Ringvorlesung 2025 "KI, Klima und Energie - Innovationen für eine nachhaltige Zukunft"
The Sensitivity of AI-Based Load-Forecasting Methods on Data
Referent: Prof. Dr. Ing. Hermann de Meer (Universiät Passau) - Vortrag auf Englisch
Kurzbeschreibung:
The Sensitivity of AI-Based Load-Forecasting Methods on Data
Energy load forecasting allows grid managers to plan their operations in advance, balancing consumption, and generation. It is crucial service used to utilize the full capacity of renewable energy sources in the modern energy systems. State-of-the-art methods for time series forecasting allow operators to get precise predictions. However, data availability and data quality pose significant challenges for the practical application of the forecasting models. In this talk, the common data quality issues and the effects of the training data on the performance of energy load forecasting are investigated. The talk concludes with proposals on how the identified challenges can be addressed in real-life applications.