Author:

Abstract:

In space heating systems equipped with thermal energy storage, model predictive control can be used to support the electrical grid through active demand response. There are however several formulations possible for the underlying optimal control problem, the most accurate being non-convex and thus more difficult to solve as the control horizon lengthens. The current paper presents a non-convex and several simplified formulations for the general optimal control problem which are compared towards attained energy cost, constraint violation and computation time. A novel iterative linear program is proposed to increase controller robustness and to be used in dynamic simulations. The novel formulation results in equal overall energy costs while computation time is reduced by a factor 10 compared to the best performing nonconvex problem. This allows simulations for energy performance assessment including model predictive control to be executed on a modern personal computer.