When the wind drops or clouds block the sun, Europe’s electricity system needs other sources of energy that can respond to changing demand. Hydropower can play an important role here. Hydropower plants can increase or lower output as demand shifts, while some can store water to generate electricity later, when it is needed most.
For Alkiviadis Tromaras, a manufacturing and mechanical engineer and researcher at the Centre for Research and Technology Hellas (CERTH) in Greece, making the most of hydropower’s flexibility is not simply a matter of producing as much power as possible. It is about skillfully balancing many different variables.
Operators must balance changing water levels, weather forecasts, equipment maintenance and environmental requirements, including managing water conditions that can affect aquatic life.
Tromaras led a three-year EU-funded research project called Di-Hydro that developed AI and digital technologies to help plants make better operating decisions while taking environmental constraints into account.
“We want to make hydropower plants smarter, more digitised and more flexible,” he said.
Why hydropower matters
Hydropower is already an important part of Europe’s renewable electricity mix. In 2025, it was the EU’s third-largest source of renewable electricity after wind and solar, accounting for 25.9% of renewable electricity generation, according to Eurostat.
In her State of the Union speech on 16 September, European Commission President Ursula von der Leyen called for Europe to “double down on our affordable, homegrown, clean energy” and to “develop storage.” Hydropower is well placed to answer both calls, offering the flexibility and storage needed to integrate variable sources such as wind and solar into the electricity system.
“
We want to make hydropower plants smarter, more digitised and more flexible.
But hydropower faces its own challenges. The water that powers the plants is becoming harder to predict. Climate change is altering rainfall and snowmelt patterns, and increasing the risk of both droughts and floods.
“In general, the amount of water may be the same, but it is distributed in a different way across the year,” said Nicola Crippa, a product manager for hydropower optimisation software at Italian firm MAS Consulting.
This makes accurate forecasting increasingly important. Knowing how much water is likely to arrive – and when – can help operators decide when to generate electricity and when to conserve water for later.
“It allows us to produce energy when it’s more needed, and cover the load demand in the hours of the day when more energy is required,” Crippa said.
Hydropower must also operate within environmental limits. Its effects on wildlife, river systems and water conditions are already addressed by EU environmental legislation, making it important to understand what is happening in and around a plant.
Rehearsing decisions digitally
One way researchers are tackling these challenges is by creating a digital version of a hydropower plant known as a digital twin.
The virtual model can be fed with information from the real plant and used to simulate different situations. Machine learning algorithms can then help identify patterns and predict what might happen under different weather, water flow or operating conditions.
“They offer information to the user – the hydropower operator – and allow them to run different scenarios in a risk-free digital environment,” said Tromaras. “You can leverage the information that every tool is able to produce, so in the end it allows you to take more informed decisions.”
The idea is not to replace the people running the plants. Instead, the digital tools can give operators a better picture of what may lie ahead, allowing them to compare different options before making a decision.
For example, if a large amount of water is expected to arrive, or if the coming months are likely to be unusually dry, operators can modify how they run the plant in response.
According to Tromaras, testing showed that the decision-making platform could produce more efficient operating decisions, while also considering environmental constraints and regulations.
Sensors for water and machinery monitoring
The researchers also developed sensors that can provide a more detailed picture of conditions around a hydropower plant.
Some monitor water quality and environmental conditions, including dissolved oxygen, pH, conductivity, temperature, ammonia, algae and E. coli. Changes in any of these can signal problems for life in the water.
“All of these parameters can affect aquatic organisms, so it can give an ‘early warning’ of what is happening in the reservoir,” said Tromaras.
The researchers combined sensor readings with historical information to develop systems for monitoring water quality upstream and downstream of hydropower plants. The aim is to give operators a clearer picture of environmental conditions and help anticipate changes.
“
It will take a lot of effort to digitise all the systems, but it needs to be done.
Other sensors turn their attention to the machinery itself. Wireless devices can continuously monitor factors such as acoustic emission, vibration, pressure and temperature, helping to identify early signs of wear or damage.
That could allow maintenance to be carried out before a small problem becomes a major failure, potentially reducing downtime and helping plants operate more reliably.
These approaches were tested in different settings, including hydropower facilities in Greece, Italy and Serbia. The sites allowed researchers to explore different challenges, from structural and machinery monitoring to water quality and water flow forecasting.
Making old plants smarter
One of the challenges is that Europe’s hydropower fleet is far from uniform. Plants vary considerably in their age, technology and existing level of digitalisation.
“The digitisation level across Europe is very different,” said Crippa. “It will take a lot of effort to digitise all the systems, but it needs to be done.”
That is why the researchers have focused on tools that can be adapted to existing hydropower plants rather than assuming every facility can start from the same technological base.
Bringing together researchers, technology developers and hydropower operators was also important, said Tromaras, because it allowed the digital tools to be developed and tested against real-world needs.
“Bringing all of these partners together from the different countries was very important because we just have more expertise together,” he said.
The work, which ran until September 2026, has now reached the point where the researchers are looking beyond the initial test sites and seeking further opportunities to put the tools into practice.
The wider goal is not to change what hydropower does, but to help it respond more intelligently to a changing energy system – while keeping an eye on the rivers and ecosystems on which it depends.
As wind and solar take up a growing share of Europe’s electricity mix, that ability to balance power, water and nature could become increasingly important.
Research in this article was funded by the EU’s Horizon Programme. The views of the interviewees don’t necessarily reflect those of the European Commission. If you liked this article, please consider sharing it on social media.
