etaGPT – LLM-Based Data Systems for Energy Data Understanding
Published:
Role: Project Lead
Project: etaGPT – LLM-Based Data Systems for Energy Data Understanding Institutions: German Research Center for Artificial Intelligence (DFKI) and TU Darmstadt (Germany)
Funding: Federal Ministry for Economic Affairs and Energy (BMWE) (2025–2028)
🎯 Objective
etaGPT focuses on building large language model-based data systems that can understand, interpret, and analyze complex energy data using expert-level language.
The project addresses the challenge that real industrial environments produce heterogeneous data in many formats, including sensor time series, knowledge graphs, inspection images, thermal data, maintenance logs, PDFs, tables, and operational reports.
🧩 Key Contributions
- LLM-based interpretation of industrial energy data
- Natural language interfaces for querying complex data sources
- Automatic generation of interactive dashboards
- Visualization of trends, anomalies, and performance patterns
- Energy data analysis across heterogeneous industrial data formats
- Support for optimization by identifying inefficiencies and suggesting operational or maintenance actions
🔗 Collaboration
etaGPT is conducted at the German Research Center for Artificial Intelligence (DFKI) within the Systems AI for Decision Support (SAIDE) research department. The project connects research on large language models, data systems, energy-aware AI, and industrial decision support. It is done in-collaboration with academic and industry partners - etalytics, PTW - TU Darmstadt, Merck, Siemens Energy, Bionech, and DARZ.
🧪 Technologies & Concepts
- Large Language Models
- Energy Data Systems
- Multimodal Data Analysis
- Sustainable AI
- Industrial AI