Deep learning architectures for inference of AC-OPF solutions

TL;DR

This paper systematically compares various neural network architectures, including graph-based models, for efficiently inferring solutions to AC-OPF problems, demonstrating improved performance and computational benefits.

Contribution

It introduces the use of network topology in neural network models for AC-OPF inference, enhancing accuracy and efficiency over baseline fully connected networks.

Findings

Graph neural networks outperform fully connected networks in accuracy.

Topology-aware models reduce computation time.

Both regression and classification approaches are effective.

Abstract

We present a systematic comparison between neural network (NN) architectures for inference of AC-OPF solutions. Using fully connected NNs as a baseline we demonstrate the efficacy of leveraging network topology in the models by constructing abstract representations of electrical grids in the graph domain, for both convolutional and graph NNs. The performance of the NN architectures is compared for regression (predicting optimal generator set-points) and classification (predicting the active set of constraints) settings. Computational gains for obtaining optimal solutions are also presented.