Probabilistic generation of topology configurations for reliable expansion planning of distribution grids

The increasing integration of distributed energy resources (DERs) in low voltage grids presents new challenges for distribution system operators (DSOs), such as voltage imbalances and network congestion. To address these challenges, DSOs need to engage in proactive long-term expansion planning. Changes in the switch configuration of the network and lack of documentation contribute to the topology uncertainty in low voltage grids, which must be considered in the planning. To handle the uncertainty, we propose using a set of alternate switch configurations to make the planning process more robust and reliable. Multiple switch states (i.e. topologies) are generated using an efficient Markov Chain Monte Carlo (MCMC) approach and are then jointly analysed to find congestion hotspots. The congestion detection procedure relies on the overload probability of the examined assets (e.g. lines or substations) and on an approach to determine how to achieve the largest number of congestion-free topologies with less interventions. The approach is applied to a real 3032-bus low voltage network in Germany. The results demonstrate that considering multiple topologies can provide valuable insights for long-term planning, instead of using only one topology.