The bottom line must be zero – this applies to power input and consumption at all times. To establish this equilibrium, power grid operators calculate load flows in advance for the next few hours. In this process, the load at the transition points between different electrical grids represents a particular challenge.

An innovative approach could now increase the level of automation and reliability, because AI can be used to make more precise estimates of the grid topology. The Vertical Load Forecasting project applies deep learning using a method based on a long short-term memory (LSTM) ANN.