Effects from large-scale employment of model-predictive control in district heating substations

H. Håkansson, M. Önnheim, E. Gustavsson, M. Jirstrand. Kaiserslautern Applied and Industrial Mathematics Days (KLAIM2023), 25 - 27 September 2023, Abstract No. A20, p.33.


With the introduction of advanced methods for controlling the heat demand of district heating substations, such as model predictive control (MPC), different objectives of the control can be varied while still ensuring a comfortable indoor climate to very good precision. This new control regime increases the importance of the supplier’s price model, which implicitly dictates the incentives of the control and may thus affect the system heat load even at an hourly level. We study the system effects of varying implementations of price models by simulating optimal heat control for populations of dwellings and workplaces, where a comfortable climate is required only within certain hours for the latter. The results suggest that both the consumption and the peak load of the entire system may be reduced by going from traditional weather-compensated control to MPC, especially if incentives given by the pricing model are appropriately tuned. We also show that by employing one common MPC for multiple buildings simultaneously, the accumulated consumption in the system and the peak loads can be reduced even further with optimal scheduling.

Photo credits: Nic McPhee