The project transpAIrent.energy – Transparent AI Forecasts for Green Energy in Austria aims to generate probabilistic real-time forecasts for energy system-relevant variables such as electricity prices and CO2 intensities in Austria using generative AI methods, and to develop a transparent platform that makes these forecasts publicly accessible.
Within the project, these forecasts are further used to develop an optimization strategy for the operation of flexible renewable energy systems under various environmental and system-related constraints. This “multi-objective” optimization approach simultaneously enables economic benefits - creating incentives - while ensuring lower CO2 intensity, thereby supporting the transition to a more environmentally sustainable energy system.
Factsheet#
| Short name | transpAIrent.energy |
| Title | transpAIrent.energy - Transparent AI Forecasts for Green Energy in Austri |
| Duration | 01.05.2024 – 30.04.2027 |
| Partners | 4 (show all) |
| Project type | Co-funded research project |
| Project lead AIT | Klara Maggauer & Stefan Strömer |
Overview#
Within the project, two main objectives are pursued: first, the development of an innovative generative AI-based algorithm for producing probabilistic forecasts of energy system-relevant variables, as well as their real-time publication on a transparent platform. Second, the use of these forecasts to optimize flexible renewable energy systems in order to make their operation both more economically efficient and more sustainable.
Findings#
Generative AI enables more accurate forecasts#
The use of generative AI allows for the creation of probabilistic live forecasts for key energy system variables such as electricity prices and CO₂ intensities, thereby improving the decision-making basis within the energy system.
Data and modelling form the core foundation#
Data collection, processing, and the development of AI-based forecasting algorithms form the basis for generating reliable predictions and their application.
Validation ensures practical applicability#
The validation of the developed methods through simulations and live tests ensures that the solutions function under real-world conditions.
Dissemination enhances project impact#
The targeted dissemination of project results increases their visibility and promotes their use by relevant stakeholders. This supports knowledge transfer and strengthens the project’s impact beyond its duration.
Activities#
WP1: Project Management
Lead: AIT
Project management and coordination of the project team, as well as monitoring of progress, costs, quality, and timelines. The work package also includes reporting, resource planning, communication, and the preparation of a final report.
WP2: Data Collection and Processing
Lead: AIT
Design and implementation of a data processing pipeline, including a database, as well as automated data collection and data cleaning. In addition, missing data is complemented using generative AI, CO2 intensities are calculated, and weather forecasts are developed.
WP3: Development of AI-based Forecasting Algorithms
Lead: AIT
Review, implementation, and validation of generative AI forecasting algorithms, with a focus on transformer-based architectures, GANs, and diffusion models. Application of the models to time series forecasting and evaluation based on benchmarks and KPIs.
WP4: Platform Development and Implementation
Lead: B-SEC
Design of the platform architecture as well as specification of functional requirements and user stories. Implementation of software components using existing systems, as well as deployment and operation of the platform on a cloud-based infrastructure.
WP5: Proof of Concept
Lead: PBEG
Execution of stochastic optimization based on AI forecasts, as well as development of a digital twin of the test sites. Validation through simulations and experimental live tests at multiple locations.
WP6: Dissemination and Exploitation
Lead: AIT
Development of strategies for disseminating the platform and increasing the visibility of the project. Involvement of stakeholders to ensure applicability, as well as promotion of scientific excellence, collaboration, and technology transfer.
Further information#
Deliverables#
- D1.1 Interim report 1
- D2.1 Documentation of data validation
- D2.2 Publication of algorithmic methodologies and complete data pipeline
- D3.1 Algorithm review and implementation documentation
- D3.2 Algorithm validation result documentation
- D4.1 Application requirements and use case specification
- D4.2 Application software development report
- D5.1 Digital twin optimization result documentation
Publications#
Project partners#
- AIT Austrian Institute of Technology GmbH (Project coordination)
- B-SEC better secure GmbH & Co KG
- Projektplanungs- Beratungs- und Entwicklungs GmbH
- UBIMET GmbH
Funding#
This project is conducted within the framework of the “AI for Green 2023” call by the Federal Ministry for Climate Action, Environment, Energy, Mobility, Innovation, and Technology (BMK). It is administered on behalf of the BMK by the Austrian Research Promotion Agency (FFG). More information can be found in the FFG project database