Agent-based distributed demand response in district heating systems

Current district heating systems are moving towards 4th generation district heating in which end-users play an active role in system operation. Research has shown that optimising buildings' heating demands can release network congestion and contribute to reducing primary energy usages. Coupled with end-user privacy concerns, an approach in which buildings jointly optimise their heating demands while preserving privacy needs to be investigated.

In view of this need, we have developed a distributed demand response approach based on exchange ADMM to support distributed agent-based heating demand optimisation for the district heating system with minimal private information exchanges. This paper summarises mathematical derivation, simulation and implementation of the proposed approach.

The results show that the proposed approach obtained the same results as its centralised counterpart proposed in the existing literature and the sensible information exchanges were substantially reduced. An implementation at multiple spatial scales and time scales on micro-controllers and a communication system validates the proposed approach in a practical context.

In conclusion, the proposed approach is suitable for real-world implementation in a large-scale district heating system.