Identifying drivers and mitigators for congestion and redispatch in the German electric power system with explainable AI

Wind power is the primary driver of grid congestion in Germany, but the study reveals this is only part of the story—hydropower and cross-border electricity trading also significantly influence congestion levels, suggesting solutions require addressing multiple factors rather than focusing on wind alone.

The explainable AI approach makes grid operations more transparent: By using machine learning that can be interpreted by non-experts, the researchers identified specific drivers and their quantified impacts on redispatch needs, demonstrating how AI can support better decision-making in critical infrastructure rather than operating as a "black box."

Solar power surprisingly offers no congestion relief, despite being a renewable energy source, which challenges assumptions about all clean energy sources having equal benefits for grid stability and suggests different renewable technologies require different management strategies.

Current market design may be inefficient: The findings suggest that changing how electricity markets operate could reduce the need for costly interventions like redispatch and countertrading, potentially lowering energy costs and system complexity as Germany transitions to renewable energy.

Data-driven analysis of grid operations can inform policy: This research demonstrates how analyzing real operational data with AI can reveal which factors actually matter for grid stability, providing evidence-based insights to guide infrastructure planning and regulatory decisions during the energy transition.