Multidimensional Big Data Modeling for Reliable Electricity Supply
Weather patterns cause significant fluctuations in weather in Europe and in the potential supply of energy from renewable sources. Weather conditions can jeopardize the long-term system adequacy of a low-carbon energy system. A robust and reliable energy supply requires advanced models capable of predicting and managing this variability. To choose the right combination of investments and measures, network operators are supported by various computational models. These models are enhanced in this research with big data analytics techniques to analyze processes in climate, weather, energy production, and demand in an integrated manner. In June, we spoke with Machteld van den Broek, Professor of Energy System Integration at TU Delft.
The Partners
The project ‘multidimensional big data modeling’ is a collaboration between Tennet, KNMI, and Utrecht University, focusing on improving computational models with big data analytics techniques. “Tennet naturally considers it very important that we have a reliable electricity system,” says Machteld. “KNMI has a lot of information about all weather data, which is essential for our work. Together, we formed a small consortium and created a good combination. The mix of expertise in mathematics, computer science, and energy systems was incredibly interesting and useful.”
The reliability of the energy system is of vital importance, especially given the variability in renewable energy sources. “We always look at moments when the energy system cannot meet the demand,” says Machteld. “These insights help network operators like Tennet in taking measures to ensure reliability, such as strengthening networks or adding backup or storage capacity.”
Weather Data
With the increase in solar and wind energy, there is greater dependence on weather conditions. This brings new challenges. “In the future, there will be much more solar and wind energy, so you need a lot of weather data from KNMI's climate models,” explains Machteld. “Big data plays a crucial role in analyzing processes in climate, weather, energy production, and demand. By analyzing these processes in an integrated manner, the requirements for a reliable electricity supply can be better estimated. Our new models consider what is actually the cheapest for the entire system.”
Innovative Modeling Approach
An important aspect of the project is the improvement of existing models by applying new algorithms. Rogier Wuijts focused on programming models that are more efficient and faster. His work has resulted in models that can now be computed in minutes, a significant improvement over the hours it previously took. “Rogier optimized the existing models, allowing us to work much faster now. This is a real breakthrough for us,” says Machteld proudly.
Additionally, Rogier conducted a systematic analysis to understand the impact of constraints within the energy system. He examined how different constraints affect the system and which constraints should or should not be included in the models.
Working with Climate Data
Simulations often focus on a single year, such as a ‘good year’ or a ‘bad year,’ which does not do justice to the large variability of weather. Moreover, it is challenging to define what constitutes a ‘good year’ due to the many different factors involved. Therefore, Laurens Stoop worked on processing vast datasets with historical and future climate data. He developed techniques to process and integrate this data from weather models more efficiently into the models.
Machteld explains: “Laurens looked more at what you can do with all that climate data now. How can we use it smartly? One of his major contributions was the cumulative calculation of energy production and demand over longer periods. This enabled the team to analyze not just snapshots but also better understand long-term trends and patterns. Through the development of his methods, we can identify challenging periods for the future energy system.”
Challenges
Despite the progress, challenges remain. Applying models to future climate years proved time-consuming. “Applying it to all those future climate years from different climate models worldwide takes more time because everything has to be converted to solar and wind availability,” says Machteld. Additionally, there is a need for more detailed data at the regional level, which is currently difficult to obtain.
Besides solar and wind energy, the role of hydropower was also discussed. Hydropower provides a stable and reliable source of renewable energy, which is essential to compensate for the variability of other renewable sources. “By smartly integrating hydropower into the European energy system, network operators can better respond to fluctuations in supply and demand, thereby further strengthening the reliability of the energy supply. We are eager to conduct more extensive research in this area,” says Machteld.
Conclusion
The multidimensional big data modeling project led by Machteld and her team offers valuable insights for the future of energy supply. By integrating advanced data analytics techniques and collaboration between different disciplines, the challenges of renewable energy can be better addressed. Machteld concludes: “Our findings provide energy companies and network operators with tools to ensure that we also have an adequate energy supply in the future.”
