Heating, ventilation, and air-conditioning (HVAC) equipment faults and operational errors result in comfort issues and waste of energy in buildings. An Automatic Fault Detection and Diagnosis (AFDD) tool could help facility managers fix comfort and energy issues more efficiently, by identifying the most probable root causes. Existing AFDD methods mostly focus on equipment-level fault detection and diagnostics ; almost no attention is given to building level fault diagnosis, considering inter-dependency between equipment through the energy distribution chain. In this work we propose a methodology to automatically derive a Bayesian network from HVAC system topology description such as Haystack. This Bayesian network models and estimates the state of all elements in the system, helping users to identify the most probable root fault. As it is able to ingest evidence from any source (field data, operators, or other models) and is capable of updating its estimates when new evidence is delivered, such a tool could have a great potential to be used interactively on the field. We applied the proposed methodology on simulated and real-world buildings and present in this paper one specific case.

Building operation is responsible for about 30% of total CO2 emissions. In the path toward a more sustainable world, there are several building level solutions that can be considered, including building control, electrification of heating, electrical vehicle charging and addition of renewable and battery. Upon the technical solutions, the energy management will make the best use of system flexibility to achieve goals like maximizing self-consumption or minimizing energy bill in the context of variable energy rates and demand charges.Unfortunately, the building market is very fragmented, and any strategy toward net zero can meet skepticism if not supported by strong facts, especially when the proposed approach combines multiple solutions. To assess strategy impacts on energy, carbon and cost savings, as well as other Key Performance Indicators (KPI) for different types of buildings in different geographies, a first method consists in compiling results for various studies and trying to aggregate them. However, this approach is not robust as there is little luck that the different sources are sufficiently aligned in hypothesis such that their results can be directly compared in a global analysis. So, for the purpose of producing more robust analysis, we propose to use a simulation-based framework that will aggregate all the required baseline models and hypothesis, simulate the proposed energy solutions and systematically recompute all economics and environmental results.The computation part of the framework involved calls to several simulations. At this stage, we are using Energy+ to simulate energy consumption, and the in-house Energy Management Simulation Framework to optimize the size of distributable energy resources. Modeling of electric vehicle charging station has recently been added. It uses its own smart control, upon which the microgrid control is achieved with a Model Predictive Control based approach able to address multi-objective energy management targets.In this paper, we will present the different components of the Energy Management Simulation Framework, the Data sources used and how a single study is defined and run.

Energy management solutions for microgrids typically rely on advanced control/optimization methods that can efficiently tackle a complex set of goals and constraints. Simulation tools can greatly contribute in the development and deployment of such solutions. Unfortunately, most of the existing software are lacking one or more of the following required features: capacity to integrate the real energy management algorithm, open environment for research development and continuous integration, and simple to use graphical user interface for the most repetitive cases. The simulation environment proposed in this contribution is a Matlab/Simulink based framework for the development of district level models and validation of real energy management algorithms (Software-In-the-Loop test bench), that has been completed by post-processing methods and a GUI to produce a design tool. Initiated in an European project, the framework has been successfully industrialized and is now used at different levels (sales teams, engineering teams, research and development), maximizing synergies and agile transfer between repetitive cases and new business opportunities.

Data-driven automatic fault detection and diagnostics (AFDD) have gained a lot of research attention in recent years. Many existing solutions need to learn from the fault operation data to be able to diagnose the faults. However, these data are usually not available in buildings. In this study we present a data-driven AFDD solution for Air Handling Units (AHUs). The solution consists of three levels of fault detection that require different levels of data availability: the first level is daily energy benchmarking; the second level is control performance evaluation; and the third level is data-driven modelling of mechanical systems. The method is applied to two case studies: experimental data from ASHRAE project 1312-RP, and real-life operation data of an office building in France. These tests show that the solution is able to isolate control faults and mechanical faults of individual components, by learning from normal operation data only.

We present the results of a case-study analysis for optimal sizing of a battery energy storage system (BESS), photovoltaic and/or genset (or cogeneration unit) using a recently developed microgrid design tool (MGDT) which integrates advanced energy management algorithm, including MPC (Model Predictive Control) approach embedded into the optimization engine. MPC algorithm is based on the resolution of an optimization problem that uses the variable electric tariff rates (for both energy and demand), the predicted load, and distributed energy resources production profiles to minimize the cost function over a time horizon (typically 24-hours) with respect to optimal energy profile results. This Matlab Simulink based tool was able to produce comparative results to indicate battery-autonomy and how the battery design impacts the cost when the microgrid operates in grid connected mode. The analyzed KPIs were: renewable penetration ratio, yearly cost for utility grid, cash flow on the full project lifetime.

Building Energy Simulation (BES) is typically used in the design phase, but predicted performances are often quite different from the real building performances later measured. In recent years, a number of studies have focused on using BES in the construction and operat ion phases (Sterling, 2015). Use in these phases involv es complex data management, advanced algorithms and targets for non-simulation experts. This paper addr esses innovative calibration based on the identification of zone level Key Parameter Modifiers (Azar, 2015), fault detection methods using Kernel Principal Component Analysis, and the development required to make them usable by building energy managers.

This paper describes the methodology used for selecting the most influential parameters on the energy performance of a building, using limited computing power. Detailed building energy performance ...

In the context of the EU FP7 AMBASSADOR project, a simulation platform has been developed, as a support for the development and the deployment of energy management systems at the district level. Such simulation platform is called District Simulation Platform (DSP), and includes both the models of the physical components from the district, and the models of the energy management algorithms. The DSP is a support for Software In the Loop (SIL) validation: the same algorithm code is first developed and tested on the DSP before being deployed for real-time operation of the district. The DSP can take into account various district configurations through user-defined configuration files. These configurations cover the field tests of the AMBASSADOR project, including the Lavrion experimental site from NTUA close to Athens (Greece), and the INCAS experimental district platform from CEA close to Chambery (France). The last one is detailed in this paper.

In this chapter, a hierarchical model predictive control framework is presented for a network of subsystems that are submitted to general resource sharing constraints. The method is based on a primal decomposition of the centralized open-loop optimization problem over several subsystems. A coordinator is responsible of adjusting the parameters of the problems that are to be solved by each subsystem. A distributed-in-time feature is combined with a bundle method at the coordination layer that enables to enhance the performance and the real-time implementability of the proposed approach. The scheme performance is assessed using a real-life energy coordination problem in a building involving 20 zones that have to share a limited amount of total power.

Poorly functioning and tuned control systems are a frequent source of building underperformance. Simulation can be an excellent method to study building controls, but a number of practical obstacles often interfere. The mapping of control functions from a physical control system to a simulation model is often error prone and contains gross simplifications. A major reason for this is that many simulation tools simply do not support modeling of realistic controls. However, even with a simulator that does allow complex controls, the practical mapping of actual to simulator supported control mechanisms is non-trivial, especially when addressing building or zone level supervisory controls typically embedded in building management systems (BMS) or room controllers.This paper proposes a simulation architecture that will help overcome some of these problems. It attempts to standardize some trivial choices, so that at least these will not lead to unnecessary complications and misunderstandings in this critical and error prone issue. The control concepts are collected from real building controls, and a simulation model that incorporates these is developed to prove the applicability in a whole-building, full year simulator context.

In this paper, a distributed model predictive control is proposed to manage the whole set of actuators (heating/cooling, ventilation, lighting, shading) in a multi-zone building to control comfort parameters (temperature, indoor CO2 level and indoor illuminance). The control process is performed in a distributed fashion and handles variable prices as well as resource limitation in a context of a multi energy source building. To this end, we firstly present a zonal nonlinear model predictive controller which is concerned by zonal decision making -part 1 of the paper-.We then provide a coordination scheme based on a primal decomposition to address the resource allocation problem which occurs in the presence of global constraints on power consumption and/or global shared storage capability in the building -part 2 of the paper-Lamoudi et al. [2012]. We finally provide some simulation results attesting the fast convergence of control algorithm and the benefit of the controller.

In this second part of the paper dedicated to energy management in buildings a Distributed Model Predictive Control strategy is proposed in order to tackle the control problem of a large building submitted to global power limitations and disposing of a storage device (electrical battery). The proposed scheme is based on previously designed Model Predictive Controllers responsible of managing the comfort quantities at the zone level. The proposed framework addresses the case of power specific limitations and dynamically varying prices. Numerical simulations are proposed for a realistic buildings model including 20 zones in order to assess the efficiency and the real time implementability of the proposed framework.

This paper presents a distributed Model Predictive Control framework based on a primal decomposition and a bundle method to control the indoor environmental conditions in a multisource/multizone building. The control aims to minimize the total energy cost under restrictions on global power consumption and local constraints on comfort and saturations on actuators. Moreover, each power source is supposed to have a time varying tarification. The distributed Model Predictive Control algorithm is based on two layers: a zone layer which is responsible of local zone decisions and a coordination layer that handles decisions that go beyond the scope of the zone. Simulation results are finally provided for a three zones building with a local power production and a changing price grid power. A computational study is also provided in order to assess the effectiveness and the real-time implementability of the proposed control method.

In this paper, a methodology for interfacing and assessing a Model Predictive Control strategy in a building simulation tool (SIMBAD) is presented. Firstly, a system identification is performed in order to derive a suitable embedded model for the predictive controller from the simulation tool. Secondly, we assess the performance of this control strategy by introducing uncertainties on forecasted weather conditions and occupancy. Finally, we provide some simulation results in order to analyse the robustness of the controller in presence of uncertainties on forecast.

Within the scope of the HOMES programme, five pilot sites (real buildings) were chosen to study the benefits of active energy efficiency on building energy performance. This article deals with using simulation to assess control functions impact on energy consumption and comfort. Model’s data came from audit report, expert knowledge but also from the use of site monitoring. Main goal for this first step was to compare the actual building performance with a similar building equipped with HOMES control solution.

In this paper, the problem of minimizing energy consumption of a building zone under pre-assigned multi-variable comfort conditions and changing energy rate is addressed. The solution involves the use of a parameterized multi-variable Nonlinear Model Predictive Control (NMPC) that manages the actuation of heating/cooling, ventilation, lighting and blinds devices. Simulations of the resulting closed-loop in winter and summer seasons under varying rate profile are proposed to assess its efficiency. Moreover, a sensitivity analysis is conducted to show how the comfort level assignment impacts the level of energy consumption.