Grey Box Models

Grey Box Models

Innovation Type: Software

What is this innovation about?

Grey box models have been developed to perform model predictive control to optimize the performance of SPEN neighbourhoods. These models will be based on existing research on using grey box models performed by other project partners.

The syn.ikia grey box models will consider physical characteristics of buildings and energy systems, weather forecasts, and monitoring data based on occupant behaviour and actual weather data from the demonstration sites.

What kind of challenges it will solve if deployed widely

While current solutions typically focus on a single component or a single unit, such as a house or an apartment, within syn.ikia, the Grey Box models of the different units are Interconnected and interdependent models. In this way, we enable a SPEN-level optimisation.

The following users will find this innovation useful:

This innovation will probably be interesting for both universities and for AI-based ESCOs. The learnings and developments of such models (this innovation) could become a game changer in controlling SPENs.

How has syn.ikia project impacted the development of this solution/ process/ tool?

Working on many demonstration cases and buildings in syn.ikia and the requirement for developing grey-box models from white-box models have enabled us to develop a step-by-step procedure for identifying different grey-box models from the collected data. This procedure also classifies the models and selects the best model using statistical tests. The results are published in (Tohidi, et al., 2022). The procedure helps provide an automated method for finding grey-box models from the data collected from white-box models.

How can this innovation be used beyond syn.ikia project?

At research level, the development of Neighbourhood Scale Grey Box Models (GBN) can lead to further developments with optimization beyond the classical building level borders, e.g. Including EV charging and discharging and the use of stationary batteries. Also, it provides a framework for automated grey-box modelling from the data.
At a commercial level, spin-off companies of the universities could bring GBN to the market.

Contact person:
Davide Cali, DTU