There are numerous papers and methods out there that include "flow per exterior area", "flow per exterior wall area", "air changes per hour", etc. There are also different recommended schedules with different fractions of maximum flow rate for when the supply fans are on/off. How do you know the best way to model infiltration in your building?
PNNL released a report that discusses this question in detail: "Infiltration Modeling Guidelines for Commercial Building Energy Analysis"
ArchEnergy has another paper that explored this topic further:
Determining which schedule to include is determined on if the space is positively pressurized by a mechanical system or if not. If it is positively pressurized the infiltration schedule should be reduced during those times.
Stack effect impacts infiltration for very tall buildings but for most building with a small amount of floors the key consideration is tightness of the shell and windows which can properly modeled using the flow per exterior wall area. Refer to either of the two links for further information.
The PNNL report is a good guideline, I have used the 0.12 cfm/ft2 number on a few projects and it seems to get reasonable results. The advantage of using a per ft2 number is that you can start out setting it as a default for all zones and only the exterior zones will correctly show infiltration. You can get into trouble if you do that with a ACH number. The PNNL paper also suggests a schedule where the infiltration multiplier is 1 when fans are off and 0.25 when fans are on.
I too recommend the PNNL report as a starting point. Keep in mind that even if you use an infiltration model that uses wind speed variation to modify infiltration rates, you are very rarely using the right wind speeds. While the temperature and solar insolation might not differ much from your building to the weather station, the instantaneous wind speed and direction will often change greatly. So, the infiltration using the detailed model won't be right. I feel (only "feel", I can't prove) that you are just as accurate, for monthly and yearly calcs, to use a static infiltration rate that is only modified by the pressurization.
For additional information search for or find a copy of a December 2001 ASHRAE Journal article by Wagdy A.Y. Anis, AIA titled " The Impact of Air Tightness of System Design." A couple of other resources are: NISTIR 7238 June 2005 titled "Investigation of the Impact of Commercial Building Envelope Airtightness on HVAC Energy Use," and a research paper titled "Multizone Airflow Models for Calculating Infiltration Rates in Commercial Reference Buildings" by Lisa C Ng, Amy Musser, et. al. dated 25 November 2012 published by Elsevier: http://www.elsevier.com/locate/enbuild.
Ralph, I beg to differ a bit. Although wind direction can certainly vary a lot, there are well-known curves for adjusting wind speeds depending on surface roughness and height above the ground. Especially for a high-rise building, I think it would be better to apply such an adjustment to the weather file wind speed and then use a physics-based infiltration model that adjusts for wind speed and temperature differences rather than a fixed ACH. The point is not to get it completely right, but to capture the general trends, such as less wind at night, more during the windy season (:-)), etc.
The other thought I want to toss out is that I've also heard recommendations that the infiltration rate should be set to 0, since commercial buildings are pressurized. That never made much sense to me, since the exfiltration has to be made up somehow, doesn't it? But ...
Joe, I absolutely agree with your comment about exfiltration needing to be made up. Pressurization can help reduce the required heating/cooling energy because you can use heat recovery to reduce the energy required to heat/cool the intake air, but if you set infiltration to zero, your system will never intake that extra air required to maintain pressurization and you will underestimate the intake air and the energy required to condition it.
The PNNL paper provides guidelines for EnergyPlus users and is applicable for most new commercial building modeling. The biggest challenge in modeling infiltration is that it is tempting to use this as a fudge factor when calibrating existing building energy modeling. If fan pressurization test data or other reliable ACH data is available, then I would recommend using these.
But you'd still have the issue that a constant infiltration rate does not reflect actual variation with weather and system operation. See my article in the July issue of ASHRAE Journal for calculating coefficients to use in EnergyPlus that varies infiltration with weather and system operation.
The Gowri approach (PNNL) is a wind only model and so clearly suspect especially in colder climates where stack effect is important. NIST has published an approach, AN IMPROVED METHOD OF MODELING INFILTRATION IN COMMERCIAL BUILDING ENERGY MODELS by Ng et. al. that purports to improve assuming E+ defaults and on the Gowri approach. It looks interesting but their method 2 which would be the practical way to implement the method is not totally convincing. I was looking for conversation on this approach and found this thread.
One issue with pressurization and infiltration reduction is that the pressure differential at the floor, when the building is warmer, is what matters. If the floor pressure is zero on all four sides of the building then infiltration is indeed zero. Substantial pressurization air can be added to a building without reaching this condition. In residential buildings, zero infiltration was found to occur only after fan flows were twice the infiltration flow that would occur without the fan. http://aceee.org/files/proceedings/19...
While I know blower door tests are the standard ways to look at infiltration on existing buildings I'm wondering if anyone has every tried or considered looking at thermal vs. pressure differential to study infiltration. The approach I'm thinking of would be to bring a number of air and surface temperature data loggers into a space or spaces for a few hours, ideally when the inside and outside temperatures have a big differential, and probably when the building is mostly unoccupied, and when there is little to no solar gain through windows. The approach would be to shut off the mechanical system and watch the temperatures float over a few hours, or however much time is needed. After data collection is done you would hard code the surface temperatures into the model, let infiltration float, and try to optimize to the recorded air temperatures.
Too much trouble? Not accurate enough to be useful? or will there be too many variables that still aren't controlled enough isolate infiltration? What originally got me thinking about this was hearing about some connected residential smart thermostats that would try to provide diagnostics if it saw a problem. For example if given the current weather conditions if the house wasn't holding temperatures as expected it may suggest a window might be open.
I think you wouldn't be able to isolate infiltration from thermal mass and conduction. A tracer gas test would accomplish what you are suggesting.
Best results probably can be achieved by computational fluid dynamics (CFD) simulations. In combination with EnergyPlus, CFD could be used for determining pressure coefficients around a building. How this can be done is demonstrated e.g. in a youtube video by ODS Engineering:
Asked: 2014-09-01 23:51:28 -0600
Seen: 3,259 times
Last updated: Mar 23 '18
