No description available.

Cette thèse fournit de nouvelles stratégies de contrôle pouvant s'attaquer aux phénomènes hétérogènes et non-linéaires qui décrivent la régulation de la température dans les bâtiments afin d'obtenir un compromis entre le confort et l'efficacité énergétique. Nous nous intéressons donc au contrôle robuste de systèmes coopératifs avec perturbations bornées. Nous résolvons d'abord ce problème grâce à la notion d'intervalle invariant contrôlé robuste, décrivant un ensemble dans lequel l'état peut être maintenu quelle que soit la valeur des perturbations. Une seconde approche décrit des méthodes symboliques pour la synthèse d'un contrôleur discret sur une abstraction finie du système, réalisant une spécification de sûreté associée à l'optimisation des performances. Nous présentons d'abord une méthode symbolique centralisée utilisant les dynamiques du système correspondant au modèle physique. Pour résoudre ses limitations en termes de passage à l'échelle, nous considérons une approche compositionnelle où les méthodes symboliques d'abstraction et de synthèse sont appliquées à des descriptions partielles du système, sous des obligations de type assume-guarantee supposant que la sûreté est satisfaite pour tous les états non-contrôlés. Dans la dernière partie, les contrôleurs présentés sont combinés et évalués dans le cadre d'une régulation de température pour un bâtiment expérimental équipé de la solution UnderFloor Air Distribution. / This thesis provides new control strategies that deal with the heterogeneous and nonlinear dynamics describing the temperature regulation in buildings to obtain a tradeoff between comfort and energy efficiency. We thus focus on the robust control of cooperative systems with bounded disturbances. We first solve this problem with the notion of robust controlled invariant interval, which describes a set where the state can be maintained for any value of the disturbances. A second approach provides dedicated symbolic methods to synthesize a discrete controller on a finite abstraction of the system, realizing safety specifications combined with a performance optimization. We first present a centralized symbolic method using the system dynamics provided by the physical model. To address its limitation in terms of scalability, a compositional approach is considered, where the symbolic abstraction and synthesis methods are applied to partial descriptions of the system under the assume-guarantee obligation that the safety specification is realized for all uncontrolled states. In the final part, the proposed controllers are combined and evaluated on the temperature regulation for an experimental building equipped with UnderFloor Air Distribution.

Dans le présent travail de thèse, nous souhaitons approfondir l’étude des systèmes à commutation pour des problèmes aux dérivées partielles en explorant de nouvelles pistes d’investigation, incluant notamment la question de la synthèse de contrôle garanti par décomposition de l’espace des états, la synthèse de contrôle nécessitant la réduction de modèle, le contrôle des différentes sources d’erreur sur des quantités d’intérêt, et la mesure des incertitudes sur les états et les paramètres du modèle. Nous envisageons l’utilisation de méthodes de calcul ensemblistes associées à des méthodes de réduction de modèle, ainsi que l’utilisation d’observateurs d’état pour l’estimation en ligne du système. / In this thesis, we focus on switched control systems described by partial differential equations, and investigate the issues of guaranteed control of such systems using state-space decomposition methods. The use of state-space decomposition methods requires model order reduction, control of the different sources of error for quantities of interest, and measure of uncertainties on the states and parameters of the system. We are considering using set-based computation methods, in association with model order reduction techniques, along with the use of state-observers for on-line estimation of the system.

Conceptual design is an early stage in the design process, in which functional requirements of a design problem are transformed into solution concepts for satisfying the requirements. It is regarded as a crucial step in design, because decisions made in this stage will strongly affect all the subsequent stages of the design process. Research evidence suggests that inspiration is useful for exploration and discovery of new solution spaces, and exploration of a wide variety of concepts increases the chances of developing more novel, and hence more creative solutions. There are various approaches to providing inspiration, e.g., creativity techniques such as trigger word technique, biomimetics such as Idea-Inspire, and computational synthesis approaches such as compositional synthesis. Computational synthesis tools are used for automated generation of concepts, which can be offered to the designer as triggers for inspiring ideation. The advantage of using solutions from computational synthesis as triggers are the following: the solutions can be produced in a relatively unbiased manner, allowing a variety of directions to be explored, and the solutions are exhaustive within the constraints of the databases or rules used, allowing a multitude of possibilities to be offered. However, computational synthesis has been traditionally used for automating solution generation, rather than creating triggers for designers’ ideation. Notwithstanding their potential for inspiring ideation, current computational synthesis approaches rarely focused on this task. One exception is FuncSION, a compositional synthesis tool, which can automatically synthesize solution concepts for mechanical devices, where a set of input and output characteristics i.e. functional requirements are provided by the user and the computer generates solutions by combining building blocks from a library to satisfy the requirements; these solutions are then used as stimuli for ideation by designers. The focus of this thesis is on evaluating and improving the effectiveness of computational synthesis in triggering ideation during conceptual design, in terms of improving the fluency and variety of the concept space produced. FuncSION has been used as the example synthesis approach on which the work has been focused. In order to evaluate the effectiveness of FuncSION in terms of fluency and variety, a method for assessing variety of a concept space is proposed, and a tool for supporting the assessment process has been developed. However, compositional synthesis research has always assumed that the building blocks are given, and has confined its focus on the process of combining the building blocks. It has not been investigated as to how such building blocks can be automatically identified. If new building blocks can be automatically identified, the resulting change in the library of building blocks would have a substantial effect on the outcomes of compositional synthesis, i.e. the triggers that can be offered to the designers for ideation, with a resulting effect on the concepts generated by the designers. Therefore, in this thesis, an automated method for building blocks synthesis has been proposed, and has been implemented as a computational tool.