An Operation Reliability Analysis Method of Distribution Networks Based on Data and Physical Fusion Models

摘要

To ease the hardness of real-time reliability assessment on distribution networks, a data and physical fusion model based on Principal Components Analysis (PCA) and Random Matrix Theory (RMT) was proposed for the operation reliability analysis of distribution networks. Using real-time data collected from substation feeders and transformers, constructing time and space joint matrix, the inner statistical characteristics of data was obtained based on RMT and its augmented matrix technology. Then we used a multi-layer physical index system and PCA to obtain main physical indicators which senses high-risk parts. The K-Means clustering method was used to unite two models and establish the proposed operation reliability analysis method. An IEEE-33 simulation example proved the effectiveness of the method.