Willem Leterme, Mike Barnes, Dirk Van Hertem. Signal processing for fast fault detection in HVDC grids[J]. CSEE Journal of Power and Energy Systems, 2018, 4(4): 469-478.
DOI:
Willem Leterme, Mike Barnes, Dirk Van Hertem. Signal processing for fast fault detection in HVDC grids[J]. CSEE Journal of Power and Energy Systems, 2018, 4(4): 469-478. DOI: 10.17775/CSEEJPES.2016.00300.
Signal processing for fast fault detection in HVDC grids
For multiterminal or meshed Voltage Source Converter (VSC) High-voltage Direct Current (HVDC) systems
high speed protection against DC faults is essential
as power electronic components cannot withstand the rapidly increasing fault currents which would otherwise result. Recently proposed DC fault detection methods were developed based on time domain simulations in EMT-type software
which requires considerable modeling and computational efforts and results in methods specifically designed for the HVDC grid under study. To simplify the initial design of DC fault detection methods
this paper proposes general guidelines based on fundamental theory and offers a reduced modeling approach. Furthermore
the impact of non-ideal measurements is investigated and a method to choose the filters that optimally discriminate these fault signals from noise
is proposed. The approach was evaluated in a case study on fault detection in a realistically dimensioned HVDC grid. The paper shows that the initial design of fast fault detection methods can be based on the relatively simple proposed guidelines and reduced models. The paper furthermore shows that a sufficiently high sampling frequency and a filter matched to the fault signal enable fault detection within hundreds of microseconds and discrimination of DC faults from transients not related to DC faults.