Data-Driven Methods for Efficient Operation of District Heating Systems

In this chapter, data-driven methods for the efficient operation of DHSs are described. DHSs are inherently non-linear and time-varying systems as the heating demand is highly influenced by non-linear dependencies on the weather conditions as well as the occupancy behaviour. Furthermore, the dependency on flow and temperature in delivering the needed heat demand using the district heating network gives a non-linear dependency on these two signals. This chapter presents several data-driven models to handle the non-linear and time-varying phenomena in order to ensure an efficient operation. First, we introduce forecasts that are used to reach an optimal operation as forecasts are needed for both control and production planning, e.g. heat demand and electricity price forecasts. Second, temperature control of a DHN will be introduced with a focus on how the physical characteristics of the network can be incorporated into a control scheme. A special focus will be on how to ensure that the temperatures in the network are high enough to ensure the needed heat supply for the attached buildings in the entire district heating network is met. We shall also briefly look at the role of smart buildings integrated into a DHN that can be used to enhance the efficiency and flexibility of a DHS.