“Artificial Intelligence plays an increasingly important role for us”, says Huijsman. “It has to, because of the energy transition. It creates a completely new situation and many new challenges. Our work was already complex, and we often do projects that last up to ten years, but the energy transition makes this even more unpredictable.”
“Until now, the energy supply was regulated based on the demand, but we are moving towards a situation where the demand has to be adapted to the supply. After all, with the closing of old power plants, that supply is no longer sufficient or even continuous; wind and solar energy are not always available. The energy transition also means that the demand for our services has exploded; all those solar parks and wind turbines need to be connected. It’s an interesting time, in other words.”
Manual checks
Artificial intelligence, data analytics and data science can help with asset management and predictive maintenance, according to Huijsman. “Some assets like circuit breakers, separators and transformers do break down sometimes. Many quality checks are still done manually right now, and this takes up a lot of time from people we desperately need to expand the grid. We are looking at how we can use smart technology to gain insight into the condition of these assets so that we can predict the probability of failures and carry out more targeted maintenance. Right now, we are still making a distinction between good and bad assets on an annual basis. We want to go to: ‘this particular component needs to be replaced now because it will very likely break down in three weeks’.”
“We need much more data about the current condition of the assets to make this possible. Sensor data is indispensable for this. Many assets are already equipped with sensors for controlling and protecting the grid. However, these are completely shielded from the outside world. People are scared to death of hackers. We will therefore share such sensitive failure data via a laser diode. This means that information can only go in one direction, so you cannot use the system to get into the network. We expect to realize this in the coming year. The application of AI within TenneT is now in full swing, in that regard.”
Control room
The same goes for the control room, the beating heart of the energy supply, where operators ensure that the energy balance in the country is maintained. This keeps the lights on for everyone. “The energy transition has had a lot of impact there, too. Energy flows are becoming more volatile, and more and more control is needed. Power has to be diverted, or we have to ask customers to temporarily reduce their consumption in exchange for compensation because the grid would get overloaded otherwise. This is happening more and more frequently, and regulating this will become too complex to handle and demand more speed than a human can offer. That is why we are working on the Control Room of the Future, where we combine the best human and computer abilities. Technology will first provide decision support: a system that suggests certain actions. It will eventually reach a certain level of automation.”
The intelligent use of data was born out of necessity, but it also creates new opportunities. “We can better utilize existing power grids, for example, by combining more data. Here’s why. The power lines that everyone will have seen hanging around in the landscape get hotter when more current flows through them. This causes the lines to sag. There is a limit to this, of course: you don’t want the power lines to be on the ground, but they also don’t need to be ten meters above it. But the amount of current that makes cables sag too far is different in summer than in winter. Cold and windy weather makes such a line cool down much faster, and more current can flow through it before it becomes too heavy. And it is precisely on these cold days that we need more power. In other words, we can make better use of the grid by including weather information in the power distribution across the grid. The grid will keep becoming smarter because of this.”
“A global competition is being organized where control operators develop these kinds of applications. Many different universities are participating, but so is Google, for example, and parties from China. They use a fictitious network to see how advanced the technology is, and they try out all kinds of innovations. We do not wait until everything has been perfected; we look for components that we can already apply. We call this practical innovation. Testing in real-life situations is still necessary. Developments like these have to prove themselves in practice; only then can you build up the confidence that is needed.”
Research
AI developments are being boosted TenneT-wide through an AI Center of Excellence. “With best practices, training and education. We collaborate with other parties such as Radboud University and the TU Delft. We have to - a Control Room of the Future like this one brings fundamental challenges and uncertainties that need to be researched. We do not have the knowledge and ambition to solve these ourselves. We also work together with French grid operator RTE: they do research, we make it practically applicable. So we are not only concerned with the future when it comes to AI; we also work on today.”
The willingness to work with AI is increasing, according to Huijsman. “We make AI tangible through proof of concepts and test its applicability in real work situations. That is often the moment when the opinions of users change from reluctance to enthusiasm. We also try to explain AI in the most understandable way possible, which is crucial for acceptance. AI sounds modern and fancy, of course, but it is ultimately just a computer that makes calculations.”
This is not only the case within TenneT, says Huijsman. “The challenge of the energy transition affects all grid operators. We are well represented here in the region in that regard. It is important to show that this is a place where you can use your talents concerning AI, also in the energy sector. You do not have to go to the west for that; you can do very nice things in the eastern part of the Netherlands as well, as an AI student or graduate.”
Do you want to know more about AI applications in the energy sector or about developments at TenneT? Contact us.
