The FFG-funded “POWERCAST” project has a clear objective: to increase the cost and supply efficiency of electricity grids. To achieve this, it relies on the prediction and optimization of electricity loads using AI-supported forecasting models. RISC Software GmbH is contributing its expertise in the development of new AI models and is leading the consortium.
Urgent need for accurate forecasts
The project team’s main aim is to facilitate the integration of large quantities of volatile renewable energies such as wind and solar energy into the Austrian electricity grid. To this end, POWERCAST is developing an adaptive AI model. This model combines several sources of information, which are continuously updated, to provide precise load forecasts and insights into key influencing factors.
The demand for such solutions is high. Errors in load forecasts have increased significantly in recent years. One reason for this is that conventional modeling techniques react too rigidly to new influencing factors such as photovoltaic systems or electromobility. As these technologies are changing rapidly, the gap between forecasts and real conditions is growing. This results in short-term adjustments that cost a lot of money. They are not only a burden for energy suppliers, but ultimately also for end customers, especially in energy-intensive sectors.
POWERCAST’s goals: grid stability and energy transition
The project pursues two central objectives:
- Firstly, the consortium wants to provide more accurate short-term forecasts. This will enable grid operators to carry out daily security checks, particularly with regard to the n-1 criterion. These forecasts also increase the economic efficiency of the grids, as they can avoid high costs for balancing energy.
- Secondly, POWERCAST is responding to the growing number of photovoltaic applications. Many systems for residential buildings are currently not approved because they endanger grid stability when the feed-in is high. However, an intelligent early warning system could dynamically regulate the feed-in until the critical situation has passed. This would mean that no PV systems would have to be rejected. This would accelerate the expansion of renewable energies and help Austria achieve its climate targets.
A new paradigm for load forecasting
The project team is introducing a new paradigm to load forecasting. Forecast models should adapt dynamically to the rapid changes in generators and consumers. They must also be transferable to new but similar scenarios. POWERCAST specifically uses adaptive AI methods to achieve this.
Reduce energy costs and increase sustainability
The new methods can reduce costs in the overall system and facilitate the integration of renewable energies. At the same time, the project contributes to the sustainability of the energy system. It is in line with the EU’s goals of significantly increasing the share of renewable energy and achieving climate neutrality by 2050. POWERCAST thus makes an important contribution to mastering current and future challenges of grid optimization.
Partners in the project
The consortium brings together strong partners: RISC Software GmbH, LINZ NETZ GmbH, Austrian Power Grid AG, the Energy Institute at the Johannes Kepler University Linz and HAKOM Time-Series GmbH. Together, they combine expertise in data management, the energy industry and AI. The project is represented by the Federal Ministry for Climate Protection, Environment, Energy, Mobility, Innovation and Technology and the Österreichische Forschungsgesellschaft mbH as part of the “AI for Green 2023” tender.er Grid AG, Energieinstitut an der Johannes Kepler Universität Linz and HAKOM Time-Series GmbH are pooling their expertise and resources to ensure the success of the project and are represented by the Federal Ministry for Climate Protection, Environment, Energy, Mobility, Innovation and Technology and financed by the Österreichische Forschungsgesellschaft mbH as part of the AI for Green 2023 tender. The consortium combines expertise and previous experience in various areas, such as data management, data analysis, energy economics and management.
