Methodological tools for energy performance guarantee – OMEGA

Methods in the OMEGA project make an extensive use of the simulation model of the building, throughout the whole energy performance contracting process.
In the design phase, an a priori model of the building is used for:

• Conducting a sensitivity analysis which enables to qualify the influence of building envelope and systems parameters, as well as outdoor (weather) and indoor (occupancy, temperature set-point) on energy consumption
• Propagate the uncertainties associated with these parameters and conditions in order to provide a probability distribution of energy consumptions, based on design hypotheses,
• Optimize the measurement and verification plan in terms of its precision/cost ratio

In the operation phase, the building simulation model can be used with building parameters and indoor/outdoor conditions coming from in-situ measurements. OMEGA methods in this phase are mostly data science methods:
• Analysis of variance on the residual between measurements and simulations
• Occupancy-related measurements interpretation using machine learning methods
• Fault detection and diagnosis using Bayesian networks or formal analysis
• Data visualization

The work being done in OMEGA enables:

• to compute a performance target and to evaluate the risk that it could not be met,
• to help with the selection of contractual adjustment variables,
• to help with the design of the measurement and verification plan,
• to guide the reception stage by identifying the most probable root causes of gaps between measurements and simulations,
• to estimate a priori the variability of occupancy,
• to understand a posteriori how the building is actually occupied and operated,
• to make occupants aware of their habits and suggest alternative, more energy-preserving habits,
• to help detect and diagnose faults and performance drifts,
• to visualize live data in ways that make relevant information available to each type of stakeholder.

This work also enabled each participating team to create or ameliorate software tools. The startup company KOCLIKO has created the AMAPOLA tool which implements the sensitivity and uncertainty analysis methods designed at ARMINES during the project. Also, methods using sensitivity and uncertainty analysis for target performance computation have been implemented in the workflows and tools for energy performance contracting at Engie Axima.
All teams have sent and presented communications in journals and conferences, mostly at the international level. These communications address the two main research topics of the OMEGA project: sensitivity and uncertainty analysis on one side, data analysis for occupancy detection and fault detection & diagnosis on the other side.

The OMEGA project is an applied research project coordinated by CSTB. It also involves CEREMA and Engie Axima, as well as the research laboratories I2M, LARIS, G-SCOP and ARMINES. It started in January 2014 and finished 48 months after. It received a € 934 253 funding from ANR, for a total cost of about € 2.5 M

.

French environmental guidelines ("Grenelle de l'environnement") have defined ambitious energy consumption and greenhouse gases emission reduction goals, imposing a short-term generalization of high energetic performance new and existing buildings. This deep mutation has already been started by switching from a best efforts obligation at the building conception step to an obligation to achieve results along the building lifecycle. High expectations are then put on energy performance guarantee methods: the whole building sector must be secured by providing ways for the final user to limit the overall yearly consumption of a building in normal use conditions.

In this context, the OMEGA research project focuses on actual energy performance guarantee for new buildings or high refurbished buildings to achieve a very low consumption. It aims to bring innovative methodological tools to seize, anticipate and manage the risks associated with an energy performance guarantee contract. The project more precisely focuses on designing solutions to accompany the whole performance guarantee process from conception to operation.

To accompany the conception phase, methodological tools are proposed, through sensitivity (determination of Sobol indices) and uncertainty analysis methods (Monte Carlo methods and chaos polynomials), to determine the most influent parameters and define an energy performance guarantee target for which a commitment can be made and a confidence level on this target can be estimated. The project will also work on finding a good compromise between precision and cost for the definition of a measurement and verification protocol in order to monitor actual performance. The most influent parameters determined by the sensitivity analysis will also be used to enrich simulation models with experimental data in order to help monitor operating performance. In the context of this work, occupant models will be developed in order to integrate statistically relevant predictions of occupant behavior.

To accompany the operation phase, monitoring real building performance to assess the conformity to the contract and maintain it over time is crucial. The OMEGA project aims to achieve it through default detection and diagnosis methods (through model-based and Bayesian data-based methods) able to distinguish technical defaults from consumption drifts due to occupant behavior, and visualization tools enabling to efficiently monitor building performance. This work will be supported by the development of virtual sensors based on the fusion of physical sensor measurements and able to estimate occupant presence and activity, which will also allow to design graphical user interfaces showing visual insights into the building usage.

Field experiments will be used to evaluate the whole methodological tools developed during the project.