ElectricGrid.jl – Automated modeling of decentralized electrical energy grids

Abstract:
The transition of a centralized energy grid towards a decentralized, cellular energy system is a difficult task in terms of grid stability and energy conversion efficiency. This is due to the increasing system complexity and uncertainty when adding more and more regenerative power plants while large and robust synchronous generators are becoming fewer. Data-driven control approaches, such as reinforcement learning, come along with a model-free and a self-adaptive controller design allowing to learn controlling such complex and heterogeneous systems. However, due to safety and availability aspects, these machine learning-driven control methods cannot be applied directly in the field, but need to be improved and evaluated on the basis of synthetic data in a closed simulation environment first, requiring a fast ad hoc synthetic data generation. This paper introduces ElectricGrid.jl, an open-source simulation tool developed for the programming language Julia. It provides a simulation environment with fast and automated configuration capabilities for dynamic models of heterogeneous microgrids, allowing faster than real-time data generation for future empirical control investigations.