Air Leaks Happen at the Surface, Not in the Volume
So why do we still report blower-door results in ‘air changes per hour’?
During the Westford Symposium on Building Science in 2010,* I was watching the tweets from the people who were there. At one point, I saw this one: “@EFL_Guy: ‘Air leaks through surfaces, not volume’ Joe Lstiburek.” I'd been meaning to blog about this issue for a while, so I wrote an article about it. Now, a couple of years later, it's time for a little update. Also, between my original article and this update, Lstiburek wrote an excellent article on what blower doors are good at: Just Right and Airtight. You'll want to read that article, too.
If you know nothing about the field of building science, Joe's statement above probably seems so obvious that you wonder why anyone would even utter it. Had it only been so obvious to a few more people back in the early days of testing buildings for infiltration, we wouldn't be in the mess we're in now.
How do we compare air leakage in homes of different sizes?
The reason we're talking about this, of course, is that we want to be able to compare the blower door test results for houses of different sizes. If I told you that I have two friends who both weigh 200 pounds, you'd have no idea if either is overweight. If I then tell you that one is 6'4" and the other is 5'2", you now have a better picture of the situation. The former is at a normal weight for his height, while the latter is obese. We can “normalize” the results by taking their heights into consideration.
We generally talk about the size of houses in terms of the square footage of conditioned floor area (CFA). Fortunately, no one decided to institutionalize a blower door metric using that quantity because, although leakage would scale with CFA, it doesn't yield an accurate picture. As you already know from the title of this article and from Joe's statement, air leakage occurs at the surface of the building enclosure, which doesn't keep the same proportion to the CFA for different houses. A 3,000-square-foot ranch house, for example, will have more enclosure area than a 3,000-square-foot colonial, because more of its floor and ceiling areas are part of the enclosure.
These days, everyone who learns how to use a blower door gets indoctrinated into the cult of the ACH (air changes per hour). You take the raw number from your test — cubic feet per minute (cfm) of air flow with a specified pressure difference between the house and the outside (usually 50 Pascals), also called the cfm50 — and you divide it by the volume. (You also multiply by 60 to change minutes to hours.)
So we normalize our blower door results using the home's volume for this metric.
Volume is the wrong quantity to normalize blower door results
Why volume?! It's perfectly obvious that if you want to normalize a number, you divide by something that makes sense. Would it make sense to measure the efficiency of your car in miles per square foot of gasoline? That's a similarly absurd ratio, because you pay for gallons of gasoline, and a given volume of gasoline can have different surface areas depending on the shape of the container. (That reminds me of the joke about the spherical cow.)
It's the same with houses. Just as two homes with the same CFA can have different surface areas, two houses with the same volume can have different surface areas. And, as Joe said, air leaks through the surface, not through the volume, so not only do we not care about volume when we're talking about air infiltration, it's misleading.
Normalizing to volume also builds in a bias toward larger homes. Since surface area is proportional to the square of the radius and volume is proportional to the cube of the radius, the volume increases faster than the surface area as a house grows in size. So, large houses benefit when dividing by volume instead of surface area.
You want a fair comparison
The whole reason to divide the cfm50 by anything is so that you can compare infiltration rates in different size houses. If you can't get a fair comparison between small houses and large houses if we use volume, why are we using volume?
Here's another example for you. Let's say you want to paint your house, so you go to Dummies.com, where they tell you that it takes about one gallon per 350 square feet of surface area you're going to paint. What if over at Elstupido.com, however, they told you that you'll need one gallon for each thousand cubic feet of volume? Whose advice are you going to take? Right! Dummies are better than El Stupidos. You're not painting the volume; you're painting the surface.
Air leaks into and out of the house through all the surfaces that make up the building enclosure: floors, ceilings, and walls. If you wanted to reduce your home's air leakage, you wouldn't seal the volume. You'd seal the leaks through the surfaces. (Just don't seal those weep holes at the bottom of your brick walls!)
An alternative to air changes per hour
As I commented in Martin Holladay's musing on blower doors a while back, there's an easy solution to this problem. Let's use the quantity that Southface uses in the EarthCraft House program: what they call the Envelope Leakage Ratio, or ELR (although I'd call it the Enclosure Leakage Ratio). Divide the cfm50 by the surface area of the building enclosure. It's simple. It makes sense.
Despite my protestations here, I do use ACH50 all the time. It's pretty well established in energy codes and programs, and it's probably not going away. Just understand that when you use ACH50, it's not the best metric and it's biased toward larger houses.
The quantity that I really dislike, however, is ACHnat, which supposedly takes your blower door result and then tells you how leaky the house will be under “natural” conditions, when the blower door is turned off. Next week I'll tackle that one.
* Yes, that was the one that inspired one of my most popular articles, I Don't Need No Stinkin' Building Science Summer Camp.
- Energy Vanguard
Wed, 12/05/2012 - 06:26
Wed, 12/05/2012 - 09:08
Wed, 12/05/2012 - 10:04
Wed, 12/05/2012 - 11:07
Wed, 12/05/2012 - 14:17
Wed, 12/05/2012 - 14:21
Wed, 12/05/2012 - 15:50
Wed, 12/05/2012 - 16:36
Wed, 12/05/2012 - 16:50
Thu, 12/06/2012 - 10:52
Fri, 12/07/2012 - 13:53
Tue, 12/11/2012 - 22:04