Water behind dams: keep it or leave it?

Researchers from the Western Norway University of Applied Sciences (HVL) and Norwegian University of Science and Technology (NTNU) have developed a method for making better decisions in hydropower production. They believe that this will play a critical role in ensuring the transition to a greener energy system.  

Reza Argandeh (HVL), Hossein Farahmand (NTNU) and their team have been looking at how hydropower producers could make better use of natural resources and work with the market at all times. More specifically, they have developed methods under the IntHydro project which use artificial intelligence to calculate how the producers should regulate how much water they allow into the water reservoirs.

“If hydropower producers were able to make decisions that were only one percent better than they used to be, that would amount to a difference of billions of kroner and help to mitigate the energy crisis,” says Arghandeh.

The future European energy system – based primarily on renewable energy sources – will be much more weather-dependent than the current energy system. The two researchers believe that consumption patterns will also change due to climate change. All this helps to create energy supply uncertainty, which makes decisions much more complicated. They wish to reduce this uncertainty, making it easier to guarantee access to energy.

 Electricity: vital for our society

Arghandeh’s team at HVL, including Mojtaba Yousefi, has found a way of monitoring meteorological data, hydrological data (about how much water is going into the reservoirs) and topographical data (about the shape of the landscape), and interpreting these with the aid of artificial intelligence. Farahmand then uses these AI models in combination with data about the electricity market. Together they have created a model that takes into consideration uncertainty in the market and uncertainty regarding wind and weather.

“Electricity is not just an ordinary product, but something that allows our society to function, almost like oxygen. That’s why it’s incredibly important to create affordable, reliable and sustainable access to electricity,” says Arghandeh.

Accurate and reliable inflow forecasting has always been a challenge in the production of hydropower. The reservoirs start to fill up around Easter (April) in Norway. The snow melts on the peaks, and water from the mountains starts to run into dammed, regulated lakes and rivers all over the country. Between Easter and the return of winter, usually in November, the producers release water at regular intervals in order to supply the market. But they must also make sure that they leave enough water in the reservoirs to last through the winter season, when the water supply dries up.

Complicated mathematics

But when and how much water should they release? This is a complex optimization problem. The answer depends on numerous factors, such as weather conditions, topography, volume of precipitation, winter weather profile, the electricity market, and the political situation in Europe. Since the energy system in Norway is connected to Europe, it makes it even more difficult to make proper decisions – something that was amply demonstrated by the incidents following the war in Ukraine. When Russia limited selling gas to Europe as a result of the European support for Ukraine, the effect both on access to energy and the price of energy was dramatic.

When there are so many factors to consider simultaneously, hydropower scheduling becomes incredibly complicated. In Norwegian hydropower industry, producers have until now used classical mathematical models to calculate how they should regulate the water level in the reservoirs.

“We wish to work with the current methods to improve them. With artificial intelligence, we can speed up the calculations and also obtain more precise answers,” says Arghandeh.

Big ambitions for offshore wind

Although much of what is involved in hydropower production is unpredictable, this type of renewable energy has huge benefits – one of the greatest being that you can store the water and produce the electricity when there is demand for it. This makes it a highly flexible and a clean type of energy.

“However, there are some limitations that we have to take into consideration. Hydropower reservoirs are limited in size, and water inflow is irregular,” says Farahmand from NTNU.

By using artificial intelligence to help to reduce decision-making uncertainty in the industry, Farahmand believes that hydropower will make even more of a contribution to securing our energy supply than it does now. Norway is working to reduce its use of fossil energy and become carbon neutral by 2050. An important step towards this is the introduction of offshore wind. There are plans to develop 30 gigawatts of offshore wind capacity on the Norwegian Continental Shelf by 2040. This equals the capacity we have in hydropower today.

“The new, green forms of energy such as wind and solar are far more unpredictable than hydropower. The wind does not always blow, nor does the sun always shine,” says Farahmand.

“Since wind conditions change so quickly, we will need models that can calculate hour by hour how wind production affects the big picture in terms of our access to energy.”

Working alongside energy producers

The researchers on the project have been working with energy producers Lyse Produksjon AS and Østfold Energi. They are now trying out the newly developed artificial intelligence tools in practice. In the long run, they hope that the big players in the hydropower industry in Norway will collaborate with them and adopt their AI methods.

“In the transition to a society which will be more heavily reliant on sun and wind as energy sources, sustainable hydropower supply will become even more important to us. We are very fortunate in this country, as we possess tremendous opportunities to create clean energy,” says Farahmand.

“Water and wind are gifts of nature to Norway. We must make every effort to manage it in the best possible way for us and for our future generations,” says Arghandeh.