Generative adversarial imbalanced learning for DC microgrid anomaly detection

Jieqi Rong

doi:10.1016/j.eswa.2026.131811

Generative adversarial imbalanced learning for DC microgrid anomaly detection

2026-03-03

0

PRE

AI

DOI：10.1016/j.eswa.2026.131811

Original

Original request for help

Save

Abstract

En 中文

• LSTM-based GAN generates realistic sequential anomaly data for DC microgrids. • Hybrid time-frequency features are selected through formal correlation analysis. • A Grid Search-optimized Random Forest enhances detection accuracy and robustness. • The integrated method shows significant gains over advanced detection baselines.

Keywords:

LSTM-based GAN

hybrid time-frequency features

Grid Search-optimized Random Forest

DC microgrid anomaly detection

imbalanced learning

Generative adversarial imbalanced learning for DC microgrid anomaly detection

Abstract

Journal

Organization