Bartoli, Nathalie
Overview
Works:  15 works in 18 publications in 2 languages and 132 library holdings 

Roles:  Author, Opponent, Thesis advisor, Other, Contributor 
Publication Timeline
.
Most widely held works by
Nathalie Bartoli
Simulation et algorithmes stochastiques : une introduction avec applications by
Nathalie Bartoli(
Book
)
3 editions published in 2001 in French and held by 107 WorldCat member libraries worldwide
3 editions published in 2001 in French and held by 107 WorldCat member libraries worldwide
Simulation and Algorithmes Stochastiques by
Nathalie Bartoli(
)
1 edition published in 2001 in French and held by 9 WorldCat member libraries worldwide
1 edition published in 2001 in French and held by 9 WorldCat member libraries worldwide
Improving kriging surrogates of highdimensional design models by Partial Least Squares dimension reduction by Mohamed Amine Bouhlel(
)
1 edition published in 2015 in English and held by 2 WorldCat member libraries worldwide
1 edition published in 2015 in English and held by 2 WorldCat member libraries worldwide
Linear regressionbased multifidelity surrogate for disturbance amplification in multiphase explosion by M. Giselle FernándezGodino(
)
1 edition published in 2019 in English and held by 2 WorldCat member libraries worldwide
1 edition published in 2019 in English and held by 2 WorldCat member libraries worldwide
MODELES POUR LA DIFFRACTION D'ONDES PAR DES OBSTACLES REVETUS DE COUCHES MINCES. RESOLUTION DE PROBLEMES DE DIFFRACTION D'ONDES
VIA UNE FORMULATION INTEGRALE DE TYPE POINT SELLE by
Nathalie Bartoli(
Book
)
2 editions published in 2000 in French and held by 2 WorldCat member libraries worldwide
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
2 editions published in 2000 in French and held by 2 WorldCat member libraries worldwide
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
Optimisation bayésienne sous contraintes et en grande dimension appliquée à la conception avion avant projet by
Rémy Priem(
)
1 edition published in 2020 in French and held by 1 WorldCat member library worldwide
Nowadays, the preliminary design in aeronautics is based mainly on numericalmodels bringingtogether many disciplines aimed at evaluating the performance of the aircraft. These disciplines,such as aerodynamics, structure and propulsion, are interconnected in order to take into accounttheir interactions. This produces a computationally expensive aircraft performance evaluationprocess. Indeed, an evaluation can take from thirty seconds for low fidelity models to severalweeks for higher fidelity models. In addition, because of the multidisciplinarity of the processand the diversity of the calculation tools, we do not always have access to the properties or thegradient of this performance function. In addition, each discipline uses its own design variablesand must respect equality or inequality constraints which are often numerous and multimodal.We ultimately seek to find the best possible configuration in a given design space.This research can be mathematically translated to a blackbox optimization problem under inequalityand equality constraints, also known as mixted constraints, depending on a large numberof design variables. Moreover, the constraints and the objective function are expensive to evaluateand their regularity is not known. This is why we are interested in derivativefree optimizationmethods and more specifically the ones based on surrogatemodels. Bayesian optimization methods,using Gaussian processes, are more particularly studied because they have shown rapid convergenceon multimodal problems. Indeed, the use of evolutionary optimization algorithms orother gradientbased methods is not possible because of the computational cost that this implies:too many calls to generate populations of points, or to approach the gradient by finite difference.However, the Bayesian optimization method is conventionally used for optimization problemswithout constraints and of small dimension. Extensions have been proposed to partially take thislock into account. On the one hand, optimization methods have been introduced to solve optimizationproblems with mixed constraints. However, none of them is adaptable to the largedimension, to the multimodal problems and to mixed constraints. On the other hand, nonlinearoptimization methods have been developed for the large dimension up to a million design variables.In the same way, these methods extend only with difficulty to the constrained problemsbecause of the computing time which they require or their random character.A first part of this work is based on the development of a Bayesian optimization algorithmsolvingunconstrained optimization problems in large dimensions. It is based on an adaptive learningstrategy of a linear subspace carried out in conjunction with the optimization. This linear subspaceis then used to perform the optimization. This method has been tested on academic testcases.A second part of this work deals with the development of a Bayesian optimization algorithm tosolve multimodal optimization problems under mixed constraints. It has been extensively comparedto algorithms in the literature on a large battery of academic tests.Finally, the second algorithm was compared with two aeronautical test cases. The first testcase is a classic medium range aircraft configuration with hybrid electric propulsion developedby ONERA and ISAESupaero. The second test case is a classic business aircraft configuration developedat Bombardier Aviation. This test case is based on an optimization at two levels of fidelity.A conceptual fidelity level and a preliminary fidelity level for which the problem is evaluated inthirty seconds and 25 minutes, respectively. This last study was carried out during an internationalmobility at Bombardier Aviation in Montreal (CA). The results showed the interest of theimplemented method
1 edition published in 2020 in French and held by 1 WorldCat member library worldwide
Nowadays, the preliminary design in aeronautics is based mainly on numericalmodels bringingtogether many disciplines aimed at evaluating the performance of the aircraft. These disciplines,such as aerodynamics, structure and propulsion, are interconnected in order to take into accounttheir interactions. This produces a computationally expensive aircraft performance evaluationprocess. Indeed, an evaluation can take from thirty seconds for low fidelity models to severalweeks for higher fidelity models. In addition, because of the multidisciplinarity of the processand the diversity of the calculation tools, we do not always have access to the properties or thegradient of this performance function. In addition, each discipline uses its own design variablesand must respect equality or inequality constraints which are often numerous and multimodal.We ultimately seek to find the best possible configuration in a given design space.This research can be mathematically translated to a blackbox optimization problem under inequalityand equality constraints, also known as mixted constraints, depending on a large numberof design variables. Moreover, the constraints and the objective function are expensive to evaluateand their regularity is not known. This is why we are interested in derivativefree optimizationmethods and more specifically the ones based on surrogatemodels. Bayesian optimization methods,using Gaussian processes, are more particularly studied because they have shown rapid convergenceon multimodal problems. Indeed, the use of evolutionary optimization algorithms orother gradientbased methods is not possible because of the computational cost that this implies:too many calls to generate populations of points, or to approach the gradient by finite difference.However, the Bayesian optimization method is conventionally used for optimization problemswithout constraints and of small dimension. Extensions have been proposed to partially take thislock into account. On the one hand, optimization methods have been introduced to solve optimizationproblems with mixed constraints. However, none of them is adaptable to the largedimension, to the multimodal problems and to mixed constraints. On the other hand, nonlinearoptimization methods have been developed for the large dimension up to a million design variables.In the same way, these methods extend only with difficulty to the constrained problemsbecause of the computing time which they require or their random character.A first part of this work is based on the development of a Bayesian optimization algorithmsolvingunconstrained optimization problems in large dimensions. It is based on an adaptive learningstrategy of a linear subspace carried out in conjunction with the optimization. This linear subspaceis then used to perform the optimization. This method has been tested on academic testcases.A second part of this work deals with the development of a Bayesian optimization algorithm tosolve multimodal optimization problems under mixed constraints. It has been extensively comparedto algorithms in the literature on a large battery of academic tests.Finally, the second algorithm was compared with two aeronautical test cases. The first testcase is a classic medium range aircraft configuration with hybrid electric propulsion developedby ONERA and ISAESupaero. The second test case is a classic business aircraft configuration developedat Bombardier Aviation. This test case is based on an optimization at two levels of fidelity.A conceptual fidelity level and a preliminary fidelity level for which the problem is evaluated inthirty seconds and 25 minutes, respectively. This last study was carried out during an internationalmobility at Bombardier Aviation in Montreal (CA). The results showed the interest of theimplemented method
Optimisation autoadaptative en environnement d'analyse multidisciplinaire via les modèles de krigeage combinés à la méthode
PLS by
Mohamed Amine Bouhlel(
)
1 edition published in 2016 in French and held by 1 WorldCat member library worldwide
Aerospace turbomachinery consists of a plurality of blades. Their main function is to transfer energybetween the air and the rotor. The bladed disks of the compressor are particularly important becausethey must satisfy both the requirements of aerodynamic performance and mechanical resistance.Mechanical and aerodynamic optimization of blades consists in searching for a set of parameterizedaerodynamic shape that ensures the best compromise solution between a set of constraints.This PhD introduces a surrogatebased optimization method well adapted to highdimensionalproblems. This kind of highdimensional problem is very similar to the Snecma's problems. Ourmain contributions can be divided into two parts : Kriging models development and enhancementof an existing optimization method to handle highdimensional problems under a large number ofconstraints. Concerning Kriging models, we propose a new formulation of covariance kernel which is able toreduce the number of hyperparameters in order to accelerate the construction of the metamodel.One of the known limitations of Kriging models is about the estimation of its hyperparameters.This estimation becomes more and more difficult when the number of dimension increases. Inparticular, the initial design of experiments (for surrogate modelling construction) requires animportant number of points and therefore the inversion of the covariance matrix becomes timeconsuming. Our approach consists in reducing the number of parameters to estimate using the Partial LeastSquares regression method (PLS). This method provides information about the linear relationshipbetween input and output variables. This information is integrated into the Kriging model kernelwhile maintaining the symmetry and the positivity properties of the kernels. Thanks to this approach,the construction of these new models called KPLS is very fast because of the low number of newparameters to estimate. When the covariance kernel used is of an exponential type, the KPLS methodcan be used to initialize parameters of classical Kriging models, to accelerate the convergence of theestimation of parameters. The final method, called KPLS+K, allows to improve the accuracy of themodel for multimodal functions. The second main contribution of this PhD is to develop a global optimization method to tacklehighdimensional problems under a large number of constraint functions thanks to KPLS or KPLS+Kmethod. Indeed, we extended the self adaptive optimization method called "Efficient Global Optimization,EGO" for highdimensional problems under constraints. Several enriching criteria have been tested. This method allows to estimate known global optima on academic problems up to 50 inputvariables. The proposed method is tested on two industrial cases, the first one, "MOPTA", from the automotiveindustry (with 124 input variables and 68 constraint functions) and the second one is a turbineblade from Snecma company (with 50 input variables and 31 constraint functions). The results showthe effectiveness of the method to handle industrial problems.We also highlight some importantlimitations
1 edition published in 2016 in French and held by 1 WorldCat member library worldwide
Aerospace turbomachinery consists of a plurality of blades. Their main function is to transfer energybetween the air and the rotor. The bladed disks of the compressor are particularly important becausethey must satisfy both the requirements of aerodynamic performance and mechanical resistance.Mechanical and aerodynamic optimization of blades consists in searching for a set of parameterizedaerodynamic shape that ensures the best compromise solution between a set of constraints.This PhD introduces a surrogatebased optimization method well adapted to highdimensionalproblems. This kind of highdimensional problem is very similar to the Snecma's problems. Ourmain contributions can be divided into two parts : Kriging models development and enhancementof an existing optimization method to handle highdimensional problems under a large number ofconstraints. Concerning Kriging models, we propose a new formulation of covariance kernel which is able toreduce the number of hyperparameters in order to accelerate the construction of the metamodel.One of the known limitations of Kriging models is about the estimation of its hyperparameters.This estimation becomes more and more difficult when the number of dimension increases. Inparticular, the initial design of experiments (for surrogate modelling construction) requires animportant number of points and therefore the inversion of the covariance matrix becomes timeconsuming. Our approach consists in reducing the number of parameters to estimate using the Partial LeastSquares regression method (PLS). This method provides information about the linear relationshipbetween input and output variables. This information is integrated into the Kriging model kernelwhile maintaining the symmetry and the positivity properties of the kernels. Thanks to this approach,the construction of these new models called KPLS is very fast because of the low number of newparameters to estimate. When the covariance kernel used is of an exponential type, the KPLS methodcan be used to initialize parameters of classical Kriging models, to accelerate the convergence of theestimation of parameters. The final method, called KPLS+K, allows to improve the accuracy of themodel for multimodal functions. The second main contribution of this PhD is to develop a global optimization method to tacklehighdimensional problems under a large number of constraint functions thanks to KPLS or KPLS+Kmethod. Indeed, we extended the self adaptive optimization method called "Efficient Global Optimization,EGO" for highdimensional problems under constraints. Several enriching criteria have been tested. This method allows to estimate known global optima on academic problems up to 50 inputvariables. The proposed method is tested on two industrial cases, the first one, "MOPTA", from the automotiveindustry (with 124 input variables and 68 constraint functions) and the second one is a turbineblade from Snecma company (with 50 input variables and 31 constraint functions). The results showthe effectiveness of the method to handle industrial problems.We also highlight some importantlimitations
Aeroelastic similarity of a flight demonstrator via multidisciplinary optimization by
Joan Mas Colomer(
)
1 edition published in 2018 in English and held by 1 WorldCat member library worldwide
The search for more efficient aircraft configurations leads designers to explore new concepts such as the blended wing body, the strutbraced wing, or the box wing. Unlike the classical wingfuselage configuration, which is well known and understood, few is known about the inflight behavior of these new aircraft concepts. In that context, the design, construction, and testing of unmanned aeroelastically scaled models presents itself as a lowrisk means of acquiring experimental knowledge on these new concepts. An aeroelastically scaled model exhibits the same scaled aeroelastic behavior as the fullscale reference aircraft. Typically, the same aeroelastic behavior implies matching the displacements for some given scaled airflow conditions, as well as the scaled flutter or static divergence speeds. To address the similarity problem, we divide the approach in three parts. In the first one we deal with the aeroelastic similarity problem when the aerodynamic flow scaling conditions can be completely preserved. In that situation, the problem is reduced to simply matching the scaled modal dynamic response of the wing through optimization of the structure and mass properties. In the second part, we focus on the wing planform design optimization to match the flutter response when the airflow scaling parameters cannot be achieved
1 edition published in 2018 in English and held by 1 WorldCat member library worldwide
The search for more efficient aircraft configurations leads designers to explore new concepts such as the blended wing body, the strutbraced wing, or the box wing. Unlike the classical wingfuselage configuration, which is well known and understood, few is known about the inflight behavior of these new aircraft concepts. In that context, the design, construction, and testing of unmanned aeroelastically scaled models presents itself as a lowrisk means of acquiring experimental knowledge on these new concepts. An aeroelastically scaled model exhibits the same scaled aeroelastic behavior as the fullscale reference aircraft. Typically, the same aeroelastic behavior implies matching the displacements for some given scaled airflow conditions, as well as the scaled flutter or static divergence speeds. To address the similarity problem, we divide the approach in three parts. In the first one we deal with the aeroelastic similarity problem when the aerodynamic flow scaling conditions can be completely preserved. In that situation, the problem is reduced to simply matching the scaled modal dynamic response of the wing through optimization of the structure and mass properties. In the second part, we focus on the wing planform design optimization to match the flutter response when the airflow scaling parameters cannot be achieved
Méthode de Galerkin discontinue isogéométrique avec domaines dépendants du temps by
Stefano Pezzano(
)
1 edition published in 2021 in English and held by 1 WorldCat member library worldwide
Timedependent domains are encountered in a vast category of fluid mechanics applications. Such problems are often characterised by complex geometries and physical phenomena. The aim of the present dissertation is the development of an accurate and reliable methodology to investigate compressible flows with timedependent geometries. To this end, we combine ideas from Isogeometric Analysis (IGA) and highorder numerical schemes for fluid dynamics, specifically Discontinuous Galerkin (DG) methods. We start by discussing the implementation details of the Isogeometric DG method. To this end, we introduce the DG scheme for conservation laws and the fundamentals of IGA. The mathematical representation of Computer Aided Design (CAD) is explained and some important geometry manipulation algorithms are presented. We then show how a DGcompatible description can be extracted starting from a CAD object, allowing us to employ a geometrically exact DG discretisation to solve the equations of fluid mechanics. Finally, we conduct two simple numerical experiments to illustrate the impact of the boundary representation on flow simulations.We then proceed to extend the Isogeometric DG methodology to deformable domains using the Arbitrary LagrangianEulerian (ALE) formalism. After an analysis of the existing ALEDG schemes, we propose an ALE formulation using the Isogeometric framework. The CAD basis functions are also adopted to deform the mesh, leading to a fully unified description of the simulation variables. Besides, we show how the proposed grid velocity algorithm can be seamlessly coupled with Adaptive Mesh Refinement (AMR). The presented approach is firstly validated with two analytical problems, showing optimal convergence rates. Then, we simulate the flows past an oscillating cylinder and a pitching airfoil to prove the robustness of the methodology. We conclude by demonstrating the capabilities of the developed ALEAMR coupling algorithm using the pitching airfoil benchmark. Secondly, we propose to employ sliding grids for flows with rotating components, as an alternative to the deformationbased ALE technique. Using the CAD basis functions, circular interfaces can be exactly represented. Therefore, it is possible to develop a highorder and fully conservative sliding mesh formulation. We begin by detailing the treatment of the sliding interface geometry and the computation of the numerical fluxes. Then, we show the optimal convergence of the implemented approach using an analytical problem. We further characterise the behaviour of the sliding interface considering the flow past a pitching ellipse. We show that the sliding mesh technique and the deformationbased ALE methodology are equivalent in terms of accuracy. Finally, a vertical axis wind turbine is simulated to present a more topologically complex test case. Lastly, we apply the proposed ALE formulation to a flow control problem. In particular, we consider a dynamic trailing edge morphing actuation to control the separation of the laminar boundary layer around a NACA 6series airfoil. Using the proposed mesh deformation algorithm, we develop a highorder accurate trailing edge morphing model. In order to find the optimal set of control parameters, the flow solver is coupled with a Bayesian optimisation algorithm. Thanks to the strong coupling between the geometry description and the numerical schemes, the resulting design chain is fully automated. The capabilities of the proposed framework are explored considering both singleobjective and multiobjective optimisation problems
1 edition published in 2021 in English and held by 1 WorldCat member library worldwide
Timedependent domains are encountered in a vast category of fluid mechanics applications. Such problems are often characterised by complex geometries and physical phenomena. The aim of the present dissertation is the development of an accurate and reliable methodology to investigate compressible flows with timedependent geometries. To this end, we combine ideas from Isogeometric Analysis (IGA) and highorder numerical schemes for fluid dynamics, specifically Discontinuous Galerkin (DG) methods. We start by discussing the implementation details of the Isogeometric DG method. To this end, we introduce the DG scheme for conservation laws and the fundamentals of IGA. The mathematical representation of Computer Aided Design (CAD) is explained and some important geometry manipulation algorithms are presented. We then show how a DGcompatible description can be extracted starting from a CAD object, allowing us to employ a geometrically exact DG discretisation to solve the equations of fluid mechanics. Finally, we conduct two simple numerical experiments to illustrate the impact of the boundary representation on flow simulations.We then proceed to extend the Isogeometric DG methodology to deformable domains using the Arbitrary LagrangianEulerian (ALE) formalism. After an analysis of the existing ALEDG schemes, we propose an ALE formulation using the Isogeometric framework. The CAD basis functions are also adopted to deform the mesh, leading to a fully unified description of the simulation variables. Besides, we show how the proposed grid velocity algorithm can be seamlessly coupled with Adaptive Mesh Refinement (AMR). The presented approach is firstly validated with two analytical problems, showing optimal convergence rates. Then, we simulate the flows past an oscillating cylinder and a pitching airfoil to prove the robustness of the methodology. We conclude by demonstrating the capabilities of the developed ALEAMR coupling algorithm using the pitching airfoil benchmark. Secondly, we propose to employ sliding grids for flows with rotating components, as an alternative to the deformationbased ALE technique. Using the CAD basis functions, circular interfaces can be exactly represented. Therefore, it is possible to develop a highorder and fully conservative sliding mesh formulation. We begin by detailing the treatment of the sliding interface geometry and the computation of the numerical fluxes. Then, we show the optimal convergence of the implemented approach using an analytical problem. We further characterise the behaviour of the sliding interface considering the flow past a pitching ellipse. We show that the sliding mesh technique and the deformationbased ALE methodology are equivalent in terms of accuracy. Finally, a vertical axis wind turbine is simulated to present a more topologically complex test case. Lastly, we apply the proposed ALE formulation to a flow control problem. In particular, we consider a dynamic trailing edge morphing actuation to control the separation of the laminar boundary layer around a NACA 6series airfoil. Using the proposed mesh deformation algorithm, we develop a highorder accurate trailing edge morphing model. In order to find the optimal set of control parameters, the flow solver is coupled with a Bayesian optimisation algorithm. Thanks to the strong coupling between the geometry description and the numerical schemes, the resulting design chain is fully automated. The capabilities of the proposed framework are explored considering both singleobjective and multiobjective optimisation problems
Enhancement of the conceptual aircraft design process through certification constraints management and full mission simulations by
Peter Schmollgruber(
)
1 edition published in 2018 in English and held by 1 WorldCat member library worldwide
The design of a new aircraft is initiated at the conceptual design phase. In an initial step, aircraftdesigners, disciplinary and subsystems experts identify a set of potential concepts that could fulfill thecustomer requirements. To select the most promising candidates, aircraft designers carry out the sizingprocess through a Multidisciplinary Design Analysis. Nowadays, in the field of civil transport aircraft,environmental constraints set challenging goals in terms of fuel consumption for the next generationsof airplanes. With the “tube and wing” configuration offering low expectations on furtherimprovements, disruptive vehicle concepts including new technologies are investigated. However,little information on such architectures is available in the early phases of the design process. Thus, inorder to avoid mistakenly selecting or eliminating a wrong concept, a key objective in Aircraft Designresearch is to add knowledge in the Multidisciplinary Design Analysis.Nowadays, this objective is achieved with different approaches: implementation of MultidisciplinaryDesign Optimization, addition of accuracy through high fidelity analyses, introduction of newdisciplines or systems and uncertainty management. The role of the aircraft designer is then tocombine these options in a multidisciplinary design process to converge to the most promising conceptmeeting certification constraints. To illustrate this process, the optimization of a transport aircraftfeaturing ground based assistance has been performed. Using monolithic optimization architecture andadvanced structural models for the wing and fuselage, this study emphasized the impact ofcertification constraints on final results. Further review of the regulatory texts concluded that aircraftsimulation capabilities are needed to assess some requirements. The same need has been identified inthe field of Air Traffic Management that provides constraints for aircraft operations. This researchproposes then to add knowledge through an expansion of the Multidisciplinary Design Analysis andOptimization with a new Certification Constraint Module and full simulation capabilities.Following the development of the Certification Constraint Module (CCM), its capabilities have beenused to perform four optimization problems associated to a conventional civil transport aircraft basedon the ONERA / ISAESUPAERO sizing tool called FAST. Facilitated by the Graphical UserInterface of the CCM, the setup time of these optimizations has been reduced and the results clearlyconfirmed the necessity to consider certification constraints very early in the design process in order toselect the most promising concepts.To achieve full simulation capabilities, the multidisciplinary analysis within FAST had to beenhanced. First, the aerodynamics analysis tool has been modified so that necessary coefficients for a6 DegreesofFreedom model could be generated. Second, a new module computing inertia propertieshas been added. Last, the open source simulator JSBSim has been used including different controllaws for stability augmentation and automated navigation. The comparison between flight trajectoriesobtained with FAST and real aircraft data recorded with ADSB antenna confirmed the validity of theapproach
1 edition published in 2018 in English and held by 1 WorldCat member library worldwide
The design of a new aircraft is initiated at the conceptual design phase. In an initial step, aircraftdesigners, disciplinary and subsystems experts identify a set of potential concepts that could fulfill thecustomer requirements. To select the most promising candidates, aircraft designers carry out the sizingprocess through a Multidisciplinary Design Analysis. Nowadays, in the field of civil transport aircraft,environmental constraints set challenging goals in terms of fuel consumption for the next generationsof airplanes. With the “tube and wing” configuration offering low expectations on furtherimprovements, disruptive vehicle concepts including new technologies are investigated. However,little information on such architectures is available in the early phases of the design process. Thus, inorder to avoid mistakenly selecting or eliminating a wrong concept, a key objective in Aircraft Designresearch is to add knowledge in the Multidisciplinary Design Analysis.Nowadays, this objective is achieved with different approaches: implementation of MultidisciplinaryDesign Optimization, addition of accuracy through high fidelity analyses, introduction of newdisciplines or systems and uncertainty management. The role of the aircraft designer is then tocombine these options in a multidisciplinary design process to converge to the most promising conceptmeeting certification constraints. To illustrate this process, the optimization of a transport aircraftfeaturing ground based assistance has been performed. Using monolithic optimization architecture andadvanced structural models for the wing and fuselage, this study emphasized the impact ofcertification constraints on final results. Further review of the regulatory texts concluded that aircraftsimulation capabilities are needed to assess some requirements. The same need has been identified inthe field of Air Traffic Management that provides constraints for aircraft operations. This researchproposes then to add knowledge through an expansion of the Multidisciplinary Design Analysis andOptimization with a new Certification Constraint Module and full simulation capabilities.Following the development of the Certification Constraint Module (CCM), its capabilities have beenused to perform four optimization problems associated to a conventional civil transport aircraft basedon the ONERA / ISAESUPAERO sizing tool called FAST. Facilitated by the Graphical UserInterface of the CCM, the setup time of these optimizations has been reduced and the results clearlyconfirmed the necessity to consider certification constraints very early in the design process in order toselect the most promising concepts.To achieve full simulation capabilities, the multidisciplinary analysis within FAST had to beenhanced. First, the aerodynamics analysis tool has been modified so that necessary coefficients for a6 DegreesofFreedom model could be generated. Second, a new module computing inertia propertieshas been added. Last, the open source simulator JSBSim has been used including different controllaws for stability augmentation and automated navigation. The comparison between flight trajectoriesobtained with FAST and real aircraft data recorded with ADSB antenna confirmed the validity of theapproach
Fiabilité des conteneurs de stockage de déchets radioactifs de haute activité by Augustin Persoons(
)
1 edition published in 2020 in French and held by 1 WorldCat member library worldwide
Le projet Cigéo vise à répondre à la problématique de la gestion à long terme des déchets radioactifs de haute activité et de moyenne activité à vie longue. Le concept de ce projet (le stockage géologique profond) repose sur le confinement passif de la radioactivité par une couche argileuse ayant des propriétés adaptées. Les infrastructures de stockage sont soumises à un processus de vieillissement sur plusieurs siècles. Ces échelles de temps induisent des incertitudes sur leur évolution et leur durée de vie. Afin de mener à bien ce projet, il est nécessaire de faire la preuve de sa fiabilité, et donc de prendre en compte ces incertitudes dans l'évaluation de la durée de vie des structures. Les travaux présentés proposent une première application des méthodes de fiabilité pour répondre à ces problématiques. Ils concernent plus particulièrement les infrastructures de stockage des déchets de haute activité, et visent à estimer l'évolution de la probabilité de défaillance du conteneur de stockage ainsi que sa sensibilité aux paramètres d'entrée. Pour ce faire l'approche fiabiliste est déployée dans son ensemble comprenant : le développement d'un modèle mécanique du système, la modélisation des incertitudes en entrée, le choix d'un critère de défaillance, l'implémentation et l'exploitation d'une méthode de fiabilité, et une étude de fiabilité
1 edition published in 2020 in French and held by 1 WorldCat member library worldwide
Le projet Cigéo vise à répondre à la problématique de la gestion à long terme des déchets radioactifs de haute activité et de moyenne activité à vie longue. Le concept de ce projet (le stockage géologique profond) repose sur le confinement passif de la radioactivité par une couche argileuse ayant des propriétés adaptées. Les infrastructures de stockage sont soumises à un processus de vieillissement sur plusieurs siècles. Ces échelles de temps induisent des incertitudes sur leur évolution et leur durée de vie. Afin de mener à bien ce projet, il est nécessaire de faire la preuve de sa fiabilité, et donc de prendre en compte ces incertitudes dans l'évaluation de la durée de vie des structures. Les travaux présentés proposent une première application des méthodes de fiabilité pour répondre à ces problématiques. Ils concernent plus particulièrement les infrastructures de stockage des déchets de haute activité, et visent à estimer l'évolution de la probabilité de défaillance du conteneur de stockage ainsi que sa sensibilité aux paramètres d'entrée. Pour ce faire l'approche fiabiliste est déployée dans son ensemble comprenant : le développement d'un modèle mécanique du système, la modélisation des incertitudes en entrée, le choix d'un critère de défaillance, l'implémentation et l'exploitation d'une méthode de fiabilité, et une étude de fiabilité
Methodology for sizing and optimising a Blended WingBody with distributed electric ducted fans by
Alessandro Sgueglia(
)
1 edition published in 2019 in English and held by 1 WorldCat member library worldwide
The increase of air traffic in the last decades and its projections pose akey challenge towards the carbon neutral growth objective. To cope with this societal goal,there is a need for disruptive air transport aircraft concepts featuring new technologies withlow environmental impact. Such future air vehicle relies on the various interactions betweensystems, disciplines and components. This Ph.D. research thus focuses on the developmentof a methodology dedicated to the exploration and performance evaluation of unconventionalconfigurations using innovative propulsion concepts. The use case to be considered is the optimisationat conceptual level of a Blended WingBody with distributed electric propulsion, apromising concept which combines high aerodynamic performances and benefits from electricpropulsion.The optimisation process based on FAST, the ISAESUPAERO / ONERA aircraft sizingtool, has been implemented within OpenMDAO, the NASA opensource multidisciplinaryanalysis and optimisation framework. With the idea of a progressive enhancement of themultidisciplinary design analysis and a better capture of the different effects, the two pioneeringelements have been studied separately. First, the classical process has been revisedto take into account the new hybrid powerplant. Second, a methodology has been revisedto consider a radically new airframe design. Last, a design process featuring both innovativeaspects has been developed to investigate a Blended Wing Body concept with distributedelectric propulsion.Concerning the design process, results show that the use of gradients in the optimisationprocedure speeds up the process against a gradientfree method up to 70%. This is an importantgain in time that facilitates designer's tasks. For the disruptive concept performances,results have been compared to the ones obtained for a conventional A320 type aircraft basedon the same top level requirements and technological horizon. Overall, the hybrid electricpropulsion concept is interesting as it allows zero emissions for Landing/TakeOff operations,improving the environmental footprint of the aircraft: fuel can be saved for missions below acertain range. This limitation is associated to the presence of batteries: indeed they introduceindeed a relevant penalty in weight that cannot be countered by benefits of electrification forlonger range. Additional simulations indicate that a Blended WingBody concept based on aturboelectric only architecture is constantly performing better than the baseline within thelimits of the assumptions
1 edition published in 2019 in English and held by 1 WorldCat member library worldwide
The increase of air traffic in the last decades and its projections pose akey challenge towards the carbon neutral growth objective. To cope with this societal goal,there is a need for disruptive air transport aircraft concepts featuring new technologies withlow environmental impact. Such future air vehicle relies on the various interactions betweensystems, disciplines and components. This Ph.D. research thus focuses on the developmentof a methodology dedicated to the exploration and performance evaluation of unconventionalconfigurations using innovative propulsion concepts. The use case to be considered is the optimisationat conceptual level of a Blended WingBody with distributed electric propulsion, apromising concept which combines high aerodynamic performances and benefits from electricpropulsion.The optimisation process based on FAST, the ISAESUPAERO / ONERA aircraft sizingtool, has been implemented within OpenMDAO, the NASA opensource multidisciplinaryanalysis and optimisation framework. With the idea of a progressive enhancement of themultidisciplinary design analysis and a better capture of the different effects, the two pioneeringelements have been studied separately. First, the classical process has been revisedto take into account the new hybrid powerplant. Second, a methodology has been revisedto consider a radically new airframe design. Last, a design process featuring both innovativeaspects has been developed to investigate a Blended Wing Body concept with distributedelectric propulsion.Concerning the design process, results show that the use of gradients in the optimisationprocedure speeds up the process against a gradientfree method up to 70%. This is an importantgain in time that facilitates designer's tasks. For the disruptive concept performances,results have been compared to the ones obtained for a conventional A320 type aircraft basedon the same top level requirements and technological horizon. Overall, the hybrid electricpropulsion concept is interesting as it allows zero emissions for Landing/TakeOff operations,improving the environmental footprint of the aircraft: fuel can be saved for missions below acertain range. This limitation is associated to the presence of batteries: indeed they introduceindeed a relevant penalty in weight that cannot be countered by benefits of electrification forlonger range. Additional simulations indicate that a Blended WingBody concept based on aturboelectric only architecture is constantly performing better than the baseline within thelimits of the assumptions
Développement d'une méthode d'assimilation de données pour la calibration et la mise à jour en continu de modèles fidèles
d'éoliennes by
Adrien Hirvoas(
)
1 edition published in 2021 in English and held by 1 WorldCat member library worldwide
In the context of energy transition, wind power generation is developing rapidly. Meanwhile, in the framework of digitalization of the industry, the exploitation of collected data can be optimized by combination with numerical models. Such models can be complex and costly as they involve dynamic equations coupled with different physics. Furthermore, some of their input parameters related to the model properties as well as the external conditions can be badly known. These uncertainties affect the predictions obtained from model simulations and thus can impact the components lifetime for example. This dissertation focuses consequently on quantifying and reducing the input parameter uncertainties involved in an aeroservoelastic wind turbine model. Nevertheless, the widely used methods in uncertainty quantification are not suitable in the present industrial context because of the stochastic nature of the external solicitation and the time consuming behavior of the simulator. Our main contributions are twofold.Firstly, we want to quantify the impact of the uncertainties on the fatigue behavior of a wind turbine. We propose a global sensitivity analysis (GSA) methodology, based on the socalled Sobol' indices, for stochastic computer simulations. Such techniques, which often refer to the probabilistic framework and Monte Carlo (MC) methods, require a lot of calls to the numerical model. The uncertain input parameters are modeled by independent random variables gathered into a random vector and characterized by their probability distribution function (pdf). Variancebased GSA for time consuming deterministic computer models is usually performed by approximating the model by a surrogate regression. Among the different surrogates, we focus on Gaussian process (GP) regression characterized by its mean and covariance functions. One advantage of the GP regression metamodeling is to provide both a prediction of the numerical model and the associated uncertainty. In order to take into account the inherent randomness from stochastic simulations, we propose as a surrogate for the mean of the output of interest a GP regression with heteroscedastic noise. Then, this surrogate model is used to perform a sensitivity analysis based on classical MC estimation procedure.Secondly, we propose a Bayesian inference framework to carry out the calibration of influential input parameters from in situ measurements. It uses some measurements to update some prior pdfs on the unknown input parameters through the Bayes' theorem. Recent decades have been marked by a simultaneous development of sensor technologies and internet of things capabilities. Thus, our research efforts have been directed toward inference techniques where the data are sequentially processed when new observations become available. In this context, model parameter inference can be carried out using data assimilation methods. We carry out the calibration using an ensemble Kalman filter (EnKF). Nevertheless, unlike the model properties having a static or slow timevariant behavior, the parameters related to the external conditions have a dynamic aspect. Thus, we propose to carry out the inference problem using an EnKF coupled with an analog forecasting strategy based on nearest neighbors to model the underlying dynamic model. However, such problems can be solved assuming that several conditions of wellposedness and identifiability are achieved. We exploit the relationship between nonidentifiability of input parameters and total Sobol' indices. Indeed, for each measure output, we compute total Sobol indices associated to input parameters. If all the total Sobol' indices associated to a prescribed input parameter are "small", it means that this parameter is nonidentifiable. Due to the functional nature of the measurements, we rely on a dimension reduction preliminary step through principal component analysis and then we compute an aggregated Sobol' index for each model parameter
1 edition published in 2021 in English and held by 1 WorldCat member library worldwide
In the context of energy transition, wind power generation is developing rapidly. Meanwhile, in the framework of digitalization of the industry, the exploitation of collected data can be optimized by combination with numerical models. Such models can be complex and costly as they involve dynamic equations coupled with different physics. Furthermore, some of their input parameters related to the model properties as well as the external conditions can be badly known. These uncertainties affect the predictions obtained from model simulations and thus can impact the components lifetime for example. This dissertation focuses consequently on quantifying and reducing the input parameter uncertainties involved in an aeroservoelastic wind turbine model. Nevertheless, the widely used methods in uncertainty quantification are not suitable in the present industrial context because of the stochastic nature of the external solicitation and the time consuming behavior of the simulator. Our main contributions are twofold.Firstly, we want to quantify the impact of the uncertainties on the fatigue behavior of a wind turbine. We propose a global sensitivity analysis (GSA) methodology, based on the socalled Sobol' indices, for stochastic computer simulations. Such techniques, which often refer to the probabilistic framework and Monte Carlo (MC) methods, require a lot of calls to the numerical model. The uncertain input parameters are modeled by independent random variables gathered into a random vector and characterized by their probability distribution function (pdf). Variancebased GSA for time consuming deterministic computer models is usually performed by approximating the model by a surrogate regression. Among the different surrogates, we focus on Gaussian process (GP) regression characterized by its mean and covariance functions. One advantage of the GP regression metamodeling is to provide both a prediction of the numerical model and the associated uncertainty. In order to take into account the inherent randomness from stochastic simulations, we propose as a surrogate for the mean of the output of interest a GP regression with heteroscedastic noise. Then, this surrogate model is used to perform a sensitivity analysis based on classical MC estimation procedure.Secondly, we propose a Bayesian inference framework to carry out the calibration of influential input parameters from in situ measurements. It uses some measurements to update some prior pdfs on the unknown input parameters through the Bayes' theorem. Recent decades have been marked by a simultaneous development of sensor technologies and internet of things capabilities. Thus, our research efforts have been directed toward inference techniques where the data are sequentially processed when new observations become available. In this context, model parameter inference can be carried out using data assimilation methods. We carry out the calibration using an ensemble Kalman filter (EnKF). Nevertheless, unlike the model properties having a static or slow timevariant behavior, the parameters related to the external conditions have a dynamic aspect. Thus, we propose to carry out the inference problem using an EnKF coupled with an analog forecasting strategy based on nearest neighbors to model the underlying dynamic model. However, such problems can be solved assuming that several conditions of wellposedness and identifiability are achieved. We exploit the relationship between nonidentifiability of input parameters and total Sobol' indices. Indeed, for each measure output, we compute total Sobol indices associated to input parameters. If all the total Sobol' indices associated to a prescribed input parameter are "small", it means that this parameter is nonidentifiable. Due to the functional nature of the measurements, we rely on a dimension reduction preliminary step through principal component analysis and then we compute an aggregated Sobol' index for each model parameter
Modèles pour la difraction d'ondes par des obstacles revêtus de couches minces : résolution de problèmes de diffraction
d'ondes via une formulation by
Nathalie Bartoli(
Book
)
1 edition published in 2000 in French and held by 1 WorldCat member library worldwide
1 edition published in 2000 in French and held by 1 WorldCat member library worldwide
Intégration d'information a priori dans la régression de processus Gaussiens : Applications à l'ingénierie aéronautique by
Ankit Chiplunkar(
)
1 edition published in 2017 in English and held by 1 WorldCat member library worldwide
In this thesis, we propose to build better Gaussian Process (GP) modelsby integrating the prior knowledge of Aircraft design with experimental data. Due tothe high cost of performing experiments on physical systems, models become an efficientmeans to designing physical systems. We demonstrate how to create efficient models byincorporating the prior information from engineering design, mainly by changing the covariancefunctions of the GP.We propose GP models to detect onset of nonlinearity, detectmodal parameters and interpolate position of shock in aerodynamic experiments. Similarly,physical laws between multiple outputs can be enforced by manipulating the covariancefunctions, we propose to integrate flightmechanics to better identify loads using thesemodels. For each application we compare the proposed model with the stateoftheartmodel and demonstrate the cost or performance gains achieved
1 edition published in 2017 in English and held by 1 WorldCat member library worldwide
In this thesis, we propose to build better Gaussian Process (GP) modelsby integrating the prior knowledge of Aircraft design with experimental data. Due tothe high cost of performing experiments on physical systems, models become an efficientmeans to designing physical systems. We demonstrate how to create efficient models byincorporating the prior information from engineering design, mainly by changing the covariancefunctions of the GP.We propose GP models to detect onset of nonlinearity, detectmodal parameters and interpolate position of shock in aerodynamic experiments. Similarly,physical laws between multiple outputs can be enforced by manipulating the covariancefunctions, we propose to integrate flightmechanics to better identify loads using thesemodels. For each application we compare the proposed model with the stateoftheartmodel and demonstrate the cost or performance gains achieved
more
fewer
Audience Level
0 

1  
General  Special 
Related Identities
 Del Moral, Pierre Other
 Institut supérieur de l'aéronautique et de l'espace (Toulouse / 2007....). Degree grantor
 École doctorale AéronautiqueAstronautique (Toulouse) Other
 Morlier, Joseph (1977....). Opponent Thesis advisor Contributor
 Équipe d'accueil doctoral Modélisation et ingénierie des systèmes (Toulouse, HauteGaronne) Other
 SpringerLink (Online service) Other
 Gourinat, Yves (1960....). Other Opponent Thesis advisor
 Institut Clément Ader (Toulouse / 2009....). Other
 Le Riche, Rodolphe Other Opponent
 Gayton, Nicolas (1975....). Other Opponent