(119a) Predictive Strategies for Coordinating Passive and Active Thermal Energy Storage in Buildings

Residential and Commercial buildings account for over 70% of the electricity demand in the US [1], and a significant portion of this is used towards maintaining indoor comfort via heating, ventilation, and air-conditioning (HVAC). Building electricity use fluctuates during the day and exhibits a peak in the afternoon. This peak is reflected in a demand maximum at the grid level, which is in general costly and inefficient for grid operators and energy producers to satisfy [2]. Ideally, this variation in energy demand on the grid should be addressed using energy storage. Buildings constitute an ideal candidate to this end, owing to their inherent capability to store thermal energy in their structural elements. This feature, which we refer to as “passive” thermal energy storage can be augmented using active TES systems consisting, e.g., of water or ice tanks [3].

TES systems can be charged at off-peak hours, when energy costs are low [4], and the stored energy can be released/discharged during peak hours. In this presentation, we focus on devising optimal strategies for coordinating charging and discharging events with the operation of the grid, while, on the other hand, meeting comfort constraints imposed by the building occupants. The dynamics of the storage systems are typically slow and are dictated by the thermal inertia of the building and, respectively, by the dynamics of the active storage capacity. On the other hand, the building is subject to relatively fast disturbances from changes in occupancy, insolation, and ambient temperature. Thus, coordinating the operation of buildings with TES is a multi-scale problem. Additional complications arise from the complexity of the dynamic models of each subsystem (building, storage tank, etc.) and from the need to account for discrete decisions related to the operating mode (charge, discharge) of the TES.

In this work, we propose a novel approach, where we decompose the optimal operation problem into (i) a scheduling problem for the active TES in a slow time scale, and (ii) a MPC with a shorter horizon in the fast time scale to address the use of passive TES [5]. We show that this proactive decision making framework is superior to heuristics for controlling the TES in the smart grid environment. Moreover, a Monte Carlo simulation demonstrates that, when implemented across many buildings, the proposed approach has a leveling effect on grid demand from HVAC systems. In a broader context, this leads to a reduction in the environmental impact of buildings by minimizing the on-time of inefficient “peaking plants” that are used to satisfy peak demand.

[1] U.S. Energy Information Administration, Annual Energy Outlook 2013

[2] J. Siirola, T. Edgar, Process energy systems: control, economic, and sustainability objectives, Comput. Chem. Eng. 47 (2012) 134–144.

[3] G. P. Henze, C. Felsmann, G. Knabe, Evaluation of optimal control for active and passive building thermal storage, Int. J. Therm. Sci. 43 (2) (2004) 173–183.

[4] M. H. Albadi and El-Saadany E. F. Demand response in electricity markets: An overview. Proc. IEEE Power Eng. Soc. Gen. Meet., pages 1–5, 2007.

[5] C. Touretzky, M. Baldea, Integrating scheduling and control for economic MPC of buildings with energy storage, J. Proc. Contr. , 24 (8) (2014) 1292-1300.