Search

Machine learning and data mining techniques have been widely used recently in several smart buildings applications. This is mainly due to the huge amount of data generated continuously by the smart sensors and meters deployed in new buildings. These data can be used to extract important knowledge about the building human part, for instance. Examples include the number of occupants and their activities which may provide crucial clues to deliver innovative end-user services to empower the building occupants by putting them in the loop of energy usage efficiency and supporting them to achieve their objectives by pointing out the impact of their actions. Estimating the number of occupants and recognizing their activities are also important inputs to develop advanced energy management systems (EMMS). Motivated by the various smart buildings and smart homes applications, the goal of this chapter is to overview several machine learning algorithms and provide a comprehensive review of such methods in activity recognition. Moreover, a case study which main goal is to estimate occupancy is detailed. Presenting this study is motivated by the fact that it could be extended and improved further for activity recognition.

Future renewable energy communities will reshape the paradigm in which we design and control efficient power systems at the district level. In this manner, the focus will be fundamentally shifted towards sustainable related concepts such as self-consumption, self-sufficiency and net energy exchanged with the grid. In this context, the paper presents a novel approach for optimally designing and controlling the photovoltaic plant and energy storage systems for a metro station in order to increase collective self-consumption and self-sufficiency at the district level. The methodology considers a community of several households connected to a subway station and focuses on the interaction between energy sources and consumers. Furthermore, the optimal solution is determined by using a Mixed Integer Linear Programming Approach, and the impact of different configurations on the overall district benefit is investigated by using several simulation scenarios. The work presents a detailed case study to underline the benefits and flexibility offered by the energy storage system in comparison with a scenario involving only a photovoltaic plant.

A general approach is proposed to determine occupant behavior (occupancy and activity) in offices and residential buildings in order to use these estimates for improved energy management. Occupant behavior is modelled with a Bayesian network in an unsupervised manner. This algorithm makes use of domain knowledge gathered via questionnaires and recorded sensor data for motion detection, power, and hot water consumption as well as indoor CO2 concentration. Different case studies have been investigated with diversity according to their context (available sensors, occupancy or activity feedback, complexity of the environment, etc.). Furthermore, experiments integrating occupancy estimation and hot water production control show that energy efficiency can be increased by roughly 5% over known optimal control techniques and more than 25% over rule-based control while maintaining the same occupant comfort. ispartof: Journal of Sensors vol:2019 status: published

WOS:000509819200044; International audience; Data quality assesment is a key component for many real applications, since it can drive better modelling. In this work a methodology to asses data quality (Qscore) is proposed and discussed. The validation of Qscore is performed via an interactive learning experiment related to occupancy estimation. Interactive learning has been shown to be crucial to consider and integrate occupant behavior in smart buildings. Indeed, valuable feedback and information can be collected from the occupants by involving them and by improving their consciousness about energy management systems. Users should feel involved to keep developing highly energy-efficient buildings. To reach this goal, occupants should be aware of the building features to feel more in control. This paper proposes a framework to interact with occupants to estimate building occupancy. This framework is based on an enhanced supervised learning approach that involves interaction with occupants, when necessary, to keep collecting training data. The training data consist of the measurements (i.e. features) collected from common sensors, for instance, motion detection, power consumption, and CO2 concentration, and the label (i.e. number of occupants) provided by the occupants during interactions. The considered learning machine in our experiments is the Multi-layer Perceptron regressor (MLP), although other approaches could be easily integrated within the proposed framework. In order to avoid useless interaction with users a new concept is introduced, called spread rate, to measure the quality of the data to decide if an interaction with the user is necessary or not. Extensive simulations have shown the merits of the proposed approach. (C) 2019 Elsevier B.V. All rights reserved.

International audience; In the field of building energy efficiency, researchers generally focus on building performance and how to enhance it. The objective of this work is to empower the building occupants by putting them in the loop of efficient energy use, supporting them to achieve their objectives by pointing out how far their actions are from an optimal set of actions. Different levels of explanation are investigated. Indicators measuring the distance to optimality are, firstly, proposed. An algorithm that generates deeper explanations is then presented to determine how changing some actions impacts comfort. The paper emphasizes the importance of explanations with a real case study. It identifies the type and level of explanations needed for different occupants. The concept of replay is presented. An occupant can replay his past actions and learn from them.

Power management in homes and offices requires appliance usage prediction when the future user requests are not available. The randomness and uncertainties associated with an appliance usage make the prediction of appliance usage from energy consumption data a non-trivial task. A general model for prediction at the appliance level is still lacking. This work proposes to improve learning algorithms with expert knowledge and proposes a general model using a knowledge driven approach to forecast if a particular appliance will start during a given hour or not. The approach is both a knowledge driven and data driven one. The overall energy management for a house requires that the prediction is done for the next 24 h in the future. The proposed model is tested over the IRISE data and using different machine learning algorithms. The results for predicting the next hour consumption are presented, but the model works also for predicting the next 24 h.

This paper proposes a formulation of the global energy management problem of dwellings, which consists in a dynamic predictive control system able to generate optimized controls taking into account the model of the concerned dwellings, i.e. homes or offices. It focuses on the adjustment of the electric energy consumption and production in order to maximize energy usage efficiency, which is seen as a compromise between energy cost and overall comfort. To reach this objective, the concept of service is introduced: basically, a service participates to the comfort and may consume energy. The available flexibilities of the services provided by domestic appliances are used to compute anticipative optimal plans for appliance controls based on a mixed integer linear programming (MILP) algorithm. A reactive mechanism based on a list algorithm is added to face unforeseen events. The paper focuses on the computation of the anticipative plans. Different MILP models of common services are proposed. Application examples are given.

This paper depicts a formulation of the global home electricity management problem, which consists in adjusting the electric energy consumption to the cost and availability variations of the power supply. The available flexibilities of the services to inhabitants are used to compute optimal plans for appliances control. Two optimization criteria are addressed: the user's comfort and the economic cost. A global multi-layer solving approach is depicted that includes first the computation of consumption/production coordination plans and then the dynamic matching to the actual consumption/production data. The paper focuses on the computation of the anticipative plan through a MILP formulation of the energy management problem. Application examples are given.

Purpose – The use of energy storage devices helps the consumers to utilize the benefits and flexibilities brought by smart networks. One of the major energy storage solutions is using electric vehicle batteries. The purpose of this paper is to develop an optimal energy management strategy for a consumer connected to the power grid equipped with Vehicle-to-Home (V2H) power supply and renewable power generation unit (PV). Design/methodology/approach – The problem of energy flow management is formulated and solved as an optimization problem using a linear programming model. The total energy cost of the consumer is optimized. The optimal values of decision variables are found using CPLEX solver. Findings – The simulation results demonstrated that if the optimal decisions are made regarding the V2H operation and managing the produced power by solar panels then the total energy payments are significantly reduced. Originality/value – The gap that the proposed model is trying to fill is the holistic determination of an optimal energy procurement portfolio by using various embedded resources in an optimal way. The contributions of this paper are in threefold as: first, the introduction of mobile storage devices with a periodical availability depending on driving schedules; second, offering a new business model for managing the generation of PV modules by considering the possibility of grid injection or self-consumption; third, considering Real Time Pricing in the suggested formulation.

The Demand-Side-Load Management will change the way people behave. Different authors have proposed energy management algorithms for Smart Home that either integrates or not renewable energy. All these researches have the same general objective: minimizing the daily energy cost without affecting the comfort of occupants. This paper deals with the performance analysis of a Global Model Based Anticipative Building Energy Management System (GMBA-BEMS) managing household energy. This GMBA-BEMS is able to optimize a compromise between user comfort and energy cost taking into account occupant expectations and physical constraints like energy price and power limitations. To validate the GMBA-BEMS, the model of a building has been developed in MATLAB/Simulink. This work analyzes GMBA-BEMS application that manages appliances such as heating, washing machine and dishwasher from a grid point of view.

International audience; This paper contributes to the design of intelligent electrical network system. A distributed management and control approach, realized through the Multi Agent Systems (MAS) technology, is proposed for Smart Electrical network, due to the MAS advantages: scalability, openness, reliability and communications efficiency. Each agent is embedded into a principal actor and cooperates and coordinates its actions in order to find acceptable near optimal solution. This system proposes multi-level control architecture composed of two main complementary mechanisms: adjustment and anticipative. This paper details the anticipation level that computes the best future set-points according to energy needs and production capacity. These set points are then adjusted in real time by the adjustment mechanism. The platform JADE is used in order to simulate the behavior of the different network actors.

This paper describes an approach aiming to automatically transform a model depicting a high level physical behavior into two different optimized building energy management application models. Basically, each application requires specific formalism and information therefore it is necessary to rewrite an initial physical behavior model to obtain application models. Up to now, these two application models should be rewritten manually, which is representing an significant work and possibly a source of error. In order to automatize this process, the MDE approach seems to be an appropriate solution. This paper presents core specifications of transformation of hinge model into application models. To illustrate this approach, transformations into both an acausal anticipative model based on a mixed integer linear programming problem and a nonlinear causal model for fast simulated annealing optimization are shown. These models are used for energy management of a smart building platform named “Monitoring and Habitat Intelligent” PREDIS in Grenoble, France.

The Demand-Side-Load Management applied to household is a control strategy which will change the way we consume and produce energy. Several works give different forms of management algorithms of Smart Home integrated or not renewable energy and especially the photovoltaic. All these researches have the same general objective: minimize the daily energy consumption level to reduce the electricity bill without harming the comfort of residents. This paper deals with a Building Energy Management (BEM) System controlling household energy capable both to satisfy the maximum available electrical energy constraint and maximize user comfort criteria. It is composed of three layers. This household energy control system is based on simple models. To validate them, a virtual building modeled more finely using MATLAB/Simulink is used. This work presents the application of optimal control to a thermal zone of a virtual building. The objective is to control energy “Smart Homes” consumption through predicted energy price.

Data centers energy efficiency has become an issue given the rising of energy prices and the increase of Web applications, which are hosted, mainly in data centers. In order to understand the opportunities for improving data center energy efficiency, it is necessary to determinate the real energy requirements for one computer application. This article will examine the power chain starting from the facility level and ending at the processor level. In this paper a data center energy model is proffered along with two important indicators that address energy efficiency: the first one is addressing Facility Efficiency and the second one is addressing IT Efficiency. The energy model is taking into account a global overview of the data center (power distribution units, cooling system and the IT system). The aim of the research is to find effective solutions to make data center reduce power consumption while keeping the desired quality of service or service level agreement.

The global dwelling energy management problem can be formalized as an optimization problem of energy consumption/ production. An optimal solution for the home energy management problem is usually solved by centralized solvers. The solver gets the totality of the thermal model of the dwelling but also each appliance composing the system. Nevertheless, this centralized resolution has some limits due to some particular appliances. For example: the appliances with a non-sharable model because of the manufacturer, the appliances that need some precisions that cannot be included in their standard representation used by the solver, the appliances which require specific solvers and the appliances that are managed by user-defined specific heuristic rules. This work proposes to combine the centralized solving approach for energy management problem in dwellings with a multi-agent solving system. The multi agent system provides the possibility of integrating specific models in the global solving of the problem. The proposed system is a mixed centralized/decentralized approach for the solving of global energy management problem. In this paper we present the system and the results of the perfamance tests realized. This work is part of the ANR project ReActivHome.

At the most fundamental level, smart buildings deliver useful building services. They aim to give occupants a good level of comfort (e.g., thermal comfort, air quality, and more) at low cost and low environmental impact. Most approaches and solutions to this problem promote automated and autonomous systems that take full control of the decisions. Although these systems relieve users from setting the temperature in offices and homes, the level of thermal comfort chosen by the technology may not meet the individuals’ requirements. Automating control is an obvious solution, but it keeps the human out of the loop. And occupant activities play a significant role in building energy consumption. A gap has been created between EMSs (Energy Management Systems) and occupants but cooperation is necessary. Neither the occupant nor the EMS can solve the problem alone: the occupant knows his/her intentions and goals; the EMS knows the physics that can help the user to fulfill his/her requirements. To make this possible requires not only developing intelligent energy systems but also needs the involvement of occupants in the loop together with the EMS and the building environment. Rather than “doing instead-of” (automation), we promote a doing together approach (cooperation) where the inhabitant (i.e., user) is involved in the final decisions and actions. However, decisions and actions require that users understand the EMS functioning and rationale. The embedded knowledge in the EMSs cannot be easily shared with occupants because they are designed and made by engineers with the main objective of optimizing the energy consumption. To assist the occupants to better understand their EMSs, this work will promote the explanations as a tool to enhance the cooperation between the EMSs and occupants. Explanations are a powerful and intuitive tool to transfer knowledge from the EMS to the end user. The explanation module needs to be able to automatically extract the knowledge from the EMSs to allow occupants to understand which criteria the EMS is acting upon when making decisions. Additionally, it should provide occupants with feedback about the impact of their actions and decisions on energy consumption and their comfort criteria. The main objective is to enable the user to make informed choices and decisions. In this chapter, we will investigate two approaches to generate causal explanations from the environment from EMSs and sensors information.

This chapter is an introductory chapter that states the main issues appearing in building energy management based on practical examples. Each issue is going to be deeper investigated in the next chapters. It starts by stating the problem in proposing modeling concepts and a global formalism independent of the type of chosen model for a building system including models related to envelopes, appliances, controls, and human practices. Then the different types of models are discussed together with their respective advantages and drawbacks. It covers the model parameter estimation issues, which are often underestimated in the scientific literature: either the model is assumed to be known, which is usually unrealistic in real life, and it is used to generate datasets, or simple model like rule-based model is used for real applications. Energy management algorithms using complex models are difficult to solve in practice. Different ways for generating optimized energy strategies are presented and discussed and a promising case-based optimized approach is presented: it makes it possible to easily solve real applications. The latter approach can be advantageously combined with input-output model based approaches.

Energy consumption in buildings is affected by various factors including its physical characteristics, the appliances inside, and the outdoor environment, etc. However, inhabitants’ behaviour that determines the global energy consumption must not be forgotten. In most of the previous works and simulation tools, human behaviour is modelled as occupancy profiles. In this chapter, the focus is more on detailed behaviour representation, particularly the cognitive, reactive, and deliberative mechanisms. The inhabitants’ dynamic behaviour is modelled and co-simulated together with the physical aspects of a building and an energy management system. The analysis of different household appliances has revealed that energy consumption patterns are highly associated with inhabitants’ behaviours. Data analysis of inhabitants’ actions and appliances’ consumptions is used to derive a model of inhabitants’ behaviour that impacts the energy consumption. This model represents the cognitive mechanisms that provide causes that motivate the actions, including the communication with other inhabitants. An approach based on multi-agent systems is developed along with a methodology for parameter tuning in the proposed behaviour model. These tools are used to co-simulate, not only the physical characteristics of the building, the reactive behaviour that is sensitive to physical data, and deliberative behaviour of the inhabitants, but also the building energy management system. The energy management system allows the direct adjustment of the building parameters or simply giving advice to the inhabitants. The impact of different types of inhabitants’ behaviours, with and without the inclusion of an energy management system is analysed. This work opens new perspectives not only in the building simulation and in the validation of energy management systems but also in the representation of buildings in the smart grid where signals can be sent to end users advising them to modulate their consumption.

Data centers provide the global community withan indispensable service: nearly unlimited access to almost any kind of information we can imagine. With the increasing demand of information, the high energy costs has raised the issue of power consumption. The main goal in achieving data center energy efficiency can be translated into the ensuring that IT equipment uses, as maximum power as possible from the power entering the data center to compute useful work but at the same time the service level agreement between costumers and managers has to be satisfied. In this paper, different concepts of efficiency that are linked with multiple methods of increasing energy efficiency into a data center, will be presented, as well as different principles of energy management. A comprehensive overview of the currently available energy metrics will be provided, followed by an in-depth analysis of their commonalities and differences. An analyze about the degree of freedom encountered into the process of energy management, will be afforded. This paper presents a global overview of the energy efficiency management process into a data center.

This paper presents an optimal power management mechanism for grid connected photovoltaic (PV) systems with storage. The objective is to help intensive penetration of PV production into the grid by proposing peak shaving service at the lowest cost. The structure of a power supervisor based on an optimal predictive power scheduling algorithm is proposed. Optimization is performed using Dynamic Programming and is compared with a simple ruled-based management. The particularity of this study remains first in the consideration of batteries ageing into the optimization process and second in the “day-ahead” approach of power management. Simulations and real conditions application are carried out over one exemplary day. In simulation, it points out that peak shaving is realized with the minimal cost, but especially that power fluctuations on the grid are reduced which matches with the initial objective of helping PV penetration into the grid. In real conditions, efficiency of the predictive schedule depends on accuracy of the forecasts, which leads to future works about optimal reactive power management.

This paper focuses on Energy Management System (EMS) applied to the residential sector. The EMS architecture is proposed. It is composed on three layers: anticipation, reactive and device (local) layers. However, to validate global control algorithms, a simulator capable of interoperating with energy management systems must be used. This work presents a solution for real-time simulation with a networked system that manages energy flows. It makes it possible to perform fine simulations of the physics, including fast dynamics of electrical appliances imbedded on the buildings. This paper deals with computing the starting times of some services in one hand and in determining set points of others on other hand; the maximal power constraints and predicted events are taken into account. A Satisfaction Problem formulation has been proposed. So, the purpose is to maximize user comfort and minimize energy costs. An application example is presented.

This paper presents a multi-agent approach dedicated to power management in buildings. This approach, defines a multi-layer control mechanism: reactive and anticipative mechanisms. In a first part, the paper presents the power system and its problems. Then, it presents the agents modeling of the proposed system, called MAHAS. In the second part of this paper, we show how a multi-agent system, well adapted to solve problems spatially distributed and opened, can dynamically adapt the consumption of energy to various constraints by exploiting the flexibilities of the services provided by domestic devices (services shifting, energy accumulating).

National audience; This paper contributes to the design of a smart monitoring system. The objective of this paper is to show that by using iPOPO framework, as an embedded application for devices, it is possible to improve the con guration/recon guration of monitoring system. The proposed experimentation platform, called PREDIS/MHI, based on iPOPO framework, is characterised by its openness, its scalability and its capability to manage diversity. In this paper, we show how application based on iPOPO framework inspired from OSGi, well adapted to represent problems spatially distributed and opened, can dynamically be adapted to various contexts of environments. This paper presents an energy management application for the laptops batteries : optimal storage life management of laptops batteries according to power supplied by photovoltaic panels and by the grid; Cet article contribueà la conception d'un système de surveillance intelligent. L'ob-jectif de cet article est de montrer qu'en s'appuyant sur la plate-forme iPOPO inspirée d'OSGi, comme une application embarquée pour des equipements, il est possible d'améliorer la configuration / reconfiguration du système de surveillance. La plate-forme d'expérimentation proposée, appelé PREDIS/MHI, basée sur la plate-forme iPOPO, se caractérise par son ouverture, son extensibilité et sa capacitéà appréhender la diversité. Dans cet article, nous montrons comment l'application basée sur la plate-forme iPOPO, bien adapté pour représenter des problèmes ouverts et distribués, peutêtre adaptée dynamiquementà différents contextes d'environnements. Cet article présente une application de gestion de l'énergie pour les batteries des ordinateurs portables : la gestion de stockage optimale des batteries des ordinateurs portables en fonction de l'énergie fournie par les panneaux photovoltaïques et par le réseau d'électricité.

International audience; The goal of this paper is to propose a real time solution to validate energy smart-home strategies that support Demand Side Management done by direct control. The validation concerns the appliances under control. An Energy Management System globally controls the appliances of a house, satisfying the subscription power limits. A hardware-in-the-loop test bench, based on a real-time simulator called RT_LAB, is proposed to exemplify the validation mean: it is used to assess the performances of an energy management strategy for a thermal zone based on the control of a heater. Physical devices are introduced into the simulated system to check whether they can support the control strategies.

Currently, indoor discomfort in dwellings is one of the crucial issues along with the building energy consumption. Indeed, people spend 60-90% of their lives in buildings. Indoor comfort plays a vital role in occupants health, productivity, and well-being. However, various optimization and rule-based anticipative or predictive building strategies have been proposed to achieve the perceived comfort taking into account the energy consumption. However, in practice, anticipation or plans are far from the reality. Usually, anticipative plans are synchronized with one-hour anticipation period and do not consider the various sources of discrepancies as well as current envelope configurations. Unbeknownst to many, discrepancies from different sources could cause big penalty over cost and comfort. To tackle this issue, building management system needs to be designed as reactive or almost with no planning so that it can respond to all discrepancies re-actively. To address this problem, a multi-scale Anticipative Reactive Diagnosing- Building Management System (ARD-BMS) is proposed in this dissertation. ARD-BMS is an internal management and performs three important actions i.e., Discrepancydetection, Cause isolation, and finally Corrective actions. ARD-BMS follow the short-time resolution i.e., 10-minute to analyze the fault trends and current the building dynamics and take necessary corrective actions to maintain the desired level of comfort. This thesis proposes a fast dynamics simplified reactive model that can be used to estimate the current status of the building. Modern buildings are very sophisticated system with a large number of sensors, controllers, and HVACs. Most of the building facilities are using a scheduled preventive maintenance services derived from periodic operations of the buildings. These preventive actions do not take into account the other inadmissible issues such as unplanned situations, weather prediction failures etc. These unplanned issues could cause unaccountable impacts over occupant’s comfort during the 24-hour operation cycle. Diagnosability of short-term discomfort causes is still a challenging job at whole building operation level. Furthermore, to analyze this situation the thesis proposes a diagnostic methodology for detection and isolation of cause (faults) in buildings. The proposed methodology includes a rule-based HAZOP (Hazard and Operability analysis) and model-based approach. Further, in order to oversee unplanned discomforts, a short-term reactive optimization has been proposed., Actuellement, l'inconfort intérieur dans les bâtiments est l'une des questions cruciales, ainsi que la consommation énergétique du bâtiment. En effet, les gens passent 60 à 90% de leur vie dans les bâtiments. Le confort intérieur est indispensable en termes de bienfaits sur la santé, la productivité et le bien-être des occupants. C'est pourquoi diverses stratégies d'optimisation fondées sur des règles, anticipatives ou prédictives ont été proposées pour atteindre le confort perçu en tenant compte de la consommation d'énergie. Dans la pratique, il existe un écart entre l'anticipation et la réalité. Habituellement, les plans anticipatifs sont synchronisés avec une période d'anticipation d'une heure et ne tiennent pas compte des différentes sources de conflits ainsi que des configurations d'enveloppes des bâtiments possibles. Il peut en résulter des conséquences négatives sur le coût et le confort. Pour résoudre ce problème, le système de gestion du bâtiment doit être conçu aussi comme réactif, de sorte à ce qu'il puisse répondre à toutes sortes de divergences par rapport au plan anticipatif, de manière réactive. Dans ce but, un système multi-échelle d'analyse de diagnostic réactif anticipatif (ARD- BMS) est proposé dans cette thèse. ARD-BMS est une gestion interne qui effectue trois actions importantes, c'est-à-dire la détection de défaut, l'isolement des causes et, enfin, les actions correctives. ARD-BMS permet ainsi une gestion réactive à court terme, à savoir 10 minutes pour analyser les tendances des défauts et la mise à jour de la dynamique du bâtiment et prendre ainsi les mesures correctives nécessaires pour maintenir le niveau de confort désiré. Cette thèse propose donc un modèle réactif à dynamique rapide simplifiée qui peut être utilisé pour estimer l'état actuel du bâtiment. Les bâtiments modernes sont un système très sophistiqué avec un grand nombre de capteurs, de contrôleurs et de CVC. La plupart des installations de construction utilisent des services prévus de maintenance préventive provenant des opérations périodiques des bâtiments. Ces problèmes imprévus puce causer des répercussions inexplicables sur le confort de l'occupant pendant le cycle de fonctionnement de 24 heures. Ces problèmes ne sont pas inadmissibles tels que les situations imprévues, les pannes de prévisions météorologiques. Le diagnostic des causes d'inconfort à court terme est encore un problème difficile au niveau de l'opération de construction intégrale. En outre, pour analyser cette situation, proposez une méthodologie diagnostique pour la détection et l'isolement des causes (fautes) dans les bâtiments. La méthodologie proposée comprend une HAZOP fondée sur les règles (analyse des risques et de l'optimisation) et une approche basée sur un modèle.

As buildings become more energy-efficient, the effect of human activities on their global consumption increases. To limit this impact, a first attempt was made to increase the level of automation in buildings. However, from surveys led in residential and office buildings, Carassus and Sidler conclude that automation does not reduce the gap between predicted and effective consumption. This work reveals that static rules can not be sufficient. In order to reach the goal of low energy consumption in the operation phase of buildings, occupants must then be involved in the energy management process. The proposed work is part of a research project focusing on a system embedding energy services with a high level of interaction to empower users with energy services. The role of these energy services is to allow users to define their own objective and consequently generate information and advice regarding the best strategies to apply on building systems (HVAC systems, sun-shading devices, windows opening,…) and envelope configuration. An analysis of the sociological literature was made in order to implement relevant services for the users, favouring real and durable changes in their behaviours. Some of these services require thermal and aeraulic behavioural models easy to set up and little expert knowledge. This thesis focuses on the development and implementation of a model relevant for energy management end-user services. It involves to deal with the definition of relevant model structures together with parameter estimation methods to tune the parameter values. The model must be simple enough to ease its set up with estimation and prediction capabilities. The limits in terms of complexity and computational time are discussed. The energy management services should be configured by occupants thus it must not require much expert knowledge. Consequently, some sensors are needed to estimate the thermal and aeraulic model. Tests have been performed to limit the needed instrumentation. Relevant model structures are discussed because they should be generic: it should adapt to both mono-zone and multi-zone case studies and it should be based on a minimal number of sensors and require little expert knowledge., Les bâtiments devenant de plus en plus performants énergétiquement, l'impact de l'activité humaine devient prépondérante. Afin de limiter cet impact, les bâtiments ont tout d'abord vu leur niveau d'automatisation augmenter. Toutefois, des études menées dans des bâtiments résidentiels et de bureaux ont mis en évidence que l'automatisation ne réduisait pas l'écart entre la consommation d'énergie prévue en phase de conception la consommation d'énergie réellement constatée. Afin d'atteindre une faible consommation en phase d'exploitation du bâtiment, les occupants doivent être impliqués dans la gestion d'énergie. Le travail proposé fait partie d'un projet de recherche visant à mettre en place un gestionnaire énergétique redonnant le pouvoir aux occupants via des services énergétiques. Ces services permettront à l'utilisateur de définir ses propres objectifs et le conseillera sur les meilleures actions à mettre en place pour les atteindre. Une analyse de la littérature sociologique est menée afin de permettre l'implémentation de services pertinents et favorisant de réels et durables changements de comportement. La mise en place de ces services requiert un modèle thermique et aéraulique du bâtiment faciles à configurer et ne nécessitant aucune donnée experte. Le travail présenté dans cette thèse se concentre sur l'implémentation d'un tel modèle. Cela implique de définir dans un premier temps la structure de modèle adéquate : permettant d'atteindre à la fois les précisions requises à la prédiction à horizon 24 heures tout en ne nécessitant que peu d'informations expertes. Dans un deuxième temps, il s'agit d'étudier les différentes méthodes d'estimation paramétrique répondant au cahier des charges. Des validations du modèle sont faites sur deux cas d'études réel : un bureau et un appartement. Cette thèse décrit la méthodologie nécessaire à la détermination de structures de modèles pertinentes. Ces structures doivent être génériques afin de s'adapter à la fois à des cas d'études mono et multi-zones. De plus, elles doivent se baser sur le moins de capteurs et de connaissance experte possible.

Energy management for building has become a major issue this last decade because of its energy impact. Building energy management reduces energy wastes and enable a better matching between energy needs and grid capabilities. Different types of energy management systems are proposed in scientific literature, most of them with anticipation capacities.The first results do not really consider the complexity issue coming from the number of modeling elements and also coming from the diversity of energy management applications.This thesis proposes elements of solution to the complexity problem. The work started by analyzing the energy management system 'GHomeTech' and its adaptation to the complex building prototype CANOPEA. The issue of composition from elementary models is explored. A solution is proposed; it enables the reusability of elementary models. Aggregation and transformation into mixed integer linear programming optimization models is presented. The resulting tool has been validated on the CANOPEA project.Energy management is not limited to MILP optimization. Different types of applications are also used to provide other services: parametric estimation models to simplify the configuration of energy management systems, simulation for validation and prediction depending on pre-defined scenarios for example. This other dimension of complexity is discussed in a second part of the manuscript. Solutions for automatic rewriting of models are detailed. It relies on symbolic manipulations in different types of processing. Several examples of applications illustrating the automatic generation of models are presented., Du fait de son impact énergétique, la gestion énergétique dans le bâtiment est devenu un enjeu majeur ses dernières années, qu'il s'agisse d'encourager la sobriété énergétique ou de s'adapter aux besoins des réseaux énergétiques. Différents travaux de recherche ont conduit à des gestionnaires énergétiques souvent dotés de capacités d'anticipation. Les premiers résultats, bien qu'encourageants, n'envisagent pas la complexité tant du fait nombre d'éléments que de la diversité des applications de gestion énergétique.Cette thèse propose d'apporter des éléments de solution à la problématique de complexité. Les travaux ont démarré par l'analyse du gestionnaire énergétique GHomeTech et son adaptation au prototype de bâtiment complexe CANOPEA. Les problématiques de composition sont explorées. Une solution favorisant la réutilisabilité d'éléments de modèles, l'agrégation et la transformation vers des modèles d'optimisation de type programmation linéaire en nombres entiers (PLNE), est proposée. L'outil résultant a été validé sur le projet CANOPEA.La gestion énergétique ne se limite pas à l'optimisation PLNE. Différentes natures d'applications permettent d'offrir d'autres services : l'estimation paramétrique de modèles pour simplifier la configuration des gestionnaires énergétiques, la simulation pour la validation et la prédiction fonction de scénarios définis par exemple. Cette autre dimension de la complexité est abordée dans un second volet du manuscrit. Des solutions de réécriture automatique de modèles sont proposées grâce à des manipulations symboliques permettant différents types de transformation. Plusieurs exemples de génération automatique de modèles applicatifs sont présentés.

Energy consumption in buildings is affected by various factors including its physical characteristics, the appliances inside, and the outdoor environment, etc. However, inhabitants’ behaviour that determines the global energy consumption must not be forgotten. In most of the previous works and simulation tools, human behaviour is modelled as occupancy profiles. In this thesis the focus is more on detailed behaviour representation, particularly the cognitive, reactive, and deliberative mechanisms. The inhabitants’ dynamic behaviour is modelled and co-simulated together with the physical aspects of a building and an energy management system. The analysis of different household appliances has revealed that energy consumption patterns are highly associated with inhabitants’ behaviours. Data analysis of inhabitants’ actions and appliances’ consumptions is used to derive a model of inhabitants’ behaviour that impacts the energy consumption. This model represents the cognitive mechanisms that provide causes that motivate the actions, including the communication with other inhabitants. An approach based on multi-agent systems is developed along with a methodology for parameter tuning in the proposed behaviour model. These tools are used to co-simulate, not only the physical characteristics of the building, the reactive behaviour that is sensitive to physical data, and deliberative behaviour of the inhabitants, but also the building energy management system. The energy management system allows the direct adjustment of the building parameters or simply giving advice to the inhabitants. The impact of different types of inhabitants’ behaviours, with and without the inclusion of an energy management system is analyzed. This work opens new perspectives not only in the building simulation and in the validation of energy management systems but also in the representation of buildings in the smart grid where signals can be sent to end users advising them to modulate their consumption., La consommation énergétique dans le secteur bâtiment dépend de diverses facteurs parmi lesquels ses caractéristiques physique, ses équipements, l’environnement extérieur, etc… mais il ne faut pas oublier le comportement des habitants qui est déterminant pour la consommation énergétique globale. Or, la plupart des travaux et outils représentent les occupants par des profils d’occupation. Cette thèse s’intéresse à la représentation plus détaillée du comportement des occupants, en particulier les mécanismes cognitifs, réactifs et délibératifs. Le comportement dynamique des occupants est modélisé et co-simulé avec les aspects physiques et des éventuels systèmes de gestion énergétique. L’analyse de la consommation de différents équipements électroménagers met en évidence que le consommation énergétique est très dépendante des comportements des occupants. L’analyse des consommations et des actions des habitants permet d’élaborer un modèle du comportement des occupants impactant la consommation énergétique. Le modèle représente des mécanismes cognitifs, qui représente les causes qui motivent les actions, incluant des échange avec d’autres acteurs humains. Une approche à base d’agents logiciels a été développée. Outre les aspects techniques, une méthodologie de réglage des paramètres des modèles de comportement est proposée. Ces outils sont utilisés pour réaliser une co-simulation représentant la physique du bâtiment, le comportement réactif, c’est-à-dire sensible aux données physiques, et délibératif des habitants mais aussi un système de gestion énergétique qui peut ajuster directement la configuration du logement ou simplement conseiller ces occupants. L’impact de différents types de comportements, avec et sans gestionnaire énergétique est analysé. Ces travaux ouvrent de nouvelles perspectives dans la simulation bâtiment, dans la validation de gestionnaires énergétiques mais aussi dans la représentation des bâtiments dans les réseaux d’énergie dits intelligents, dans lesquels des signaux peuvent être envoyés aux utilisateurs finaux pour les inviter à moduler leur consommation.

This PhD focuses on the global energy management based on energetical flux models ofbuildings. The objective of the energy management system is to help the inhabitants to manage theirbuilding, by scheduling the consumption/production of the different appliances, taking into accountenergy costs, availability et inhabitants’ preferences. The PhD objective is to propose a resolution approach with several solvers inspired by the multi agent systems.A multi-phasis service representing several phasis has been developped. Each phasis is defined by its own consumption level. This type of service presents a better precision than singlephasis approaches to model some appliances. A new optimization approach has been developped. Itcombines different solvers embedded into software agents. The results is an hybrid approach forthe optimization based on the multi-agent system, using MILP algorithms and meta-heuristics.; Cette thèse concerne à la gestion globale à base de modèle des flux énergétiques dans le bâtiment. L'objectif des systèmes de gestion de l’énergie proposés est d’aider les occupants à gérer leur système bâtiment en planifiant la consommation/production des différents équipements présents en fonction des tarifs, de la disponibilité de l'énergie et des usages de l'occupant. L’objectif de cette thèse est de proposer une approche de résolution multi-solveurs s'appuyant sur le paradigme multi-agent.Pour y parvenir, nous avons développé, dans un premier temps, un service multi-phase, qui représente plusieurs phases flexibles avec des niveaux de consommation différents, offrant un niveau de précision supérieure pour des équipements comme une machine à laver. Dans un second temps, notre travail a consisté à proposer une nouvelle approche d’optimisation combinant différents solveurs embarqués dans des agents logiciels. Le résultat est une approche d’optimisation hybride à base d’agents, s’appuyant sur des algorithmes PLNE et des méta-heuristiques.

This PhD focuses on the global energy management based on energetical flux models ofbuildings. The objective of the energy management system is to help the inhabitants to manage theirbuilding, by scheduling the consumption/production of the different appliances, taking into accountenergy costs, availability et inhabitants’ preferences. The PhD objective is to propose a resolution approach with several solvers inspired by the multi agent systems.A multi-phasis service representing several phasis has been developped. Each phasis is defined by its own consumption level. This type of service presents a better precision than singlephasis approaches to model some appliances. A new optimization approach has been developped. Itcombines different solvers embedded into software agents. The results is an hybrid approach forthe optimization based on the multi-agent system, using MILP algorithms and meta-heuristics., Cette thèse concerne à la gestion globale à base de modèle des flux énergétiques dans le bâtiment. L'objectif des systèmes de gestion de l’énergie proposés est d’aider les occupants à gérer leur système bâtiment en planifiant la consommation/production des différents équipements présents en fonction des tarifs, de la disponibilité de l'énergie et des usages de l'occupant. L’objectif de cette thèse est de proposer une approche de résolution multi-solveurs s'appuyant sur le paradigme multi-agent.Pour y parvenir, nous avons développé, dans un premier temps, un service multi-phase, qui représente plusieurs phases flexibles avec des niveaux de consommation différents, offrant un niveau de précision supérieure pour des équipements comme une machine à laver. Dans un second temps, notre travail a consisté à proposer une nouvelle approche d’optimisation combinant différents solveurs embarqués dans des agents logiciels. Le résultat est une approche d’optimisation hybride à base d’agents, s’appuyant sur des algorithmes PLNE et des méta-heuristiques.

This PhD dissertation concerns the power management in buildings. The problem of power management is modeled as a mixed linear program. The work presented in this thesis aims to take into account the uncertainties in the optimization of energy flow in the buildings. The uncertainties are analyzed and two types of uncertainty are identified: parametric uncertainties concerning the vagueness of the parameters (weather forecasts, energy demand forecast ...) and the occurrence uncertainties that are related to uncontrollable actions of the user. A robust optimization approach based on a formulation presented by Bertsimas and Sim for robust linear programming is proposed to take into account the parametric uncertainties. An optimization procedure in two stages, based on stochastic programming, is proposed to answer the occurrence uncertainties. This procedure allows to take into account the energy consumption that is not driven by the management system. The proposed methods have been illustrated on various examples of dwellings. Different energy pricing are addressed., Le travail présenté dans ce mémoire de thèse concerne la gestion de la consommation et de la production d'énergie électrique dans les bâtiments. Le problème de gestion d'énergie est modélisé sous forme de programme linéaire mixte. Le travail présenté dans ce mémoire propose des outils qui permettent de prendre en compte les incertitudes dans l'optimisation des flux d'énergie dans l'habitat. Dans un premier temps les incertitudes à prendre en compte sont étudiées. Nous distinguons 2 types d'incertitudes : les incertitudes paramétriques qui concernent le caractère imprécis des coefficients du modèle (prévisions météorologiques, paramètres des modèles, demande prévisionnelle d'énergie…) et les incertitudes d'occurrence qui sont liées aux actions directes de l'usager sur sa consommation d'énergie. Une approche d'optimisation robuste s'appuyant sur une formulation présentée par Bertsimas et Sim pour la programmation linéaire robuste est proposée pour prendre en compte les incertitudes paramétriques. Une procédure d'optimisation en deux étapes, basée sur la programmation stochastique, est proposée pour anticiper les possibilités de démarrage des services pilotés par l'usager. Cette procédure apporte une réponse aux incertitudes d'occurrence en permettant de prendre en compte les consommations d'énergie qui ne sont pas pilotées par le système d'optimisation. Différents exemples d'appartements sont utilisés pour illustrer la validité des méthodes proposées. Différents scénarios de tarification de l'énergie sont également étudiés.

This PhD dissertation concerns the power management in buildings. The problem of power management is modeled as a mixed linear program. The work presented in this thesis aims to take into account the uncertainties in the optimization of energy flow in the buildings. The uncertainties are analyzed and two types of uncertainty are identified: parametric uncertainties concerning the vagueness of the parameters (weather forecasts, energy demand forecast ...) and the occurrence uncertainties that are related to uncontrollable actions of the user. A robust optimization approach based on a formulation presented by Bertsimas and Sim for robust linear programming is proposed to take into account the parametric uncertainties. An optimization procedure in two stages, based on stochastic programming, is proposed to answer the occurrence uncertainties. This procedure allows to take into account the energy consumption that is not driven by the management system. The proposed methods have been illustrated on various examples of dwellings. Different energy pricing are addressed.; Le travail présenté dans ce mémoire de thèse concerne la gestion de la consommation et de la production d'énergie électrique dans les bâtiments. Le problème de gestion d'énergie est modélisé sous forme de programme linéaire mixte. Le travail présenté dans ce mémoire propose des outils qui permettent de prendre en compte les incertitudes dans l'optimisation des flux d'énergie dans l'habitat. Dans un premier temps les incertitudes à prendre en compte sont étudiées. Nous distinguons 2 types d'incertitudes : les incertitudes paramétriques qui concernent le caractère imprécis des coefficients du modèle (prévisions météorologiques, paramètres des modèles, demande prévisionnelle d'énergie…) et les incertitudes d'occurrence qui sont liées aux actions directes de l'usager sur sa consommation d'énergie. Une approche d'optimisation robuste s'appuyant sur une formulation présentée par Bertsimas et Sim pour la programmation linéaire robuste est proposée pour prendre en compte les incertitudes paramétriques. Une procédure d'optimisation en deux étapes, basée sur la programmation stochastique, est proposée pour anticiper les possibilités de démarrage des services pilotés par l'usager. Cette procédure apporte une réponse aux incertitudes d'occurrence en permettant de prendre en compte les consommations d'énergie qui ne sont pas pilotées par le système d'optimisation. Différents exemples d'appartements sont utilisés pour illustrer la validité des méthodes proposées. Différents scénarios de tarification de l'énergie sont également étudiés.