Improving Industrial Air Quality: Modeling, Monitoring, and Use of Low-Cost Sensors

REN GARITY, Jianing Bao, Andrew Metcalf, Clemson University

     Abstract Number: 341
     Working Group: Aerosol Science of Infectious Diseases: Lessons and Open Questions on Models, Transmission and Mitigation

Abstract
While there are several techniques for quantifying indoor air quality using low-cost sensors, there can be a disconnect between the results from measurements and from full-scale fluid modeling that accurately predicts indoor air quality over a range of operating parameters. This disconnect is especially common in industrial environments. Although monitoring may be more prevalent here than in other indoor spaces, such as classrooms, the measurements are often focused on individual exposures rather than on understanding how the overall facility is performing, leading to gaps in understanding.

To bridge this gap, we propose an air quality modeling, monitoring, and sampling plan that we then implement at an industrial test facility. With the combined use of short-term, high-resolution sampling, modeling, and monitoring, there is a high chance of being able to create a ventilation plan that is adaptable to the needs of a changing industrial facility. For example, revisions can be made, and the model reevaluated, if machinery is moved or added or if ventilation capacity is changed.

In this talk, we discuss the feasibility of this verified modeling approach. We first compare data from a prior indoor air quality study in classrooms at Clemson University to results from NIST’s CONTAM model. In the prior study, low-cost sensors were used to measure the spatial distribution of particulate matter. We discuss the challenges in reaching agreement between the measurements and the CONTAM model, owing to effects such as the accuracy of the low-cost sensors, ventilation rate, and filtration efficiency in the HVAC system. Finally, we expand our study to an industrial space, with new measurements and modeling to determine potential exposures to factory workers. Ultimately, this type of verified modeling approach will greatly benefit environmental, health, and safety professionals in the workplace who want to make data-driven decisions to protect workers’ health.