Distribution network topology identification based on gradient boosting decision tree and attribute weighted naive Bayes

Abstract

Topology identification is important to ensure a safe and stable distribution network operation, especially in the case of high proportion of new energy access to the distribution network. It could provide structural information for distribution network management and the foundation of distribution network system analysis. In consideration of the influence of different power flow parameters on topology identification results, this paper proposes a topology identification method based on gradient boosting decision tree (GBDT) and attribute weighted naive Bayes (AWNB). Firstly, a gradient boosting decision tree was used to calculate the importance of different power flow parameters to reduce the data dimension and computational complexity. Then each attribute is given a weight based on the result of the importance calculation. Secondly, the weighted set was used as input to train the AWNB classifier to make the model more realistic. Finally, the IEEE 30-node model was used to verify the performance of the proposed method. The results indicate that the proposed method has high accuracy and good robustness, which could maintain a high success rate of topology identification under different noise environments.

Keywords

• Distribution network
• Topology identification
• Gradient boosting decision tree
• Attribute weighted naive Bayes