A market-clearing mechanism based on hierarchical-spectral clustering with multiple similarity measures for flexibility spot markets

Modern power systems face increasing congestion issues due to rising electricity demand and the integration of distributed energy resources. Flexibility markets offer a promising solution to deal with congestion by enabling system operators to incentivize consumption or production shifts through real-time spot trades. However, market mechanisms must maintain clearing tractability despite the expanding number of flexibility providers and the complexity of bid structures. This paper presents a hierarchical-spectral clustering approach with affinity matrix aggregation to improve market-clearing performance. The proposed method forms network-aware clusters by aggregating affinity matrices based on flexumers’ geographical location, electrical proximity, and behavioral preferences. The simulation results on the IEEE 33-bus and a large-scale distribution grid, based on the 9241-bus PEGASE test system, demonstrate a clear trade-off between computational burden and economic outcomes. Notably, the framework can reduce market-clearing times while maintaining acceptable levels of economic efficiency, providing valuable insights for the design of scalable flexibility markets.