SMART GRID AI TWIN TO SUPPORT THE ENERGY TRANSITION IN LUXEMBOURG
Full-fledged AI Twin for the Luxembourg electricity grid.
Creos Luxembourg
Context
The energy transition - driven by the need to integrate renewable energies, like solar panels or wind energy, a growing number of electric vehicles and heat pumps - is putting an increasing pressure on today's electrical grids, especially on low voltage grids. While most people tend to take electric energy for granted, managing its effective distribution is in fact a constant technical challenge. To prepare for the ongoing energy transition, Creos, the largest distribution system operator in Luxembourg, deployed nearly 350.000 smart meters for almost every household across the country over the last years. On top of this new infrastructure deployment, Creos was searching for relevant approaches to analyze efficiently a massive amount of data related to consumption and production of electricity across the low voltage network and - to some extent - envisage the implementation of a Digital Twin of the grid that would enable a unified view of several databases, systems, and functionalities. Eventually, such Digital Twin would allow a better understanding of the behavior of the electricity grid to ultimately enable precise live monitoring, simulations, and predictions capabilities.
Our solution
- Precise live monitoring and improved understanding of the behavior of the grid.
- Leveraging smart meter data by extremely fast data processing of massive datasets.
- Fast and accurate predictions per point of distribution, taking contextual data into consideration.
- Accurate grid calculations based on actual usage data and advanced predictions.
- Advanced ad-hoc simulations to plan grid intervention in seconds.
- Improvements of data quality that payed off.
- Enhanced efficiency in operating the low- medium- and high voltage grid.
- Getting ready for the energy transition.
In this context, we have developed Kopr in close collaboration with Creos Luxembourg. Kopr is a full-fledged AI Twin of the Luxembourg electricity grid. This digital counterpart of the physical grid and processes can be trained in near real-time – with the ever-increasing amount of available data – to serve as operational decision helper. Kopr aggregates, visualizes, analyzes, and learns data from various systems, e.g., GIS, SAP, metering infrastructures, real-time sensors, and much more. Kopr is built on top of our Greycat technology that allows us to scale to millions of grid elements and to billions of metering measurement points per year.
Benefits / returns
Kopr helps Creos to operate and plan its low-, medium-, and high voltage grids more efficiently. Today, Creos relies on a static, rather worst-case scenario oriented approach for grid calculations. Given a lines' capacity and its potential usage, the available capacity is estimated. However, for an effective intervention planning more accurate estimations are required. There is a need to improve the accuracy of grid calculations by utilizing actual usage data and more appropriate prediction- and simulation algorithms. This need is further reinforced by the ongoing energy transition that puts more and more pressure in terms of production (e.g., solar panels, wind turbines, risk of supply shortage) and consumption (e.g., increased number of heat pumps, electric vehicles) on the grid. To meet this need, Creos deployed the Kopr AI Twin solution. Kopr enabled Creos to manage its electricity grid more efficiently by integrating data from various systems and extremely fast data processing of massive datasets. By aggregating data from various different systems, the Kopr AI Twin not just enables a better understanding of the behavior of the electricity grid and precise live monitoring but also to validate the correctness of data. This helped Creos to significantly improve the data quality in their source systems. Starting with the low voltage grid, Kopr will now also be used for the medium- and high voltage grid. The future roadmap of Kopr foresees to transition from a monitoring and operational decision-helper AI Twin towards a system that also interacts with the grid, e.g., via actively switching relays in smart meters.
Yves Reckinger
People tend to take electric energy for granted. But managing its effective distribution is, in fact, a constant technical challenge, especially when the grid is experiencing more and more turbulences in terms of production (e.g., solar panels, wind turbines) and consumption (e.g., increased number of heat pumps, electric vehicles). Kopr allows us to manage our electricity grid more efficiently by integrating data from various systems. It enables a better understanding of the behavior of the electricity grid and ultimately enables precise live monitoring, simulations, and prediction capabilities. Kopr also helps us to continuously improve our data quality by pinpointing data inaccuracies. We are convinced that Kopr is an important part in the preparation towards the energy transition.