Nathalie Bartoli - Academia.edu (original) (raw)

Papers by Nathalie Bartoli

Journal of Intelligent & Robotic Systems, 2010

Presented is a novel framework for performing flexible computational design studies at preliminar... more Presented is a novel framework for performing flexible computational design studies at preliminary design stage. It incorporates a workflow management device (WMD) and a number of advanced numerical treatments, including multi-objective optimization, sensitivity analysis and uncertainty management with emphasis on design robustness. The WMD enables the designer to build, understand, manipulate and share complex processes and studies. Results obtained after applying the WMD on various test cases, showed a significant reduction of the iterations required for the convergence of the computational system. The tests results also demonstrated the capabilities of the advanced treatments as follows:

Unmanned Systems

Vertical takeoff and landing (VTOL) vehicles are among the most versatile UAVs, appropriate for v... more Vertical takeoff and landing (VTOL) vehicles are among the most versatile UAVs, appropriate for various missions. Given that there are still open challenges regarding the VTOL design, this paper presents the full development and test cycle of a tail-sitter. IMAV 2022 competition rules are used to define the mission objective and constraints. A multidisciplinary design and optimization strategy is defined with the goal of maximizing competition score considering design, manufacturing, and competition constraints. While still being a work in progress, the resulting vehicle is designed to fly at 18 m/s, for around 14 minutes while carrying 200g of payload and weighting approximately 750g.

arXiv (Cornell University), May 23, 2023

The Surrogate Modeling Toolbox (SMT) is an open-source Python package that offers a collection of... more The Surrogate Modeling Toolbox (SMT) is an open-source Python package that offers a collection of surrogate modeling methods, sampling techniques, and a set of sample problems. This paper presents SMT 2.0, a major new release of SMT that introduces significant upgrades and new features to the toolbox. This release adds the capability to handle mixed-variable surrogate models and hierarchical variables. These types of variables are becoming increasingly important in several surrogate modeling applications. SMT 2.0 also improves SMT by extending sampling methods, adding new surrogate models, and computing variance and kernel derivatives for Kriging. This release also includes new functions to handle noisy and use multi-fidelity data. To the best of our knowledge, SMT 2.0 is the first open-source surrogate library to propose surrogate models for hierarchical and mixed inputs. This open-source software is distributed under the New BSD license 1 .

Mechanics & Industry, 2022

Aeroelastic scaling theory shows that the design problem of aeroelastically equivalent scaled air... more Aeroelastic scaling theory shows that the design problem of aeroelastically equivalent scaled aircraft can be treated as a structural-only design problem if the aerodynamic shape and airflow properties of the full scale aircraft are preserved. In that case, the theory shows that it is sufficient to match the scaled natural mode shapes, frequencies and mass of the reference aircraft. In this paper, we present a new method for the dynamic scaling of flexible structures where the objective function is based on the modal assurance criterion (MAC). This criterion is used for a mode tracking strategy during the optimization process. Finally, we apply this method to a scaled version (1:5) of the uCRM wing, achieving an agreement greater than 99% on the average MAC value of the first 5 modes.

HAL (Le Centre pour la Communication Scientifique Directe), Jun 15, 2017

Cette étude présente une approche originale pour l’approximation de systèmes multidisciplinaires ... more Cette étude présente une approche originale pour l’approximation de systèmes multidisciplinaires comportant des incertitudes de modèle, basée sur la construction de métamodèles par processus gaussiens de chaque solveur disciplinaire. La solution du système d’équations aléatoires est approchée par chaos polynomial creux permettant d’approximer les indices de sensibilité de Sobol et de déterminer ainsi quel métamodèle disciplinaire enrichir. Un exemple d’aéroélasticité statique illustre la démarche.

AIAA Aviation 2019 Forum, Jun 15, 2019

Cette these porte sur la resolution numerique des problemes de diffraction d'ondes electromag... more Cette these porte sur la resolution numerique des problemes de diffraction d'ondes electromagnetiques en regime harmonique. Nous developpons deux etudes qui peuvent servir d'outils de base dans les applications d'ingenierie. La premiere utilise l'approche par equations integrales basee sur une formulation efie (electric field integral equation) pour calculer la diffraction par des obstacles metalliques recouverts d'une couche mince dielectrique. Une resolution du probleme complet donne lieu a des instabilites numeriques des que l'epaisseur de la couche dielectrique est faible. Nous proposons de prendre en compte les effets de la couche mince a l'aide d'une condition d'impedance. Nous elaborons un procede de construction et d'analyse de ces conditions. L'utilisation de la condition d'impedance permet d'eliminer les instabilites numeriques. Des conditions d'impedance d'ordre suffisamment eleve sont necessaires pour prendre en compte les effets de courbure et pour ne pas perdre de precision a plus haute frequence. Nous developpons et etudions d'un point de vue theorique et numerique, des conditions aux limites d'ordre 3 pour des obstacles arbitraires. Dans une seconde partie, nous reprenons la formulation integrale introduite par b. Despres. Cette formulation repose sur la minimisation d'une fonctionnelle quadratique et le systeme matriciel a resoudre utilise des matrices reelles et symetriques. A la difference de la formulation efie, qui aboutit a une matrice symetrique complexe mais non hermitienne, la formulation qui est presentee ici favorise l'utilisation des methodes iteratives avec des theoremes de convergence. Nous decrivons deux derivations possibles pour l'ecriture du systeme relatif au probleme de point selle que l'on doit resoudre. De nombreuses applications numeriques valident cette approche.

HAL (Le Centre pour la Communication Scientifique Directe), Feb 25, 2020

One of the challenges in Multidisciplinary Design Analysis and Optimization is raised by the fide... more One of the challenges in Multidisciplinary Design Analysis and Optimization is raised by the fidelity of the methods and tools used within each single discipline. This element, introducing uncertainty in predictions, is neglected generally. In this work, a framework is built to analyse the impact that the fidelity of the aerodynamic model has on four key figures of merit, including Maximum Takeoff Weight, maximum Lift-to-drag ratio, fuel weight and takeoff length. It is found that the aircraft configuration with minimum fuel weight computed with two aerodynamic models differs significantly, with large variations in terms of predicted fuel weight and wing planform.

Springer optimization and its applications, 2020

Journal of Aircraft, Jul 1, 2020

Hybrid electric aircraft are a potential way to reduce the environmental footprint of aviation. R... more Hybrid electric aircraft are a potential way to reduce the environmental footprint of aviation. Research aimed at this subject has been pursued over the last decade; nevertheless, at this stage a full overall aircraft design procedure is still an open issue. This work proposes to enrich the procedure for the conceptual design of hybrid aircraft found in literature, through the definition of a multidisciplinary design optimization (MDO) framework, aimed at handling design problems for such kind of aircraft. The MDO technique has been chosen because the hybrid aircraft design problem shows more interaction between disciplines than a conventional configuration, and the classical approach based on multidisciplinary design analysis may neglect relevant features. The procedure has been tested on the case study of a single-aisle aircraft, featuring hybrid propulsion with distributed electric ducted fans. The analysis considers three configurations, with 16, 32

HAL (Le Centre pour la Communication Scientifique Directe), Oct 17, 2022

This report reflects a cooperation between the Inria Project Team Acumes and the Information Proc... more This report reflects a cooperation between the Inria Project Team Acumes and the Information Processing and Systems Department (DTIS) of Onera Toulouse. We present a case study of prioritized multi-objective optimization of the flight performance of an Airbus-A320-type aircraft in terms of takeoff fuel mass (f 1), operational empty weight (f 2) and ascent-to-cruisealtitude duration (f 3). The optimization is subject to functional constraints: interval bound on static margin and upper bound on wing span. Six design variables are considered defining geometry, mean aerodynamic chord and takeoff thrust potential. Designs are evaluated by means of the FAST-OAD open-source software developed by ONERA and ISAE-SUPAERO, and the prioritized optimization is conducted by the Inria Nash-MGDA software. From a point of Pareto-optimality of the sole pair (f 1 , f 2), considered preponderant, the climb duration f 3 is reduced drastically by a one-shot procedure establishing a continuum of Nash equilibria. The application of this procedure is far less computationally demanding than is the numerical establishment of the Pareto-front associated with the three cost functions together, while the Pareto-optimality of the preponderant cost functions (f 1 , f 2) is only marginally degraded.

AIAA AVIATION 2022 Forum, Jun 20, 2022

Recently, there has been a growing interest for mixed categorical meta-models based on Gaussian p... more Recently, there has been a growing interest for mixed categorical meta-models based on Gaussian process (GP) surrogates. In this setting, several existing approaches use different strategies. Among the recently developed methods, we could cite: GP models built using continuous relaxation of the variables, Gower distance based models or GP models derived from direct estimation of the correlation matrix. In this paper, we present a kernel-based approach that extends continuous Gaussian kernels to handle mixed-categorical variables. The proposed kernel leads to a GP surrogate that generalizes continuous relaxation and Gower distance based GP models. The good potential of the proposed framework is shown on analytical mixed-categorical variables test cases. On different settings, our proposed GP models is as accurate as the state-of-the-art GP models.

The paper deals with a new approach to compute the interactions between an antenna and a surround... more The paper deals with a new approach to compute the interactions between an antenna and a surrounding structure. This method is divided into 2 steps: (1) the computation of the incident field from the antenna; (2) the computation of the scattered field. In the first step, the antenna is described by its radiation pattern (from measurement, for example) and its location point. We use a new method to perform a far field to near field transformation in order to compute the incident field on each point of the structure. In the second step, the current on the structure is computed by solving the electric field integral equation (EFIE) using a multilevel fast multipole method (FMM).

Chinese Journal of Aeronautics, May 1, 2023

AIAA AVIATION 2022 Forum, Jun 20, 2022

HAL is a multidisciplinary open access archive for the deposit and dissemination of scientific re... more HAL is a multidisciplinary open access archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. L'archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d'enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

AIAA SCITECH 2022 Forum, Jan 3, 2022

Multidisciplinary design optimization methods aim at adapting numerical optimization techniques t... more Multidisciplinary design optimization methods aim at adapting numerical optimization techniques to the design of engineering systems involving multiple disciplines. In this context, a large number of mixed continuous, integer and categorical variables might arise during the optimization process and practical applications involve a large number of design variables. Recently, there has been a growing interest in mixed variables constrained Bayesian optimization but most existing approaches severely increase the number of the hyperparameters related to the surrogate model. In this paper, we address this issue by constructing surrogate models using less hyperparameters. The reduction process is based on the partial least squares method. An adaptive procedure for choosing the number of hyperparameters is proposed. The performance of the proposed approach is confirmed on analytical tests as well as two real applications related to aircraft design. A significant improvement is obtained compared to genetic algorithms.

Journal of Intelligent & Robotic Systems, 2010

Presented is a novel framework for performing flexible computational design studies at preliminar... more Presented is a novel framework for performing flexible computational design studies at preliminary design stage. It incorporates a workflow management device (WMD) and a number of advanced numerical treatments, including multi-objective optimization, sensitivity analysis and uncertainty management with emphasis on design robustness. The WMD enables the designer to build, understand, manipulate and share complex processes and studies. Results obtained after applying the WMD on various test cases, showed a significant reduction of the iterations required for the convergence of the computational system. The tests results also demonstrated the capabilities of the advanced treatments as follows:

Unmanned Systems

Vertical takeoff and landing (VTOL) vehicles are among the most versatile UAVs, appropriate for v... more Vertical takeoff and landing (VTOL) vehicles are among the most versatile UAVs, appropriate for various missions. Given that there are still open challenges regarding the VTOL design, this paper presents the full development and test cycle of a tail-sitter. IMAV 2022 competition rules are used to define the mission objective and constraints. A multidisciplinary design and optimization strategy is defined with the goal of maximizing competition score considering design, manufacturing, and competition constraints. While still being a work in progress, the resulting vehicle is designed to fly at 18 m/s, for around 14 minutes while carrying 200g of payload and weighting approximately 750g.

arXiv (Cornell University), May 23, 2023

The Surrogate Modeling Toolbox (SMT) is an open-source Python package that offers a collection of... more The Surrogate Modeling Toolbox (SMT) is an open-source Python package that offers a collection of surrogate modeling methods, sampling techniques, and a set of sample problems. This paper presents SMT 2.0, a major new release of SMT that introduces significant upgrades and new features to the toolbox. This release adds the capability to handle mixed-variable surrogate models and hierarchical variables. These types of variables are becoming increasingly important in several surrogate modeling applications. SMT 2.0 also improves SMT by extending sampling methods, adding new surrogate models, and computing variance and kernel derivatives for Kriging. This release also includes new functions to handle noisy and use multi-fidelity data. To the best of our knowledge, SMT 2.0 is the first open-source surrogate library to propose surrogate models for hierarchical and mixed inputs. This open-source software is distributed under the New BSD license 1 .

Mechanics & Industry, 2022

Aeroelastic scaling theory shows that the design problem of aeroelastically equivalent scaled air... more Aeroelastic scaling theory shows that the design problem of aeroelastically equivalent scaled aircraft can be treated as a structural-only design problem if the aerodynamic shape and airflow properties of the full scale aircraft are preserved. In that case, the theory shows that it is sufficient to match the scaled natural mode shapes, frequencies and mass of the reference aircraft. In this paper, we present a new method for the dynamic scaling of flexible structures where the objective function is based on the modal assurance criterion (MAC). This criterion is used for a mode tracking strategy during the optimization process. Finally, we apply this method to a scaled version (1:5) of the uCRM wing, achieving an agreement greater than 99% on the average MAC value of the first 5 modes.

HAL (Le Centre pour la Communication Scientifique Directe), Jun 15, 2017

Cette étude présente une approche originale pour l’approximation de systèmes multidisciplinaires ... more Cette étude présente une approche originale pour l’approximation de systèmes multidisciplinaires comportant des incertitudes de modèle, basée sur la construction de métamodèles par processus gaussiens de chaque solveur disciplinaire. La solution du système d’équations aléatoires est approchée par chaos polynomial creux permettant d’approximer les indices de sensibilité de Sobol et de déterminer ainsi quel métamodèle disciplinaire enrichir. Un exemple d’aéroélasticité statique illustre la démarche.

AIAA Aviation 2019 Forum, Jun 15, 2019

Cette these porte sur la resolution numerique des problemes de diffraction d'ondes electromag... more Cette these porte sur la resolution numerique des problemes de diffraction d'ondes electromagnetiques en regime harmonique. Nous developpons deux etudes qui peuvent servir d'outils de base dans les applications d'ingenierie. La premiere utilise l'approche par equations integrales basee sur une formulation efie (electric field integral equation) pour calculer la diffraction par des obstacles metalliques recouverts d'une couche mince dielectrique. Une resolution du probleme complet donne lieu a des instabilites numeriques des que l'epaisseur de la couche dielectrique est faible. Nous proposons de prendre en compte les effets de la couche mince a l'aide d'une condition d'impedance. Nous elaborons un procede de construction et d'analyse de ces conditions. L'utilisation de la condition d'impedance permet d'eliminer les instabilites numeriques. Des conditions d'impedance d'ordre suffisamment eleve sont necessaires pour prendre en compte les effets de courbure et pour ne pas perdre de precision a plus haute frequence. Nous developpons et etudions d'un point de vue theorique et numerique, des conditions aux limites d'ordre 3 pour des obstacles arbitraires. Dans une seconde partie, nous reprenons la formulation integrale introduite par b. Despres. Cette formulation repose sur la minimisation d'une fonctionnelle quadratique et le systeme matriciel a resoudre utilise des matrices reelles et symetriques. A la difference de la formulation efie, qui aboutit a une matrice symetrique complexe mais non hermitienne, la formulation qui est presentee ici favorise l'utilisation des methodes iteratives avec des theoremes de convergence. Nous decrivons deux derivations possibles pour l'ecriture du systeme relatif au probleme de point selle que l'on doit resoudre. De nombreuses applications numeriques valident cette approche.

HAL (Le Centre pour la Communication Scientifique Directe), Feb 25, 2020

One of the challenges in Multidisciplinary Design Analysis and Optimization is raised by the fide... more One of the challenges in Multidisciplinary Design Analysis and Optimization is raised by the fidelity of the methods and tools used within each single discipline. This element, introducing uncertainty in predictions, is neglected generally. In this work, a framework is built to analyse the impact that the fidelity of the aerodynamic model has on four key figures of merit, including Maximum Takeoff Weight, maximum Lift-to-drag ratio, fuel weight and takeoff length. It is found that the aircraft configuration with minimum fuel weight computed with two aerodynamic models differs significantly, with large variations in terms of predicted fuel weight and wing planform.

Springer optimization and its applications, 2020

Journal of Aircraft, Jul 1, 2020

Hybrid electric aircraft are a potential way to reduce the environmental footprint of aviation. R... more Hybrid electric aircraft are a potential way to reduce the environmental footprint of aviation. Research aimed at this subject has been pursued over the last decade; nevertheless, at this stage a full overall aircraft design procedure is still an open issue. This work proposes to enrich the procedure for the conceptual design of hybrid aircraft found in literature, through the definition of a multidisciplinary design optimization (MDO) framework, aimed at handling design problems for such kind of aircraft. The MDO technique has been chosen because the hybrid aircraft design problem shows more interaction between disciplines than a conventional configuration, and the classical approach based on multidisciplinary design analysis may neglect relevant features. The procedure has been tested on the case study of a single-aisle aircraft, featuring hybrid propulsion with distributed electric ducted fans. The analysis considers three configurations, with 16, 32

HAL (Le Centre pour la Communication Scientifique Directe), Oct 17, 2022

This report reflects a cooperation between the Inria Project Team Acumes and the Information Proc... more This report reflects a cooperation between the Inria Project Team Acumes and the Information Processing and Systems Department (DTIS) of Onera Toulouse. We present a case study of prioritized multi-objective optimization of the flight performance of an Airbus-A320-type aircraft in terms of takeoff fuel mass (f 1), operational empty weight (f 2) and ascent-to-cruisealtitude duration (f 3). The optimization is subject to functional constraints: interval bound on static margin and upper bound on wing span. Six design variables are considered defining geometry, mean aerodynamic chord and takeoff thrust potential. Designs are evaluated by means of the FAST-OAD open-source software developed by ONERA and ISAE-SUPAERO, and the prioritized optimization is conducted by the Inria Nash-MGDA software. From a point of Pareto-optimality of the sole pair (f 1 , f 2), considered preponderant, the climb duration f 3 is reduced drastically by a one-shot procedure establishing a continuum of Nash equilibria. The application of this procedure is far less computationally demanding than is the numerical establishment of the Pareto-front associated with the three cost functions together, while the Pareto-optimality of the preponderant cost functions (f 1 , f 2) is only marginally degraded.

AIAA AVIATION 2022 Forum, Jun 20, 2022

Recently, there has been a growing interest for mixed categorical meta-models based on Gaussian p... more Recently, there has been a growing interest for mixed categorical meta-models based on Gaussian process (GP) surrogates. In this setting, several existing approaches use different strategies. Among the recently developed methods, we could cite: GP models built using continuous relaxation of the variables, Gower distance based models or GP models derived from direct estimation of the correlation matrix. In this paper, we present a kernel-based approach that extends continuous Gaussian kernels to handle mixed-categorical variables. The proposed kernel leads to a GP surrogate that generalizes continuous relaxation and Gower distance based GP models. The good potential of the proposed framework is shown on analytical mixed-categorical variables test cases. On different settings, our proposed GP models is as accurate as the state-of-the-art GP models.

The paper deals with a new approach to compute the interactions between an antenna and a surround... more The paper deals with a new approach to compute the interactions between an antenna and a surrounding structure. This method is divided into 2 steps: (1) the computation of the incident field from the antenna; (2) the computation of the scattered field. In the first step, the antenna is described by its radiation pattern (from measurement, for example) and its location point. We use a new method to perform a far field to near field transformation in order to compute the incident field on each point of the structure. In the second step, the current on the structure is computed by solving the electric field integral equation (EFIE) using a multilevel fast multipole method (FMM).

Chinese Journal of Aeronautics, May 1, 2023

AIAA AVIATION 2022 Forum, Jun 20, 2022

HAL is a multidisciplinary open access archive for the deposit and dissemination of scientific re... more HAL is a multidisciplinary open access archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. L'archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d'enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

AIAA SCITECH 2022 Forum, Jan 3, 2022

Multidisciplinary design optimization methods aim at adapting numerical optimization techniques t... more Multidisciplinary design optimization methods aim at adapting numerical optimization techniques to the design of engineering systems involving multiple disciplines. In this context, a large number of mixed continuous, integer and categorical variables might arise during the optimization process and practical applications involve a large number of design variables. Recently, there has been a growing interest in mixed variables constrained Bayesian optimization but most existing approaches severely increase the number of the hyperparameters related to the surrogate model. In this paper, we address this issue by constructing surrogate models using less hyperparameters. The reduction process is based on the partial least squares method. An adaptive procedure for choosing the number of hyperparameters is proposed. The performance of the proposed approach is confirmed on analytical tests as well as two real applications related to aircraft design. A significant improvement is obtained compared to genetic algorithms.