Substation Lightning Early Warning Model Considering Spatiotemporal Clustering
Effective lightning protection of substation is an important part of power grid stability.Based on single lightning detection data,this paper divides it into lightning strike point sets of continuous periods,and obtains relatively concentrated lightning strike point sets in each period through the DBSCAN clustering algorithm.For the lightning strike point sets successfully clustered in two consecutive periods,calculate the moving direction and speed of the thundercloud in order to predict the future lightning strike area.When the predicted lightning strike area covers the substation,the staff will be warned in advance.The simulation results show that the early warning model in this paper works well,and the prediction time advance can reach more than ten minutes.When the time advance is within 10 minutes,it can almost achieve accurate prediction.Different from the current commonly used hardware lightning protection facilities,the early warning model in this paper can provide lightning warnings before lightning occurs,so that the staff can take emergency measures in advance.
NETL
NSTL
万方数据
