Data Analytics-Based Multi-Objective Particle Swarm Optimization for Determination of Congestion Thresholds in LV Networks

Abstract

A growing presence of distributed energy resources (DER) and the increasingly diverse nature of end users at low-voltage (LV) networks make the operation of these grids more and more challenging. Particularly, congestion and voltage management strategies for LV grids have usually been limited to some elemental criteria based on human experience, asset oversizing, or grid reinforcement. However, with the current massive deployment of sensors in modern LV grids, new approaches are feasible for distribution network assets operation. This article proposes a multi-objective particle swarm optimization (MOPSO) approach, combined with data analytics through affinity propagation clustering, for congestion threshold determination in LV grids. A real case study from the smart grid of Smartcity Malaga Living Lab is used to illustrate the proposed approach. Within this approach, distribution system operators (DSOs) can take decisions in order to prevent situations of risk or potential failure at LV grids.

Keywords

• congestion management
• low-voltage networks
• multi-objective particle swarm optimization
• affinity propagation clustering
• optimal congestion threshold