Researchers at the Fraunhofer Institute for Manufacturing Engineering and Automation IPA in Stuttgart, Germany, have succeeded in using machine learning to predict bid prices for balancing power more reliably. Companies that flexibly adjust their electricity demand can significantly increase their revenues on the balancing energy market with this forecasting method, according to a statement from the research institute.
»Control power is traded on a so-called “pay-as-bid” market,« explains Professor Alexander Sauer, director of the Fraunhofer IPA. »This market uses a bidding process in which each supplier is paid the price at which they submitted their bid. Anyone who significantly undercuts the actual electricity price forfeits money. Anyone whose bid is above the actual price comes away empty-handed. Companies can therefore significantly increase their revenues with our newly developed forecasting method.«
In a second step, the research team supplemented the AI-supported price forecast with a specially developed offset procedure. »This is, in a sense, the post-processing of the forecast electricity price, so that the bid submitted is slightly below it,« explains Vincent Bezold from the Data-Driven Energy System Optimisation research team at Fraunhofer IPA. »This has to do with the rules of the game on the “pay-as-bid” market.
If your bid is too high, you come away empty-handed. That’s why it pays to undercut the actual electricity price in a targeted manner – and that’s exactly what we achieve with our offset method.«
According to Fraunhofer IPA, this optimised bidding process can increase revenues by up to 37 percent compared to other strategies, as the bid submitted is less likely to exceed the actual price and is therefore more likely to be accepted.
© PHOTON
