Analysis Results of Simulating KIT.TC1.TS1 FlexOffice Thermal MPC -title

Analysis Results of Simulating KIT.TC1.TS1 FlexOffice Thermal MPC

Analysis Results of Simulating KIT.TC1.TS1 FlexOffice Thermal MPC -related configuration

Analysis Results -- Related Configurations

Author / organization: A. Engelmann and T. Faulwasser / KIT / KIT

Related Simulation Results:

Analysis Results of Simulating KIT.TC2.TS1 (Electrical Peak Shaving BESS Sizing) - related documents

Analysis Results of Simulating KIT.TC1.TS1 FlexOffice Thermal MPC - result description

Analysis Result Description

We evaluated the results from the controlled FlexOffice along three criteria: The comfort level which means “how close” the room temperature stays at 21°C, the energy cost and the total energy demand. Furthermore, we evaluated the differences in these criteria with and without a concrete core activation system.

The simulation results can be summarized as follows: One of our first observations was that when model predictive controller (MPC) is used for temperature tracking with a corresponding tracking objective, there is almost no difference compared to a classical PI controller in all target metrics mentioned before. This is not surprising given the close relationship between MPC and LQR (Linear Quadratic Regulator) controllers in absence of state/input constraints.

When using a weighted sum of the before mentioned KPIs as stage cost for MPC, a different MPC system behavior can be induced. By choosing the weighting coefficients appropriately, one can lower energy demand or energy cost. Our simulation results furthermore indicate that these goals are conflicting, e.g. a low energy demand also leads to a low thermal comfort in general. As an additional result, we get that a concrete core activation system can improve the “load shifting capabilities” of the building. This means, that it is way easier to store energy in walls and ceilings then directly in the room temperature, as the energy from the concrete core dissipates only slowly into the room, i.e. one can “charge” the concrete core with very high powers when energy is cheap (which is not easily possible without CCA).

A detailed investigation of these phenomena including numerical data can be found in

P. Zwickel, A. Engelmann, L. Gröll, V. Hagenmeyer, D. Sauer, T. Faulwasser, A Comparison of Economic MPC Formulations for Thermal Building Control, IEEE PES Innovative Smart Grid Conference 2019, (2019).

As the corresponding conference proceedings are not appeared yet, we provide a preliminary version in the appendix to this document.