Experimental Implementation of Frequency Regulation Services Using Commercial Buildings

Abstract

This paper illustrates the potential of commercial buildings to act as frequency reserves providers through an experimental demonstration conducted in a multi-zone university building. The proposed control methodology is presented in detail, including the control architecture, the controller design, model identification, and hardware description. Finally, the effectiveness of the presented approach is tested by means of simulations and experiments in a controlled environment.

Experimental demonstration of buildings providing frequency regulation services in the Swiss market

“…The contribution with respect to our previous work [28] is twofold: First, we provide a method to determine the amount of regulation that can be provided taking into account the possibility to adjust power consumption on the intraday market. Second, we show how to optimally trade regulation commitment and comfort in the building during real-time operations.…”

“…It is readily seen that the problem has multiple timescales interacting: when deciding the capacity bid and day-ahead purchase, a prediction over a minimum of one day is required, while during operation, fast rate tracking is required. Similarly to [27] and [28], the control architecture proposed is therefore hierarchical with three interacting layers as depicted in Figure 1:…”

“…In addition to experiments carried out during the day in occupied offices and for extended periods of time, the main difference in the formulation with our previous work [28] is the possibility to readjust the baseline in the intraday market. We will see that this feature is paramount to the robustness of the scheme.…”

“…The server collects temperature measurements from the web database, decides on the allocation of the AGC signal among heaters and communicates the power reference to the microcomputers, through Ethernet. More details regarding the experimental setup can be found in [28].…”

“…Our previous work [28] demonstrated the applicability of the methods for calculating regulation participation described in [29] and [21] by means of laboratory-scale experiments using electric heaters. It differed from other published works by computing the optimal amount of frequency service offered rather than finding it through trial and error procedures.…”