Observability analysis of distribution networks based on robust greedy algorithm

In this paper, a robust greedy algorithm is applied to select the node with the highest weight value for fault observation in distribution networks. Except for the first observation node, the weights of the remaining nodes are calculated based on the results of the last observation node. The scheduling cycle is divided into several consecutive periods, and the network operation structure is observed for each period, which is gradually merged to finally obtain a set of solution schemes that satisfy the constraints on the number of switching actions for dynamic reconfiguration collaboration. The results show that the line fault system measurability is all in the 81.8%-100% interval. The method in this paper can directly locate 64.3% of the line faults in the 10kV distribution network, which can significantly reduce the number of traveling wave-locating devices, lower the investment cost, and reduce the fault locating error. It has a certain degree of adaptability in the case of the line N-1. Network operation observation can be found that the highest load-interval in a day is from 5:00 to 18:00. Using the method of this paper reduces 1 division period compared with the traditional method, reconfigures the structure to reduce the number of switching actions, and reduces the network loss from 812.06kW·h to 805.69kW·h.