Demo 1 in Reken, Germany

Key Figures

  • Located in Germany, North-Rhine Westphalia, area of the municipality of Reken
  • Semi-urban area with moderate continental climate conditions
  • MV grid with approx. 100 secondary substations, (7 switching modules, 11 measurement modules)
  • Ratio between maximum load and Decentralized Generation (DG) almost balanced
  • Massive increase in DG expected

Main Deliverables

  • Specification and requirements
  • Report of development for flexible MV-network operations
  • Economical assessment
  • Handbook for implementation of multi-agent system in MV-network


Reken, Germany


Overview German Demo

The high share and still massively increasing amount of distributed generation, predominantly wind and photovoltaic, set new challenges to the DSOs in Germany. In order to provide hosting capacity to integrate these resources huge investments in grid infrastructure are required. Moreover, grid operation and grid observation become more complex since power flows become less predictable. At present in Germany there are hardly any surveillance facilities or grid automation in place in medium voltage networks.  

The GRID4EU German Demo addressed these challenges with the demonstrator built up in the area of “Reken”, located in North Rhine-Westphalia. The considered grid was well selected since it showed already at the beginning of the project a balance between installed generation power and maximum demand. Further increase in renewables to be connected was forecasted. The grid focused on consists of around 100 secondary substations.

Implementation of autonomous multi-agent systems for surveillance and automated control of MV networks is a potential solution to a better utilisation of MV networks.

Due to the expected increasing integration of DER in MV networks as well as in the subordinated LV networks the medium-dated necessity of an advanced network surveillance and a network control in the MV level is foreseeable. Currently there is neither a surveillance of the MV network installed nor remote controllability for switches and transformers in RWE’s MV network.

In the field of a classical enhancement of the networks and a full-value SCADA-system in RWE-Demo an alternative and cost-efficient approach should be pursued. Autonomous working agents will be installed in a distributed way at critical locations in the grid.

The agents themselves communicate amongst each other as well as with sensors which could be e.g. smart meters, to figure out the current state of the MV network. Deriving from all available data the agents make their own decisions, e.g. how to switch or to adjust transformers or DER to operate the MV network in an optimized way. The decisions will not be made independently from other agents but will be negotiated amongst the agents.

The agents give commands to actors like remote controllable switches at disconnection points.

Decisions and commands of the agents will be sent to a control centre which fulfils surveillance purposes and which could also intervene or overrule the agents’ commands. For safety reasons the control centre is able to block particular grid-sections for the agents to adjust. Hence the control centre always knows about the current grid topology.

The added value of the proposed agent system in comparison to a conventional system is their robustness and low need for maintenance. Any central approach needs excessive maintenance in case of system modifications. In comparison to that fully distributed agents detect their environment themselves and draw their conclusions for actions from this environment. This means the agent system adapts itself to changing system conditions. This increases the robustness significantly.

To go further, we invite you to consult the posters of GRID4EU partners involved in the Demo 1:

Main objectives

Demonstrate that autonomous multi-agent systems can become an industrial solution to manage MV networks thus allowing:

  • Integrating an increasing number of DER in the MV network and underlying LV networks
  • Achieving higher reliability, shorter recovery times after grid failures, improved workforce-management 
  • Avoiding unknown overloads
  • Fulfilling the needs of surveillance and remote-control in MV networks