Efficient model-based DC fault detection and location scheme for multi-terminal HVDC systems with voltage source converter

This paper introduces an efficient model-based DC fault detection and location scheme for voltage source converter (VSC)-based multi-terminal high voltage DC (MTHVDC) systems. The main idea of the proposed approach is to use the difference signal between the real and estimated currents of the HVDC line as the signature to detect the faulty conditions. The state-space model of the HVDC line is established and used together with the Kalman filter (KF) algorithm for optimal estimation of the DC current at the sending end of the line. The observability of the developed state-space model is analytically proved. The developed model is also used to locate the fault point by post-processing of the measured data with the least-squares method. The main superiorities of the proposed approach are its high accuracy and excellent robustness against measurement noises/errors. These features are demonstrated through extensive MATLAB simulations of a typical three-terminal radial MTHVDC system. Likewise, the real-time feasibility of the proposed approach is guaranteed by some processor-in-the-loop tests. The results show that the proposed method achieves excellent fault detection and location performance, while the measurements are contaminated with noises as strong as 30 dB.