Effects on district heating networks by introducing demand side economic model predictive control

Using economic model predictive control in space heating systems, the heat demand can flexibly be adapted to varying scenarios of conflicting objectives such as thermal comfort, energy consumption, and peak demand. With a controller that minimizes the heating costs subject to thermal comfort constraints within occupancy hours, the resulting heat demand will depend on the cost mechanism. In the context of Swedish district heating networks, optimal demand side control is a multi-objective problem due to variable costs based on both energy consumption and peak demand. By simulating heat demand control using a gray-box model estimated from a Swedish space heating system, we investigate how the established price structures influence the heat load of a population of buildings with economic model predictive control. Our results suggest that by adjusting the incentives from the type of price structures commonly used today, the peak demand can often be reduced by 10-20% with a minor increase in consumption of 1-2%. We also show that by charging the peak demand for multiple buildings collectively, it is financially beneficial to cooperatively control buildings which can reduce the combined consumption and peak demand even further.