Infiltration
Infiltration is the unintentional air leakage into a building, which comes thru leaks thru small cracks in the walls, ceilings and floors (like plywood joints, but in fact leaks are almost all at the joints of two materials), places like fan vents, and seals around windows and doors. Like all air movement, infiltration only occurs when there is both an opening and a pressure difference. The amount of leakage is dependent on the area of these holes and the difference in pressure between outside and inside, which is due to three factors:
- Wind: wind raises the pressure on the windward side of the house, and lowers the pressure on the leeward side, resulting in air being both pushed and pulled thru the building.
- Stack effect: when the outdoor air is cooler than the indoor air, the indoor air wants to rise thru the ceiling via convective pressure. The greater the temperature difference, and the taller the building, the greater the pressure. At moderate temperatures, the stack effect is generally smaller than the effect of wind. At colder temperatures, the stack effect is greater unless the wind is blowing hard.
- Mechanical ventilation: typically this is exhaust fans such as dryer vents, range hoods, and especially downdraft cook tops force air outside without bringing any back in, resulting in lower pressure inside than outside, which in turn, sucks air in from outside. Likewise any supply ventilation would do the same thing, but these are less common. Likewise recirculation ducts in heating or cooling systems that leak into unconditioned space (for example a heating duct in an unheated attic) can act as exhaust fans.
In practical terms, there is almost always openings and there is almost always a pressure difference due to weather conditions, and there may be one due to mechanical ventilation as well. In fact, mechanical ventilation can make the air pressure in a house either higher or lower than outside.
Whether the air enters the building via infiltration or ventilation there is an energy penalty because the air from outside now has to be brought to room temperature. Obviously the penalty is bigger when the outside temperature is not close to indoor temperature. In the case of ventilation, the rate of air movement is essentially a constant, determined by the characteristics of the fan. If the fan is also and HRV or ERV, the energy penalty is less because the temperature of the incoming air is closer to room temperature. In the case of infiltration, the quantity of air entering varies with weather and the characteristics of the building. When its 60F outside and the wind speed is less than 10mph, the infiltration amount will be almost nothing, but when its 20F out and the wind is blowing over 15mph, there will be quite of bit of infiltration.
Measuring infiltration
Because the conditions that cause infiltration vary significantly, infiltration is measured a specific pressure difference via a device called a Blower Door. With this device placed in an exterior doorway, the house is depressurized to be 50 Pascals less than outside (which is the equivalent of a steady 20mph wind blowing at all sides of the building). Once this pressure is achieved, the devices measures the airflow it needed to produce this pressure, which of course is the same as the airflow leaking into the house thru all of its various cracks. The test is typically also run with the house pressurized to 50 Pascals, and the average is taken. The two results aren't the same because holes are often not pure holes, but places that are partially blocked. For example, a hole covered in housewrap might be quite tight under negative pressure, but leak quite a bit when the house is pressurized--in one case the housewrap is pulled tight over the hole, in the other it is pushed away.
Note on Nomenclature
The term CFM50 refers to the air leakage number in cubic feet per minute as measured at 50 pascals. The ACH50 number is just the CFM50 number converted to a ratio based on the volume of the house and done hourly instead of by the minute. If V is the volume of the house, then:
ACH50=CFM50*60/V
So when you see 3ACH50, that means that the building has 3 air changes per hour at 50 pascals; likewise when you see 100CFM50 it means that the building leaks 100CFM at 50 pascals.
Occasionally air leakage is measured at a pressure of other than 50 pascals, and so if the pressure used was 30 pascals instead, the values would be written as CFM30 and ACH30. You will rarely see any number besides CFM50 mentioned.
When these numbers are converted to the average, the value is referred to as the natural ventilation, and usually designated by ACHnat or CFMnat. This average is over the entire year, and although somewhat useful in determining whether a building needs mechanical ventilation, its not necessarily a good measure of yearly heat loss: it essentially depends on whether there are more windy cold days than windy "not so cold" days.
In addition to CFM and ACH, leakage is also sometimes described by an equivalent leakage area, which is the effective size of the hole. While useful for visualization, it is not used in calculations.
This measured value, CFM50, can then be converted to an average amount, called ACHnat, over the course of the year by using a statistical technique that takes into account local weather, site exposure and building height, so for example buildings in windier climates will have higher numbers. While the technique is beyond the scope of this document, its result is simple: you look up a fudge factor for your location and building height in a table, and divide the blower door result by that number. Initially, the factor was 20 for the whole US, then a more sophisticated model was created such that the factor varies with location somewhat, with warmer, less windy location having a lower factor than colder, windier ones. Currently there are 4 zones in the US, with fudge factors that range from a low or 14 to a high of 26.1 See this from the Jan/Feb 1994 issue of Home Energy magazine for background on this conversion.
Unfortunately this number, being an average doesn't help us with knowing whether our required ventilation amount is met, nor does it necessarily give us a good value to use in our heat loss calculation. To do either will require some guessing.
If your goal is an estimate of the annual heating use, ACHnat is probably a reasonable estimate, although considering that the heat loss when its cold is greater than when its not cold, using the average infiltration rate will result in a larger error on cold days than on moderate days. For example assume on cold days (20F, ΔT=50F) that infiltration rate is CFM50/15 and on moderate days (40F,ΔT=30F) its CFM50/25, and CFMnat is CFM50/20, ie the average of the two. First, the error for the cold case is larger than the error for the moderate case, because CFM50/25-CMF50/20 will always be a smaller number than CFM50/20-CFM50/15. For example, assume as CFM50 value of 400 --that 2ACH50 for a 1500ft2 house. So now 400/15=26.67, 400/20=20, 400/25=16, so in the cold case the error is 26.67-20=6.67, and on the moderate case, the error is 20-16=4. Since the heat loss is function of ΔT, the error is now compounded, ie in the cold case we're off by 6.67*50, while in the warm case were off in the opposite direction by 4*30. Clearly they don't balance out unless the ACHnat value is closer to what happens on a cold day than it is on a moderate day. This doesn't mean that using ACHnat isn't a reasonable estimate, only that is its also reasonably likely to be inaccurate. As always, caveat emptor, and needless to say if you're calculating the typical worst case heat loss, ACHnat is likely not a good value.
Comparing Tightness
The CFM50 number that the blower door measures is the absolute air leakage number. When we convert that number to the equivalent ACH50 number, that value is a relative leakage amount in that it is now adjusted for the size of the house. For example, a 1000ft2 building measured at 2ACH50 and a 3000ft2 building also at 2ACH50 both replace all their air twice an hour at 50 pascals pressure, so hence have the same relative leakage rate. However, the 3000ft2 building still have three times the absolute leakage rate (800CFM50) as the 1000ft2 building (267CFM50) because its three times as large2.
A building that is less than 1ACH50 is considered extremely tight, one that is less than 2ACH50 is tight, one that is less than 3ACH50 is moderately tight, 3-5ACH50 would be typical, and anything more than 5ACH50 is leaky. Its not that unusual for older building to be 10ACH503, or even to be so leaky that the blower door can't pump enough air to reach 50 Pascals. Of course what is tight, very tight etc is pretty subjective, and not everyone will agree on the classification here.
How tight should it be?
This depends on a lot of factors. Tightening up until near the point your need mechanical ventilation clearly saves energy and after that it depends on how you provide that mechanical ventilation and how the ventilation is controlled. If the mechanical ventilation is provided by a fan, the heat loss per CFM is the same as for infiltration, so the energy savings only comes if the tighter building's total CFM is on average less, which then depends largely on how the fan is controlled. If the mechanical ventilation is via HRV, then tightening saves energy although in moderate climates the savings is not large. The following is comparison of various levels of tightness for various weather conditions.
As an example, consider a 1500SF (12,000ft3) house built to four different tightness levels: a loose house at 7.5ACH50 (which much better than most houses build before 1980, but not quite as good those built in the last 15 years), a moderately tight house at 2.5ACH50, and a very tight house at 1.5ACH50, and super tight at .6ACH50. Since infiltration rates measured in CFM vary with house size, larger or smaller homes will result in different comparisons, although if the number of occupants goes up with house size, the required ventilation rate will go up accordingly and the comparison is then very similar for all house sizes.
Because we want to compare these homes during different weather scenarios, the standard LBL model for converting the measured infiltration rate (for background, read this from Home Energy) into a "natural" rate (ie under actual weather conditions) averages over all weather conditions, so the charts are constructed using a range of conversions: 1/20 for moderate weather (40F), 1/15 for cold weather (20F), and 1/10 for very cold weather (0F).
For mechanical ventilation, the assumption is that an 80CFM fan uses 25W, and a 100CFM HRV uses 100W and is 80% efficient. A required total ventilation rate of 60CFM is assumed (as per the standard: that's 45CFM for the 1500SF, and 15 CFM for two occupants) and the ventilation rate is constant over the heating season, and determined by how much is needed during moderate weather. This means that the building is over-ventilated during cold weather. If there were better controllers the ventilation rate would vary with the infiltration rate so that the total amount of fresh air was constant, but this is currently not an option.
All of these are all ballpark numbers and some have been rounded to simplify calculations. Needless to say, if you make different assumptions, you will get somewhat different results.
| House type | Infilt. rate (cfm) | Infil. loss (btu/hr) | Vent rate | Fan energy (btu/hr) | Exhaust heat loss (btu/hr) | HRV energy (btu/hr) | HRV exhaust loss (btu/hr) | Total loss - Fan | Total Loss - HRV |
| Very Cold (0F, ACHnat=1/10 ACH50) | |||||||||
| Super tight | 12 | 907 | 54 | 58 | 4082 | 184 | 816 | 5047 | 1908 |
| Very tight | 30 | 2268 | 45 | 48 | 3402 | 154 | 680 | 5718 | 3102 |
| Tight | 50 | 3780 | 35 | 37 | 2646 | 119 | 529 | 6463 | 4429 |
| Leaky | 150 | 11340 | 0 | 0 | 0 | 0 | 0 | 11340 | 11340 |
| Cold (20F, ACHnat=1/15 ACH50) | |||||||||
| Super tight | 8 | 432 | 54 | 58 | 2916 | 184 | 583 | 3406 | 1200 |
| Very tight | 20 | 1080 | 45 | 48 | 2430 | 154 | 486 | 3558 | 1720 |
| Tight | 33 | 1800 | 35 | 37 | 1890 | 119 | 378 | 3727 | 2297 |
| Leaky | 100 | 5400 | 0 | 0 | 0 | 0 | 0 | 5400 | 5400 |
| Moderate (40F, ACHnat=1/20 ACH50) | |||||||||
| Super tight | 6 | 194 | 54 | 58 | 1750 | 184 | 350 | 2002 | 729 |
| Very tight | 15 | 486 | 45 | 48 | 1458 | 154 | 292 | 1992 | 931 |
| Tight | 25 | 810 | 35 | 37 | 1134 | 119 | 227 | 1981 | 1156 |
| Leaky | 75 | 2430 | 0 | 0 | 0 | 0 | 0 | 2430 | 2430 |
The first obvious conclusion from the chart is that building tight saves energy up to a point (the case labeled tight, which is 2.5ACH50), but that once you start using mechanical ventilation, if you use a fan, then further tightening results in very little savings until its very cold outside, and even then the additional savings is small. This conclusion is encapsulated in the common saying in green building "Built tight, ventilate right".
The second obvious conclusion is that using an HRV generally saves energy because the reduced heat loss versus a fan more than compensates the additional fan energy used. In fact this is the case even if you use double the fan energy (ie 200W) than was assumed. The tighter the house and the colder the weather, the greater the HRV savings. However, HRVs are often wired to run 24/7 and at higher than necessary ventilation rates. Needless to say, this is not a good way to save energy.
The third obvious conclusion is that going to super-tight saves more energy when its very cold then when its not.
What is not obvious is that if the fan/HRV runs all year and passive ventilation would suffice for some sizable chunk of the year (say 4-6 month from spring till fall), then on a yearly basis the equation is changed because your using energy for those fans and getting no value all. The cost is significant: for the super-tight (passive house) case you're using 235btu/hr times 24 hours times say 20 weeks. That's a about 800kbtu/year, and add to that any cooling energy needed if the air brought in is too warm. While not a big value compared to typical overall energy use, its not a small value either, so clearly you want to turn off the system whenever you can open windows.
Note as the building gets tighter, the energy use due to fans is greater than infiltration energy. This makes sense because we've substituted mechanical ventilation for infiltration.
A larger house (say 3000SF, ie double this one), would have double the infiltration rate, while a 750SF house would have half the infiltration rate. As a result, the larger house doesn't need ventilation unless its tighter or there are proportionally more occupants while the smaller house starts needing ventilation even when its not that tight, and hence it make more sense to tighten a large house than a small one, at least if your concern is absolute energy use rather than relative.
Really, what can you get?
Ignoring the above theoretical exercise, getting under 3ACH50 mostly involves just foaming penetrations, around windows, and maybe some joints in the framing. To get down to 2ACH50 you just do more of the same, but to get below that you have to either use air-tight drywall or tape all sheathing seams, or use a liquid applied coating. To get lower than 1ACH50 you have to do all of the above, and also reduce the number of penetrations to the exterior, including vent fans. For more details see the construction section
Resources
For more on blower door tests, see
http://www.energysavers.gov/your_home/energy_audits/index.cfm/mytopic=11190
OR this article from Home Energy
http://www.homeenergy.org/archive/hem.dis.anl.gov/eehem/94/940110.html
Notes
1: As far as I can this number is for the entire year, not just the heating (or cooling season).
2: Its easy to get carried away here with value judgments; the point isn't to do that but to be aware that comparisons are difficult when the building size isn't the same and hence ACH has its limitations. This often comes up in passive house arguments because a 3000SF .6ACH50 passive house leaks as much air as a moderately tight 1000SF 1.8ACH50 house, so should the smaller house really have to be .6ACH50? What about a 300SF house? Partly it depends on whether you want to meet the passive house standard or not.
3: from what I've heard old houses are all over the map; some much worse than 10ACH50, some a bit better. If you have a really loose one, you likely know it.