Search

Heating, ventilation, and air conditioning (HVAC) systems consume nearly 40% of the total energy consumption in developed countries. Traditional techniques such as rule based control (RBC) fail to control these systems in an optimal way. Model predictive control (MPC) has been widely explored in literature as well but it doesn't represent a practical solution due to the complexity of buildings' dynamics that it relies on. Recently, deep reinforcement learning (DRL) has shown great success in the domain of optimal control such as robotics and gaming. In this paper, we develop two model-free DRL approaches to optimize the energy consumption of an office while maintaining thermal comfort and good indoor air quality through controlling the radiator and the opening/closing of a window and a door existing in the office. The two DRL approaches belong to deep-Q network (DQN): the first approach represents a DQN agent with no knowledge of the environment and the second approach represents a DQN agent with initial knowledge of the environment: A hybrid approach DQN+RBC. The goal of having external knowledge in DQN agent is to boost convergence by exploiting the RBC rules. We evaluate the performance of these two approaches against an RBC approach through simulations using a physical model of the office's dynamics. Experiments show that the two DRL approaches succeeded to maintain better thermal comfort and better indoor air quality compared with RBC approach while consuming nearly the same energy. In addition, experiments demonstrate that the DQN with knowledge outperforms the DQN with no knowledge in the beginning and converges faster to the optimal value.

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.

Similar to other engineering systems, buildings are also complex systems. The whole building system is highly vulnerable to different kinds of faults, failures and human misbehavior that could affect the building performance. The available fault diagnosis tools and methodologies particularly rely on rule or pure model-based approaches. It is assumed that the model could be applied to any situation without taking into account actual contexts. Setting a complete model with all contextual representations is a challenging job. On the other hand, unreliable instrumentation and non-formalization of the expert's knowledge are also challenging ones. The objective of this paper is to highlight the challenges that lie in this problem domain, as well as to provide a strategy about how to solve them.A real case study has been investigated to exemplify the new trends in buildings: a classroom in the University of Southern Denmark.

The recent trends in designing sustainable power systems emphasise the importance of self-consumption (SC) both at individual and community level. This new paradigm changes the way in which we design photovoltaic facilities for residential houses and for various municipality services as well. In this context, the paper aims to formulate several optimisation problems using criteria such as self-consumption, self-sufficiency (SS) and net present value (NPV) as objectives to provide an optimal photovoltaic (PV) plant size for a singular power system - a subway station. By using this multi-objective approach, the work emphasises how each criteria impacts the profitability and value of the overall investment, involving possible shareholders in the design process by choosing a desired solution from the Pareto-efficient set of configurations. Moreover, a global optimal solution is provided by formulating an optimisation problem through a single-objective Mixed Integer Linear Programming (MILP) approach involving an equivalent metric, the net-energy exchanged with the grid (NEEG). The proposed methodology is validated in a case study on the power system of a subway station in Bucharest (Romania), thus identifying a configuration that focuses on self-consumption maximisation and a solution that reduces the yearly energy bill of the respective power system by 25%.

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.

According to experts, the average temperature of the planet has increased at an unprecedented and alarmingly high rate over the last fifty (50) years. Carbon emissions have been found to be a major catalyst for climate change and the energy sector one of the highest emitters globally. Thus any significant reduction in energy related emissions would have a significant impact on global carbon emissions and consequently global warming. UN-Habitat estimates that approximately 56% of energy produced in most African nations is consumed in buildings. There is a need for energy efficiency and possibly conservation in buildings since they represent the single largest consumer of energy on the continent. Net Zero Energy Buildings (NZEBs), a possible solution for reducing the energy footprint of buildings, represents the evolution of buildings in the near future. The Zero energy concept has a major impact on the design and construction of future buildings. This paper focuses on the review and development of existing Load Match Indicators for zero energy buildings. Four indicators are provided and discussed (i.e. self-consumption, self-production, loss of load probability, and coverage rate indicators). For the purpose of this paper, Predis-MHI (a platform of G2ELab) was used as a case study. Data was collected from the platform’s living lab and was used in the calculation and evaluation of these indicators. The results indicate the relevance of each indicator in evaluating the energetic performance of a building and also highlight the practical difficulties faced in evaluating the platform.

Building system is highly vulnerable to different kinds of faults and human misbehaviors. Energy efficiency and user comfort are directly targeted due to abnormalities in building operation. The available fault diagnosis tools and methodologies particularly rely on rules or pure model-based approaches. It is assumed that model or rule-based test could be applied to any situation without taking into account actual testing contexts. Contextual tests with validity domain could reduce a lot of the design of detection tests. The main objective of this paper is to consider fault validity when validate the test model considering the non-modeled events such as occupancy, weather conditions, door and window openings and the integration of the knowledge of the expert on the state of the system. The concept of heterogeneous tests is combined with test validity to generate fault diagnoses. A combination of rules, range and model-based tests known as heterogeneous tests are proposed to reduce the modeling complexity. Calculation of logical diagnoses coming from artificial intelligence provides a global explanation consistent with the test result. An application example shows the efficiency of the proposed technique: an office setting at Grenoble Institute of Technology.

The research presented in this paper is a first investigation of the application of online unsupervised learning techniques in the case of smart buildings. An online learning approach is proposed to estimate the number of people (within a range) in a room based on data collected from standard sensors. This estimated range is changing and depends on the highest occupancy faced in a training dataset for instance. Occupant behavior is modeled using an online learning approach. The resulting algorithm makes use of recorded sensor data for motion detection, power consumption, and door position as well as acoustic pressure from a microphone. This paper presents an online algorithm for a mixture model-based clustering which has been applied to estimate the occupancy in an office context, with an excellent estimation accuracy.

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

International audience; A general approach is proposed to determine the common sensors that shall be used to estimate and classify the approximate number of people (within a range) in a room. The range is dynamic and depends on the maximum occupancy met in a training data set for instance. Means to estimate occupancy include motion detection, power consumption, CO 2 concentration sensors, microphone or door/window positions. The proposed approach is inspired by machine learning. It starts by determining the most useful measurements in calculating information gains. Then, estimation algorithms are proposed: they rely on decision tree learning algorithms because these yield decision rules readable by humans, which correspond to nested if-then-else rules, where thresholds can be adjusted depending on the living areas considered. In addition, the decision tree depth is limited in order to simplify the analysis of the tree rules. Finally, an economic analysis is carried out to evaluate the cost and the most relevant sensor sets, with cost and accuracy comparison for the estimation of occupancy. C45 and random forest algorithms have been applied to an office setting, with average estimation error of 0.19-0.18. Over-fitting issues and best sensor sets are discussed.

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.

This paper proposes a new general approach based on Bayesian networks to model the human behaviour. This approach represents human behaviour withprobabilistic cause-effect relations based not only on previous works, but also with conditional probabilities coming either from expert knowledge or deduced from observations. The approach has been used in the co-simulation of building physics and human behaviour in order to assess the CO 2 concentration in an office., IBPC 2015 Turin , Jun 2015, Turin, Italy. 2015

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.

http://www.sciencedirect.com/science/article/pii/S0378778812004987; International audience; Inhabitant's decisions and actions have a strong impact on the energy consumption and are an important factor in reducing energy consumption and in modelling future energy trends. Energy simulations that take into account inhabitants' behaviour are benchmarked at office buildings using controlled activity profiles and predefined scenarios. In this paper we have proposed a co-simulation environment for energy smart homes that takes into account inhabitants' dynamic and social behaviour. Based on this kind of complex behaviour, the setpoints for different controllers are adjusted in the physical simulator. In this platform, human behaviour is modelled using the Brahms environment and the thermal model and controllers for different appliances are modelled as a physical simulator. The thermal model computes the temperature decrease/increase in a room based on the contextual information resulting from the behaviour simulator. This information is then given to the controller to act upon.

International audience; A general approach is proposed to determine the occupancy in a room using sensor data and knowledge coming respectively from observation and questioning are determined. Means to estimate occupancy include motion detections, power consumption and and acoustic pressure rewarded by a microphone. The proposed approach is inspired from machine learning. It starts by determining the most useful measurements in calculating information gains. Then, a non supervised estimation algorithm is proposed: it relies on Bayesian Network algorithms to model a human behaviour with probabilistic cause-effect relations based on knowledge and questioning. In addition, knowledge has been extracted from supervised learning algorithm. Bayesian Network (BN) based approach has been applied to an office setting, with an average estimation error of 0.09 and an accuracy of 90%. This approach avoids the usage of a camera to determine the actual occupancy required for supervised learning.

International audience; As energy demand increased and production means diversified, conventional approaches of looking into distri- bution grids need to evolve. The Smart Grid paradigm introduces new possibilities of real-time market sensing and interaction models between producers and consumers. In particular, by understanding the types of con- sumers and their potential willingness to adapt their energy demand with price incentives, innovative pricing strategies in the Smart Grid are expected to lead to better production management, profit maximization and end consumers satisfaction levels. In this work we propose a novel framework and a simulation scenario of a global energy network with heterogeneous types of producers and consumers from which different types of behaviors and interactions can be studied.

This paper presents methods for prediction of energy consumption of different appliances in homes. The aim is to predict the next day electricity consumption for some services in homes. Historical data for a set of homes in France was used. Two basic predictors are tested and a stochastic based predictor is proposed. The performance of the predictors is studied and it shows that the proposed predictor gives better results than other approaches. Two processings are proposed to improve the performance of the predictor, segmentation and aggregation of data. Application results are provided.

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.

Two algorithms of the SPH Lagrangian numerical method, the first weakly compressible, the second truly incompressible, are presented and applied to two free-surface three-dimensional flows. The first (schematic) case consists of a water column collapsing in a rectangular tank with a central rectangular obstacle, and allows the comparison and validation of both algorithms. It appears that the incompressible method is superior to predict the total strength experienced by the obstacle, while the weakly compressible method shows weaknesses under this criterion. The second application case, very close to an industrial study, represents a “ski-jump” spillway connecting the reservoir of a river dam to a valley with complex bottom shape. The global flow pattern is compared to laboratory observations from a physical model, leading to satisfactory conclusions which prove SPH has the potential to be a promising method for the design of complex waterworks.

The display of space filling data is still a challenge for the community of visualization. Direct volume rendering (DVR) is one of the most important techniques developed to achieve direct perception of such volumetric data. It is based on semitransparent representations, where the data are accumulated in a depth-dependent order. However, it produces images that may be difficult to understand, and thus several techniques have been proposed so as to improve its effectiveness, using for instance lighting models or simpler representations (e.g., maximum intensity projection). In this article, we present three perceptual studies that examine how DVR meets its goals, in either static or dynamic context. We show that a static representation is highly ambiguous, even in simple cases, but this can be counterbalanced by use of dynamic cues (i.e., motion parallax) provided that the rendering parameters are correctly tuned. In addition, perspective projections are demonstrated to provide relevant information to disambiguate depth perception in dynamic displays.

International audience; In the recent years, surveys and studies have established the importance of occupant's behaviour on energy consumption in buildings. Therefore, inclusion of in-habitants' behaviours is compulsory for the assessment of building energy management system's (BEMS) strategies, which highly depends on human behaviour. The purpose of modelling the inhabitants behaviour is to see how their choices and control of household appliances can impact the energy consumption. In this paper, a co-simulation approach is presented where the inhabitants' behaviours are co-simulated with the SIMBAD-MOZART thermal model of a reference house and BEMS. The realization of all the different kinds of inhabitant behaviours into energy co-simulations will help to improve the smart grid technology and hence provide inhabitants with better services to save energy and cost while maintaining their comfort levels.

International audience; A general approach is proposed to estimate the number of occupants in a zone using different kinds of measurements such as motion detection, power consumption or CO2 concentration. The proposed approach is inspired from machine learning. It starts by determining among different measurements those that are the most useful by calculating the information gains. Then, an estimation algorithm is proposed. It relies on a C4.5 learning algorithm that yields human readable decision trees using measurements to estimate the number of occupants. It has been applied to an office setting.

A new approach for the design of parity relations for linear dynamic systems with additive and multiplicative uncertainties is presented. Instead of cancelling uncertainties following the example of the so-called robust approaches, uncertain parity relations take uncertainties into account as bounded variables. The method is based on the analysis of zonotopes representing the uncertainties. It leads both to Boolean detection results and to an indicator representing the distance to the opposite decision.

This paper presents a diagnostic methodology relying on a set-membership approach for fault detection and on a causal model for fault isolation. Set-membership methods are a promising approach to fault detection because they take into account a priori knowledge of model uncertainties and measurement errors. Every uncertain model parameter and/or measurement is represented by a bounded variable. In this paper, detection consists of verifying the membership of measurements to an interval. First order discrete time models are used and their output is explicitly computed with interval arithmetic. Fault isolation relies on a causal analysis and the exoneration principle, which allows focusing the consistency tests on simple local models. The isolation strategy consists of two steps: performing minimal tests found with the causal graph and determining on line additional relevant tests that reduce the final diagnosis. An application for a nuclear process is used in order to illustrate the method's efficiency.

This paper presents a fault isolation method, which is able to handle large scale distributed industrial plants. The method relies on a logical approach, which has been improved in order to soundly take into account doubts in the decisions provided by detection tests. Formal logic of consistency based diagnosis has been extended into a fuzzy logical. These techniques have been implemented within the EC project named MAGIC and applied on hydraulic looper.

After some remarks on terminology, this paper introduces a general method for fault diagnosis in complex dynamic systems, which takes advantage of the results on analytical redundancy methods from the Automatic Control community and on logical reasoning from the Artificial Intelligence community. The proposed method tackles both the problem of diagnosing complex dynamic systems with models of normal and abnormal behavior including differential equations, and the problem of providing logically sound diagnosis. Moreover, it is shown how diagnoses can be sorted.

The variety of diagnostic methods proves that none can pretend to be much better than the others. A significant improvement of industrial applications can only be achieved when FDI problems are dealt with and solved in a framework of integrated use of different FDI methods. This paper presents a study about the possible cooperation between detection methods. Two different isolation methods, which analyze symptoms provided by observer-based and signal-based detection algorithms are compared using a two water tank system: one method is based on signature tables and the other one is based on a logical approach. Strategic aspects are also stressed.

International audience

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.

This paper focuses on a fault detection method taking model uncertainties into account. Every uncertain parameter is represented by a bounded variable. The proposed strategy for fault detection and diagnosis consists in modeling the uncertainties, in generating output envelopes and then in testing their coherency with measurements. An application of the method was carried out on a nuclear fuel reprocessing plant simulator. A fault scenario will be presented in order to enlighten the methodology

In this paper, an extended model of induction machines bearing short-circuits on a single stator phase is proposed. It is shown that a model based on five differential equations whose state vector contains only stator variables expressed in the stator Park base is more relevant to represent such failures. Moreover a simpler model, valid for small short-circuits, is obtained by linearization. It is further extended in order to take into account resistor disturbances. Finally, a structural analysis is developed to point out; that the stator currents are sufficient to generate residuals for discriminating failures under abnonnal operation.

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.

International audience; NUMA nodes are potentially powerful but taking benefit of their capabilities is challenging due to their architec- ture (multiple computing cores, advanced memory hierarchy). They are nonetheless one of the key components to enable processing the ever growing amount of data produced by scientific simulations. In this paper we study the parallelization of patterns commonly used in VTK algorithms and propose a new multi- threaded plugin for VTK that eases the development of parallel multi-core VTK filters. We specifically focus on task-based approaches and show that with a limited code refactoring effort we can take advantage of NUMA node capabilities. We experiment our patterns on a transform filter, base isosurface extraction filter and a min/max tree accelerated isosurface extraction. We support 3 programming environments, OpenMP, Intel TBB and X-KAAPI, and propose different algorithmic refinements according to the capabilities of the target environment. Results show that we can speed execution up to 30 times on a 48-core machine.

International audience; Energy waste due to inhabitants' behaviour in residential buildings has emerged as a potential research area due to the increasing worldwide population and growing energy needs. However, existing approaches for simulating energy consumption are mainly limited to office buildings and are based on static profiles. In this paper we propose a 4-step co-simulation methodology to assess how inhabitants' interactions with household appliances affect energy consumption. The approach is validated using a case study showing how human activities influence the energy consumption patterns of a refrigerator. The fridge was specifically chosen because it is a high energy-consuming appliance that is strongly affected by inhabitants' behaviours. In addition, modelling the fridge is nontrivial, and in choosing this appliance we show that it is possible to apply the approach to less complex appliances. A co-simulation approach is adopted with the fridge being physically modelled in Matlab and with human behaviour being modelled in the Brahms language and simulation environment. The consumption distribution from the simulated scenario is compared with the actual distribution (using data from a consumption database), to find optimum values of tuning parameters with less than 10% variation. This methodology enables us to simulate how human behaviours affect energy appliance consumption.

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.

Energy consumption is a major issue nowadays. The importance of forecasting energy consumption from end-user to power system operator becomes more obvious than ever. The consumption in the residential sector represents a significant percentage in the total electricity demand in Europe and all over the world and it is expected to grow. So, the prediction of energy consumption becomes a key component in the management (e.g. power flow) of the electrical grid. This paper presents different methods for prediction of energy consumption of electrical appliances used in dwellings. A stochastic approach is used since forecasting the consumption for a single appliance is more difficult that predicting the overall consumption. Different basic predictors are presented and a stochastic predictor is proposed and tested according to a prediction precision criterion. The enhancement of forecast precision is done by segmentation and aggregation of data. Several experiments are conducted for different appliances in the house and the results are discussed.

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.