iBECOME project started in June 2020 and will run until November 2023 for a total of 42 months. The first year of the project has just ended and a lot of work has been performed in this period from all the project partners.
More specifically the highlights of our first year are listed below.
IEQ Virtual Sensors
We have developed computational methods to predict the Indoor Environmental Quality (IEQ) within a room, by using hybrid models that combines physics-based simulations and real data processing using machine learning techniques. The models are able to predict thermal comfort, illuminance, glare probability and pollutants concentration within a room with, accuracy ranging between 60-95%. In comparison, the PMV method predicts thermal comfort with accuracy ranging between 30-40%.
Automation of Building Energy Model Calibration
We have improved our process to produce accurate digital twins of buildings, essential for the vBMS to perform predictions of the future conditions. Up until now, the calibration methodology was a desktop study that includes an intensive manual iterative process, followed on desktop application, and relies mostly on the experience of the energy modeller. The Apache Engine, which is the core simulation engine of the IES Virtual Environment can now operate on the cloud and enables the automation of many of the manual steps required in the calibration process. This way, we achieved better accuracy and similar computational requirements with Reduced Order Models. After testing the new method on a case study building we found that the duration of calibration process was reduced by a remarkable 98%, while the accuracy was improved by 27% in average.
Co-Simulation
Further to the cloudification of the Apache Engine, we managed to enable the co-simulation capability. The physics-based simulation engine, which, is now capable to run in parallel with a machine learning model and to interact with it in every timestep of the simulation. This way we combine real and virtual sensors towards the most accurate predictions of the future building conditions. This will be used to inform a number of control functions for optimising building operations at each timestep.
Energy modelling and insights of case study buildings
To support all the developments above, we collected data from case study buildings from iBECOME partners portfolio and contacts. We managed to calibrate building energy models of various building types and generate very useful datasets in the development of algorithms required by our data analysis toolbox. We have prepared calibrated building energy models for the Helios Building, CEA offices in France, an office and a warehouse owned by Schneider Electric, a manufacturing site by Prior Medical, and finally a Clean Room by Seagate. All buildings were studied in detail and reports with insights on how to save energy have been prepared and shared with the building owners. We will continue testing both methods in more case study buildings, including our demo sites.
Work in Progress and Next Steps
The work will continue in the next six months to finalise the software and hardware architecture of the vBMS, and the development of innovative computational methods to implement automated control of building energy systems in order to maximise comfort and minimise energy use. Furthermore, innovative solutions for Fault Detection, Predictive Maintenance, M&V and Demand Response services will be developed. Additional services will also be established including healthcare management for the elderly, a car sharing app and optimised EV charging. In September 2021 we expect to start the work to integrate all the above solutions into the iBECOME vBMS software as a service, that will serve both buildings with or without a BMS and with minimum on-site equipment.
In parallel, the work has already started in our retrofit demonstration sites in Italy and France with engagement with occupants, retrofits and audits. Also in progress is the planning and installation of the iBECOME hardware and calibration of building energy models.
In our updated “about page” on our website you can find the concept of the virtual BMS explained in detail, as well as the progress against our objectives. If you wish to read more on our results, please visit our downloads section (https://ibecome-project.eu/downloads/) to get a copy of our public reports.
Quite a busy period, a lot of exciting results are coming up, so stay tuned and get in touch if you are interested in receiving our newsletter or follow the project as a stakeholder.