Intelligent AI monitoring protects Germany’s power grid from cyber-attacks

The topic of cybercrime has become even more important in recent years. Many companies have fallen victim to so-called ransomware attacks. Data is encrypted and only released again after payment of a large sum of money. Experts have therefore been recommending more investments in cybersecurity for some time. However, this does not only apply to private companies, but also to government institutions. The focus is mainly on vital infrastructure. For example, last year an analysis concluded that many of the waterworks in Germany are vulnerable to digital attacks. The electricity grid is also generally considered to be risky. The Federal Ministry of Economy and Climate Protection, therefore, funded a research project of the Fraunhofer Institute that explicitly dealt with the safety of power lines. The result of the work has now been presented: an AI-based monitoring system should detect cyber-attacks at an early stage before damage occurs.

Power Grid
Power Grid

Network control rooms are an attractive target for hackers

The experts identified the so-called network control rooms as the gateway. Numerous data and information are collected here, from which actions are subsequently derived that ensure the safety of the electricity grid. For example, if a tree falls on a power line, this is recognized in the network control rooms. The experts there then ensure that the power supply is guaranteed as well as possible. The same goes for potentially dangerous frequency fluctuations. The network control rooms, therefore, play a decisive role in protecting the electricity grids. This is exactly what makes them an attractive target for hacker attacks. For example, if it is possible to manipulate the incoming data, this can lead to incorrect switching actions. In the worst-case scenario, there would be massive blackouts. Such a scenario does not come completely out of the blue: in 2016, after such an attack, the power went out in hundreds of Ukrainian cities.

The AI ​​must detect deviations in time

In Germany, the electricity grid is organized in a much more decentralized way with about 800 distribution system operators. Hacker attacks can also wreak havoc here on a regional level. The Fraunhofer researchers have therefore developed an intelligent digital surveillance system based on artificial intelligence. This monitors the incoming data and the communication between the electricity grid and the control system. From this, in turn, to a certain extent, the normal state is derived. Based on this, anomalies can then be identified and analyzed as quickly as possible. The idea behind it: Cyber ​​attacks are often not recognized until they are already doing damage. But then countermeasures are difficult to implement. On the other hand, if the attacks are detected early, they can be repelled without affecting the end customer. The researchers are now looking for network operators interested in integrating intelligent monitoring software into their systems.