Mixed integer linear programming optimization of heat distribution within a local heating system

At substations of district heating systems, water is heated and transferred by pumps through piping networks to the buildings of each local heating system. When they were designed, the operational parameters and system characteristics of the local heating system were selected to provide thermal comfort in all the buildings served by the local heating system. After several years of operation, however, these parameters and characteristics change and, as a result, thermal comfort may be expected to deteriorate. This study investigates one strategy to mitigate this problem; that of retrofitting the local heating system with additional circulation pumps to provide greater system control and, thereby, improved thermal comfort. Using a steady-state, bottom-up approach, heat transfer in the local heating system is described by a system of equations that are then linearized. Mixed integer linear programming is then applied to these equations, implemented in the program PUMP, and used to optimize system characteristics. The results indicate that pump-placement optimization strategy is more effective in overcoming changes in hydraulic resistance than variations in radiator and building heat transmittance resulting from aging of the local heating system.