<p align='justify'>This thesis is devoted to the micromechanical study of the size-dependent strengthening in Transformation Induced Plasticity (TRIP) steels. Such grades of advanced high-strength steels are compelling for the automotive industry, due to their improved mechanical properties. Among others, they combine a good strength versus ductility balance. In this context, many research works have been carried out to study these grades of steels. In particular, from a numerical point of view, earlier studies within the framework of classical plasticity do not properly reproduce the strengthening levels characterizing TRIP steels and obtained experimentally.</p> <p><p align='justify'>In this study, the strain gradient plasticity theory presented by Fleck and Hutchinson (2001) is chosen to account for the strengthening effect resulting from the phase transformation. A two-dimensional embedded cell model of a simplified microstructure composed of small cylindrical metastable austenitic inclusions, partially undergoing the phase transformation, within a ferritic matrix is used.</p><p><p align='justify'>First, the single-parameter version of the strain gradient plasticity theory under small strain assumption is used for the simulations. The impact of the higher order boundary conditions is assessed. It is shown that, when the plastic flow is unconstrained at the elasto-plastic boundaries, the transformation strain has no significant impact on the overall strengthening. The strengthening is essentially coming from the composite effect with a marked inclusion size effect resulting from the appearance during deformation of new boundaries (at the interface between parent and product phases) constraining the plastic flow.</p><p><p align='justify'>Second, the multi-parameter version of the strain gradient plasticity theory, incorporating separately the rotational and extensional gradients in the formulation, is employed under small strain assumption. The effect of the plastic strain gradients resulting from the transformation strain is better captured. In particular, the results show a significant influence of the shear component of the transformation strain. An implicit confinement effect is revealed at the elasto-plastic boundaries which is partly responsible for the transformation strain effect. Size effects on the overall strengthening are also revealed, due to a combined size dependent effect of the transformation strain and of the evolving composite structure.</p><p align='justify'>Third, the extension of the strain gradient plasticity theory to a finite strain description is applied. A significant effect of the transformation strain is obtained with the multi-parameter version of the theory as well as an optimal austenite grain size improving the damage resistance of the martensite, in agreement with the typical grain size of the current TRIP-assisted steels (Jacques et al. 2007).</p> / Doctorat en Sciences de l'ingénieur / info:eu-repo/semantics/nonPublished

<p align="justify">In a context of restoration of historical masonry structures, it is crucial to properly estimate the residual strength and the potential structural failure modes in order to assess the safety of buildings. Due to its mesostructure and the quasi-brittle nature of its constituents, masonry presents preferential damage orientations, strongly localised failure modes and damage-induced anisotropy, which are complex to incorporate in structural computations. Furthermore, masonry structures are generally subjected to complex loading processes including both in-plane and out-of-plane loads which considerably influence the potential failure mechanisms. As a consequence, both the membrane and the flexural behaviours of masonry walls have to be taken into account for a proper estimation of the structural stability.</p><p><p align="justify">Macrosopic models used in structural computations are based on phenomenological laws including a set of parameters which characterises the average behaviour of the material. These parameters need to be identified through experimental tests, which can become costly due to the complexity of the behaviour particularly when cracks appear. The existing macroscopic models are consequently restricted to particular assumptions. Other models based on a detailed mesoscopic description are used to estimate the strength of masonry and its behaviour with failure. This is motivated by the fact that the behaviour of each constituent is a priori easier to identify than the global structural response. These mesoscopic models can however rapidly become unaffordable in terms of computational cost for the case of large-scale three-dimensional structures.</p><p><p align="justify">In order to keep the accuracy of the mesoscopic modelling with a more affordable computational effort for large-scale structures, a multi-scale framework using computational homogenisation is developed to extract the macroscopic constitutive material response from computations performed on a sample of the mesostructure, thereby allowing to bridge the gap between macroscopic and mesoscopic representations. Coarse graining methodologies for the failure of quasi-brittle heterogeneous materials have started to emerge for in-plane problems but remain largely unexplored for shell descriptions. The purpose of this study is to propose a new periodic homogenisation-based multi-scale approach for quasi-brittle thin shell failure.</p><p><p align="justify">For the numerical treatment of damage localisation at the structural scale, an embedded strong discontinuity approach is used to represent the collective behaviour of fine-scale cracks using average cohesive zones including mixed cracking modes and presenting evolving orientation related to fine-scale damage evolutions.</p><p><p align="justify">A first originality of this research work is the definition and analysis of a criterion based on the homogenisation of a fine-scale modelling to detect localisation in a shell description and determine its evolving orientation. Secondly, an enhanced continuous-discontinuous scale transition incorporating strong embedded discontinuities driven by the damaging mesostructure is proposed for the case of in-plane loaded structures. Finally, this continuous-discontinuous homogenisation scheme is extended to a shell description in order to model the localised behaviour of out-of-plane loaded structures. These multi-scale approaches for failure are applied on typical masonry wall tests and verified against three-dimensional full fine-scale computations in which all the bricks and the joints are discretised.</p> / Doctorat en Sciences de l'ingénieur / info:eu-repo/semantics/nonPublished

Chaque année, d'innombrables collisions se produisent entre des avions en vol et des oiseaux. L'impact aviaire, menace redoutée par les pilotes, concerne tant l'aviation civile que son pendant militaire. Les statistiques démontrent que, même si fort heureusement le nombre d'accidents graves reste limité, les incidents sont de plus en plus nombreux.<p>Parmi les acteurs qui luttent contre ce danger, les constructeurs d'avions jouent un rôle prépondérant. Contraints par des réglementations internationales, ils s'attachent à produire des éléments de structure qui résistent à l'impact d'oiseaux.<p>Dans la mise au point de leur produits, les avionneurs démontrent cette résistance à l'aide d'essais d'impact :on accélère un simulant d'oiseau jusqu'à la vitesse voulue -- de l'ordre de la vitesse de croisière nominale de l'avion -- et on le projette sur un aileron ou un morceau de fuselage. <p><p>La présente thèse doctorale, co-dirigée par les professeurs Philippe Bouillard de l'ULB et Marc Pirlot de l'ERM, contribue doublement à l'amélioration de ces techniques d'essais dynamiques.<p><p><p>D'une part, elle réalise la mise au point et la validation d'un lanceur pyrotechnique à double étage pour l'accélération du simulant d'oiseau. Un canon de calibre 20 mm est combiné avec un accélérateur de calibre 160 mm. La combustion d'un mélange de poudre propulsive contenu dans une douille adaptée génère les gaz à haute température et à haute pression nécessaires à l'accélération d'un simulant d'oiseau dûment confiné dans un conteneur de protection. Un dispositif de séparation arrête le conteneur afin que seul le simulant d'oiseau percute l'élément d'avion à l'essai. La solution pyrotechnique à double étage mise au point est validée par de nombreux tirs instrumentés en vitesse, en accélération et en pression ;elle se révèle conforme aux exigences de sécurité et de reproductibilité. Le lanceur pyrotechnique présente par rapport aux solutions pneumatiques, utilisées à notre connaissance dans tous les autres centres d'essais, des avantages indéniables de compacité ainsi que de rapidité et de souplesse de mise en oeuvre.<p><p><p>D'autre part, la migration des alliages métalliques vers les matériaux composites est amorcée depuis plusieurs années déjà dans le monde de la construction aéronautique. Pour optimiser les structures, une connaissance des caractéristiques de ces matériaux est indispensable. Les modes de rupture font partie des caractéristiques encore mal connues. La mesure du déplacement hors-plan lors du tir sur panneaux plans est une des manières de quantifier le comportement du matériau sous l'action d'un impact. Cette mesure s'opère généralement de manière statique, après le tir. Une méthode de mesure dynamique a été mise au point, basée sur l'emploi de techniques de stéréoscopie par corrélation numérique d'images. Cette technique a été validée au moyen d'une méthode métrologique indépendante d'extensométrie laser. <p><p><p>ABSTRACT<p><p>Countless collisions occur each year between airplanes and birds. Bird strike is a concern to both civilian and militay aircraft. Statistics show that, although the number of serious accidents fortunately remains low, the number of incidents keeps increasing.<p>Amongst the actors tackling this issue, aircraft manufacturers play an important role. In compliance with international regulations, they have to produce structural elements that withstand bird impact. During the development of their products, aircraft manufacturers have to demonstrate this resistance through bird impact trials :a bird surrogate is accelerated to the required velocity - often close to the nominal cruise speed of the aircraft - and launched onto a flap or a piece of fuselage.<p><p>This PhD thesis has been co-supervised by Professor Philippe Bouillard (ULB) and Professor Marc Pirlot (ERM-KMS). Its contribution to the improvement of the aforementioned dynamic trials is twofold.<p><p><p>One one hand, a two-stage pyrotechnical launcher for bird surrogates has been developed and assessed. A 20 mm caliber gun is connected to a 160 mm diameter launcher. The combustion of a propellant mixture in a cartridge case generates high pressure, high temperature gases which accelerate a bird surrogate protected by a cylindrical container. A stripper refrains the container from hitting the target pane. <p>The pyrotechnical solution has been assessed through an important number of firings where pressure, velocity and acceleration have been measured. The solution has proven compliance with both the safety requirements and the repeatability specifications. Its advantages compared to the pneumatic solutions used, as far as we know, in all other test centres, include compactedness as well as flexibility and high firing rate.<p><p><p>On the other hand, migration towards composite materials has been initiated years ago in the area of aeronautical constructions.<p>To optimize structures, a thorough knowledge of these new materials is required. Failure modes belong to the still badly known features of carbon reinforced plastics. Measuring the out-of-plane deformation when firing on a flat pane is one way of quantifying the material's behaviour under impact. This measurement is most frequently made in a static way, after completion of the firing. A dynamic measuring method has been developed, based upon stereoscopic digital image correlation techniques. This technique has been validated by means of an independent laser extensometer measuring method. <p> / Doctorat en Sciences de l'ingénieur / info:eu-repo/semantics/nonPublished

<p align='justify'>L’adaptation des surfaces pour des fonctions prédéterminées par le choix des matériaux métalliques ou des couches minces ayant des propriétés mécaniques avancées peut potentiellement permettre de réaliser des nouvelles applications à petites échelles. Concevoir de telles applications utilisant des nouveaux matériaux nécessite en premier lieu la connaissance des propriétés mécaniques des matériaux ciblés à l’échelle microscopique et nanoscopique. Une méthode souvent appliquée pour caractériser les matériaux à petites échelles est la nanoindentation, qui peut être vue comme une mesure de dureté à l’échelle nanoscopique.</p><p><p align='justify'>Ce travail présente une contribution relative à l'interprétation des résultats de la nanoindentation, qui fait intervenir un grand nombre de phénomènes physiques couplés à l'aide de simulations numériques. A cette fin une approche interdisciplinaire, adaptée aux phénomènes apparaissant à petites échelles, et située à l’intersection entre la physique, la mécanique et la science des matériaux a été utilisée. Des modèles numériques de la nanoindentation ont été conçus à l'échelle atomique (modèle discret) et à l'échelle des milieux continus (méthode des éléments finis), pour étudier le comportement du nickel pur. Ce matériau a été choisi pour ses propriétés mécaniques avancées, sa résistance à l'usure et sa bio-compatibilité, qui peuvent permettre des applications futures intéressantes à l'échelle nanoscopique, particulièrement dans le domaine biomédical. Des méthodes avancées de mécanique du solide ont été utilisées pour prendre en compte les grandes déformations locales du matériau (par la formulation corotationelle), et pour décrire les conditions de contact qui évoluent au cours de l'analyse dans le modèle à l'échelle des milieux continus (traitement des conditions de contact unilatérales et tangentielles par une forme de Lagrangien augmenté).</p><p><p align='justify'>L’application des modèles numériques a permis de contribuer à l’identification des phénomènes qui gouvernent la nanoindentation du nickel pur. Le comportement viscoplastique du nickel pur pendant nanoindentation a été identifié dans une étude expérimentale-numérique couplée, et l'effet cumulatif de la rugosité et du frottement sur la dispersion des résultats de la nanoindentation a été montré par une étude numérique (dont les résultats sont en accord avec des tendances expérimentales).</p> <p><p align='justify'>Par ailleurs, l’utilisation de l’outil numérique pour une autre application à petites échelles, la manipulation des objets par contact, a contribué à la compréhension de la variation de l’adhésion électrostatique pendant micromanipulation. La déformation plastique des aspérités de surface sur le bras de manipulateur (en nickel pur) a été identifiée comme une source potentielle d’augmentation importante de l'adhésion pendant la micromanipulation, qui peut potentiellement causer des problèmes de relâche et de précision de positionnement, observés expérimentalement.</p><p><p align='justify'>Les résultats présentés dans cette thèse montrent que des simulations numériques basées sur la physique du problème traité peuvent expliquer des tendances expérimentales et contribuer à la compréhension et l'interprétation d'essais couramment utilisé pour la caractérisation aux petites échelles. Le travail réalisé dans cette thèse s’inscrit dans un projet de recherche appelé "mini-micro-nano" (mµn), financé par la Communauté Française de Belgique dans le cadre de "l'Action de Recherche Concertée", convention 04/09-310.</p> / Doctorat en Sciences de l'ingénieur / info:eu-repo/semantics/nonPublished

In the global framework of improving vibro-acoustic numerical prediction quality together with the need to decrease the number of prototyping stages, this manuscript focuses on achieving greater accuracy for acoustic numerical simulations by making use of a parametric updating technique, which enables tuning the model parameters inside physically meaningful boundaries. The improved model is used for the next prototyping stages, allowing more accurate results within reduced simulation times. The updating technique is based on recent works dealing with the constitutive relation error method (CRE) applied to acoustics. The updating process focuses on improving the acoustic damping matrix related to the absorbing properties of the materials covering the borders of the acoustic domain. / Doctorat en sciences appliquées / info:eu-repo/semantics/nonPublished

Though lots of numerical methods have been proposed in the literature to optimize me-chanical structures at the final stage of the design process, few designers use these tools since the first stage. However, a minor modification at the first step can bring significant change to the global performances of the structure. Usually, during the initial stage, models are based on theoretical and empirical equations, which are often characterized by mixed variables: continuous (e.g. geometrical dimensions), discrete (e.g. the cross section of a beam available in a catalogue) and/or integer (e.g. the number of layers in a composite material). Furthermore, the functions involved may be non differentiable, or even discontinuous. Therefore, classical algorithms based on the computation of sensi-tivities are no more applicable. <p><p>Consequently, to solve these problems, the most wide-spread meta-heuristic methods are evolutionary algorithms (EAs), which work as follows: the best individuals among an initial population of randomly generated potential solutions are favoured and com-bined (by specific operators like crossover and mutation) in order to create potentially better individuals at the next generation. The creation of new generations is repeated till the convergence is reached. The ability of EAs to explore widely the design space is useful to solve single-objective unconstrained optimization problems, because it gener-ally prevents from getting trapped into a local optimum, but it is also well known that they do not perform very efficiently in the presence of constraints. Furthermore, in many industrial applications, multiple objectives are pursued together. <p><p>Therefore, to take into account the constrained and multicriteria aspects of optimization problems in EAs, a new method called PAMUC (Preferences Applied to MUltiobjectiv-ity and Constraints) has been proposed in this dissertation. First the user has to assign weights to the m objectives. Then, an additional objective function is built by linearly aggregating the normalized constraints. Finally, a multicriteria decision aid method, PROMETHEE II, is used in order to rank the individuals of the population following the m+1 objectives. <p><p>PAMUC has been validated on standard multiobjective test cases, as well as on the pa-rametrical optimization of the purge valve and the feed valve of the Vinci engine, both designed by Techspace Aero for launcher Ariane 5.<p>\ / Doctorat en sciences appliquées / info:eu-repo/semantics/nonPublished

<p align="justify">The conservation of structures of the historical heritage is an increasing concern nowadays for public authorities. The technical design phase of repair operations for these structures is of prime importance. Such operations usually require an estimation of the residual strength and of the potential structural failure modes of structures to optimize the choice of the repairing techniques.</p> <p><p align="justify">Although rules of thumb and codes are widely used, numerical simulations now start to emerge as valuable tools. Such alternative methods may be useful in this respect only if they are able to account realistically for the possibly complex failure modes of masonry in structural applications.</p><p><p align="justify">The mechanical behaviour of masonry is characterized by the properties of its constituents (bricks and mortar joints) and their stacking mode. Structural failure mechanisms are strongly connected to the mesostructure of the material, with strong localization and damage-induced anisotropy.</p><p><p align="justify">The currently available numerical tools for this material are mostly based on approaches incorporating only one scale of representation. Mesoscopic models are used in order to study structural details with an explicit representation of the constituents and of their behaviour. The range of applicability of these descriptions is however restricted by computational costs. At the other end of the spectrum, macroscopic descriptions used in structural computations rely on phenomenological constitutive laws representing the collective behaviour of the constituents. As a result, these macroscopic models are difficult to identify and sometimes lead to wrong failure mode predictions.</p><p><p align="justify">The purpose of this study is to bridge the gap between mesoscopic and macroscopic representations and to propose a computational methodology for the analysis of plane masonry walls. To overcome the drawbacks of existing approaches, a multi-scale framework is used which allows to include mesoscopic behaviour features in macroscopic descriptions, without the need for an a priori postulated macroscopic constitutive law. First, a mesoscopic constitutive description is defined for the quasi-brittle constituents of the masonry material, the failure of which mainly occurs through stiffness degradation. The mesoscopic description is therefore based on a scalar damage model. Plane stress and generalized plane state assumptions are used at the mesoscopic scale, leading to two-dimensional macroscopic continuum descriptions. Based on periodic homogenization techniques and unit cell computations, it is shown that the identified mesoscopic constitutive setting allows to reproduce the characteristic shape of (anisotropic) failure envelopes observed experimentally. The failure modes corresponding to various macroscopic loading directions are also shown to be correctly captured. The in-plane failure mechanisms are correctly represented by a plane stress description, while the generalized plane state assumption, introducing simplified three-dimensional effects, is shown to be needed to represent out-of-plane failure under biaxial compressive loading. Macroscopic damage-induced anisotropy resulting from the constituents' stacking mode in the material, which is complex to represent properly using macroscopic phenomenological constitutive equations, is here obtained in a natural fashion. The identified mesoscopic description is introduced in a scale transition procedure to infer the macroscopic response of the material. The first-order computational homogenization technique is used for this purpose to extract this response from unit cells. Damage localization eventually appears as a natural outcome of the quasi-brittle nature of the constituents. The onset of macroscopic localization is treated as a material bifurcation phenomenon and is detected from an eigenvalue analysis of the homogenized acoustic tensor obtained from the scale transition procedure together with a limit point criterion. The macroscopic localization orientations obtained with this type of detection are shown to be strongly related to the underlying mesostructural failure modes in the unit cells.</p> <p><p align="justify">A well-posed macroscopic description is preserved by embedding localization bands at the macroscopic localization onset, with a width directly deduced from the initial periodicity of the mesostructure of the material. This allows to take into account the finite size of the fracturing zone in the macroscopic description. As a result of mesoscopic damage localization in narrow zones of the order of a mortar joint, the material response computationally deduced from unit cells may exhibit a snap-back behaviour. This precludes the use of such a response in the standard strain-driven multi-scale scheme.</p> <p><p align="justify">Adaptations of the multi-scale framework required to treat the mesostructural response snap-back are proposed. This multi-scale framework is finally applied for a typical confined shear wall problem, which allows to verify its ability to represent complex structural failure modes.</p><p> / Doctorat en sciences appliquées / info:eu-repo/semantics/nonPublished

L’apprentissage statistique moderne démontre des résultats impressionnants, où les or- dinateurs viennent à atteindre ou même à excéder les standards humains dans certaines applications telles que la vision par ordinateur ou les jeux de stratégie. Pourtant, malgré ces avancées, force est de constater que les applications fiables en déploiement en sont encore à leur état embryonnaire en comparaison aux opportunités qu’elles pourraient apporter. C’est dans cette perspective, avec une emphase mise sur la théorie de décision séquentielle et sur les recherches récentes en apprentissage automatique, que nous démontrons l’applica- tion efficace de ces méthodes sur des cas liés au réseau électrique et à l’optimisation de ses acteurs. Nous considérons ainsi des instances impliquant des unités d’emmagasinement éner- gétique ou des voitures électriques, jusqu’aux contrôles thermiques des bâtiments intelligents. Nous concluons finalement en introduisant une nouvelle approche hybride qui combine les performances modernes de l’apprentissage profond et de l’apprentissage par renforcement au cadre d’application éprouvé de la recherche opérationnelle classique, dans le but de faciliter l’intégration de nouvelles méthodes d’apprentissage statistique sur différentes applications concrètes. / While modern statistical learning is achieving impressive results, as computers start exceeding human baselines in some applications like computer vision, or even beating pro- fessional human players at strategy games without any prior knowledge, reliable deployed applications are still in their infancy compared to what these new opportunities could fathom. In this perspective, with a keen focus on sequential decision theory and recent statistical learning research, we demonstrate efficient application of such methods on instances involving the energy grid and the optimization of its actors, from energy storage and electric cars to smart buildings and thermal controls. We conclude by introducing a new hybrid approach combining the modern performance of deep learning and reinforcement learning with the proven application framework of operations research, in the objective of facilitating seamlessly the integration of new statistical learning-oriented methodologies in concrete applications.

La reconstruction de bâtiments à partir de photographies aériennes et d’autres données spatiales urbaines multi-sources est une tâche qui utilise une multitude de méthodes automatisées et semi-automatisées allant des processus ponctuels au traitement classique des images et au balayage laser. Dans cette thèse, un système de relaxation itératif est développé sur la base de l'examen du contexte local de chaque bord en fonction de multiples sources d'entrée spatiales (masques optiques, d'élévation, d'ombre et de feuillage ainsi que d'autres données prétraitées, décrites au chapitre 6). Toutes ces données multisource et multirésolution sont fusionnées de manière à extraire les segments de ligne probables ou les arêtes correspondant aux limites des bâtiments. Deux nouveaux sous-systèmes ont également été développés dans cette thèse. Ils ont été conçus dans le but de fournir des informations supplémentaires, plus fiables, sur les contours des bâtiments dans une future version du système de relaxation proposé. La première est une méthode de réseau de neurones à convolution profonde (CNN) pour la détection de frontières de construction. Le réseau est notamment basé sur le modèle SRCNN (Dong C. L., 2015) de super-résolution à la pointe de la technologie. Il accepte des photographies aériennes illustrant des données de zones urbaines densément peuplées ainsi que leurs cartes d'altitude numériques (DEM) correspondantes. La formation utilise trois variantes de cet ensemble de données urbaines et vise à détecter les contours des bâtiments grâce à une nouvelle cartographie hétéroassociative super-résolue. Une autre innovation de cette approche est la conception d'une couche de perte personnalisée modifiée appelée Top-N. Dans cette variante, l'erreur quadratique moyenne (MSE) entre l'image de sortie reconstruite et l'image de vérité de sol (GT) fournie des contours de bâtiment est calculée sur les 2N pixels de l'image avec les valeurs les plus élevées. En supposant que la plupart des N pixels de contour de l’image GT figurent également dans les 2N pixels supérieurs de la reconstruction, cette modification équilibre les deux catégories de pixels et améliore le comportement de généralisation du modèle CNN. Les expériences ont montré que la fonction de coût Top-N offre des gains de performance par rapport à une MSE standard. Une amélioration supplémentaire de la capacité de généralisation du réseau est obtenue en utilisant le décrochage. Le deuxième sous-système est un réseau de convolution profonde à super-résolution, qui effectue un mappage associatif à entrée améliorée entre les images d'entrée à basse résolution et à haute résolution. Ce réseau a été formé aux données d’altitude à basse résolution et aux photographies urbaines optiques à haute résolution correspondantes. Une telle différence de résolution entre les images optiques / satellites optiques et les données d'élévation est souvent le cas dans les applications du monde réel. / Building reconstruction from aerial photographs and other multi-source urban spatial data is a task endeavored using a plethora of automated and semi-automated methods ranging from point processes, classic image processing and laser scanning. In this thesis, an iterative relaxation system is developed based on the examination of the local context of each edge according to multiple spatial input sources (optical, elevation, shadow & foliage masks as well as other pre-processed data as elaborated in Chapter 6). All these multisource and multiresolution data are fused so that probable line segments or edges are extracted that correspond to prominent building boundaries.Two novel sub-systems have also been developed in this thesis. They were designed with the purpose to provide additional, more reliable, information regarding building contours in a future version of the proposed relaxation system. The first is a deep convolutional neural network (CNN) method for the detection of building borders. In particular, the network is based on the state of the art super-resolution model SRCNN (Dong C. L., 2015). It accepts aerial photographs depicting densely populated urban area data as well as their corresponding digital elevation maps (DEM). Training is performed using three variations of this urban data set and aims at detecting building contours through a novel super-resolved heteroassociative mapping. Another innovation of this approach is the design of a modified custom loss layer named Top-N. In this variation, the mean square error (MSE) between the reconstructed output image and the provided ground truth (GT) image of building contours is computed on the 2N image pixels with highest values . Assuming that most of the N contour pixels of the GT image are also in the top 2N pixels of the re-construction, this modification balances the two pixel categories and improves the generalization behavior of the CNN model. It is shown in the experiments, that the Top-N cost function offers performance gains in comparison to standard MSE. Further improvement in generalization ability of the network is achieved by using dropout.The second sub-system is a super-resolution deep convolutional network, which performs an enhanced-input associative mapping between input low-resolution and high-resolution images. This network has been trained with low-resolution elevation data and the corresponding high-resolution optical urban photographs. Such a resolution discrepancy between optical aerial/satellite images and elevation data is often the case in real world applications. More specifically, low-resolution elevation data augmented by high-resolution optical aerial photographs are used with the aim of augmenting the resolution of the elevation data. This is a unique super-resolution problem where it was found that many of -the proposed general-image SR propositions do not perform as well. The network aptly named building super resolution CNN (BSRCNN) is trained using patches extracted from the aforementioned data. Results show that in comparison with a classic bicubic upscale of the elevation data the proposed implementation offers important improvement as attested by a modified PSNR and SSIM metric. In comparison, other proposed general-image SR methods performed poorer than a standard bicubic up-scaler.Finally, the relaxation system fuses together all these multisource data sources comprising of pre-processed optical data, elevation data, foliage masks, shadow masks and other pre-processed data in an attempt to assign confidence values to each pixel belonging to a building contour. Confidence is augmented or decremented iteratively until the MSE error fails below a specified threshold or a maximum number of iterations have been executed. The confidence matrix can then be used to extract the true building contours via thresholding.

Le phénomène de la Précarité Energétique (PE) ne cesse de gagner du terrain, il concerne 20,4% de la population françaises en 2016 contre 18,4% en 2013 (ONPE ; 2016). Ce constat oblige une attention particulière à la question de la performance énergétique des bâtiments.Malheureusement, cette performance est limitée aujourd’hui à la seule question de l’efficacité énergétique, or, il faudra prendre en compte la sobriété énergétique et le comportement des acteurs pour atteindre une vraie performance. Pour ce faire, nous avons mobilisé le parc de la copropriété qui est un parc avec une gouvernance complexe où la prise de décision commune pour voter les travaux est difficile à atteindre. Afin de répondre à ce problème complexe, une démarche délibérative établissant un dialogue entre les acteurs de la copropriété (syndic, conseil syndical, copropriétaires résidents et copropriétaires bailleurs), mais aussi avec les autres parties prenantes portant des connaissances sur la rénovation énergétique des bâtiments (architecte, thermicien, chargés d’opérations ...etc.), permettrait d’une part d’accompagner les copropriétaires à comprendre et à s’approprier les défis de la performance énergétique et d’une autre part de faire une évaluation pour comprendre comment les copropriétaires prennent des décisions sur des choix de performance énergétique dans une situation de précarité énergétique.Le défi de cette démarche est qu’à travers ce dialogue et évaluation, les copropriétés puissent faire des décisions performantes pour les travaux de rénovation même quand plusieurs de ces copropriétaires sont en situation de PE. / The phenomenon of fuel poverty continues to grow, it is affecting 20.4% of the French population in 2016 against 18.4% in 2013 (ONPE, 2016). This situation requires special attention to the issue of energy performance in buildings energy.Unfortunately, this performance, nowadays is only limited to the question of energy efficiency, it will be necessary to consider the energy saving and the behavior of the actors to achieve true performance. In order to do this, we have mobilized the park of the condominium housing which is a park with complex governance where common decision-making votes for needed tasks is difficult to achieve. To address this complex problem, a deliberative approach establishing a dialogue between co-owners (the managing office, union council, resident co-owners and lessor co-owners) but also with other stakeholders with knowledge on buildings energy renovation (Architect, heat engineer, operation manager…etc.), would enable the co-owners to understand and appropriate the challenges of energy performance and to make an assessment to understand how the co-owners take decisions on energy performance choices in a situation of fuel poverty.The challenge of this approach is that through this dialogue and evaluation, the condominiums housing can make efficient decisions for renovations even when several of its co-owners are in a fuel poverty situation.