Load Flow–Based Probabilistic Power System Clustering Using Power System Distribution Factor

ABSTRACT Unlike conventional methods that minimise power flow between independent sub‐systems, this study proposes a novel load flow–based clustering methodology to segment the power system into source (generation‐rich) and sink (load‐rich) areas. The system is separated based on the iterative identification and removal of the maximum power flow lines (bottleneck lines), which is justified for planning and market applications. Furthermore, a probabilistic approach is applied to reflect future system variability. The probabilistic source/sink characteristics of the final clusters are then determined through the cluster aggregation probability between the buses. The clustering procedure is expressed mathematically using the power transfer and line outage distribution factors, and it helps simplify the iterative process in the Monte–Carlo simulation that is used for probabilistic analysis. To verify the feasibility of the proposed method, the simulation is conducted using modified IEEE 39‐bus systems.