Adaptive distribution topology learning on distributed source energisation and islanding

Abstract

Monitoring and controlling power sources in the distribution system can be challenging, especially when integrating distributed energy sources (DERs). The presence of multiple DERs introduces fluctuation and complexity, which can result in entangled power flow directions. The direction of power flow represents a pivotal signal in this study to understand the DER behaviours regarding their power injection and intermittent characteristics. Therefore, the paper introduces directional connectivity through graph analysis to tackle uncertainty from DER interconnections where islanding detection and restoration rely on acyclic and unidirectional energy flows. We propose the topology imbalance concept to manage directional power flow, loops, and interconnections. Our model employs phase signals to track topology changes and build grid structures without prior configuration information. The process is explored using radial subsystems with multi-directional energy supply scenarios. The findings demonstrate that the model can create diverse network configurations by integrating DER interconnections and islanding in steady state radial systems. The study explores the relationship between energisation and source injections, focusing on the back-feeding behaviour of DERs. Test results indicate index ranges of up to 198 % for imbalance and 179 % for energisation, reflecting the locations of DER sources and the energy injected.