Way less infiltration when Air Flow Network is used

Hi all

I have a question regarding the drop in amount of infiltration when an AFN is used.

This is the balance without AFN

Same Model with AFN

What is the reason for the change using AFN.
Is it because without AFN each room face has an infiltration calculated?
With AFN only the outside building faces are part of the air exchange?


Yes, usually the infiltration rate in AFN models are about an order of magnitude less than what is typically assumed in the “Infiltration rate per exterior area” input. There’s a lot of differences between the two that contribute to the discrepency, but the main reason is the vastly different approximations of pressure difference through the envelope that drives infiltration flow rates. The AFN dynamically computes the evolution of the envelope pressure difference through time, whereas the “Infiltration rate per exterior area” is based on empirical airtightness measurements of different buildings assuming a constant 4 Pa pressure difference across the envelope.

The problem is that the AFN pressure differences are usually quite different from the 4 Pa assumption, you can check this by outputting the AFN Linkage Node 1 to Node 2 Pressure Difference from the EP results. If they had the same pressure difference, then the “Infiltration rate per exterior area” value and AFN values would agree.

Alternatively, you can get HB to recalculate AFN crack/opening parameters to result in an infiltration value equal to the “Infiltration rate per exterior area”, by changing the latter’s assumption of 4 Pa to whatever the actual AFN is computing.That’s why there’s a _delta_pressure_ input in the AFN: you can add a envelope pressure difference that is closer to the version dynamically calculated from the AFN, which allows both methods to match.

That’s the simple explaination, there’s some more (optional) complexity in terms of deriving the _delta_pressure_, and a more thorough explaination about the underlying physics here: