Bi-LSTM based fault diagnosis scheme having high accuracy for Medium-Voltage Direct Current systems using pre- and post-processing
Diagnosing system faults is essential for ensuring the safety and reliability of Medium-Voltage Direct Current (MVDC) systems. In this regard, this study proposes a highly accurate Artificial Intelligence (AI)-based fault diagnosis scheme for MVDC systems. The proposed scheme pre-processes the measu...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-08-01
|
| Series: | International Journal of Electrical Power & Energy Systems |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S0142061525003412 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Diagnosing system faults is essential for ensuring the safety and reliability of Medium-Voltage Direct Current (MVDC) systems. In this regard, this study proposes a highly accurate Artificial Intelligence (AI)-based fault diagnosis scheme for MVDC systems. The proposed scheme pre-processes the measured voltage and current data using a Discrete Wavelet Transform (DWT), considering a 60 × 100 2D window size. Subsequently, a bi-directional long short-term memory (Bi-LSTM) network is employed to diagnose and classify fault types and locations accurately. A stack method is applied in the data post-processing stage to achieve 100 % fault diagnosis accuracy. The effectiveness of the proposed fault diagnosis scheme was verified by comparing its accuracy in 4-terminal MVDC system with that of existing schemes that employ other AI algorithms, such as CNN and LSTM. The proposed fault diagnosis scheme shows improved accuracy by 1.6 %, 3.8 %, and 2.9 %, 2.4 %, respectively, compared to existing schemes such as Bi-LSTM without stack method, LSTM, and CNN, GRU. Moreover, the scalability of the fault diagnosis scheme was verified by training and testing the scheme on a 5-terminal system and 4-terminal system, respectively. To a limited extent, the results demonstrate that the proposed fault diagnosis scheme improves accuracy even when the training and testing systems differ. |
|---|---|
| ISSN: | 0142-0615 |