Ensuring the security of renewable energy in Europe from above

Scientists are using artificial intelligence and satellite technologies to improve the reliability of the continent’s renewable energy infrastructure.

In May this year, solar panels and wind turbines in Greece had to be briefly turned off because the national power grid could not absorb the amount of energy being produced.

For electronics engineer Effie Makri, this highlights one of the main challenges in switching to renewable energy sources.

“As renewable energy sources become more widely used, the closure of renewable energy infrastructure will become more common,” Makri said.

Europe is working to massively increase its use of renewable energy. Under the EU’s REPowerEU plan, 45% of energy is to come from renewable sources by 2030.

One of the problems with renewable energy sources, however, is that they are intermittent. There are times when the sun doesn’t shine or the wind doesn’t blow. Or, as in Greece, there are times when the energy produced is too much for the energy infrastructure to handle—overloading the system.

Forecasting energy production and demand

Makri leads a research project called RESPONDENT, which received EU funding to harness the combined power of AI, machine learning and Earth observation (EO) data from European satellites to improve energy supply and demand forecasting and management. The project, which is managed by the EU Space Programme Agency (EUSPA) on behalf of the European Commission, started in 2022 and will run until April 2025.

Being able to predict in advance how much energy will be produced and consumed could prevent future outages like the one in Greece.

Makri’s team of researchers from Athens-based Future Intelligence, a Greek company, has teamed up with energy and communications specialists in Greece, Ireland and Spain to explore ways to use cutting-edge technologies to address the challenge of providing Europe with energy supplies that are both secure and sustainable.

The approach developed by the RESPONDENT team is based on data from the Copernicus programme, the EO component of the EU space programme and Galileo services. Makri says it could help Europe become more energy independent, as the models will use European satellites.

EO adds value

The work being carried out is funded by EUSPA, which promotes the use of European EO data in practical value-added services such as those proposed by RESPONDENT. This is a fast-growing market that is expected to grow from €3.4 billion in 2023 to almost €6 billion in 2033.

Combining that data with new AI and machine learning capabilities allows them to build a model that can account for a wide range of factors that affect energy supply, including changing weather patterns and different user profiles based on three distinct types of users: residential, industrial and commercial.

“Each of these types of users has their own consumption patterns,” Makri said.

This information can be used to better distribute electricity across the grid. For example, on a cold day, it would be better to direct more electricity to homes that have a greater need for heating.

However, the relationship between distributed energy and generated power is important – effective distribution may look different on a sunny day than on a cloudy day.

Pilot demonstrations of the proposed solutions are scheduled to take place this year and in 2025 at a photovoltaic park in Athens, Greece, and at a distribution system operator in Barcelona, ​​Spain.

Extreme weather forecast

During the research, Makri and her team realized that extreme weather events could also affect the results of their calculations. Although it was not initially planned, they decided to include this factor in their model.

In this case, they join the concerns of Francesco Parisio, a software engineer from Italy who led a two-year EU-funded research project at Berlin-based start-up LiveEO, which specialises in using EO data to deliver actionable insights to businesses, policymakers and managers of critical infrastructure, including energy suppliers.

“Extreme weather events are becoming more frequent and intense, and medium- and long-term forecasts predict that this trend will worsen,” he said. “This is putting a huge strain on our critical infrastructure.”

In 2022, LiveEO was among 74 companies awarded a European Innovation Council Accelerator grant to develop a real-time monitoring service for infrastructure networks using a combination of satellite data, artificial intelligence and machine learning algorithms.

Their research project, called EOinTime, successfully concluded in March of this year, resulting in an advanced monitoring service that can quickly assess infrastructure damage after a storm and even predict areas of potential weakening in the future.

Seeing through the clouds

The service uses a special type of satellite data called synthetic aperture radar (SAR), which can gather detailed information about a surface, such as structure and moisture. One of the main advantages of SAR imaging is that, unlike optical technology, it can “see” through darkness, clouds and rain, detecting changes that might otherwise be difficult to spot.

Typically, a human would have to manually assess the integrity of energy infrastructure, either on foot or in the air using a helicopter or plane. Parisio says satellites can provide better results.

“We were able to see damage that the helicopter couldn’t,” he said. In addition to quickly analyzing storm damage, they can also look at how vegetation might affect infrastructure, such as detecting tree health issues that could threaten power lines in the future. To validate their models, they sent experts to assess trees identified by the system as problematic and found that their predictions were correct.

The system can also have a very wide scope of operation – EOinTime already monitors over 100,000 km of infrastructure for vegetation threats in all European countries for the German multinational energy company E.ON.

By conducting pilots on most continents, in countries such as Germany, Australia, the US and Indonesia, LiveEO was able to reduce the response time to detect an issue and inform the customer from days to hours.

The next steps will be to further reduce this phenomenon in cooperation with large industry partners such as Deutsche Bahn and E.ON.

“Any downtime is costly, dangerous and has a major impact on society,” Parisio said. “When there’s a blackout, every minute counts.”