Deep Active Learning for Solvability Prediction in Power Systems

TL;DR

This paper introduces a deep active learning framework for predicting power system solvability regions, significantly reducing labeling effort and improving accuracy compared to passive learning methods.

Contribution

It presents a novel active learning approach with specific acquisition functions for power system solvability prediction, validated on the IEEE 39-bus system.

Findings

Active learning reduces labeled data requirements.

The proposed method outperforms passive learning in accuracy.

Effective sampling strategies improve prediction quality.

Abstract

Traditional methods for solvability region analysis can only have inner approximations with inconclusive conservatism. Machine learning methods have been proposed to approach the real region. In this letter, we propose a deep active learning framework for power system solvability prediction. Compared with the passive learning methods where the training is performed after all instances are labeled, the active learning selects most informative instances to be label and therefore significantly reduce the size of labeled dataset for training. In the active learning framework, the acquisition functions, which correspond to different sampling strategies, are defined in terms of the on-the-fly posterior probability from the classifier. The IEEE 39-bus system is employed to validate the proposed framework, where a two-dimensional case is illustrated to visualize the effectiveness of the sampling method followed by the full-dimensional numerical experiments.