Giacomazzi, ElenaElenaGiacomazziHaag, FelixFelixHaagHopf, KonstantinKonstantinHopf0000-0002-5452-06722023-08-072023-08-072023https://fis.uni-bamberg.de/handle/uniba/89942Recent developments related to the energy transition pose particular challenges for distribution grids. Hence, precise load forecasts become more and more important for efective grid management. Novel modeling approaches such as the Transformer architecture, in particular the Temporal Fusion Transformer (TFT), have emerged as promising methods for time series forecasting. To date, just a handful of studies apply TFTs to electricity load forecasting problems, mostly considering only single datasets and a few covariates. Therefore, we examine the potential of the TFT architecture for hourly short-term load forecasting across diferent time horizons (day-ahead and week-ahead) and network levels (grid and substation level). We fnd that the TFT architecture does not offer higher predictive performance than a state-of-the-art LSTM model for day-ahead forecasting on the entire grid. However, the results display signifcant improvements for the TFT when applied at the substation level with a subsequent aggregation to the upper grid-level, resulting in a prediction error of 2.43% (MAPE) for the best-performing scenario. In addition, the TFT appears to ofer remarkable improvements over the LSTM approach for week-ahead forecasting (yielding a predictive error of 2.52% (MAPE) at the lowest). We outline avenues for future research using the TFT approach for load forecasting, including the exploration of various grid levels (e.g., grid, substation, and household level).engShort-Term Load ForecastingArtifcial Neural NetworksTemporal Fusion Transformer (TFT)Long-Term Short-Term Memory (LSTM)004333.7conferenceobjecturn:nbn:de:bvb:473-irb-899421