Project update: long-term power and heat system adequacy
Smarter and neater computing
"The goal of the project is to develop an energy system model in which a lot of renewable energy is used," explains Rogier. "This implies that the weather is an important factor in this model: renewable energy is, after all, strongly dependent on the wind and the sun. Calculating simulations with different weather scenarios still takes a lot of time: depending on the time horizon and the level of detail, this can take days, sometimes weeks. In this project, I am trying to design algorithms in a smarter way in order to speed things up. Another important question is how to include something complex like the weather in our energy system models: that's what Laurens is working on."
"We want to include the weather dependencies of today and in the future in our energy system models in a neater way," Laurens explains. "Simulations still often look at a single year, for example a 'good year' or a 'bad year', and that doesn't do justice to the great variability of the weather. Moreover, it is difficult to say what a 'good year' is, because so many different factors come into play. I am working on improving the use of data from weather models in order to convert as many of these factors as possible into energy variables. In effect, I'm making a large amount of weather data small enough to fit into the energy system model."
The enhanced use of weather data has another purpose, which is to predict the impact of weather. For example, the researchers use this to see what degree of variability in the weather the energy system can handle. Climate model outcomes are also used for this purpose: climate change affects weather patterns and this has consequences for the expected impact of the weather and the risk assessments linked to it.
Collaborating on datasets Project partner TenneT, the operator of the Dutch high-voltage grid, played an important role in making the right datasets available: for commercial reasons these are not easily accessible. Climate datasets are publicly accessible, but this presented another challenge, says Laurens. "Climate datasets are gigantic. They are so large that they are stored on magnetic tape - a kind of large cassette tape, in fact. Hard disks would be too expensive and consume an unnecessary amount of energy. Retrieving relevant parts of such a climate dataset can take a very long time. Fortunately, thanks to the help of project partner KNMI (the Royal Netherlands Meteorological Institute), we were able to get faster access."
Still, data challenges remain, Rogier says. "As researchers, we like to look at issues around energy systems on a European level. The tricky thing is that you then have to deal with multiple high-voltage grid operators (there are four in Germany alone), all of whom store their data in a different way. That makes it complicated. Moreover, the data is not public."
At the European level
Although there are no European agreements on data storage as yet, the European high-voltage grid operators (more precisely, the Transmission System Operators) are united in the European network ENTSO-E. Among other things, this organization tries to estimate how adequate European energy systems will be in the future. Laurens unexpectedly discovered an opportunity to improve these estimates by using newer data sets: "Together with other experts, we were able to contribute to this improvement from ACDC-ESM. I think that's a great example of direct implementation."
Moreover, ACDC-ESM can further contribute to the work of ENTSO-E in the future, Rogier adds. "ENTSO-E works with scenarios. From analyses with the energy system models we are making in this project, ways to improve those scenarios may be identified. This could have implications for the measures European high-voltage grid operators take to reduce risks and keep the European grid adequate in the future."
Next steps in the project
Optimizing power system models basically requires solving a mathematical unit commitment problem. "The electricity demand of the Netherlands can be met by several generators, which generate electricity using gas, wind and solar power, among other things," Rogier explains. "The question is then how the total demand can best be distributed among those generators. That, in a nutshell, is a unit commitment problem. Although there is existing software that can solve this kind of problems, I developed my own program for ACDC-ESM, based on modeling techniques from the literature. That way we have all the freedom to make improvements."
At the time of writing (summer 2021), Laurens and Rogier's work is coming together: the energy system models are as good as finished and the first energy variables are ready to be implemented in these models. After some initial testing, the real calculation work can begin: the extensive calculation of various scenarios. The results will help high-voltage grid operators to ensure that we will have an adequate energy network in the future.
Sign up for our newsletter and stay informed of the latest news Commit2data