Model Predictive Control Formulation: A Review with Focus on Hybrid GEOTABS Buildings

Model predictive control (MPC) has been demonstrated to be a potential method for optimal control applied to building energy systems that provides both energy savings and enhanced thermal comfort compared to the traditional rule-based controllers (RBC). Nevertheless, an appropriate and correct formulation of the optimization problem is necessary to achieve this improved controller performance. One of the key aspects of this formulation is the objective/cost function, which usually consists of multiple objectives, typically energy use and thermal discomfort. Depending on the objective formulation, the trade-off between energy savings and occupants’ thermal comfort will be affected. Furthermore, the objective function should match the final user’s preferences, e.g. a “greener” user will prefer to sacrifice some money to reduce the building’s carbon footprint. To this end knowledge of the primary energy sources used allows converting energy use into CO2 emissions, which could represent another objective to be minimized. When smart meters are available higher economic benefits can be realized by including dynamic prices in the objective function. Moreover, some MPC formulations tend to have a non-robust (bang-bang) behavior which can be avoided by modifying the objective formulation. Furthermore, thermal ground balance (which affects geothermal heat pump performance) can be guaranteed by including a penalization for heat injection-extraction imbalance or ground temperature drift in the objective function.status: publishe