Cette thèse porte sur la conception des méthodes robustes de réduction d’ordre de modèles numériques de type Éléments Finis (EF) avec contrôle de la précision. La réduction d’ordre est en général nécessaire pour réduire drastiquement les temps de calcul et permettre ainsi une analyse paramétrique, une étude de faisabilité ou de performance de système (avion, unité de production, procédé complexe, etc). Dans cette étude, la technique de décomposition orthogonale aux valeurs propres (POD) sera utilisée pour construire des modèles réduits locaux. Informatiquement parlant, le “modèle” sera considéré comme une base de données de résultats de calcul avec capacité d’extrapolation et d’interpolation locale. Une stratégie adaptative pour stocker et accéder à la base de données est étudiée en étendant l’algorithme In situ Adaptive Tabulation (ISAT) proposé initialement par Pope. En fonction de l’usage et des exigences en précision des résultats, la base de données est enrichie en ligne (online) par des appels au modèle fin en respectant une précision spécifiée jusqu’à couvrir le domaine paramétrique entier, après quoi l’évaluation d’une solution devient très peu couteuse. L’approche couplée POD-ISAT proposée dans cette thèse fournit une méthode de réduction de modèle EF très performante. La méthodologie est évaluée sur un cas réel de conditionnement d’air en régime stationnaire de cabine d’avion dépendant de plusieurs paramètres de conception (température et vitesse d’entrée d’air, mode de ventilation personnalisée, conductivité thermique du fuselage, etc.). Pour les problèmes d’évolution en temps, nous explorons une piste de chainage de modèles et d’utilisation d’algorithme de parallélisation en temps tel que l’algorithme pararéel initialement proposé par Lions, Maday et Turinici (2001). Nous proposons ici une variante quasi-Newton de l’algorithme pararéel que nous appelons algorithme Broyden-pararéel. Il est appliqué au calcul de la diffusion d’un gaz dans la cabine d’avion. Cette thèse s’insère dans le cadre du projet CSDL (Complex System Design Lab, Fond Unique Interministériel) visant à développer une plate-forme logicielle multidisciplinaire pour la conception de systèmes complexes. / In this thesis, an efficient Reduced Order Modeling (ROM) technique with control of accuracy for parameterized Finite Element solutions is proposed. The ROM methodology is usually necessary to drastically reduce the computational time and allow for tasks like parameter analysis, system performance assessment (aircraft, complex process, etc.). In this thesis, a ROM using Proper Orthogonal Decomposition (POD) will be used to build local models. The “model” will be considered as a database of simulation results store and retrieve the database is studied by extending the algorithm In Situ Adaptive Tabulation (ISAT) originally proposed by Pope (1997). Depending on the use and the accuracy requirements, the database is enriched in situ (i.e. online) by call of the fine (reference) model and construction of a local model with an accuracy region in the parameter space. Once the trust regions cover the whole parameter domain, the computational cost of a solution becomes inexpensive. The coupled POD-ISAT, here proposed, provides a promising effective ROM approach for parametric finite element model. POD is used for the low-order representation of the spatial fields and ISAT for the local representation of the solution in the design parameter space. This method is tested on a Engineering case of stationary air flow in an aircraft cabin. This is a coupled fluid-thermal problem depending on several design parameters (inflow temperature, inflow velocity, fuselage thermal conductivity, etc.). For evolution problems, we explore the use of time-parallel strategies, namely the parareal algorithm originally proposed by Lions, Maday and Turinici (2001). A quasi-Newton variant of the algorithm called Broyden-parareal algorithm is here proposed. It is applied to the computation of the gas diffusion in an aircraft cabin. This thesis is part of the project CSDL (Complex System Design Lab) funded by FUI (Fond Unique Interministériel) aimed at providing a software platform for multidisciplinary design of complex systems.

Identifying the parameters with the largest influence on the predicted outputs of a model revealswhich parameters need to be known more precisely to reduce the overall uncertainty on themodel output. A large improvement of such models would result when uncertainties in the keymodel parameters are reduced. To achieve this, new experiments could be very helpful,especially if the measurements are taken at the spatio-temporal locations that allow estimate the parameters in an optimal way. After evaluating the methodologies available for optimal sensor location, a few observations were drawn. The method based on the Gram determinant evolution can report results not according to what should be expected. This method is strongly dependent of the sensitivity coefficients behaviour. The approach based on the maximum angle between subspaces, in some cases, produced more that one optimal solution. It was observed that this method depends on the magnitude of outputs values and report the measurement positions where the outputs reached their extrema values. The D-optimal design method produces number and locations of the optimal measurements and it depends strongly of the sensitivity coefficients, but mostly of their behaviours. In general it was observed that the measurements should be taken at the locations where the extrema values (sensitivity coefficients, POD modes and/or outputs values) are reached. Further improvements can be obtained when a reduced model of the system is employed. This is computationally less expensive and the best estimation of the parameter is obtained, even with experimental data contaminated with noise. A new approach to calculate the time coefficients belonging to an empirical approximator based on the POD-modes derived from experimental data is introduced. Additionally, an artificial neural network can be used to calculate the derivatives but only for systems without complex nonlinear behaviour. The latter two approximations are very valuable and useful especially if the model of the system is unknown.

In this thesis, the method of Proper Orthogonal Decomposition (POD) is implemented to construct approximate, reduced order models (ROM) of mesh movement methods. Three mesh movement algorithms are implemented and comparatively evaluated, namely radial basis function interpolation, mesh optimization and elastic deformation. POD models of the mesh movement algorithms are constructed using a series of system observations, or snapshots of a given mesh for a set of boundary deformations. The scalar expansion coefficients for the POD basis modes are computed in three different ways, through coefficient optimization, Galerkin projection of the governing set of equations and coefficient interpolation. It is found that using only coefficient interpolation yields mesh movement models that accurately approximates the full order mesh movement, with CPU cost savings in excess of 99%. We further introduce a novel training procedure whereby the POD models are generated in a fully automated fashion. The technology is applicable to any mesh movement method and enables potential reductions of up to four orders of magnitude in mesh movement related costs. The proposed model can be implemented without having to pre-train the POD model, to any fluid-structure interaction code with an existing mesh movement scheme. Copyright / Dissertation (MEng)--University of Pretoria, 2010. / Mechanical and Aeronautical Engineering / unrestricted

Cette thèse de doctorat est consacrée à l'étude des problèmes directe et inverse en électrophysiologie cardiaque. Comme les équations qui décrivent l'activité électrique du coeur peuvent être très couteuses en temps de calcul, une attention particulière est apportée aux méthodes d'ordre réduit et à leur applications aux modèles de l'électrophysiologie.Dans un premier temps, nous introduisons les modèles mathématiques et numériques de l'électrophysiologie cardiaque. Ces modèles nous permettent de réaliser des simulations numériques que nous validons à l'aide de plusieurs critères qualitatifs et quantitatifs trouvés dans la littérature médicale. Comme notre modèle prend en compte les oreillettes et les ventricules, nous sommes capables de reproduire des cycles complets d'électrocardiogrammes (ECG) à la fois dans des conditions saines et dans des cas pathologiques.Ensuite, plusieurs méthodes d'ordre réduit sont étudiées pour la résolution des équations de l'électrophysiologie. La méthode Proper Orthogonal Decomposition (POD) est appliquée pour la discrétisation des équations de l'électrophysiologie dans plusieurs configurations, comme par exemple la simulation d'un infarctus du myocarde. De plus, cette méthode est utilisée pour résoudre quelques problèmes d'identification de paramètres comme localiser un infarctus à partir de mesures d'un électrocardiogramme ou simuler une courbe de restitution. Pour contourner les limitations de la POD, une nouvelle méthode basée sur des couples de Lax approchés (Approximated Lax Pairs, ALP) est utilisée. Cette méthode est appliquée aux problèmes directe et inverse. Pour finir, un nouvel algorithme, basé sur les méthodes ALP et l'interpolation empirique discrète, est proposé. Cette nouvelle approche améliore significativement l'efficacité de l'algorithme original ALP et nous permet de considérer des modèles plus complexes utilisés en électrophysiologie cardiaque. / This PhD thesis is dedicated to the investigation of the forward and the inverse problem of cardiac electrophysiology. Since the equations that describe the electrical activity of the heart can be very demanding from a computational point of view, a particular attention is paid to the reduced order methods and to their application to the electrophysiology models. First, we introduce the mathematical and numerical models of electrophysiology and we implement them to provide for simulations that are validated against various qualitative and quantitative criteria found in the medical literature. Since our model takes into account atria and ventricles, we are able to reproduce full cycle Electrocardiograms (ECG) in healthy configurations and also in the case of several pathologies. Then, several reduced order methods are investigated for the resolution of the electrophysiology equations. The Proper orthogonal Decomposition (POD) method is applied for the discretization of the electrophysiology equations in several configurations, as for instance the simulation of a myocardial infarction. Also, the method is used in order to solve some parameters identification problems such as the identification of an infarcted zone using the Electrocardiogram measures and for the efficient simulation of restitution curves. To circumvent some limitations of the POD method, a new reduced order method based on the Approximated Lax Pairs (ALP) is investigated. This method is applied to the forward and inverse problems. Finally, a new reduced order algorithm is proposed, based on the ALP and the Discrete Empirical Interpolation methods. This new approach significantly improves the efficiency of the original ALP algorithm and allow us to consider more complex models used in electrophysiology.

In this document we review the status of existing techniques for nonlinear model order reduction by investigating how well these techniques perform for typical industrial needs. In particular the TPWL-method (Trajectory Piecewise Linear-method) and the POD-approach (Proper Orthogonal Decomposion) is taken under consideration. We address several questions that are (closely) related to both the theory and application of nonlinear model order reduction techniques. The goal of this document is to provide an overview of available methods together with a classification of nonlinear problems that in principle could be handled by these methods.

No description available.

No description available.

This dissertation contains work that addresses both theoretical and numerical aspects of reduced order models (ROMs). In an under-resolved regime, the classical Galerkin reduced order model (G-ROM) fails to yield accurate approximations. Thus, we propose a new ROM, the data-driven variational multiscale ROM (DD-VMS-ROM) built by adding a closure term to the G-ROM, aiming to increase the numerical accuracy of the ROM approximation without decreasing the computational efficiency. The closure term is constructed based on the variational multiscale framework. To model the closure term, we use data-driven modeling. In other words, by using the available data, we find ROM operators that approximate the closure term. To present the closure term's effect on the ROMs, we numerically compare the DD-VMS-ROM with other standard ROMs. In numerical experiments, we show that the DD-VMS-ROM is significantly more accurate than the standard ROMs. Furthermore, to understand the closure term's physical role, we present a theoretical and numerical investigation of the closure term's role in long-time integration. We theoretically prove and numerically show that there is energy exchange from the most energetic modes to the least energetic modes in closure terms in a long time averaging. One of the promising contributions of this dissertation is providing the numerical analysis of the data-driven closure model, which has not been studied before. At both the theoretical and the numerical levels, we investigate what conditions guarantee that the small difference between the data-driven closure model and the full order model (FOM) closure term implies that the approximated solution is close to the FOM solution. In other words, we perform theoretical and numerical investigations to show that the data-driven model is verifiable. Apart from studying the ROM closure problem, we also investigate the setting in which the G-ROM converges optimality. We explore the ROM error bounds' optimality by considering the difference quotients (DQs). We theoretically prove and numerically illustrate that both the ROM projection error and the ROM error are suboptimal without the DQs, and optimal if the DQs are used. / Doctor of Philosophy / In many realistic applications, obtaining an accurate approximation to a given problem can require a tremendous number of degrees of freedom. Solving these large systems of equations can take days or even weeks on standard computational platforms. Thus, lower-dimensional models, i.e., reduced order models (ROMs), are often used instead. The ROMs are computationally efficient and accurate when the underlying system has dominant and recurrent spatial structures. Our contribution to reduced order modeling is adding a data-driven correction term, which carries important information and yields better ROM approximations. This dissertation's theoretical and numerical results show that the new ROM equipped with a closure term yields more accurate approximations than the standard ROM.

This thesis describes several approaches for stabilizing POD-ROMs (that is, reduced order models based on basis functions derived from the proper orthogonal decomposition) for both the CDR (convection-diffusion-reaction) equation and the NSEs (Navier-Stokes equations). Stabilization is necessary because standard POD-ROMs of convection-dominated problems usually display numerical instabilities. The first stabilized ROM investigated is a streamline-upwind Petrov-Galerkin ROM (SUPG-ROM). I prove error estimates for the SUPG-ROM and derive optimal scalings for the stabilization parameter. I test the SUPG-ROM with the optimal parameter in the numerical simulation of a convection-dominated CDR problem. The SUPG-ROM yields more accurate results than the standard Galerkin ROM (G-ROM) by eliminating the inherent numerical artifacts (noise) in the data and dampening spurious oscillations. I next propose two regularized ROMs (Reg-ROMs) based on ideas from large eddy simulation and turbulence theory: the Leray ROM (L-ROM) and the evolve-then-filter ROM (EF-ROM). Both Reg-ROMs use explicit POD spatial filtering to regularize (smooth) some of the terms in the standard G-ROM. I propose two different POD spatial filters: one based on the POD projection and a novel POD differential filter. These two new Reg-ROMs and the two spatial filters are investigated in the numerical simulation of the three-dimensional flow past a circular cylinder problem at Re = 100. The numerical results show that EF-ROM-DF is the most accurate Reg-ROM and filter combination and the differential filter generally yields better results than the projection filter. The Reg-ROMs perform significantly better than the standard G-ROM and decrease the CPU time (compared against the direct numerical simulation) by orders of magnitude (from about four days to four minutes). / Ph. D.

<p>Train transportation is a vital part of the transportation system of today anddue to its safe and environmental friendly concept it will be even more impor-tant in the future. The speeds of trains have increased continuously and withhigher speeds the aerodynamic effects become even more important. One aero-dynamic effect that is of vital importance for passengers’ and track workers’safety is slipstream, i.e. the flow that is dragged by the train. Earlier ex-perimental studies have found that for high-speed passenger trains the largestslipstream velocities occur in the wake. Therefore the work in this thesis isdevoted to wake flows. First a test case, a surface-mounted cube, is simulatedto test the analysis methodology that is later applied to a train geometry, theAerodynamic Train Model (ATM). Results on both geometries are comparedwith other studies, which are either numerical or experimental. The comparisonfor the cube between simulated results and other studies is satisfactory, whiledue to a trip wire in the experiment the results for the ATM do not match.The computed flow fields are used to compute the POD and Koopman modes.For the cube this is done in two regions of the flow, one to compare with a priorpublished study Manhart & Wengle (1993) and another covering more of theflow and especially the wake of the cube. For the ATM, a region containing theimportant flow structures is identified in the wake, by looking at instantaneousand fluctuating velocities. To ensure converged POD modes two methods toinvestigate the convergence are proposed, tested and applied. Analysis of themodes enables the identification of the important flow structures. The flowtopologies of the two geometries are very different and the flow structures arealso different, but the same methodology can be applied in both cases. For thesurface-mounted cube, three groups of flow structures are found. First groupis the mean flow and then two kinds of perturbations around the mean flow.The first perturbation is at the edge of the wake, relating to the shear layerbetween the free stream and the disturbed flow. The second perturbation isinside the wake and is the convection of vortices. These groups would then betypical of the separation bubble that exists in the wake of the cube. For theATM the main flow topology consists of two counter rotating vortices. Thiscan be seen in the decomposed modes, which, except for the mean flow, almostonly contain flow structures relating to these vortices.</p> / QC 20100518 / Gröna Tåget