Real-world performance of HVAC filters/air cleaners in commercial and institutional buildings

Most filters are evaluated under controlled laboratory conditions for a short duration (e.g., ASHRAE Standard 52.2). Although these tests are useful for standardization and product specification, the actual performance of any filtration product depends on building characteristics, HVAC operation, filter age, and other dynamic factors that are impossible to standardize because they differ across environments and vary over time within the same environment. This study evaluated the real-world performance of four HVAC filter types (MERV 8, MERV 8 electret, MERV 13 electret, and active polarization) and a portable air cleaner across four commercial and five institutional air handling units using low-cost monitors. Each filter was tested for a week in each environment to capture a variety of real-world dynamic factors. Filtration performance was estimated based on the absolute PM reduction, comparing indoor PM2.5 concentrations between a screen (a very low-efficiency filter) and each filter/air cleaner during occupied periods after adjusting for covariates such as outdoor PM in an analysis of covariance (ANCOVA) model. The amount of PM removed by a filter varied from one environment to another depending on space volume and particle sources and loss rates. For example, absolute PM2.5 reduction ranged from 0.34–3.87 µg/m3 for MERV 8 and 2.29–9.99 µg/m3 for MERV 13 electret filters, with larger differences between filters observed in environments with substantial indoor or outdoor PM sources. Although higher-efficiency filters caused 1.1 to 1.8 times more pressure drop, airflow rate through the HVAC system often remained similar across filters, likely because the adjustable fan controls compensated to maintain consistent airflow. Overall, this study contributes to the development of an in-situ test approach that will help facilitate the adoption, use, and evaluation of real-world performance for air cleaning devices across diverse built environments.