Evaluating Distribution System Reliability with Hyperstructures Graph Convolutional Nets

TL;DR

This paper introduces Hyper-GCNNs, a novel graph convolutional neural network framework utilizing hyperstructures, to improve distribution grid planning by enhancing accuracy, resilience, and computational efficiency in power system reliability assessments.

Contribution

It presents the first application of hyperstructures graph convolutional networks for distribution grid planning, demonstrating significant efficiency and performance improvements over existing models.

Findings

Hyper-GCNNs outperform seven state-of-the-art deep learning models.

The approach significantly reduces computational time in grid planning.

Hyper-GCNNs improve the accuracy and resilience of distribution system reliability evaluations.

Abstract

Nowadays, it is broadly recognized in the power system community that to meet the ever expanding energy sector's needs, it is no longer possible to rely solely on physics-based models and that reliable, timely and sustainable operation of energy systems is impossible without systematic integration of artificial intelligence (AI) tools. Nevertheless, the adoption of AI in power systems is still limited, while integration of AI particularly into distribution grid investment planning is still an uncharted territory. We make the first step forward to bridge this gap by showing how graph convolutional networks coupled with the hyperstructures representation learning framework can be employed for accurate, reliable, and computationally efficient distribution grid planning with resilience objectives. We further propose a Hyperstructures Graph Convolutional Neural Networks (Hyper-GCNNs) to capture hidden higher order representations of distribution networks with attention mechanism. Our numerical experiments show that the proposed Hyper-GCNNs approach yields substantial gains in computational efficiency compared to the prevailing methodology in distribution grid planning and also noticeably outperforms seven state-of-the-art models from deep learning (DL) community.