Abstract
The fault location method based on the natural frequency of the fault travelling wave has outstanding superiority in high-voltage dc system, because of high operation speed, acceptable sampling rate, and so on. However, in flexible dc grid, the natural-frequency characteristic is quite different from that in conventional high-voltage dc system, meaning the existing fault location method cannot be directly used. This paper analyses the distance-frequency characteristics in different stages of fault period (from fault occurrence to fault isolation) in flexible dc grid, to find a suitable stage for fault location. On the basis of this, the improved fault location method based on clustering and iterating algorithms is proposed. Compared with the traditional method which directly uses the multiple signal classification algorithm to estimate the natural frequency for fault location, the proposed method can effectively eliminate the large error caused by signal dimension fluctuation and false spectrum peak of multiple signal classification algorithm. Finally, simulation cases are carried out to prove the feasibility and superiority of the proposed method.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have