The effect of indoor air filtration on contaminant distribution and occupant exposure

Poor indoor air quality, driven by the presence of airborne contaminants, can have a significant effect on human health particularly given that we typically spend upwards of 80% of our time indoors. A potential solution for improving indoor air quality is to filter the air inside buildings, thereby removing contaminants from the air. There is a range of commercially available options, such as air-cleaning devices and filtered mechanical ventilation, that claim to improve air quality. Little scientific evidence exists, however, to support how the release of filtered air into a room can be optimised to reduce occupant exposure to contaminants. This thesis determines the key parameters that govern an individual’s exposure to contaminants and forms guidance for the users of air cleaners and mechanical ventilation to reduce exposure. By considering the case of a localised release of contaminant in a room for which the boundaries are treated as perfectly reflecting, the factors that primarily govern how contaminant disperses in a room are identified as: the characteristics of the filtrated flow supplied to the room, the geometry of the room, and the position of release of contaminant. Motivated by the need for rapid predictive capability and the success of simplified models used during the COVID-19 pandemic, a mathematical modelling approach is adopted based upon turbulent eddy diffusivity models and the theory of images to formulate solutions of the diffusion equation. An appropriate non-dimensionalisation based on physical arguments allows for a wide range of cases to be captured and flow of filtered air to be described by a single parameter. Surprisingly, despite similar models having been used in previous studies, this is the first to conduct a thorough fidelity study and justify the accuracy of the results by demonstrating the boundary conditions have been satisfied prior to examining the solutions and their implications. We therefore develop a rapid predictive capability for evaluating the concentration of contaminant and occupant exposure with confidence in the solutions used. We evaluate what may be considered two extremes of filtration strategy, first the ‘well- mixed strategy’ for which the release of filtered air rapidly mixes the air in the room to a uniform concentration of contaminant, and then the ‘passive-filtration strategy’ for which contaminant disperses from a localised release whilst being simultaneously filtered. Comparing the passive-filtration strategy to the well-mixed strategy reveals that one single filtration strategy does not best serve the entire room, and it is deduced that a different strategy should be adopted depending on the proximity of an occupant to the contaminant release position. Through the introduction of a novel concept, the ‘proximity threshold’, we establish where in a room an occupant’s exposure is reduced by employing one strategy or the other. Further to this, we demonstrate that a reduction in exposure can be achieved without increasing the filtration rate, but by selecting the most appropriate strategy. The reflections of contaminant by the boundaries of the room determine the spatial distribution of contaminant. Crucially, the relative bounding of the shortest to the longest boundaries determines the spatial variation in concentration, and therefore, exposure. As such, the geometry of the room plays a governing role in the exposure and ultimately guides the decision as to which filtration strategy the user of a device should select. The reflections of contaminant from boundaries dull concentration gradients leading the room ultimately to a well-mixed state. We demonstrate that the time for a localised source to be mixed throughout the entirety of the room based upon the geometry of the room. Releases of contaminant near the boundaries of a room lead to significantly different distributions compared to those near the centre of a room. If the contaminant is not released at the centre of the room, the position of maximum concentration moves from the release position towards a boundary of the room. We calculate a predictive relationship for the time scales of this behaviour and show it leads to the exposure on the near ‘boundary side’ side of a release being the higher. Evaluation of different release positions in the room reveal the unforeseen result that in highly occupied rooms, releases near the boundaries are likely to lead to a worse outcome for exposure than a release near the centre of the room. For the first time we have been able to offer objective, scientific-based advice on how to operate an air cleaning device (or mechanical ventilation) to reduce occupant exposure. Based upon the analysis developed herein we are able to advise the users of a portable air-cleaning device as to the optimum location within the room thereby improving indoor air quality based on scientific evidence. We reveal that reflections of contaminant from the boundaries of a room, which remarkably have been previously overlooked, govern the spatial distribution of contaminant in a room. The culmination of this research is the advancement of an analytical modelling approach beyond the simple well-mixed approach, which was adopted as standard in the prediction of exposure to COVID-19 during the recent pandemic. Application of the methodology and thresholding concepts developed herein have the potential to enhance the prediction of exposure to such viral aerosols as well as other airborne contaminants that pose a risk to human health.

Name dc:type.qualificationname

Doctor of Philosophy (PhD)

Level dc:type.qualificationlevel

Doctoral

Grantor dc:publisher.institution

University of Cambridge

Year dc:date.issued

2024

Author dc:creator

• Handy, Alice Ella Edith

Advisor dc:contributor.advisor

• Hunt, Gary

• Exposure
• Filtration
• Indoor air quality

Licence

• https://www.repository.cam.ac.uk/bitstreams/ec96af12-ab5b-456d-b4d5-bb7a46e59378/download
• https://www.rioxx.net/licenses/all-rights-reserved/

Language dc:language

eng

DOI dc:identifier.doi

https://doi.org/10.17863/CAM.112790

OAI identifier oai:identifier

oai:www.repository.cam.ac.uk:1810/374906