PLC-Enabled Low Voltage Distribution Network Topology Monitoring

Keeping track of low voltage (LV) topology has always been a challenge; tracking the connecrivity would traditionally involve relying on field reports, which is both expensive and error prone. Recent developments in the area tend to either be costly to implement or assume a lack of unregistered consumers. In this paper, the authors propose a novel, robust method of tracking and identifying LV topology changes in power distribution networks containing unmetered loads. The approach employs machine learning (ML). The performance of the method is demonstrated using energy and topology data from an actual advanced metering infrastructure (AMI) located in Toruń, Poland.