.
  • Vicente, Javier de
  • Lacasa, Lucas
  • Valero, Eusebio
  • Rubio, Gonzalo
Fatigue life prediction is essential in both the design and operational phases of any aircraft, and in this sense safety in the aerospace industry requires early detection of fatigue cracks to prevent in-flight failures. Robust and precise fatigue life predictors are thus essential to ensure safety. Traditional engineering methods, while reliable, are time consuming and involve complex workflows, including steps such as conducting several Finite Element Method (FEM) simulations, deriving the expected loading spectrum, and applying cycle counting techniques like peak-valley or rainflow counting. These steps often require collaboration between multiple teams and tools, added to the computational time and effort required to achieve fatigue life predictions. Machine learning (ML) offers a promising complement to traditional fatigue life estimation methods, enabling faster iterations and generalization, providing quick estimates that guide decisions alongside conventional simulations. In this paper, we present a ML-based pipeline that aims to estimate the fatigue life of different aircraft wing locations given the flight parameters of the different missions that the aircraft will be operating throughout its operational life. We validate the pipeline in a realistic use case of fatigue life estimation, yielding accurate predictions alongside a thorough statistical validation and uncertainty quantification. Our pipeline constitutes a complement to traditional methodologies by reducing the amount of costly simulations and, thereby, lowering the required computational and human resources., The authors acknowledge funding from project TIFON (PLEC2023-010251) funded by MCIN/AEI/10.13039/501100011033, Spain. LL acknowledges partial support from projects MISLAND (PID2020-114324GB-C22), project CSxAI (PID2024-157526NB-I00) funded by MICIU/AEI/10.13039/501100011033/FEDER, UE, project MdM Seal of Excellence (CEX2021-001164-M) funded by MICIU/AEI/10.13039/501100011033 and from the European Commission Chips Joint Undertaking project No. 101194363 (NEHIL)., With funding from the Spanish government through the "Maria de Maeztu Centre of Excellence" accreditation (CEX2021-001164-M)., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Otros (Other)

A Certifiable Machine Learning-Based Pipeline to Predict Fatigue Life of Aircraft Structures

  • Ladrón, Ángel
  • Sánchez-Domínguez, Miguel
  • Rozalén, Javier
  • Sánchez, Fernando R.
  • Vicente, Javier de
  • Lacasa, Lucas
  • Valero, Eusebio
  • Rubio, Gonzalo
34 pages, 17 figures., Fatigue life prediction is essential in both the design and operational phases of any aircraft, and in this sense safety in the aerospace industry requires early detection of fatigue cracks to prevent in-flight failures. Robust and precise fatigue life predictors are thus essential to ensure safety. Traditional engineering methods, while reliable, are time consuming and involve complex workflows, including steps such as conducting several Finite Element Method (FEM) simulations, deriving the expected loading spectrum, and applying cycle counting techniques like peak-valley or rainflow counting. These steps often require collaboration between multiple teams and tools, added to the computational time and effort required to achieve fatigue life predictions. Machine learning (ML) offers a promising complement to traditional fatigue life estimation methods, enabling faster iterations and generalization, providing quick estimates that guide decisions alongside conventional simulations. In this paper, we present a ML-based pipeline that aims to estimate the fatigue life of different aircraft wing locations given the flight parameters of the different missions that the aircraft will be operating throughout its operational life. We validate the pipeline in a realistic use case of fatigue life estimation, yielding accurate predictions alongside a thorough statistical validation and uncertainty quantification. Our pipeline constitutes a complement to traditional methodologies by reducing the amount of costly simulations and, thereby, lowering the required computational and human resources., The authors acknowledge funding from project TIFON (PLEC2023-010251) funded by MCIN/AEI/10.13039/501100011033, Spain. LL acknowledges partial support from projects MISLAND (PID2020-114324GB-C22), project CSxAI (PID2024-157526NB-I00) funded by MICIU/AEI/10.13039/501100011033/FEDER, UE, project MdM Seal of Excellence (CEX2021-001164-M) funded by MICIU/AEI/10.13039/501100011033 and from the European Commission Chips Joint Undertaking project No. 101194363 (NEHIL)., With funding from the Spanish government through the "Maria de Maeztu Centre of Excellence" accreditation (CEX2021-001164-M)., No
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Otros (Other)

On the Role of Consistency Between Physics and Data in Physics-Informed Neural Networks (v1)

  • Becerra-Zuniga, Nicolás
  • Lacasa, Lucas
  • Valero, Eusebio
  • Rubio, Gonzalo
24 pages, 7 Figures, 3 Tables., Physics-informed neural networks (PINNs) have gained significant attention as a surrogate modeling strategy for partial differential equations (PDEs), particularly in regimes where labeled data are scarce and physical constraints can be leveraged to regularize the learning process. In practice, however, PINNs are frequently trained using experimental or numerical data that are not fully consistent with the governing equations due to measurement noise, discretization errors, or modeling assumptions. The implications of such data-to-PDE inconsistencies on the accuracy and convergence of PINNs remain insufficiently understood. In this work, we systematically analyze how data inconsistency fundamentally limits the attainable accuracy of PINNs. We introduce the concept of a consistency barrier, defined as an intrinsic lower bound on the error that arises from mismatches between the fidelity of the data and the exact enforcement of the PDE residual. To isolate and quantify this effect, we consider the 1D viscous Burgers equation with a manufactured analytical solution, which enables full control over data fidelity and residual errors. PINNs are trained using datasets of progressively increasing numerical accuracy, as well as perfectly consistent analytical data. Results show that while the inclusion of the PDE residual allows PINNs to partially mitigate low-fidelity data and recover the dominant physical structure, the training process ultimately saturates at an error level dictated by the data inconsistency. When high-fidelity numerical data are employed, PINN solutions become indistinguishable from those trained on analytical data, indicating that the consistency barrier is effectively removed. These findings clarify the interplay between data quality and physics enforcement in PINNs providing practical guidance for the construction and interpretation of physics-informed surrogate models., The authors acknowledge funding from project TIFON (PLEC2023-010251) funded by MCIN/AEI/10.13039/501100011033, Spain. This research has received funding from the European Union (ROSAS, project number 101138319). This research has received funding from the European Research Council Executive Agency under TRANSDIF-FUSE project, Grant Agreement No. 101167322. Views and opinions expressed are, however, those of the authors only and do not necessarily reflect those of the European Union. Neither the European Union nor the granting authority can be held responsible for them. LL acknowledges partial support from project CSxAI (PID2024-157526NB-I00) funded by MICIU/AEI/10.13039/501100011033/FEDER, UE, from a Maria de Maeztu Seal of Excellence grant (CEX2021-001164-M) funded by the MICIU/AEI/10.13039/501100011033, and from the European Commission Chips Joint Undertaking project No. 101194363 (NEHIL)., With funding from the Spanish government through the "Maria de Maeztu Centre of Excellence" accreditation (CEX2021-001164-M)., No
UPCommons. Portal del coneixement obert de la UPC
Publication Contribución de congreso (ConferenceOutput)

Towards the usage of geometry agnostic ß-variational-autoencoders for model order reduction of turbulent flows

  • Eiximeno Franch, Benet
  • Miró Jané, Arnau|||0000-0002-2772-6050
  • Kumar, Vishal
  • Rodríguez Pérez, Ivette María|||0000-0002-3749-277X
  • Lehmkuhl Barba, Oriol|||0000-0002-2670-1871
The present work introduces a novel data compression framework for model order reduction of turbulent flows. The compression algorithm relies on the Geometry Agnostic Variatonal-autoencoders Integration (GAVI) technique. GAVI decouples the geometry and mesh information from the process of reducing a dataset via a CNN-based autoencoder via the QR factorization of the snapshot matrix. The framework is tested on a channel flow at Ret = 950. After reducing 9 times the amount of storage required, the root mean square of the velocity fluctuations and the budgets of the Reynolds stresses are recovered with excellent agreement when compared to the original DNS., The research leading to this work has been partially funded by the project TIFON with reference PLEC2023-010251/ AEI/10.13039/501100011033. B. Eiximeno’s work was funded by a contract from the Subprograma de Ayudas Predoctorales given by the Ministerio de Ciencia e Innovación (PRE2021- 096927). O. Lehmkuhl has been partially supported by a Ramon y Cajal postdoctoral contract (Ref: RYC2018-025949-I). V. Kumar acknowledges his AI4S fellowship within the Generación D initiative by Red.es, Ministerio para la Transformación Digital y de la Función Pública, for talent attraction (C005/24-ED CV1), funded by NextGenerationEU through PRTR. The authors acknowledge the support of the Departament de Recerca i Universitats de la Generalitat de Catalunya through the research group Large-scale Computational Fluid Dynamics (ref.: 2021 SGR 00902) and the Turbulence and Aerodynamics Research Group (ref.: 2021 SGR 01051)., Peer Reviewed, 9 - Indústria, Innovació i Infraestructura
UPCommons. Portal del coneixement obert de la UPC
Publication Artículo científico (JournalArticle)

On the integration of geometry agnostic variational-autoencoders into large-scale SVD based models

  • Eiximeno Franch, Benet
  • Miró Jané, Arnau|||0000-0002-2772-6050
  • Kutz, Nathan
  • Rodríguez Pérez, Ivette María|||0000-0002-3749-277X
  • Lehmkuhl, Oriol
This manuscript introduces GAVI (Geometry Agnostic Variational-autoencoder Integration), a scalable technique for model order reduction of terabyte-level state spaces. Such methodology is agnostic to the spatial location of the data points, hence, it can handle any complex geometry regardless of the grid type in which it is represented. The dimensionality reduction of GAVI is performed on two separate steps. First, a parallel QR factorization is applied to the snapshot matrix to decompose, then a variational autoencoder learns a latent representation of the reduced R matrix. The final step can be done at laptop level regardless of the size of the original data. The methodology is used to compress the streamwise velocity component of the flow around a circular cylinder at 𝑅𝑒𝐷 = 𝑈∞𝐷∕𝜈 = 104 , the pressure coefficient of the flow around the Windsor body at 𝑅𝑒𝐿 = 𝑈∞𝐿∕𝜈 = 2.9 × 106 and the streamwise velocity of the realistic urban flow in the Zona Universitària neighborhood located in Barcelona. The latter example is the most demanding one as the 1032 snapshots, which are represented on an unstructured grid of 335 million points, have a total size of 2.05 Tb. GAVI can compress the full dataset into 6 latent vectors that recover up to the 95.29% of the energy., The research leading to this work has been partially funded by the project TIFON with reference PLEC2023-010251/ AEI/10.13039/ 501100011033 and by the Ministerio de Ciencia e Innovación of Spain (PID2023-150408OB-C21/C22). B. Eiximeno’s work was funded by a contract from the Subprograma de Ayudas Predoctorales given by the Ministerio de Ciencia e Innovación (PRE2021-096927). Oriol Lehmkuhl has been partially supported by a Ramon y Cajal postdoctoral contract (Ref: RYC2018-025949-I). The authors acknowledge the support of Departament de Recerca i Universitats de la Generalitat de Catalunya to the Research Group Large-scale Computational Fluid Dynamics (Code: 2021 SGR 00902) and the Turbulence and Aerodynamics Research Group (Code: 2021 SGR 01051). Nathan Kutz acknowledges the NSF AI Institute in Dynamics System., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Artículo científico (JournalArticle)

On deep-learning-based closures for algebraic surrogate models of turbulent flows

  • Eiximeno Franch, Benet
  • Sanchís Agudo, Marcial
  • Miró Jané, Arnau|||0000-0002-2772-6050
  • Rodríguez Pérez, Ivette María|||0000-0002-3749-277X
  • Vinuesa Motilva, Ricardo
  • Lehmkuhl Barba, Oriol|||0000-0002-2670-1871
A deep-learning-based closure model to address energy loss in low-dimensional surrogate models based on proper-orthogonal-decomposition (POD) modes is introduced. Using a transformer-encoder block with an easy-attention mechanism, the model predicts the spatial probability density function of fluctuations not captured by the truncated POD modes. The methodology is demonstrated on the wake of the Windsor body at yaw angles of d=[2.5°,5°,7.5°,10°,12.5°] , with d=7.5° as a test case, and in a realistic urban environment at wind directions of d=[-45°,-22.5°,0°,22.5°,45°] , with d=0° as a test case. Key coherent modes are identified by clustering them based on dominant frequency dynamics using Hotelling’s T2 on the spectral properties of temporal coefficients. These coherent modes account for nearly 60% and 75% of the total energy for the Windsor body and the urban environment, respectively. For each case, a common POD basis is created by concatenating coherent modes from training angles and orthonormalising the set without losing information. Transformers with different size on the attention layer, (64, 128 and 256), are trained to model the missing fluctuations in the Windsor body case. Larger attention sizes always improve predictions for the training set, but the transformer with an attention layer of size 256 slightly overshoots the fluctuation predictions in the Windsor body test set because they have lower intensity than in the training cases. A single transformer with an attention size of 256 is trained for the urban flow. In both cases, adding the predicted fluctuations close the energy gap between the reconstruction and the original flow field, improving predictions for energy, root-mean-square velocity fluctuations and instantaneous flow fields. For instance, in the Windsor body case, the deepest architecture reduces the mean energy error from 37% to 12% and decreases the Kullback–Leibler divergence of velocity distributions from DKL=0.2 to below DKL=0.026., The research leading to this work has been partially funded by the project TIFON with reference PLEC2023-010251/ AEI/10.13039/501100011033. B. Eiximeno’s work was funded by a contract from the Subprograma de Ayudas Predoctorales given by the Ministerio de Ciencia e Innovación (PRE2021-096927). O. Lehmkuhl has been partially supported by a Ramon y Cajal postdoctoral contract (Ref: RYC2018-025949-I). The authors acknowledge the support of Departament de Recerca i Universitats de la Generalitat de Catalunya to the Research Group Large-scale Computational Fluid Dynamics (Code: 2021 SGR 00902) and the Turbulence and Aerodynamics Research Group (Code: 2021 SGR 01051). M. Sanchis-Agudo and R. Vinuesa would like to acknowledge the support from Marie Sklodowska-Curie Actions project MODELAIR, funded by the European Commission under the Horizon Europe program through grant agreement number 101072559., Peer Reviewed, 9.5 - Augmentar la investigació científica i millorar la capacitat tecnològica dels sectors industrials de tots els països, en particular els països en desenvolupament, entre d’altres maneres fomentant la innovació i augmentant substancialment, d’aquí al 2030, el nombre de persones que treballen en el camp de la investigació i el desenvolupament per cada milió d’habitants, així com la despesa en investigació i desenvolupament dels sectors públic i privat, 9 - Indústria, Innovació i Infraestructura
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Hybrid State Estimation: Integrating Physics-Informed Neural Networks with Adaptive UKF for Dynamic Systems

  • Curtò i Díaz, Joaquim de
  • Zarzà i Cubero, Irene de
This article belongs to the Special Issue Cyber-Physical Systems: Recent Developments and Emerging Trends, In this paper, we present a novel approach to state estimation in dynamic systems by combining Physics-Informed Neural Networks (PINNs) with an adaptive Unscented Kalman Filter (UKF). Recognizing the limitations of traditional state estimation methods, we refine the PINN architecture with hybrid loss functions and Monte Carlo Dropout for enhanced uncertainty estimation. The Unscented Kalman Filter is augmented with an adaptive noise covariance mechanism and incorporates model parameters into the state vector to improve adaptability. We further validate this hybrid framework by integrating the enhanced PINN with the UKF for a seamless state prediction pipeline, demonstrating significant improvements in accuracy and robustness. Our experimental results show a marked enhancement in state estimation fidelity for both position and velocity tracking, supported by uncertainty quantification via Bayesian inference and Monte Carlo Dropout. We further extend the simulation and present evaluations on a double pendulum system and state estimation on a quadcopter drone. This comprehensive solution is poised to advance the state-of-the-art in dynamic system estimation, providing unparalleled performance across control theory, machine learning, and numerical optimization domains., The work is developed under the following projects at BARCELONA Supercomputing Center: ‘TIFON’. This work has also received funding from the ‘NEXTBAT’ project funded by the EUROPEAN Union’s HORIZON EUROPE research and innovation programme under grant agreement No. 101103983. The work is also developed under UFV R&D pre-competitive project ‘OpenMaas: Open Manufacturing as a Service’., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Analysis of Transportation Systems for Colonies on Mars

  • Curtò i Díaz, Joaquim de
  • Zarzà i Cubero, Irene de
The colonization of Mars poses unprecedented challenges in developing sustainable and efficient transportation systems to support inter-settlement connectivity and resource distribution. This study conducts a comprehensive evaluation of two proposed transportation systems for Martian colonies: a ground-based magnetically levitated (maglev) train and a low-orbital spaceplane. Through simulation models, we assess the energy consumption, operational and construction costs, and environmental impacts of each system. Monte Carlo simulations further provide insights into the cost variability and financial risk associated with each option over a decade. Our findings reveal that while the spaceplane system offers lower average costs and reduced financial risk, the maglev train boasts greater scalability and potential for integration with Martian infrastructural development. The maglev system, despite its higher initial cost, emerges as a strategic asset for long-term colony expansion and sustainability, highlighting the need for balanced investment in transportation technologies that align with the goals of Martian colonization. Further extending our exploration, this study introduces advanced analysis of alternative transportation technologies, including hyperloop systems, drones, and rovers, incorporating dynamic environmental modeling of Mars and reinforcement learning for autonomous navigation. In an effort to enhance the realism and complexity of our navigation simulation of Mars, we introduce several significant improvements. These enhancements focus on the inclusion of dynamic atmospheric conditions, the simulation of terrain-specific obstacles such as craters and rocks, and the introduction of a swarm intelligence approach for navigating multiple drones simultaneously. This analysis serves as a foundational framework for future research and strategic planning in Martian transportation infrastructure., We thank the following funding sources from GOETHE-University Frankfurt am Main: ‘DePP—Dezentrale Plannung von Platoons im Straßengüterverkehr mit Hilfe einer KI auf Basis einzelner LKW’, ‘Center for Data Science & AI’, and ‘Hessian.AI-AIBiology’. The work is also developed under UFV R&D precompetitive project ‘OpenMaas: Open Manufacturing as a Service’. The work is also developed under the following projects at BARCELONA Supercomputing Center: ‘TIFON’ and ‘NEXTBAT’., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

On deep-learning-based closures for algebraic surrogate models of turbulent flows

  • Eiximeno Franch, Benet
  • Sanchís Agudo, Marcial
  • Miró Jané, Arnau
  • Rodríguez Pérez, Ivette María
  • Vinuesa Motilva, Ricardo
  • Lehmkuhl Barba, Oriol
A deep-learning-based closure model to address energy loss in low-dimensional surrogate models based on proper-orthogonal-decomposition (POD) modes is introduced. Using a transformer-encoder block with an easy-attention mechanism, the model predicts the spatial probability density function of fluctuations not captured by the truncated POD modes. The methodology is demonstrated on the wake of the Windsor body at yaw angles of d=[2.5°,5°,7.5°,10°,12.5°] , with d=7.5° as a test case, and in a realistic urban environment at wind directions of d=[-45°,-22.5°,0°,22.5°,45°] , with d=0° as a test case. Key coherent modes are identified by clustering them based on dominant frequency dynamics using Hotelling’s T2 on the spectral properties of temporal coefficients. These coherent modes account for nearly 60% and 75% of the total energy for the Windsor body and the urban environment, respectively. For each case, a common POD basis is created by concatenating coherent modes from training angles and orthonormalising the set without losing information. Transformers with different size on the attention layer, (64, 128 and 256), are trained to model the missing fluctuations in the Windsor body case. Larger attention sizes always improve predictions for the training set, but the transformer with an attention layer of size 256 slightly overshoots the fluctuation predictions in the Windsor body test set because they have lower intensity than in the training cases. A single transformer with an attention size of 256 is trained for the urban flow. In both cases, adding the predicted fluctuations close the energy gap between the reconstruction and the original flow field, improving predictions for energy, root-mean-square velocity fluctuations and instantaneous flow fields. For instance, in the Windsor body case, the deepest architecture reduces the mean energy error from 37% to 12% and decreases the Kullback–Leibler divergence of velocity distributions from DKL=0.2 to below DKL=0.026., The research leading to this work has been partially funded by the project TIFON with reference PLEC2023-010251/ AEI/10.13039/501100011033. B. Eiximeno’s work was funded by a contract from the Subprograma de Ayudas Predoctorales given by the Ministerio de Ciencia e Innovación (PRE2021-096927). O. Lehmkuhl has been partially supported by a Ramon y Cajal postdoctoral contract (Ref: RYC2018-025949-I). The authors acknowledge the support of Departament de Recerca i Universitats de la Generalitat de Catalunya to the Research Group Large-scale Computational Fluid Dynamics (Code: 2021 SGR 00902) and the Turbulence and Aerodynamics Research Group (Code: 2021 SGR 01051). M. Sanchis-Agudo and R. Vinuesa would like to acknowledge the support from Marie Sklodowska-Curie Actions project MODELAIR, funded by the European Commission under the Horizon Europe program through grant agreement number 101072559., Peer Reviewed, 9.5 - Augmentar la investigació científica i millorar la capacitat tecnològica dels sectors industrials de tots els països, en particular els països en desenvolupament, entre d’altres maneres fomentant la innovació i augmentant substancialment, d’aquí al 2030, el nombre de persones que treballen en el camp de la investigació i el desenvolupament per cada milió d’habitants, així com la despesa en investigació i desenvolupament dels sectors públic i privat, 9 - Indústria, Innovació i Infraestructura, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Spectral Properties of Mimetic Operators for Robust Fluid–Structure Interaction in the Design of Aircraft Wings

  • Curtò i Díaz, Joaquim de
  • Zarzà i Cubero, Irene de
This paper presents a comprehensive study on the spectral properties of mimetic finite-difference operators and their application in the robust fluid–structure interaction (FSI) analysis of aircraft wings under uncertain operating conditions. By delving into the eigenvalue behavior of mimetic Laplacian operators and extending the analysis to stochastic settings, we develop a novel stochastic mimetic framework tailored for addressing uncertainties inherent in the fluid dynamics and structural mechanics of aircraft wings. The framework integrates random matrix theory with mimetic discretization methods, enabling the incorporation of uncertainties in fluid properties, structural parameters, and coupling conditions at the fluid–structure interface. Through spectral and localization analysis of the coupled stochastic mimetic operator, we assess the system’s stability, sensitivity to perturbations, and computational efficiency. Our results highlight the potential of the stochastic mimetic approach for enhancing reliability and robustness in the design of aircraft wings, paving the way for optimization algorithms that integrate uncertainties directly into the design process. Our findings reveal a significant impact of stochastic perturbations on the spectral radius and eigenfunction localization, indicating heightened system sensitivity. The introduction of randomized singular value decomposition (RSVD) within our framework not only enhances computational efficiency but also preserves accuracy in low-rank approximations, which is critical for handling large-scale systems. Moreover, Monte Carlo simulations validate the robustness of our stochastic mimetic framework, showcasing its efficacy in capturing the nuanced dynamics of FSI under uncertainty. This study contributes to the fields of numerical methods and aerospace engineering by offering a rigorous and scalable approach for conducting uncertainty-aware FSI analysis, which is crucial for the development of safer and more efficient aircraft., The work is developed under the following projects at BARCELONA Supercomputing Center: ‘TIFON’ and ‘NEXTBAT’. This work has also received funding from the EUROPEAN High-Performance Computing Joint Undertaking (EuroHPC-JU) under grant agreement no. 101093393 (CEEC). The work is also developed under UFV R&D pre-competitive project ‘OpenMaas: Open Manufacturing as a Service’., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Towards the usage of geometry agnostic ß-variational-autoencoders for model order reduction of turbulent flows

  • Eiximeno Franch, Benet
  • Miró Jané, Arnau
  • Kumar, Vishal
  • Rodríguez Pérez, Ivette María
  • Lehmkuhl Barba, Oriol
The present work introduces a novel data compression framework for model order reduction of turbulent flows. The compression algorithm relies on the Geometry Agnostic Variatonal-autoencoders Integration (GAVI) technique. GAVI decouples the geometry and mesh information from the process of reducing a dataset via a CNN-based autoencoder via the QR factorization of the snapshot matrix. The framework is tested on a channel flow at Ret = 950. After reducing 9 times the amount of storage required, the root mean square of the velocity fluctuations and the budgets of the Reynolds stresses are recovered with excellent agreement when compared to the original DNS., The research leading to this work has been partially funded by the project TIFON with reference PLEC2023-010251/ AEI/10.13039/501100011033. B. Eiximeno’s work was funded by a contract from the Subprograma de Ayudas Predoctorales given by the Ministerio de Ciencia e Innovación (PRE2021- 096927). O. Lehmkuhl has been partially supported by a Ramon y Cajal postdoctoral contract (Ref: RYC2018-025949-I). V. Kumar acknowledges his AI4S fellowship within the Generación D initiative by Red.es, Ministerio para la Transformación Digital y de la Función Pública, for talent attraction (C005/24-ED CV1), funded by NextGenerationEU through PRTR. The authors acknowledge the support of the Departament de Recerca i Universitats de la Generalitat de Catalunya through the research group Large-scale Computational Fluid Dynamics (ref.: 2021 SGR 00902) and the Turbulence and Aerodynamics Research Group (ref.: 2021 SGR 01051)., Peer Reviewed, 9 - Indústria, Innovació i Infraestructura, Postprint (published version)
e_Buah Biblioteca Digital Universidad de Alcalá
Publication Artículo científico (JournalArticle)

Self-Bootstrapping Automated Program Repair: Using LLMs to Generate and Evaluate Synthetic Training Data for Bug Repair

  • Fitero Domínguez, David de
  • García Cabot, Antonio|||0000-0002-0298-3237
  • García López, Eva|||0000-0002-7598-3289
This paper presents a novel methodology for enhancing Automated Program Repair (APR) through synthetic data generation utilizing Large Language Models (LLMs). Current APR systems are constrained by the limited availability of high-quality training data encompassing diverse bug types across multiple programming languages. The proposed approach addresses this limitation through a two-phase process: a synthetic sample generation followed by a rigorous quality assessment. Multiple state-of-the-art LLMs were employed to generate approximately 30,000 paired examples of buggy and fixed code across 12 programming languages and 13 bug categories. Subsequently, these samples underwent cross-model evaluation against five criteria: correctness, code quality, security, performance, and completeness. Experimental evaluation on the VulRepair test set dataset showed statistically significant improvements in Perfect Prediction rates, with the quality-filtered synthetic dataset achieving 17.18% (Top@1) and 23.00% (Top@5) compared to the baseline?s 11.68% and 18.88% respectively, representing a 47% relative improvement in Top@1 and 22% in Top@5. The methodology was validated through rigorous statistical testing, including ANOVA and post-hoc Tukey?s Honest Significant Difference analysis. Furthermore, the best-performing configurations surpassed existing systems despite using a less computationally intensive decoding strategy. This research establishes a self-bootstrapping paradigm in which LLMs generate and evaluate their own training data, suggesting promising directions for addressing data scarcity in similar software engineering tasks and advancing the development of robust, adaptable tools for automated code maintenance., Ministerio de Ciencia, Innovación y Universidades, Junta de Comunidades de Castilla-La Mancha
e_Buah Biblioteca Digital Universidad de Alcalá
Publication Artículo científico (JournalArticle)

RePaCA: leveraging reasoning large language models for static automated patch correctness assessment

  • Fuster Peña, Marcos
  • Fitero Domínguez, David de
  • García Cabot, Antonio|||0000-0002-0298-3237
  • García López, Eva|||0000-0002-7598-3289
Automated Program Repair (APR) seeks to automatically correct software bugs without requiring human intervention. However, existing tools tend to generate patches that satisfy test cases without fixing the underlying bug. These are known as overfitting patches. To address this issue, Automated Patch Correctness Assessment (APCA) attempts to identify overfitting patches generated by APR tools. This problem can be solved using a static approach, meaning that no additional information is needed beyond the original and fixed code snippets. Current static techniques often struggle with reliability, flexibility and transparency. To address these issues, we introduce RePaCA, a reasoning-based static APCA technique that leverages Large Language Models (LLMs) specialized in thinking tasks. Our model is prompted with both buggy and fixed code snippets and guided to generate a Chain of Thought that analyzes code differences, reasons about how the patch addresses the root cause, and ultimately provides a binary classification: correct or overfitting. To enhance these reasoning capabilities for the APCA task specifically, the LLM is fine-tuned using Reinforcement Learning with the Group Relative Policy Optimization algorithm. When evaluated on a standard Defects4J-derived test, our approach achieves state-of-the-art performance, with 83.1% accuracy and an 84.8% F1-score. Furthermore, our model demonstrates robustness to domain shift when trained on different datasets, outperforming the leading technique. This reasoning capability also provides explicit rationales that can support the inspection of patch assessments. These findings underscore the considerable promise of fine-tuned, reasoning LLMs to advance static APCA by enhancing accuracy and transparency., Consejería de Fomento (JCCM), Ministerio de Ciencia e Innovación, Oficina de Transferencia de Resultados de Investigación (OTRI-UAH)
e_Buah Biblioteca Digital Universidad de Alcalá
Publication Artículo científico (JournalArticle)

Test-free fault localization through deep source code analysis using encoder-based language models

  • Hoyo Gabaldón, Jesús Ángel del
  • García Cabot, Antonio|||0000-0002-0298-3237
  • García López, Eva|||0000-0002-7598-3289
  • Fitero Domínguez, David de
Fault Localization (FL) has evolved with the emergence of the Transformer architecture. Traditionally, FL relied on generating a ranking of suspiciousness scores for each code unit using information obtained during program execution from a set of test cases. However, Language Models trained with source code have enabled approaches that avoid the need for test suites, which are, sometimes, difficult or time-consuming to obtain. In this work, we present a refined training strategy tested on DeBERTa-v3-base, leveraging an encoder-based approach, to address two key challenges in FL: (1) the accurate classification of lines of code according to their suspiciousness scores, and (2) the class imbalance inherent in this type of problem. To this end, we extract the last hidden state corresponding to the embeddings of the new line tokens to generate a suspiciousness score for each line. This score then classifies each line as defective or not. The evaluation of our approach on the ConDefects dataset, repeated over ten random seeds, shows promising results compared to existing approaches, with ROC AUC gains up to 0.27 in cross-dataset comparisons. Thereby, we offer a competitive baseline for future test-free fault localization research while using language models that are approximately 200 times smaller than those used in prior work., Consejería de Fomento (JCCM), Ministerio de Ciencia e Innovación, Oficina de Transferencia de Resultados de Investigación (OTRI-UAH), Ministerio de Ciencia, Innovación y Universidades
e_Buah Biblioteca Digital Universidad de Alcalá
Publication Artículo científico (JournalArticle)

Context-aware few-shot learning SPARQL query generation from natural language on an aviation knowledge graph

  • Hernández Camero, Inés Virginia
  • García López, Eva|||0000-0002-7598-3289
  • García Cabot, Antonio|||0000-0002-0298-3237
  • Caro Álvaro, Sergio|||0000-0002-3192-8499
Question answering over domain-specific knowledge graphs implies several challenges. It requires sufficient knowledge of the world and the domain to understand what is being asked, familiarity with the knowledge graph?s structure to build a correct query, and knowledge of the query language. However, mastering all of these is a time-consuming task. This work proposes a prompt-based approach that enables natural language to generate SPARQL queries. By leveraging the advanced language capabilities of large language models (LLMs), we constructed prompts that include a natural-language question, relevant contextual information from the domain-specific knowledge graph, and several examples of how the task should be executed. To evaluate our method, we applied it to an aviation knowledge graph containing accident report data. Our approach improved the results of the original work?in which the aviation knowledge graph was first introduced?by 6%, demonstrating its potential for enhancing SPARQL query generation for domain-specific knowledge graphs., Agencia Estatal de Investigación, Ministerio de Economía, Industria y Competitividad