Dena demonstrates energy savings through AI

The German Energy Agency (Dena) has launched a pilot project for AI-based energy optimization in non-residential buildings. This could save energy and generate income.

A pilot project by the German Energy Agency (Dena) with the Fraunhofer Institute IEE in Kassel is demonstrating the interaction of AI energy management, smart metering systems and fully dynamic prices in non-residential buildings. Buildings are responsible for around 35 percent of final energy consumption in Germany. Non-residential buildings in particular make a significant contribution to energy consumption and grid utilization.

As part of the Start-up Energy Transition (Set) Pilots, their contribution to cost-effective optimization of the energy system is being investigated, as Dena announced on 5 February. The specific use case in Kassel is an office location of the Fraunhofer Institute for Energy Economics and Energy System Technology (IEE). Digital solutions for a flexible energy system are to be tested and further developed under real conditions.

Kassel building under test

The IEE site has a charging infrastructure for electric vehicles, a photovoltaic system (PV) and a heat pump. The control system takes into account a dynamic electricity tariff, time-variable grid charges and grid-oriented control in accordance with Section 14a EnWG. Fraunhofer IEE coordinated the project and developed the AI algorithms.

Two other companies also took on these tasks as project partners: Smartplace integrated the controllable systems and implemented the AI agent on the edge component, Teleseo provided the smart meter infrastructure and secure data transmission. The pilot operation showed that the regulatory and technical infrastructure required today for AI-based, market- and grid-friendly control of building consumption works together reliably.

According to the project partners, the specially developed AI agent proved its efficiency by charging electric vehicles at times of low prices and high local PV generation while communicating effectively with the local building management system (BMS).

Advantages of AI-supported control approaches

The SET Pilot therefore demonstrates the potential of AI-supported control approaches in complex energy systems to achieve efficiency, reduce energy costs and achieve grid-friendly behavior. According to Dena, the experience gained in the project can be used for the broader use of AI-supported energy management solutions in office buildings with similar existing systems (PV, heat pump, charging stations).

The evaluation also shows that flexibility from non-residential buildings can be bundled via aggregators or virtual power plants and marketed on the intraday market. This opens up additional revenue opportunities and new roles for building operators in the energy system.

Philipp Richard, Dena Head of Digital Technologies and Start-up Ecosystem, concludes: "One prerequisite is that the regulatory framework provides appropriate incentives." He appeals to the regulators, saying that the ongoing further development of the grid fee system for non-residential buildings with high energy consumption is crucial in order to enable and incentivize flexibility.

Background to the set pilots

The set pilots provide a testing and implementation space aimed at further technological development in the context of the digitalization of the energy transition. The focus is particularly on the skills of start-ups and innovative organizations to drive innovation using smart metering systems in all areas of application in the energy sector.

The participants in the pilot projects benefit from the financing or partial financing of their pilot projects, from the resulting opportunities to test the technological approach and from the expansion of their network.
The findings from the fourth set pilot are also an important basis for the further development of the set pilots. In the next two pilot projects, dynamic grid charges and innovative storage solutions in low voltage will be tested and investigated.

Dena's AI-based energy management set is available online.

Author: Susanne Harmsen