Search

Aiming at the practical problem of the difficulty in identifying associated users when abnormal line loss occurs in the station area, a method for judging and accurately locating the line loss abnormality in the station area based on the correlation measurement algorithm is proposed. Firstly, the optimal clustering number of the data set is determined by the gap statistics-contour coefficient fusion algorithm, and on this basis, the dichotomous K-means++ is used to construct the station area line loss standard library. Secondly, the station area line loss anomaly identification is completed based on the standard library and then the abnormal time is determined. The Spearman correlation coefficient (SCC) and Euclidean-discrete Frand#233;chet distance (E-DFD) of each user's power and line loss during the abnormal time is calculated. And based on SCC and E-DFD, a comprehensive evaluation index to analyze user relevance is estabilished. Finally, the technique for order preference by similarity to an ideal solution (TOPSIS) is used to sort the comprehensive evaluation index values to achieve precise positioning of abnormally associated users. The calculation example uses real field data in a certain area to analyze, and the results show that the method proposed in this paper has better performance in clustering effectiveness, calculation time, and identification accuracy.

In order to improve the recognition accuracy of different voltage sag disturbance sources and effectively control the voltage sag, a method of voltage sag source identification based on beetle antennae search (BAS)-back propagation (BP) classifier model constructed by longicorn BAS and BP neural network is proposed. In this paper, the improved S-transform is used to extract 16 characteristic indicators to form a voltage sag source identification indicator system. In order to eliminate the influence of redundant information on the classification results, 9 indicators are selected as the input of the classifier using the combination weighting method. By optimizing the initial weights and thresholds of BP neural network by BAS, the BAS-BP classifier model is constructed to identify different types of voltage sag sources in distribution network. The simulation results show that the classifier model has certain anti-noise ability and applicability, and has a better classification than the conventional classifier model dose.