A novel fault detection, classification and location scheme in non-homogenous MTDC transmission lines

In this paper, a protection scheme for fault detection, classification, and location in non-homogenous HVDC lines is presented based on the rate of change of voltage (ROCOV) signal attained from a single end by applying a wavelet-based signal processing tool named MODWT. The fault features extracted by MODWT are applied to a general regression neural network (GRNN) to estimate fault location. The performance of the proposed scheme is evaluated against different scenarios including fault parameter variations such as type, resistance, and location, processing parameter variations e.g. sampling frequency, noise, and ultimately topologic variations of non-homogenous transmission lines. Altogether, analyzing the results approves the efficacy of the proposed scheme in case of robustness, identification time, and noisy condition.