The strong growth of renewable energy sources as well as the increasing amount of volatile energy consumptionis leading to major challenges in the electrical grid. In order to ensure safety and reliability in the elec...The strong growth of renewable energy sources as well as the increasing amount of volatile energy consumptionis leading to major challenges in the electrical grid. In order to ensure safety and reliability in the electricitygrid, a high quality of power flow forecasts for the next few hours are needed. In this paper we investigateforecasts of the vertical power flow at transformer between the medium and high voltage grid. Forecastingthe vertical power flow is challenging due to constantly changing characteristics of the power flow at thetransformer. We present a novel approach to deal with these challenges. For the multi step time series forecastsa Long-Short Term Memory (LSTM) is used. In our presented approach an update process where the model isretrained regularly is investigated and compared to baseline models. The model is retrained as soon as asufficient amount of new measurements are available. We show that this retraining mostly captures changesin the characteristic of the transformer that the model has not yet seen in the past and therefore cannot bepredicted by the model without an update process. To give more weight to recent data, we examined differentstrategies in terms of the number of epochs and the learning rate. We show that our new approach significantlyoutperforms the investigated baseline models. On average, we achieved an improvement of about 8% with theregular update process compared to the approach without update process.展开更多
基金funding from the European Union’s Horizon 2020 research and innovation program under grant agreement No 773505。
文摘The strong growth of renewable energy sources as well as the increasing amount of volatile energy consumptionis leading to major challenges in the electrical grid. In order to ensure safety and reliability in the electricitygrid, a high quality of power flow forecasts for the next few hours are needed. In this paper we investigateforecasts of the vertical power flow at transformer between the medium and high voltage grid. Forecastingthe vertical power flow is challenging due to constantly changing characteristics of the power flow at thetransformer. We present a novel approach to deal with these challenges. For the multi step time series forecastsa Long-Short Term Memory (LSTM) is used. In our presented approach an update process where the model isretrained regularly is investigated and compared to baseline models. The model is retrained as soon as asufficient amount of new measurements are available. We show that this retraining mostly captures changesin the characteristic of the transformer that the model has not yet seen in the past and therefore cannot bepredicted by the model without an update process. To give more weight to recent data, we examined differentstrategies in terms of the number of epochs and the learning rate. We show that our new approach significantlyoutperforms the investigated baseline models. On average, we achieved an improvement of about 8% with theregular update process compared to the approach without update process.
