AI Assisted Electrical Distribution System Design

Recent advances in graph neural network (GNN) research have made it possible to train neural network-based machine learning models on graph structured datasets. However, the number of such algorithms being utilized by industry within the power distribution space are limited. Additionally, we find that traditional GNNs are not well suited to learning on distribution graphs due to typical attributes of distribution systems such as: their large size and low density, their heterophilic nature, and the long paths between nodes along which information must travel. Herein we outline a novel inductive GNN architecture which has been optimized for learning on power distribution graphs and which can pass information efficiently across the graph. We also demonstrate the performance of this algorithm by applying it to ComEd’s data as part of a practical use-case: planning new fuse locations on distribution networks. We also benchmark our algorithm’s performance on this task versus other well-known GNN architectures.