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

**Authors:** Maurizio Titz, Sebastian Pütz, Dirk Witthaut  
**arXiv:** [2307.12636](https://arxiv.org/abs/2307.12636)  
**Date:** 2023-07-24  
**Summary page:** https://www.summarizepaper.com/en/arxiv-id/2307.12636v1/

## Key points

- Transmission grid loads increase due to the installation of wind and solar power sources far from consumers
- System operators often need to intervene through countertrading or redispatch for grid stability
- Authors develop an explainable machine learning model to predict volume of redispatch and countertrade
- Wind power generation is a significant driver of congestion, along with hydropower and cross-border electricity trading
- Solar power does not have a mitigating effect on grid congestion
- Change in market design could alleviate congestion issues in the German transmission grid
- Research provides valuable insights into congestion management strategies during transition to sustainable energy sources
- Use of explainable AI technology highlights potential areas for improvement in market design.

## Summary

In their article titled "Identifying drivers and mitigators for congestion and redispatch in the German electric power system with explainable AI," Maurizio Titz, Sebastian Pütz, and Dirk Witthaut address the challenges faced by electric power systems during the transition to a sustainable energy supply. They highlight that transmission grid loads increase due to the installation of wind and solar power sources far away from consumers. To ensure grid stability, system operators often need to intervene through countertrading or redispatch. To provide insights into congestion in the German transmission grid, the authors develop an explainable machine learning model. This model predicts the volume of redispatch and countertrade on an hourly basis and identifies factors that drive or mitigate grid congestion. The study reveals that wind power generation is a significant driver of congestion as expected. Additionally, hydropower and cross-border electricity trading also play essential roles in contributing to congestion. Interestingly, the analysis shows that solar power does not have a mitigating effect on grid congestion. Based on these findings, the authors suggest that a change in market design could alleviate congestion issues in the German transmission grid. This research provides valuable data-driven insights into congestion management strategies in electric power systems during the transition to sustainable energy sources. Furthermore, it allows for a deeper understanding of factors influencing grid congestion through its use of explainable AI technology which highlights potential areas for improvement in market design.