Field Implementation of a Predictive Control Strategy in District Heating Systems: A Tale of Two Demonstration Sites
Published:
District heating systems have become increasingly complex by integrating even more efficient technologies to help decarbonize the built environment. However, the full potential of such systems has yet to be reached due to inadequate controls. Predictive control has emerged as a promising solution to leverage operational data, modelling capabilities and various forecasts (weather conditions, price signals, carbon intensity) to optimize district energy system operation in real time. This paper discusses practical hurdles and lessons learned from the implementation of an artificial intelligence (AI)-based model predictive control (MPC) strategy in two Canadian district heating systems. These systems are equipped with natural gas boilers, which supply space and water heating through a steam network. This AI-based MPC strategy builds upon district heating demand forecasting models and data-driven boiler performance curves to optimize boiler thermal outputs that minimize greenhouse gas emissions. Practical hurdles include the usual suspects – data collection and preparation, communication with the control system, equipment maintenance – but also unexpected aspects such as weather forecast access issues and partial application of the recommendations. Lessons learned deal with the adoption of the proposed strategy, the potential for performance improvement of multi-boiler district heating systems, and the scalability and generalization to more complex systems.
Recommended citation: Saloux, E., Runge, J., Zhang, K. (2024). Field Implementation of a Predictive Control Strategy in District Heating Systems: A Tale of Two Demonstration Sites. In: Jørgensen, B.N., da Silva, L.C.P., Ma, Z. (eds) Energy Informatics. EI.A 2023. Lecture Notes in Computer Science, vol 14468. Springer, Cham. https://doi.org/10.1007/978-3-031-48652-4_21.