What is the purpose of testing for regulations?

The commonly perceived purpose of testing airtightness is to produce a benchmarking comparison between buildings. However, in reality, airtightness results are used to determine total ventilation rates for energy calculations in SAP within the UK (and other energy modelling tools elsewhere) and, to a lesser extent, for indoor air quality. As such, determining the correct airtightness is important for more than simple benchmarking and, if incorrect, can lead to errors in SAP/energy assessments and also incorrect HVAC system design/setup or indeed will potentially lead to unhealthy buildings.


What is the PULSE air test?

The PULSE technique works by subjecting a building envelope or enclosure to a known volume change in the form of a low pressure burst of air. An airtightness result is then derived not from where the pressure peaks or how the pulse decays but by instead measuring the pressure response as a whole. The known volume change is introduced by releasing compressed air into the test space for a period of 1.5 seconds from an air tank, which generates a flow rate through any adventitious (unintended) openings whilst, in turn, creating a pulse in the internal pressure. During this pulse, a period of quasi-steady flow is established that directly gives the leakage characteristic at a specified level of pressure. When used in buildings, the advantage of PULSE over a blower door fan is that PULSE measures air tightness at far lower pressures that are more representative of typical atmospheric pressures - ~4 Pa.

After adjusting for still air conditions and making a small correction for the effective flow rate arising from the compressibility of air, the result can be plotted or read in the same way as that from the blower door fan technique. PULSE therefore delivers a result that directly represents the leakage at natural atmospheric pressures whereas the blower door test method operates in a much higher pressure range, requiring results to be extrapolated down. This extrapolation process can and does result in uncertainties which can reportedly be as high as ±40%.


How do the results compare between a blower door and PULSE test?

In order to make a direct comparison at the same pressure level, either the blower door or PULSE result has to be extrapolated and significant uncertainties are introduced to the results as a consequence (see comments below on using a single ‘n’ value).

In deciding whether it is best to extrapolate the pulse up or the blower door down, it is necessary to consider what is the point of airtightness testing – essentially, it is to try and infer the infiltration that happens under natural conditions, for which 4 Pa is considered a typical pressure difference. Therefore, a comparison at 4 Pa is the most relevant as the use of 50 Pa is simply a compromise as there was previously no other way of measuring reliably at such low pressures. However, if it is accepted that the blower door cannot give accurate results below 10 Pa (either directly or through extrapolation) then comparing them at 4 Pa could rightly be seen as biased towards PULSE. Therefore, perhaps the fairest comparison is to see which technique is more accurate at measuring a known opening added to the building envelope. Tests conducted throughout the development of the PULSE technique suggest that PULSE is able to give better agreement with the known opening than the blower door. Also, external literature suggests using the blower door with standard infiltration models routinely over-predicts the directly measured infiltration (using the tracer gas technique). We are continuing to conduct more extensive testing of this, using controlled openings and tracer gas, to further validate how this relates to different building types.


Is there any way to directly compare the results between a blower door and PULSE test?

We fully recognise how important this is, particularly for demonstrating compliance with building codes where all existing conventions are geared towards the blower door test and a tendency for compliance values to be cited at relatively high pressures (50 Pa in the UK). Whilst we would ideally like the building industry to fully adopt an approach of accurately measuring infiltration at more natural pressures, we are also currently working on a ‘step pulse’ concept where three pulses will be fired in succession to enable results to be spread across a wider pressure range. This will give us more confidence in extrapolation and allow PULSE to cite the air leakage rate at pressures outside our current 3-10 Pa range.


How is the PUSLE equipment configured and what are the logistics of handling a PULSE machine?

We are continually looking at ways in which to develop the PULSE device in ways that make it as user friendly and ergonomic as possible. In simple terms, the three main components that form the PULSE device are a compressor, an air tank and a small electronics control box. Currently, this equipment is mounted together on a trolley, where all connections are fixed and site set-up requirements are minimised. With large rugged wheels, this configuration is ~25kg in total and well suited for un-skilled operatives using the equipment on large live construction sites. In order to make the device more portable, we are currently experimenting with disaggregating the components so that the compressor (15kg) is encased within a single carry case, the air tank (8kg) has its own carry handle and control box and accessories (2kg) are also within a separate case. This will make the equipment far more versatile for testers going between multiple sites and better suited to packing in and out of a vehicle. Another advantage of this configuration is that the compressor may operate from a vehicles 12v power supply or a power source remote from the site to be tested, thus allowing the tester to simply carry the tank and control box into the test environment. Another equipment configuration we are developing is one where the equipment may be housed in a single case that could be unobtrusively mounted within an enclosure to test airtightness on a periodic basis.


Time requirement for blower door testing versus PULSE testing

The time it takes to conduct an airtightness test can have a direct impact on the income that a tester might expect to earn, and can be equally critical for customers in terms of the level of downtime or disruption caused.

Whilst it is often claimed anecdotally that the blower door test takes only 15 minutes, the UK blower door competent person’s schemes cite 30-45 minutes, which matches our experience having witnessed numerous blower door tests. Performing the blower door test too quickly is highly likely to add considerable uncertainty (and is arguably the cause of the poor repeatability found in blind ‘round robin’ tests).

If a full blower door test takes 30 minutes, 10 minutes is used to measure and prepare the building, five minutes to set up the equipment, 10 minutes to conduct the test itself, with the remaining five minutes used to pack-up. The PULSE test, on the other-hand, can be switched on to charge (5-10 minutes depending on the tank size) whilst the tester measures up and prepares the building. A PULSE test can then be fired in a matter of seconds, with results instantly captured within the control box and the whole test procedure can be wrapped up and completed within 15 minutes - half the time of a blower door test. This time benefit increases significantly as the complexity of a given testing situation grows.

Incidentally, there are also a number of things not included in the present UK technical standard that could be done to improve the certainty of blower door results, namely: take an average external pressure reading from all sides of the building; take repeated average readings at a given pressure with a zero reading taken before and after each pressure change; take the average of both pressurisation and de-pressurisation results; when making a curve fit do so with greater weighting on higher pressure readings (otherwise errors in the lower pressure readings have a disproportionally large impact on the Q-P relationship.). Of course, implementing any of these points would lead to the blower door test taking even longer than it currently does. PULSE, however, negates the need for all of this extra work whilst giving accurate and repeatable results.


What is the calibration and maintenance requirement associated with the PULSE equipment?

The PULSE equipment configuration comprises a rocking piston compressor, a composite air tank, a reference pressure tank, two solenoid valves that open and close the tanks, a nozzle, two pressure transducers and the control box with firmware and a user interface.

The pressure transducers are the key pieces of instrumentation that require annual calibration. However, as the PULSE unit is currently not yet a formally recognised means of demonstrating compliance with building codes, we have yet to confirm the full protocol with regards to calibration/swapping-out the transducers but, it essence, the procedure should be very quick and simple. The remainder of the equipment is very low maintenance and simply requires servicing on a periodic basis. Some people ask about building site dust clogging up the air release nozzle but the main function of this device is as a pneumatic silencer, and the diffusion of air as part of the test process will keep the component sufficiently clear.


Can PULSE be used to identify leakage paths?

One advantage of the current fan-based blower door air leakage testing technique is in its ability for pressurisation or depressurisation to be sustained so that a smoke pen and/or thermal imaging camera can be used to visualise air leakage paths. PULSE cannot achieve the pressures required to undertake this type of diagnostics and therefore requires a small, cheap uncalibrated fan to be mounted in an open window in order to conduct the same test. The PULSE device will never be able to be used to pinpoint leakage paths but we are continuing to test how much information about the leakage characteristics might be inferred from the PULSE shape itself – similar in some ways to how the power law flow coefficient can indicate if there is a large unrestricted opening or lots of smaller, narrower gaps and cracks.

Our general view on the matter of leakage diagnostics is that, despite a blower door fan’s capability to trace leaks, such a test is often not conducted until the property or renovation works is completed and ready for handover. By this point it is often too late to do anything meaningful other than to use mastics and expanding foams, which only offer a temporary fix. Our hope is that PULSE will make air tightness testing far more commonplace, with PULSE tests conducted at regular intervals throughout the works phase by both skilled and unskilled operatives. In this digital age, test results for different house types at different construction phases would soon paint a very comprehensive picture as to what steps might need to be taken to either improve air tightness or reduce risks associated with under-ventilation.


How are the outputs from PULSE captured?

The brain of the PULSE device is its control box which houses a processor, memory and also a GPRS module. The control box simply requires the user to enter the building volume and envelope area and then at the push of a button the PULSE is fired and the processor automatically calculates the result. This result is shown immediately on the screen of the control box but also stored in the boxes memory and sent as a GPRS data packet on to the PULSE server. Individual or batch results may then be downloaded either directly via USB from the control boxes memory or from the server for detailed analysis and reporting. Depending on the user’s level of expertise and intended use for the data, either the full raw data from a 6 second PULSE test (comprising 120 data points) or just the coefficients and final test result figures can be communicated. In the future it is our intention to also enable results to be used for benchmarking and also remote quality checking before then allowing results to be used for demonstrating regulatory compliance.


How repeatable is the PULSE test?

A high level of repeatability is a critical factor in demonstrating a high level of accuracy. We see excellent repeatability with a relative percentage difference (RPD) from the mean falling comfortably within +/- 5% for all tests. If required, the repeatability can be improved even further, simply by using a larger tank. The PULSE operative also has far lesser impact on repeatability, with tests conducted at the single push of a button. Unfortunately, we are not aware of any studies that have looked specifically at the repeatability of the blower door test and presentation of such information to us would be very much welcomed. The Zero Carbon Hub ‘Design vs As Built’ project (reported upon in July 2014) did, however, conduct a ‘round robin’ experiment, sending five different companies to perform standard air tests on six different plots spit evenly across two development sites. Despite the tests being conducted in close succession, variations as large as 66% were recorded between testers. These errors were largely attributable the use of different measurement assumptions but other variables such as differences between pressurisation and depressurisation, results sampling and processing methodology and equipment error were also not ruled out.


Is the air pulse uniform throughout the house? How does the 4pa pulse deal with more complex building forms, for example?

We have measured the internal pressure in the middle of buildings and also at multiple points throughout them (up to three storeys high) without detecting any noticeable differences. We have also done numerical analysis (using Fluent CFD software) that shows the pressure across the test space equalises almost instantaneously and certainly by the time of the important quasi-steady analysis period. This is due to the fact that the pulse spreads at the speed of sound (approx. 340 metres per second) and is a long wavelength infrasonic frequency. Just as with a conventional blower door test, the only requirement is that internal doors are kept open in order to ensure full and equal pressurisation.


Does PULSE generate a localised pressure that gets weaker as it disperses throughout the building?

The pulse spreads at the speed of sound (approximately 340 metres per second) and is a long wavelength infrasonic frequency that simply doesn’t get absorbed or slowed down within typical building environments in the same way shorter acoustic wavelengths do. The pressure transducer currently used for measurement within the PULSE device samples at 20Hz, giving us 30 readings in the 1.5 second pulse period (120 readings in the full six seconds of the test) and it is very clear in all our data that the enclosure pressurises almost instantaneously and quickly, and consistently reaches a quasi-steady state before then decaying. For equal pressure distribution within very large buildings, we can tether multiple units evenly throughout the building and release them simultaneously.

It is worth noting that if the internal pressure tapping is too close to the nozzle outlet there is a risk of it being affected by dynamic pressure, but this risk can be removed in the design of the unit. It is a similar issue to the position of pressure tappings with the blower door fan. However, dynamic pressures within the building itself are less likely to be an issue with PULSE, as the pressures are much lower.


How is Q4 calculated?

Q4, (the air leakage flow rate at 4 Pa), is obtained in the same way the blower door test obtains the Q50 (leakage rate at 50 Pa). A curve fit is made to the plot of leakage against pressure difference, and then the leakage corresponding to 4 Pa is calculated. This can be done by using either the power law or quadratic equation, and in practice (with experimental data) is simply a case of flipping the axes to convert between them. While the quadratic equation has been shown (both numerically and experimentally) to provide a better model of low-pressure leakage, in practice it doesn’t really matter that much which is used for fitting measured data as long as it isn’t extrapolated outside the measurement range.


Can a single ‘n’ value (power law flow coefficient) be used to represent all pressures?

It is incorrect to say ‘n’ is a building property and will be constant. The flow coefficient ‘n’ is a function of both the opening geometry and the flow regime through it. As such, ‘n’ does actually change with flow rate (and therefore pressure difference). This isn’t apparent when fitting the power law to test results as it is a function of the measured points themselves. However, if you extrapolate outside the measured data range it can be expected to incur significant uncertainty (either up or down) as the flow regime is unknown.


How does the control box choose the time period for analysis?

In short, we have identified a period of quasi-steady flow between the peak of the pulse and the closing of the valve, by matching theoretical curves with experimental ones. We have also shown that, by using the same equipment configuration, we can rely on a certain time range to always give a quasi-steady flow. In practice, this can be seen by plotting the measured pulse pressure against time, as a period where the driven pressure decay is relatively small (some change in pressure is actually desirable, in order to give a range of values).

The theoretical curves are obtained using a model based on the quadratic equation. An advantage of the quadratic relationship is that you can represent the steady and unsteady elements of the flow separately, which is impossible with the power law as it assumes one flow type as representative of the total. During the analysis period, the steady flow is typically around 99% of the total flow (hence quasi-steady) for openings up to approximately 3.5 metres in length, at which point the inertia in the openings starts to increase the unsteady component.