Convolution Based Time Domain Fault Location Method for Lines in MMC-HVDC Grids With Distributed and Frequency Dependent Line Model

In this paper, a convolution based time domain model based fault location method for lines is proposed in modular multilevel converter based high voltage direct current (MMC-HVDC) grids. The proposed method accurately considers the effect of distributed and frequency dependent line parameters. Two terminal measurements with 5 ms data window are utilized to calculate the accurate voltage distribution on the whole line to locate faults with relatively low sampling rate (10 kHz). First, since the existing convolution based frequency dependent line model (FDLM) is not suitable for the fault location problem, an improved convolution based FDLM is proposed. Second, to calculate the accurate voltage distributions along the line, improved curve fitting methods for the coefficients of the FDLM are proposed. Finally, to make the algorithm more practical, the convolution based time domain fault location algorithm is specially designed to greatly decrease the calculation burden. Simulation results prove the necessity of the improved curve fitting method, and demonstrate that the proposed method can accurately consider the frequency dependency of line. The proposed method shows much improved fault location results than the Bergeron model based methods in MMC-HVDC grids.