On zero-shot learning in neural state estimation of power distribution systems

TL;DR

This paper explores zero-shot neural state estimation in power distribution systems, highlighting the potential of graph neural networks and proposing data augmentation techniques to enhance adaptability to grid changes.

Contribution

It identifies the limitations of current models in zero-shot scenarios and demonstrates how graph neural networks with data augmentation improve robustness and performance.

Findings

Graph neural networks show robustness to grid changes.

Deeper networks do not always outperform shallower ones.

Data augmentation enhances zero-shot learning performance.

Abstract

This paper addresses the challenge of neural state estimation in power distribution systems. We identified a research gap in the current state of the art, which lies in the inability of models to adapt to changes in the power grid, such as loss of sensors and branch switching, in a zero-shot fashion. Based on the literature, we identified graph neural networks as the most promising class of models for this use case. Our experiments confirm their robustness to some grid changes and also show that a deeper network does not always perform better. We propose data augmentations to improve performance and conduct a comprehensive grid search of different model configurations for common zero-shot learning scenarios.