Forecasting Vertical Power Flows at Transmission Grid Nodes characterized by high Penetration of Renewable Generation and Consumption

Rising shares of distributed and volatile generation from renewable energies as well as growing electricity trading have led to volatile weather and price dependent power flows within and between the different voltage levels. Often, the electricity grid infrastructure is stressed to the limit. Grid operators have to identify congestions as early as possible to take appropriate measures. Grid calculations are performed for their identification. Power flows from and to the underlying voltage levels are essential inputs for these calculations. In this paper a machine learning approach is presented that is able to calculate day-ahead forecasts the vertical power flows between transmission and distribution grid.