Intelligent fault detection and location scheme for modular multi‐level converter multi‐terminal high‐voltage direct current

Qingqing Yang, Jianwei Li, Ricardo Santos, Kaijia Huang, Petar Igic

Research output: Contribution to journalArticlepeer-review

9 Citations (Scopus)
28 Downloads (Pure)


In order to overcome the drawbacks of the conventional protection methods in high-voltage direct current transmission lines, a deep learning approach is proposed that directly learn the fault conditions based on unsupervised feature extraction to the detection and location decision by leveraging the hidden layer activations of recurrent neural network. The deep-recurrent neural network boosting with the gated recurrent unit compared with the long short-term memory unit is used by analysing both the signal presented in time domain and frequency domain. The proposed method is tested based on a modular multilevel converter based four-terminal high-voltage direct current system. Various faults under different conditions were simulated against fault resistance, external faults and small disturbance immunity with the validity, and the simulation verified a high accuracy, robustness and fast results because of the utilization of characteristic feature extraction.

Original languageEnglish
Pages (from-to)125-137
Number of pages13
JournalHigh Voltage
Issue number1
Early online date3 Nov 2020
Publication statusPublished - 4 Mar 2021

Bibliographical note

© 2020 The Authors. High Voltage published by John Wiley & Sons Ltd on behalf of The Institution of Engineering and Technology and China Electric Power Research Institute.

This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes.




Dive into the research topics of 'Intelligent fault detection and location scheme for modular multi‐level converter multi‐terminal high‐voltage direct current'. Together they form a unique fingerprint.

Cite this