Identification Method for Users-Transformer Relationship in Station Area Based on Local Selective Combination in Parallel Outlier Ensembles Algorithm

In the power distribution system, missing or incorrect file of users-transformer relationship (UTR) in low-voltage station area (LVSA) will affect lean management LVSA, and operation maintenance network. To effectively improve paper proposes an identification method for UTR based on Local Selective Combination Parallel Outlier Ensembles algorithm (LSCP). Firstly, voltage data is reconstructed information entropy to highlight differences between. Then, LSCP combines four base outlier detection algorithms, namely Isolation Forest (I-Forest), One-Class Support Vector Machine (OC-SVM), Copula-Based Detection (COPOD) Factor (LOF), construct model UTR. This can accurately detect users’ data, identify users with wrong Meanwhile, key input parameter determined automatically through line loss rate, influence artificial settings recognition accuracy be reduced. Finally, this verified actual LVSA where recall precision rates are 100% compared other methods. Furthermore, applicability LVSAs difficult acquisition error transmission analyzed. The proposed adopts ensemble learning framework does not need set threshold manually. And it applicable high similarity, which improves stability LVSA.