eduzhai > Applied Sciences > Engineering >

Interpretable Detection of Partial Discharge in Power Lines with Deep Learning

  • king
  • (0) Download
  • 20210506
  • Save

... pages left unread,continue reading

Document pages: 13 pages

Abstract: Partial discharge (PD) is a common indication of faults in power systems,such as generators, and cables. These PD can eventually result in costlyrepairs and substantial power outages. PD detection traditionally relies onhand-crafted features and domain expertise to identify very specific pulses inthe electrical current, and the performance declines in the presence of noiseor of superposed pulses. In this paper, we propose a novel end-to-end frameworkbased on convolutional neural networks. The framework has two contributions.First, it does not require any feature extraction and enables robust PDdetection. Second, we devise the pulse activation map. It providesinterpretability of the results for the domain experts with the identificationof the pulses that led to the detection of the PDs. The performance isevaluated on a public dataset for the detection of damaged power lines. Anablation study demonstrates the benefits of each part of the proposedframework.

Please select stars to rate!


0 comments Sign in to leave a comment.

    Data loading, please wait...