Search

The high‐resistance faults at the far end are the most important factor that causes protection failure of HVDC transmission lines. In the initial stage of the transmission lines fault, the transient information of power grid is abundant, and the sequential overlapping derivative (SOD) transform can emphasise the characteristics of the high‐frequency transient signal and its sudden change direction, amplify fault characteristics, and eliminate noise as far as possible, which is of great benefit to identify the high‐resistance faults at the far end. A fault identification method based on the SOD transform of voltage transient is proposed. In this method, the initial fault voltage travelling wave is processed by the SOD transform, and the initial fault voltage is a travelling wave is quickly detected. The set value of SOD transform is obtained according to the traversal simulation fault location, thus the identification criteria of internal and external faults are obtained. The simulation results show that the method can accurately identify the high‐resistance faults at the far end, and it is hardly affected by transition resistance and fault distance. [ABSTRACT FROM AUTHOR]

Reflected wave identification is the core of single‐ended travelling wave fault location, in which the obvious superimposed oscillation is the main obstacle for precise reflected wave identification with field data. This study proposes a reflected wave identification method based on speeded‐up robust features, which can gradually approximate from the possibly largest fault distance to the near‐end fault distance so as to exclude the reflected wave from time window containing initial wave. In particular, the time window segmentation divides initial wave and reflected wave into two inconsistent time windows, hence the reflected wave from the superimposed oscillation that usually follows the travelling wave can be effectively separated, such that the identification of reflected wave can be significantly simplified. Lastly, the superimposed oscillation is thoroughly analysed while both the field data with and without obvious superimposed oscillation are presented to evaluate the effectiveness of the proposed approach. Field data obtained from 21 substations scattered among Yunnan power grid of China validate the implementation feasibility of the proposed method. [ABSTRACT FROM AUTHOR]