The energy transition does not just change the energy generation technologies we use, but also requires the power grids to be updated. Today, instead of a dozen large-scale power plants, an enormous swarm of power plants all feed into our grid: Several thousand wind farms and over 2.6 million PV systems have been installed in Germany alone – and there is a steep upward trend. This makes our power generation weather dependent: On windless or cloudy days, the energy supply may not match the demand, leading to bottlenecks – or worse – blackouts.
To keep the grid frequency at a stable 50.2 hertz, grid operators need yield forecasts. Artificial intelligence (AI) and swarm intelligence help process this highly complex information and integrate it into the energy system.
With millions of solar installations, wall boxes and heat pumps, home storage systems and car batteries that will someday be able to feed electricity back into the grid, today’s utility grid system is already so complex that utilities and grid operators would be lost without their digitalized systems. The more they know about when, where and how much electricity is being produced, fed into the grid and demanded by consumers, the more precisely they can match their supply to the demand.
This also makes it more profitable, as making up for shortages is expensive. When generation does not cover consumption within a balancing group, grid operators are forced to pay balancing costs. The more accurate the weather and yield forecasts for wind and solar power plants, the more successful the companies will be. Individuals who want to optimize their self-consumption using batteries or controllable consumers like e-cars or heat pumps also require forecasts to determine what the electricity should be used for. The same goes for the energy management of buildings.
Weather forecast for the energy industry
At EM-Power Europe, the international exhibition for energy management and integrated energy solutions, companies from all around the world present their digitalization concepts as well as innovative technologies and services for the optimized grid of the future. “Machine learning and AI play a crucial role in the irradiation forecast,” Jan Remund, Head of the Department for Energy and Climate at Meteotest AG, a Swiss weather forecast provider, explains. The company is able to predict cloud movements by combining physical models with self-learning algorithms based on satellite images.
“The accuracy for the next few hours is quite high. The longer the time span, the less accurate the forecasts are,” says Remund. His company collects irradiation and temperature data and uses it to make analyses, profiles and forecasts that can be integrated into the respective PV monitoring or control software. Companies can use weather forecasts as a basis for calculating solar irradiation using dedicated software based on swarm intelligence, or they can buy such calculations. Grid operators rely on irradiation forecasts to balance the grid’s feed-in and feed-out based on the production and demand forecasts.
Swarm Intelligence explained
Swarm intelligence refers to collective intelligence. Biologists and natural scientists have been studying the behavior of social insects due to their efficiency of solving complex problems such as finding the shortest path between their nest and food source or organizing their nests. Insects are not capable of this individually, but they are as a swarm by interacting with each other and their environment.
Meteocontrol, a monitoring and forecast specialist company from the South of Germany makes yield calculations based on the data of different weather forecast providers. The company also investigates how aerosols like ash or sand affect cloud formation. “On March 3 and 4, 2021, our research project “PermaStrom” exemplified how important this issue is,” says Stijn Stevens, CEO of meteocontrol. “There was a lot of Saharan dust in Europe on those two days. Thanks to optimized forecasts, about three million euros were saved in balancing energy costs in Germany alone.
The integration of smart applications and AI into the utility grid system creates digital swarm intelligence. By combining the power generation and consumption forecasts with the capacity of conventional power plants, grid operators can predict when and where the grids and transformers will reach their maximum capacity. This allows them to take the necessary precautions.
Prosumers: underestimated key players
Prosumers will play a vital role in the energy systems of the future: For instance, private individuals and households can feed excess electricity from their e-cars into the grid, contributing significantly to its stability. The company Hive Power already offers software for the intermittent feedback of electricity from car batteries into the grid, called Vehicle-to-Grid. Artificial intelligence helps here, too: The software learns when vehicles are in use and when they are not, allowing it to feed electricity into the grid when there is excess capacity. Prosumers could earn up to 1,000 euros per year in this way.
Smart technology and swarm intelligence for buildings
Weather and yield forecasts are also crucial for the energy management of buildings. Heat pumps can charge the heating storage device when the sun is shining, and feed the energy into the heating circuit in the evening. With the right software, heat pumps can apply the yield forecasts provided by the energy management system to their operating schedules. This prevents heat pumps from using grid-supplied electricity to heat thermal energy storage when sufficient solar energy is expected within the next hour.
Another possibility is to time the charging of electric vehicles to maximize the share of solar power. Solar-Log’s systems are based on the same principle. They provide optimal electricity self-consumption for office buildings and industrial facilities through battery storage and flexible loads, such as heat pumps or charging stations.