On December 7, N-SIDE will have the pleasure to present an innovative topic together with Elia at the Energy Days 2020 (virtual event), organized by Anthony Papavalisiou, Associate Professor at CORE (Center for Operations Research and Econometrics) UCLouvain.
Johan Maricq, Head of Digital workplace and AICoE at ELIA, and Wolf Berwouts, Head of the Energy Analytics Incubation at N-SIDE
will demonstrate the application of reinforcement learning in finding the optimal grid topology.
Finding the optimal topology for a transmission network is a complex problem which is in practice solved by highly experienced system operators, backed by approximations and heuristic methods. Elia & N-SIDE research the application of reinforcement learning in a practical innovation project in an attempt to make the decisions faster and better, avoiding unnecessary congestion & redispatch actions. By letting a machine learning model train itself on a digital twin of the Elia network, together with the smart injection of valuable business knowledge, it can learn what optimal actions to take in each situation. Such actions could be the disconnection of lines, splitting of substations, or even acting on the maintenance plan. The ultimate goal is to provide the operators with a tool that can support them in taking these topology decisions in their day-to-day tasks.