Hi,

I've been hovering over these boards for sometime but never registered and now I have a bugging question that I hope someone can offer some detail.

I'm very interested in having an air tightness test on my home to reduce carbon emissions, preserve energy and to reduce my bills. Want I want to know is there a formula that I can use or partly use to convert an air tightness result i.e 10 m3/h/m2@50PA to how much energy and money can be saved?

Many thanks in advance

Hi Tony,

I didn't think it would be simple hence I've asked for some guidence.

I know someone who has had an air tightness test and achieved 8.5 m3/h/m2@50PA. The engineer found numerous leakage points which were addressed and a re-test done some weeks later. The result then was 5.1 m3/h/m2@50PA. I'm interested in working out what kind of saving this has achieved. Yes, there would have to be some assumptions on energy used etc but I imagine you could get consumption levels from your supplier?

For your information, my house is masonry construction, cavity insulated. And yes I'm in the rainy UK.

Many thanks

Thanks James.

Yes I thought there must be some software out there that can easily convert your result into an approximate saving given the size of the building and a few other assumptions.

The guy I know who's had this done has only had it done recently so at the moment it's very difficult to look at his consumption figures to see if theres much change.

Most air testing firms I have looked at advertise that they can test existing dwellings and that savings can be made but none of them can tell me by how much. For e.g If your air permeability result can be reduced by 20% then in a standard 3 bed semi-detached, masonry constructed house this could save x co2, y in energy and z in £££'s.

I did read somewhere that BRE did a comprehensive case study but I can't find it. I've tried emailing them with not much joy.

Many thanks for your input.

Thanks Paul, that could be just what I'm after. I will take a look.

Knowing you have holes and knowing how much air air flowing through them are two different things. How do programs convert from one to the other? Presumably they have to make assumptions of some sort?

House is 30 years old and dry lined.

Downloading the Hot2000 software as I type.

Posted By: CWattersKnowing you have holes and knowing how much air air flowing through them are two different things. How do programs convert from one to the other? Presumably they have to make assumptions of some sort?

hot2000 uses the measurements from a blower door test to calculate the energy loss through air leakage. The blower door test establishes the "equivalent leakage area" represented by the measured flow rate at various depressurizations. Knowing the inside and outside temperatures over the period of a year enables a fairly accurate calculation to be performed. The hot2000 model of my 110 year old house has been within about 5% of actual measured consumption over the past 3 years (though my consumption is going down as various improvements have continued - I probably should do another blower door test to feed back into the model).

Without a blower door test, it's just guesswork (and some programs do allow various levels of tightness to be entered such as "tight" "leaky" etc.). Only a measurement will give accurate results. I have no idea if SAP can do anything like this - at least hot2000/hot3000 use climate data (including insolation) over a whole year's cycle. I'm not sure what the granularity of the simulation is, but suspect it is probably on the level of a day or even less.

Paul in Montreal

Tas is said to give very accurate results for energy losses due to air leakage, though as Paul says you have to have a value to enter. I touched on this for a recent study by looking at 1ach v 0.25ach. This equates roughly to 20m3/m2/hr @50Pa and 5m3/m2/hr@50Pa. Current building regs is 10m3/m2/hr@50Pa.

The results are in the Green Building Bible Vol 2 pg 69-75

Posted By: Paul in MontrealThough there will still be some kind of timestep

Just to get it clear - you're talking about Hot3000 here?
Mike George, wd you agree? - apx hour timesteps in Tas?

Posted By: Paul in MontrealI'm not sure what the granularity of the (Hot3000?) simulation is

Posted By: Mike Georgenot sure if the (Tas) simulation is hourly or lesser increments

Wd be gd to know - can it be varied at will? I guess it's a bit like graphics or CAD programs - Autocad is full of legacy settings whereby you can coarsen all sorts of things to make the program run faster on yesterday's computers, but on my 3yr old fast (then) machine it flies just fine with all settings set to finest. So subject to computing power, these dynamic modellers should approach reality closer, the finer the 'granularity' (shorter the time steps - e.g. 1min or less, is my guess).

Posted By: Paul in Montrealrates of change of temperature are rarely more than 1-2C per hour

seems to me to offer ample scope for cumulative errors, with 1 hour time steps. Anyone know?

Paul, this perhaps where the notion that a bin-based calculator (e.g. Hot2000) is just a dynamic (FEA) modeller (like Hot3000) running with big time steps (or vice versa) breaks down. In fact it may be misleading to think of time steps at all with a bin based calculator because that's not really how they work?

As I understand it, a bin based calculator is just a spreadsheet of factors that have been tweaked empirically to give output that corresponds to reality, over a range of variables. The algorithms don't necessarily attempt to model physics processes, instead shortcut straight from 'if this' to 'then that'. As such, a bin based calculator is only as good as the tricks that it's been pre-programmed to do. So what starts as a quick-and-dirty way of extrapolating useful results from a resource-bank of empirical observations, becomes increasingly onerous to maintain and extend with reliability, as additional tasks are required of it.

That's quite different from a dynamic (FEA - Finite Element Analysis) modeller, which is only as good as its algorithms which model the interactions of variables within various laws of physics. There's no shortcutting - every physical/thermal change is modelled from scratch according to laws of physics. As such, a FEA modeller can model 'anything' that it has physics-law algorithms for, although novel tasks may bring into play unexpected physics processes which the modeller may not have algorithms for, hence give false results.

The result of an integral calculus calculation can be approximated by less sophisticated means, by dividing the timeline into a large number of slices of small duration, and calculating each in sequence, taking the output of the previous slice as the input of the next slice. That's what a FEA modeller does. However, it's only an approximation to the true result which integral calculus gives, in which the time slices become an infinite number of zero duration, so the output changes seamlessly and continuously along the timeline.

Thus a FEA modeller isn't the ultimate, that an 'integral calculus' modeller would be, but that's into super-computer territory. A FEA modeller's virtue is that it will run on ordinary computers. However, such time slicing can give big cumulative errors, the fewer/bigger the slices get, so FEA is a balance between speed (i.e. minimised no of iterations, afforded by small no of big timeline slices), and accuracy (i.e. lots of iterations, necessitated by large no of small timeline slices).

If a FEA program doesn't have a way for the user to choose that balance, according to the task in hand and the accuracy required, then it should have! - unless it's intelligent enough to vary and make that choice automatically.

Edit: I've copied the discussion about modeller timeslices to http://www.greenbuildingforum.co.uk/newforum/comments.php?DiscussionID=1427&page=3#Comment_37464, where I suggest it should be continued, not here.

To add to the above, for a Home built to new build specifications (PArt L1A)

I have built a SAP spreadsheet o compare up to 100 different home specifiactions in one go (ongoing project - complex so still de-bugging).
I have used it to set standard specifications on large developments (eg renovating a 70 unit Mill) so a simple spec change can be seen for the whole project.

the following is a sample output that is relevant to this topinc (and explains why I am in the air tightness and ventilation business rather than renewables, which I originaly trained in untill i did the maths and had a son who is severely allergic)

The analysis shown is for the Heat Loss parameter portion of SAP only.

The standard ' insulate to improve house' has:
Trickle ventilation and stanard air tightness (q50 = 10)
Walls U = 0.2
Floor U = 0.17
Roof U = 0.16
TRIPLE Glazed windows (U = 0.8 + curtains to 0.74)
Low E argon filled, thermal spacer glazed doors.
Accredited Construction details.

The other home is at:
Part L1A worst case standards apart from:
Roof at U = 0.16 as easy (only need U = 0.25 in reality)
Air tighness q50 =3
Decent HRV system (75%)

Amazingly, to many, they have the same Heat Loss plus in reality the AIr tight home will have lower peak heating on cold windy nights.

Also:
there are a lot of other comparisons - the key for each is:
NAT,WHV, HRV - type of ventilation
q50 = air tightness target
HLP - heat loss parameter.
Note that WHV (Whole House Ventilation) increases heat loss! ( result 4)ALso going bellow q50 = 3 SAP starts to increase the HRV throughput - so diminishing returns
(Also SAP is not HRV friendly - Passiv Haus gives more accurate results for well desinged systems - SAP assumes the british building industry are ALL incompetent so de-rates ALL HRV systems)

MORE ANALYSIS AT THE BOTTOM OF THE INSERT.

If you take thermal briding heat loss, ventilation heat loss and add it to the elemental heat loss to get a whole element heat loss what is the effective U value of the home?
(Convert these lump sump heat losses and divide them by the surface areas to get an top up U value).
- this assume equal leakage across all areas - not true , but for want of hard data a startign point.

So the Wall U value increase from 0.2 to 0.52, the roof from 0.16 to 0.33.

The latter has some related evidence behind it - the Canadian Mortgage Council found, several years ago, that the infiltration heat loss in cold roofs was the dominant heat loss.

This type of analysis is intended as a though proviking tool rather than absolute measurement but is fundamentaly sound in principle (gulp). - Think of the Leeds MeT University finding of U = 0.5 for PArt walls because of convection currents and air leakage.

I will come back and edit for clarifiaction - have not posted recently as been very busy and need a break.

The following may be of interest but may also be full of errors and i would really appreciate any corrections:

SAP estimates air leakage from a set of given assumptions or from BSEN13829 air tightness testing, carried out at 50 Pascals differential pressure. To convert measured air loss into an estimated air loss rate at ‘typical’ pressure rates, the figure is divided by 20. This assumes an annual average UK wind speed at 10 meters height of 5 meters per second. In reality wind speed and temperature varies considerably between places and times. Wind exerts a velocity pressure (P, in Pascals) in proportion to the square of its velocity (V, in m/s): P=0.6xV2,
The average UK wind velocity of 5m/s exerts 15Pa (PaulT says 4Pa - where is this figure from? Ground level? 50Pa/20=2.5Pa. So where does the /20 correction factor come from?). However, given the proportionality to the square of velocity, any wind speeds above the average are substantially more significant at increasing pressure than speeds below the average are at decreasing it, so any long-term air permeability assumptions based on pressure generated by average wind speeds will be wildly inaccurate, for the same reason that wind turbine site assessments are based on mean energy distribution, ie the average of a set of V2 is different to square of the average of a set of V. It is impossible to calculate mean wind pressure from mean wind speed alone.
Towns may have gusts up to 70mph (560Pa), and some areas in Scotland may have gusts up to 100mph (1185Pa); a city centre may exceed 10mph 5% of the time, a coastal area may exceed it 55% of the time. So if your thermal performance assumptions are based on 15Pa, you may be disappointed when at over 1000 Pa. 50Pa is exerted at 9.1 m/s (24.4mph - a “fresh breeze“ or 5 on the Beaufort scale). This is about the average wind speed for Lerwick in December, so if you are designing a building to be warm on an average winter day in Lerwick, the rule of thumb to divide by 20 is redundant, let alone if you want to make it tight against 100mph gusts.
Air permeability can be converted into u-value equivalent (W/m2K) to make a performance comparison between rate of heat loss from air permeability, and that from conductance: Ueq = air permeability (CIBSE defined, ie air infiltration volume per building area per hour)/3600 x VHC
(where VHC = volumetric heat capacity, ie Joules/m3K. VHC of air = density x SHC = 1.205kg/m3 x 1.005 kJ/kg K = just over 1200 J/m3 K at 20C.
So a building of air permeability of 10 has a Ueq of 10 / 3600 x 1200 = 3.33, ie losing around ten times more thermal energy at 50 Pa than typically lost through conductance. If we divide this by 20 to achieve the typical assumed average, it becomes 0.1665, ie around half that lost through conductance (or around the same if comparing a Passivhaus). With an air permeability of 1, conductance and permeability losses become of the same order in a fresh breeze. The question becomes: do you protect against the average of the whole year, or only against the times when you need it? The logical answer seems to be the latter. If internal temperatures are set at 20C, then a summer gale blowing at 20C will give zero ventilation heat loss, no matter how leaky the building. Thus a site characterised by warm summer winds (such as southerly sea breezes) but relatively sheltered from cold winter north-easterlies, will lose less heat to air infiltration than a site where the opposite is true yet with the same average annual wind velocity. Insulation is redundant for 6 months of the year - required u-values are assessed against comfort and savings in the winter months. This would be the sensible approach for air tightness too. If my working can be verified, simply scrapping the correction factor seems to be a reasonable start.
The effect of building shape and wind direction have not been taken into account and so may introduce a large error to my working. Even so, if actual pressure differential on the building is half the wind velocity pressure, the wind speed required produce an apparent pressure of 50Pa from a velocity pressure of 100Pa would only be increased to 12.9 m/s (31.5mph/Beufort 6/Strong breeze) - still a speed that should, arguably, be designed for.