Optimal control of a fourth generation district heating network using an integrated non-linear model predictive controller

District heating (DH) can help increasing the share of renewable and residual energy sources in the heating sector. Usually a rule-based controller (RBC) is used for controlling these networks while more advanced control strategies like model predictive control (MPC) can support the transition to a decarbonised heating sector. In this paper, an integrated non-linear MPC approach for the control of DH networks is presented. The main novelties are the consideration of the system as a whole (including the heat demand side) and the inclusion of important non-linearities and all types of flexibility in the controller model of the MPC. A simulation-based methodology and dedicated solver are used in which the MPC is applied to an existing fourth generation DH network. The results of a three-day simulation in both a winter and spring period show that the MPC outperforms the currently used RBC: the thermal discomfort is lower in the winter period and comparable in the spring period while the electrical energy use of the DH system is reduced by 3% and 17%, respectively. The MPC achieves these results by lowering the network temperatures, using its anticipatory ability, and exploiting the flexibility of the building thermal inertia.