Mechanism–Data Dual-Driven Outage Identification Framework for Distribution Networks with High Penetration of Inverter-Based Distributed Generation

INTRODUCTION: High penetration of inverter-based distributed generators (IBDGs) changes power-flow patterns and fault transients in distribution networks, reducing the effectiveness of conventional outage identification. Under the adopted low-voltage ride-through control model, IBDG fault currents are limited to 1.2–1.5 times the rated current, weakening protection discrimination and degrading classifiers trained on conventional fault features.

Objectives

To address the aforementioned challenges, this paper proposes a mechanism–data dual-driven outage identification framework for distribution networks with high levels of renewable energy penetration.

Methods

The framework integrates improved weighted Dempster–Shafer (D–S) evidence theory for multi-source fusion, constructs an IBDG-aware fault-feature library covering converter-specific transient behaviors, and designs decision rules linking post-fault analysis with early fault warning.

Results

Under the tested feeder, DG operating schemes, fault cases, and measurement-noise conditions, the proposed method achieved an F1-score of 96.8%. Its measured model-inference-and-fusion latency was 38 ms, compared with 120 ms for the reference implementation.

Conclusion

Under the tested feeder, operating schemes, fault cases, and measurement-noise conditions, the proposed method also achieved higher outage-identification performance than the implemented comparison methods.