Addressing Reviewers Comments: Explaining Voltage Control Decisions: A Scenario-Based Approach in Deep Reinforcement Learning

The electricity distribution system is evolving with increased electrification in transportation and heating, and the integration of distributed energy resources. This affects power quality in low-voltage networks. While Deep Reinforcement Learning (DRL) has been explored for voltage control, these methods lack the explainability, which reduces or even prevents their practical use. We introduce a Scenario-Based explanation (SBX) method to clarify the actions of the DRL agent in voltage control, enhancing operator understanding and ensuring system safety and reliability. Our method is based on the identification of the prototypical trajectories and the identified scenarios reflect comprehensible scenarios of the DRL agent behaviour in network management.