Conference Year

July 2018


Building-to-Grid, Ancillary services, Building flexibility, Demand management


The increased adoption of intermittent renewable energy, such as wind and solar, onto the electrical grid is increasing the need for greater demand flexibility and the development of more advanced demand management solutions. For example, in March 2017 solar and wind set record highs in California, contributing over 49% of its power supply. Furthermore, Hawaii has committed to meeting 100% of its electrical demand from renewables by 2045. This transformation requires solutions to robustly and cost-effectively manage dynamic changes on the grid while ensuring quality of service. Advanced demand response approaches are a key way of enabling this required grid flexibility. Advances in direct digital control of building systems, combined with the increased connectivity of end devices now enable greater participation. To achieve this, end devices will need to estimate the amount of grid services (flexibility) they can offer, and then automatically fulfil that commitment when called upon without noticeable loss in quality of service (e.g. indoor comfort). This paper presents data-driven methods for estimating the demand flexibility of commercial buildings and the control architecture to enable the execution of committed reserves while ensuring quality of service. In particular, we describe the methodology for 1) qualifying the HVAC system to provide three power grid ancillary services (frequency response, frequency regulation and ramping services) based on defined metrics for response and ramp time, 2) quantifying the magnitude and frequency bandwidth of the service it can provide, and 3) controlling the building’s cooling and heating demand within the specified flexibility limits to provide grid service. UTRC’s high performance building test-bed, a medium-sized commercial office building was used for the experimental study. The building testing was focused on the air-side electricity consumer - the supply air fans in the AHU. The resulting data verifies that air-side HVAC loads (ventilation fans) are sufficiently responsive to meet the requirements of frequency regulation (<5 seconds response time) and ramping services (<10 minutes response time) with ON/OFF control command, direct fan speed control, and indirect control through static pressure set-point adjustment. The proposed frequency regulation control changes the command to the AHU fan motor speed (and hence power consumption) by indirectly modifying the duct static pressure set-point to track a given regulation reference signal. This architecture was selected for equipment reliability and ease of implementation. The experimental frequency response data from static pressure set-point to AHU fan power consumption shows that each ventilation fan can provide up to 1.5 kW for frequency regulation (16.7% of its rated power) during operational hours without impacting the indoor climate or baseline controls, and the acceptable frequency range was identified as 0.0055 - 0.022 Hz based on the grid response metrics and controls requirement. The accuracy of the flexibility estimation and the performance of the frequency regulation controller were verified through closed-loop active response experiment. Moreover, we describe how a population of commercial buildings with different flexibilities can be engaged and coordinated to provide adequate and reliable frequency regulation service to the grid.