Comparison of Different Machine Learning Models for Short-Term Load Forecasting at Transformer Level with High Amounts of Photovoltaic Generation
Jungh T, Steinhagen B, Hesse M, Schulte K (2023)
In: 2023 IEEE PES Innovative Smart Grid Technologies Europe (ISGT EUROPE). Piscataway, NJ: IEEE: 1-5.
Konferenzbeitrag
| Veröffentlicht | Englisch
Download
Es wurden keine Dateien hochgeladen. Nur Publikationsnachweis!
Autor*in
Einrichtung
Abstract / Bemerkung
In this work, different models for short-term load forecasting for transformers between the low-voltage and medium-voltage level were implemented and compared. The focus was on transformers with a high proportion of photovoltaic generation in the grid, leading to backflows to higher voltage levels. These backflows can be problematic for the grid stability and should be reduced as much as possible. The forecast is needed to set up a control system. The forecast horizon will be 24 hours with an hourly resolution. In addition to past load values, weather data is also used for the forecast. Different machine learning models are used as methods (linear, XGB, CNN, LSTM, CNN-LSTM). A data set of 18 transformers from Germany was used to compare the models. No model could achieve the best result for every transformer, however CNNs achieved the highest average accuracy for all transformers.
Erscheinungsjahr
2023
Titel des Konferenzbandes
2023 IEEE PES Innovative Smart Grid Technologies Europe (ISGT EUROPE)
Seite(n)
1-5
Konferenz
2023 IEEE PES Innovative Smart Grid Technologies Europe (ISGT EUROPE)
Konferenzort
Grenoble, France
eISBN
979-8-3503-9678-2
Page URI
https://pub.uni-bielefeld.de/record/2987581
Zitieren
Jungh T, Steinhagen B, Hesse M, Schulte K. Comparison of Different Machine Learning Models for Short-Term Load Forecasting at Transformer Level with High Amounts of Photovoltaic Generation. In: 2023 IEEE PES Innovative Smart Grid Technologies Europe (ISGT EUROPE). Piscataway, NJ: IEEE; 2023: 1-5.
Jungh, T., Steinhagen, B., Hesse, M., & Schulte, K. (2023). Comparison of Different Machine Learning Models for Short-Term Load Forecasting at Transformer Level with High Amounts of Photovoltaic Generation. 2023 IEEE PES Innovative Smart Grid Technologies Europe (ISGT EUROPE), 1-5. Piscataway, NJ: IEEE. https://doi.org/10.1109/ISGTEUROPE56780.2023.10407216
Jungh, Timon, Steinhagen, Bastian, Hesse, Marc, and Schulte, Katrin. 2023. “Comparison of Different Machine Learning Models for Short-Term Load Forecasting at Transformer Level with High Amounts of Photovoltaic Generation”. In 2023 IEEE PES Innovative Smart Grid Technologies Europe (ISGT EUROPE), 1-5. Piscataway, NJ: IEEE.
Jungh, T., Steinhagen, B., Hesse, M., and Schulte, K. (2023). “Comparison of Different Machine Learning Models for Short-Term Load Forecasting at Transformer Level with High Amounts of Photovoltaic Generation” in 2023 IEEE PES Innovative Smart Grid Technologies Europe (ISGT EUROPE) (Piscataway, NJ: IEEE), 1-5.
Jungh, T., et al., 2023. Comparison of Different Machine Learning Models for Short-Term Load Forecasting at Transformer Level with High Amounts of Photovoltaic Generation. In 2023 IEEE PES Innovative Smart Grid Technologies Europe (ISGT EUROPE). Piscataway, NJ: IEEE, pp. 1-5.
T. Jungh, et al., “Comparison of Different Machine Learning Models for Short-Term Load Forecasting at Transformer Level with High Amounts of Photovoltaic Generation”, 2023 IEEE PES Innovative Smart Grid Technologies Europe (ISGT EUROPE), Piscataway, NJ: IEEE, 2023, pp.1-5.
Jungh, T., Steinhagen, B., Hesse, M., Schulte, K.: Comparison of Different Machine Learning Models for Short-Term Load Forecasting at Transformer Level with High Amounts of Photovoltaic Generation. 2023 IEEE PES Innovative Smart Grid Technologies Europe (ISGT EUROPE). p. 1-5. IEEE, Piscataway, NJ (2023).
Jungh, Timon, Steinhagen, Bastian, Hesse, Marc, and Schulte, Katrin. “Comparison of Different Machine Learning Models for Short-Term Load Forecasting at Transformer Level with High Amounts of Photovoltaic Generation”. 2023 IEEE PES Innovative Smart Grid Technologies Europe (ISGT EUROPE). Piscataway, NJ: IEEE, 2023. 1-5.
Google Scholar