At the end of February 2026, we presented the final demonstration of the OpenEnergyTwin (OET) project in the GridControl laboratory at OFFIS, together with our project partners Fraunhofer FIT and RWTH Aachen University, as well as associated partners and other interested parties, including manufacturers of grid control technology and grid operators. The screen displayed a human-machine interface (HMI) view that combines the medium-voltage and low-voltage grids on a single surface: grid topology, live measurements, condition assessment results and forecasts – consistently displayed and automatically drawn from the stored grid data. What at first glance appears to be a straightforward map view marks a fundamental change in perspective: the low-voltage level, previously largely a black box for grid operation, becomes transparent and controllable.

Why new approaches are needed

The energy transition is introducing a multitude of new players into the distribution grid: Photovoltaic systems, heat pumps, electric vehicles and storage systems are fundamentally changing the requirements for grid operation. At the same time, the digitalisation of the grid infrastructure is advancing, and sector coupling requires cross-domain operational management. Traditional grid control systems are reaching their structural limits: they are monolithic, proprietary and tied to a few providers. The integration of new functions – such as for forecasts or AI-supported assistance – leads to long and cost-intensive development cycles in such ecosystems. In addition, the cyber security requirements for critical infrastructures are growing, while a widespread shortage of skilled workers is further exacerbated by outdated user interfaces of existing systems.

What is OpenEnergyTwin?

OpenEnergyTwin is not a finished product, but rather an open platform approach for modular open-source control systems that also incorporate the low-voltage level into grid management automation. The core idea: a digital twin virtually maps the physical energy system – in real time and event-driven. Services can be added, exchanged or removed in a modular fashion without changing the platform core.

The architecture follows a three-layer model. The bottom layer translates field protocols into a uniform format. Above this is the platform core with topology service, measurement service and status estimation. The third layer – the service layer – contains applications such as HMI, alerting, data import/export and a forecasting service.

Communication is event-driven via a central Kafka broker in a publish-subscribe pattern. The Common Information Model (CIM), serialised as CIM/JSON-LD, serves as the semantic data model.

Topology data is imported in the Common Grid Model Exchange Standard (CGMES) as CIM/XML, managed in a graph database and made available platform-wide as CIM/JSON-LD. In practice, this enables clearly defined, standardised interfaces, rapid expandability and independence from individual manufacturers.

What the demonstration reveals

The demonstration provided an experience of the entire processing chain: a mosaic-based co-simulation fed data from the medium and low voltage levels – based on a grid topology for the city of Bremerhaven – into the platform via virtualised telecontrol devices and smart meter gateways.

The HMI automatically records the topology on a geographical map and displays measured, estimated and predicted values for a selected piece of equipment side by side. Time series visualise voltage curves and power flows.

A control command for dimming in accordance with Section 14a of the German Energy Industry Act (EnWG) can be issued via the HMI, translated by the protocol adapter and executed in the simulation – the effects are immediately visible. Particularly noteworthy is the ability to exchange different state estimation implementations and other components via configuration without having to make adjustments to other system components.

18 months of targeted research

The project, funded by the Federal Ministry for Economic Affairs and Energy, achieved considerable scientific and practical visibility within a year and a half. In eight bilateral stakeholder workshops and two joint sessions, requirements were identified, feedback on the architecture was obtained and transfer scenarios were identified – including cross-voltage level transparency and predictive maintenance. The results were presented at national and international conferences, including the ETG Congress, IEEE PowerTech, ISGT Europe, DACH+ Energy Informatics and the CIRED 2026 Brussels Workshop. In addition, a formal architecture description according to ISO/IEC/IEEE 42010 and a machine-readable AsyncAPI specification of the platform interfaces were created.

Open to manufacturers, grid operators and research institutions

OpenEnergyTwin is particularly attractive as a test bed for manufacturers of grid control technology, grid operators and research institutions. New processes – whether condition assessment, forecasting or AI-based assistance – can be quickly integrated via the standardised interfaces and evaluated under reproducible conditions. Manufacturers can integrate components as prototypes, test interoperability against the CIM data model and set up demo environments without having to familiarise themselves with proprietary system landscapes. The modular, containerised architecture reduces the integration hurdle and makes results comparable.

Outlook: Release as open source by the end of August 2026

The platform's source code will be freely available at github.com/OpenEnergyTwin by the end of August 2026 at the latest. In addition, a permanently accessible public demo is in preparation; a demonstration video will be published on the OFFIS YouTube channel. We expressly invite manufacturers, grid operators and research institutions to use the platform, test their own use cases and work with us on the next generation of modular control systems.