Fraunhofer presents AI tool for energy forecasts

Fraunhofer IOSB-AST has presented the Wattpredictor, a new tool for AI-supported energy forecasting for electricity, gas, heating and cooling.

The Fraunhofer Institute of Optronics, System Technologies and Image Exploitation (IOSB-AST) has developed a new forecasting tool for the energy industry. The "WattPredictor" uses artificial intelligence (AI) to create time series forecasts in electricity, gas, heating and cooling networks, according to the institute. The aim is to increase accuracy in resource planning, procurement, energy trading and marketing.

The system is integrated into the existing "EMS-EDM PROPHET" software ecosystem and works without additional third-party software. It supports both short-term and long-term forecasts. In addition to classic calendar logic, the Wattpredictor also takes external influencing factors into account and enables mass data processing. Strategies for calculating replacement values and assistance functions for operation supplement the functions.

Fields of application include load forecasts under dynamic tariffs, consumption forecasts in a district and industrial context and grid loss analyses. Electromobility and district heating can also be included.

In use at Erfurt electricity supplier

"Our many years of forecasting expertise from research and industrial projects have been incorporated into the Wattpredictor. With its analysis and forecasting functions, it delivers reliably precise results even for complex applications and customer-specific requirements," says Stefan Klaiber, Group Manager Cross-Sectoral Energy Systems at Fraunhofer IOSB-AST.

The tool is already in use at Erfurt-based electricity supplier SWE Netz GmbH. Managing Director Frank Heidemann draws a positive initial balance: "The results obtained with the introduction of the Wattpredictor make us very confident. In future, we will be able to create more accurate grid load forecasts over a longer observation horizon."

Author: Katia Meyer-Tien