Guoxuan Zhao / College of Electrical and Information Engineering Hunan University
Zijun Zhao / College of Electrical and Information Engineering Hunan University
This paper presents a protection scheme based on clustering analysis for wind farm collector. Since the short circuit current is related to the voltage sags, after the rotor Crowbar protection is put into operation during the short circuit fault of doubly-fed induction generator (DFIG). There are three components in stator fault current which are taken as the fault feature. The fault information under different operating states are collected to form historical sample set. The cluster center of historical fault data was calculated to analyze fault data under different fault conditions by using clustering algorithm. The real-time sample categories are analyzed by distance similarity to achieve protection judgment. A detailed model of wind farm is built in DIgSILENT. The simulation results show that this method can solve the problem of selective loss of coordination and speed-up of protection for collector line in wind farm.