A Secure Energy Future

The power grids are old but the energy sources are new – not an ideal mix. Fraunhofer researchers have developed innovative solutions for keeping the lights on in Germany.

Berlin, January 3, 2026: An arson attack cripples parts of the power grid, leaving 45,000 households and over 2,200 businesses in the southwest of the capital in the dark. It takes four days to restore power. France and Great Britain, January 9, 2026: Storm “Goretti” sweeps across the country overnight with winds of up to 200 kilometers per hour, leaving behind 55,000 households in Great Britain and more than 380,000 in France without power. Spain and Portugal, April 28, 2025: Voltage surges, frequency fluctuations and a complex chain reaction trigger a twelve-hour power outage in Spain and Portugal – the largest disruption to the European power grid in 20 years.

As widely varying as the causes of these power outages may be, they all have one thing in common: They show just how vulnerable our power grid is. For example, the German power grid contains numerous grid bottlenecks – transmission towers that are easy to find and even easier to damage and whose failure would have serious consequences. Fire-resistant enclosures would be a solution but are still not available. “We are in a new situation – no one anticipated attacks like this many years ago,” explains Kerstin Andreae, chair of the executive board of the German Association of Energy and Water Industries. “Physical protection of critical infrastructure must therefore be a top priority on the political agenda.” The German federal cabinet is responding with a draft bill for the KRITIS Umbrella Act, which takes an overall federal and cross-sector approach to protecting critical infrastructure. In the future, every operator must take appropriate measures to respond to the specific risks to its facilities.

This is far from the only challenge facing grid operators. “Every protection measure for a line costs money—we are trapped in a dilemma between economics, cost-effectiveness and system robustness. Expanding the grid infrastructure to include renewable energy sources should be a clear priority,” says Christof Wittwer, Head of Business Unit at the Fraunhofer Institute for Solar Energy Systems ISE. “The entire energy transition currently hinges on our aging power grid.” 

However, renewable energy sources do not yet provide the instantaneous reserve needed to act as a buffer against short-term fluctuations in the power grid. Until now, large coal-fired and nuclear power plants have helped balance out load peaks and stabilize the grid frequency by transferring rotational energy into or out of the grid. However, the new decentralized power plants lack the inertia of the heavy generators and turbines, which is why the necessary instantaneous reserve must be provided by other means. “Expansion of the power grid and the associated changes are highly complex,” says Wittwer. “Where grid operators used to have to intervene in grid operations once or twice a day, roughly 500 interventions are now necessary every day.” This means higher costs for customers.

Artificial intelligence is intended to help stabilize the grid and prevent chain reactions that could lead to power outages. In collaboration with chip manufacturer NVIDIA and grid operator TransnetBW, a team from the Fraunhofer Institute for Energy Economics and Energy System Technology IEE is currently investigating how this might work at the transmission grid level as part of the eKI4DS (Explainable AI for dynamic stability) project. The goal is to use machine learning to enable the AI model to immediately detect critical situations, alert human operators and suggest possible countermeasures. The main challenge here is to feed the AI model with current data on the state of the grid, as the model requires a large number of simulations to cover all possible scenarios. Conventional software quickly reaches its limits here. The solution? An open-source software application based on a complex, large-scale model of the German transmission grid. This is used to train and later run the AI model. The model has already demonstrated its capabilities on smaller test grids. 

AI always needs people to keep the grid stable. However, it is becoming increasingly difficult to find well-trained staff. “The training period for operators in grid control centers is one to two years,” explains Dennis Rösch, Group Manager at the Fraunhofer Institute for Optronics, System Technologies and Image Exploitation IOSB, Advanced System Technology branch AST. An AI-based agent solution that facilitates data exchange between data storage and analysis systems and automates the analysis process is intended to assist grid operators with operational decisions in the future. The agent, a kind of digital assistant, performs the usual analysis steps and provides the result. The operating staff will then make the decision based on this information. A pilot project involving three German grid operators was started in February 2026; within one or two years, the system could already be supporting human colleagues. While this won’t grant immunity to attacks such as the one in Berlin, it could help mitigate or even completely prevent power outages like those in Spain, which are primarily caused by grid fluctuations.

Of course, grid stability is not just a problem in Germany; it represents a challenge for the whole continent. After all, the power grid operates as an interconnected system in Europe. “An exchange of information within Europe is essential. We must have coordination and a shared understanding,” says Diana Strauss-Mincu, group manager at Fraunhofer IEE. This is where the EU project InterSCADA comes in, which aims to strengthen cooperation among European grid operators. The researchers involved in the project are developing an open-source platform for data collection, remote monitoring and power grid control for advanced detection and correction of anomalies throughout Europe. Contributions from the Fraunhofer team include a module for monitoring and estimating synchronous inertia. In addition, Fraunhofer is providing a decision-support module based on a digital twin that is designed to help analyze various disruptions and investigate the results of different stabilization measures. Strauss-Mincu: “InterSCADA should help improve grid stability and reduce the probability of outages in the future.”

The grid itself also offers some possibilities for ensuring stability as further renewable energy capacity is added. Grid-forming inverters offer great promise. They can actively stabilize frequency and voltage, thereby replacing the inertia of conventional power plants, which improves resilience to grid faults. The first grid-forming inverters are already on the market. Starting this year, there is also a market for instantaneous reserves on which this capability can be traded.

To test grid-forming inverters, researchers at Fraunhofer ISE have developed a measurement and analysis procedure as well as a catalog of requirements. In the SUREVIVE project, they are also investigating the degree to which grid-forming inverters stabilize the German power grid. For this purpose, project partner Schoenergie has, for the first time, directly connected a grid-forming battery system to a substation. To minimize the risks of this field test, researchers at Fraunhofer ISE first analyzed the grid-forming characteristics and behavior in their multi-megawatt lab. In the future, battery storage systems with hundreds of megawatts of storage capacity could balance the energy supply as needed while also stabilizing the grid and providing the necessary instantaneous reserve capacity that was previously supplied by conventional power plants.

As part of the joint research project ALene, researchers from the Fraunhofer Institute for Factory Operation and Automation IFF are also working on integrating grid-forming inverters in the grid as a stabilizing measure – in combination with battery systems. The project aims to develop intelligent algorithms and power electronic systems to increase grid resilience. “We want to reduce and compensate negative effects on the grid, improve the quality of supply, reduce the load on operating equipment and make optimal use of the existing grid infrastructure,” explains Christoph Wenge from Fraunhofer IFF.

An innovative online impedance measurement system enables researchers to continuously monitor the state of the grid in real time. The measurement systems immediately transmit their data to the control center, where the algorithms then analyze grid status, determine the optimal setpoint and feed this information back to the inverters so that they can work to stabilize the grid. The grid-forming inverter has already been finished, and the battery system is ready at Fraunhofer IFF, too. These will be merged and combined with the measurement system over the coming months. In the future, ALene is set to help respond more effectively in crisis situations.

How could this system have helped in Berlin? “After the attack, real-time grid status was no longer available,” says Wenge. In particular, real-time information on the extent of grid disruption and the limits of functioning areas was lacking. To a certain extent, this left us blind.” ALene expected to change this in the future. 

Berlin shows that improvement is also needed in communications in the event of a crisis. “The events in Berlin have shown that communication between grid operators, crisis management, disaster preparedness and emergency management, and fire departments can be further improved in complex situations,” says Marc Richter, Head of Department at Fraunhofer IFF. This is where the Integrated Operations Center (IOC) at Fraunhofer IFF comes into play. This is a type of control center that consolidates data on power generation, transmission, storage and consumption into one standardized system. The IOC also serves as a research and testing environment. It examines, for example, how digital assistance systems and tools can help control center staff during hectic emergency situations. The team is developing training scenarios for this purpose. The resulting data is then used to train an AI model. What makes this special is that the AI independently designs the scenarios, prioritizes them and develops solutions.

Ideally, control center staff should never have to face such stressful situations in the first place. But how can the power grid be protected against floods, storms and terrorist attacks? Researchers at the Fraunhofer institutes ISE, EMI, IEE, IEG, and IOSB are following up on this question. They have defined three fault scenarios for this: technical defect, cyberattack and natural disaster. They simulated these scenarios in various configurations on a general digital twin at the distribution grid level. The data for the fault scenarios recorded by the researchers using the digital twin is bundled, analyzed and visualized in the Resilience Monitor. A planning tool enables the identified resilience criteria to be incorporated into the planning of future grid infrastructure in order to prevent power outages as much as possible. In addition, the Resilience Monitor serves as a kind of early warning system for grid operators. 

To increase the resilience of the power grid in major outages, Wittwer and his team also investigated the use of grid-forming inverters in the distribution grid. These could form local island grids, so-called microgrids, in combination with large batteries, maintaining the power supply to affected areas. To this end, the researchers integrated a battery inverter into a digital twin of the grid. This places real test components, in this case the battery inverter, in a virtual power grid. The virtual grid simulates various faults, calculates their real-time effects on frequency and voltage and then applies these results to the real equipment. 

“Couldn’t we take energy from car batteries instead of large battery storage systems to bridge grid outages?” pondered Anton Gorodnichev and Marco Jung from Fraunhofer IEE. “Prerequisite for this is bidirectional charging, where the car’s battery can not only charge but can also feed power back into the grid,” says Jung. The idea of drawing on electric car batteries during dark lulls to balance out grid bottlenecks and to then fully recharge the batteries when there is a surplus available is not entirely new. The concept is known as vehicle-to-grid. 

In the Charging Infrastructure 2.0 project, Fraunhofer IEE researchers have successfully tested this approach in Munich and Hamburg in collaboration with automotive supplier Vitesco Technologies. 

In the CombiPower project, Fraunhofer IEE is developing the innovative Vehicle-to-Grid+ system. This is not primarily intended for long-term grid stabilization but can be used with grid-forming control to generate emergency power during power failures. In this setup, one or more vehicles form their own island grid – independent of the main power grid. A vehicle battery has a storage capacity of 90 kilowatt-hours, while the average household uses roughly 10 to 15 kilowatt-hours a day. A vehicle like this could thus power a household for a week. 

So there is hope for Berlin– and elsewhere.