Linear programming control of a group of heat pumps

For a new district in the Dutch city Meppel, a hybrid energy concept is developed based on bio-gas co-generation. The generated electricity is used to power domestic heat pumps which supply thermal energy for domestic hot water and space heating demand of households. In this paper, we investigate direct control of the heat pumps by the utility and how the large-scale optimization problem that is created can be reduced significantly. Two different linear programming control methods (global MILP and time scale MILP) are presented. The latter solves large-scale optimization problems in considerably less computational time. For simulation purposes, data of household thermal demand is obtained from prediction models developed for this research. The control methods are compared with a reference control method resembling PI on/off control of each heat pump. The reference control results in a dynamic electricity consumption with many peak loads on the network, which indicates a high level of simultaneous running heat pumps at those times. Both methods of mix integer linear programming (MILP) control of the heat pumps lead to a much improved, almost flat electricity consumption profile. Both optimization control methods are equally able to minimize the maximum peak consumption of electric power by the heat pumps, but the time scale MILP method requires much less computational effort. Future work is dedicated on further development of optimized control of the heat pumps and the central CHP