Incomplete Air Mixing May Reduce the Efficiency of Commercial Buildings Behaving as Virtual Grid Batteries

Abstract

Commercial building Heating, Ventilation, and Air Conditioning (HVAC) systems can provide flexibility to the electric grid. In previous research, HVAC systems have been modeled as virtual batteries to align with models used to plan and operate the grid. However, real-world experiments show that HVAC fan power load shifting can be inefficient in terms of roundtrip efficiency, which most virtual battery models fail to capture. These models typically use average room temperature as the system's state of charge, ignoring other factors affecting HVAC power and energy consumption, such as airflow and mixing. In this paper, we use a building model to show that incomplete mixing of supply air may be one physical mechanism that explains the inefficiency of HVAC fan power load shifting. The model qualitatively matches experimental results better than models typically used to capture the energy impacts of HVAC fan power load shifting. Through simulations, we show that poorer mixing leads to worse virtual battery efficiencies. Although air mixing is unmeasured in practice, we show that closing the control loop around fan power measurements improves virtual battery performance without requiring air mixing measurements. In the best case, roundtrip efficiency improves from 75% to 99%. These results suggest inefficiency arises from unobserved internal dynamics rather than poor control design, and that readily-available fan power measurements could restore predictable performance.