Intermittent series direct current arc fault detection in direct current more‐electric engine power systems based on wavelet energy spectra and artificial neural network
Abstract The move towards More‐Electric Aircraft (MEA) and Engine (MEE) systems has resulted in system integrators exploring the use of Direct Current (DC) for primary power distribution to both reduce energy conversion stages and enable parallelling of non‐synchronised engine off‐take generation. A...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Wiley
2022-09-01
|
| Series: | IET Electrical Systems in Transportation |
| Online Access: | https://doi.org/10.1049/els2.12047 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Abstract The move towards More‐Electric Aircraft (MEA) and Engine (MEE) systems has resulted in system integrators exploring the use of Direct Current (DC) for primary power distribution to both reduce energy conversion stages and enable parallelling of non‐synchronised engine off‐take generation. A prevalent challenge of utilising DC systems is the safe management of arc faults. Arcing imposes a significant threat to aircraft and can result in critical system damage. Series intermittent arc faults in DC systems are particularly challenging to detect due to the lack of a zero current crossing coupled with the intermittency caused by in‐flight vibrations. Additionally, several state‐of‐the‐art arc fault detection methodologies fail to address the handling of aircraft system‐specific conditions such as dynamic loading and stability requirements. This paper addresses these unique MEA/MEE issues through the proposal of a new generalised real‐time arc fault detection methodology (WaSp) based on time–frequency domain‐extracted features applied to a feed‐forward Artificial Neural Network (ANN). The paper outlines the analysis of arc fault time–frequency domain features. Simulation‐based case studies emulating a range of on‐board EPS conditions are presented and show the proposed system has the potential for highly accurate and generalised detection performance with fast detection times. |
|---|---|
| ISSN: | 2042-9738 2042-9746 |