Solar Energy Autopilot: A Digital Twin Framework for Autonomous Solar Thermal Plants

Kurzfassung

Concentrated solar thermal (CST) plants can provide dispatchable renewable heat for industrial and thermochemical processes but remain complex and costly to operate compared to fossil-based plants. Autonomous operation is a promising way to simplify CST integration, increase reliability, and reduce costs. At the German Aerospace Center (DLR), the Solar Energy Autopilot (SOLEA) has demonstrated the feasibility of autonomous CST operation using a holistic four-layer architecture and a multi-agent system (MAS) at the top automation layer.

In safety-critical plants, MAS-based control introduces transparency and connectivity challenges, as distributed agents often behave like black boxes and the agent set changes during the transition towards autonomy. To maintain oversight, ensure robust collaboration between agents and human operators, and support dynamic reconfiguration of responsibilities at runtime, DLR has integrated a digital twin framework into SOLEA. This framework closes transparency and connectivity gaps by enabling unified monitoring of agent behaviour and flexible human-agent and inter-agent collaboration.