Computational Fluid Dynamic Analysis of Covid Clean Isolation Room

Airflow analysis of a 12’width x 8’length x 7’height

Temporary Negative Pressure Isolation Unit with air-lock.

These simulations were run by the lead engineer for the

air-intake of heavy equipment engines for a multi-billion

multi-national corporation.


To access the parameters of air-flow rate, air-flow pattern, and negative pressure within our temporary negative pressure isolation (TNPI) unit we conducted several rounds of computational fluid dynamic analysis (CFD). We wanted to know what effect the design of the air-purifier (AP), filters, placement of vents, fan size, and flow would have on these parameters and the system. Our findings were interesting and suggest if we desire further refinement more analysis is needed. We conducted two tests; the first with an 8” inline fan with a flow of 800CFM, and the second with a 6” inline fan of the same design with 400CFM.


The model used an older AP design we started with but later redesigned however the most important effects happen within the TNPI so the model is just as valid. The model was built with the smaller and likely more commonly implemented 12’x8’x7’ model with a small airlock. The cones before the AP were put into place to eliminate extra un-needed variables. The supply was directed into both the main room and the air-lock while the exhaust was run to the opposite corner of the room near the patient’s head. We placed a bed in the simulation to recreate as accurate a result as we could. The filter restriction was placed into the model as well.


The 8” 800 CFM simulation showed us that a similar amount of air enters both the isolation room and the air-lock at the same rate. The flow of air in the isolation room cuts a path across the room and travels along with the patient in a highly ideal way. The negative pressure was many times higher than required what is considered ideal with a reading of 0.8 inH20. The air changes per hour (ACH) were 56 where the minimum required is 12 ACH. All of these readings were pleasantly surprising to us but overkill. We decided to dial down the power and see what resulted.

For the second test, we used a 6” inline fan with a CFM of 400 and running the fan on only the exhaust. We observed more uniform air distribution which has its pros and cons. The pro is that air is cleaned faster while the con is that dirty air will circle around anyone in the room and being upstream of the source is not possible. A recommended solution was to run the larger 8” fan at a lower level then turn up the power during staff entry.


We will need to run several more CFD and make a map of all the visual results to see how the system is likely to react to these changes.