Non-unit protection for HVDC grids: an analytical approach for wavelet transform-based schemes

Vasileios Psaras, Dimitrios Tzelepis, Dimitrios Vozikis, Grain Adam, Graeme Burt

Research output: Contribution to journalArticlepeer-review

42 Citations (Scopus)
49 Downloads (Pure)

Abstract

Speed and selectivity of DC fault protection are critical for High-Voltage DC (HVDC) grids and present significant technical and economic challenges. Therefore, this paper proposes a non-unit protection solution that detects and discriminates DC faults based on frequency domain analysis of the transient period of DC faults. The representation of a generic HVDC grid section and the corresponding DC-side fault signatures in the frequency domain form the basis of a generalized approach for analytically designing a protection scheme based on Wavelet Transform (WT). The proposed solution is adaptive within its design stage and offers general applicability and immunity to system changes, while the protection settings are configured for optimized performance. The scheme is validated through offline simulations in PSCAD/EMTDC and the technical feasibility of the algorithm in the real world is demonstrated through the use of real-time digital simulation (using RTDS) and Hardware-in-the-Loop (HIL) testing. Both offline and real-time simulations demonstrate that the scheme is able to detect and discriminate between internal and external faults at a significantly high speed, while remaining sensitive to high impedance faults and robust to external disturbances and outside noise.

Original languageEnglish
Number of pages11
JournalIEEE Transactions on Power Delivery
Early online date18 Sept 2020
DOIs
Publication statusE-pub ahead of print - 18 Sept 2020

Keywords

  • DC grid protection
  • frequency domain analysis
  • HVDC grids
  • non-unit protection methods
  • wavelet transform

Fingerprint

Dive into the research topics of 'Non-unit protection for HVDC grids: an analytical approach for wavelet transform-based schemes'. Together they form a unique fingerprint.

Cite this