Your browser is no longer supported

For the best possible experience using our website we recommend you upgrade to a newer version or another browser.

Your browser appears to have cookies disabled. For the best experience of this website, please enable cookies in your browser

We'll assume we have your consent to use cookies, for example so you won't need to log in each time you visit our site.
Learn more

Grenfell | Burning questions unanswered

Grenfell pa (30)

A simple question has to be answered. Did the burning cladding set light to the flats – or did the burning flats set light to the cladding?

So far the full-scale BS 8414 -1 fire testing of overcladding systems by materials testing body BRE on behalf of the Department for Communities and Local Government (DCLG) seems to have confirmed one reasonably predictable outcome: overcladding systems incorporating rainscreen cladding manufactured from aluminium composite material (ACM) with an unmodified polyethylene (PE) core are highly unlikely to comply with current Building Regulations.

Another likely conclusion is that the type of insulation used has little effect on the performance of the system overall if the rainscreen cladding has a PE core.

However, the tests will not, on their own, explain what happened in the early hours of 14 June. Too tight a focus on the overcladding may lead to other key factors being missed. The test programme will yield valuable data, but its limitations must be acknowledged.

The first overcladding system on the 8m high test rig was a replica of that used on Grenfell Tower. Rainscreen cladding was 4mm thick ACM, made up of two skins of 0.5mm thick prepainted aluminium with a 3mm PE core. The insulation, 100mm thick rigid polyisocyanurate (PIR) foam was the same, as was the fixing system and the cavity barriers. A crib of burning wood in the combustion chamber below the cladding simulated a fire breaking out through a window. Thermocouples monitored internal and external temperatures.

What happened during the test is described in the BRE report DCLG test 1, which is available on the DCLG website. Section 5.2 Table 1 gives a timeline of the test progression. Unfortunately, it is somewhat short on key details.

Five minutes after the crib is ignited “burning droplets” were observed, “with a self-sustained burning duration longer than 20 seconds”.

Did the burning cladding set light to the flats – or did the burning flats set light to the cladding?

After six minutes and 20 seconds there is a “continuous flow of flaming material” from the base of the first row of panels. Melting aluminium is observed after another minute in the same region. There is a “pool fire” at the base of the rig with flames 300mm high.

The test was terminated after eight minutes and 45 seconds, with flames several metres above the top of the rig. The minimum duration specified for the test is 40 minutes, and thus the overcladding system was deemed to have failed.

Tapering damage area



Photo taken after the BRE’s test shows a tapering area of damage reaching up 5m

Photos taken after the test and included in the report show a tapering area of damage reaching up 5m. At the centre a significant area of the ACM has disappeared completely, exposing the PIR insulation. This has charred rather than combusted. Both the horizontal intumescent and vertical compression cavity barriers appear to have functioned as they should.

What else can be deduced from the report? First, it seems obvious that the “burning droplets” come from the PE core. There are several variants of PE, with densities in the range 910kg/m³ to 970kg/m³ and melting points between 80°C and 180°C. PE is said to burn slowly, with a blue flame, and to drip. The combustion is self-sustaining. Presumably the “flaming material” is also PE, as is the “pool fire” at the base of the test rig.

And what of the aluminium skins? Their contribution to the flames is likely to be negligible – or even non-existent. Yes, very fine aluminium powder is often added to military explosives and solid rocket fuels – yet aluminium foil is used to cover food in very hot ovens. The apparent paradox is down to the ignition point of aluminium, – around 3,800°C. It will melt at temperatures above 660°C, but thin sheets will oxidise and crumble away rather than igniting if exposed to a high temperature flame.

Compliance assessment

The BS 8414-1 test only assesses a cladding system’s compliance or otherwise with the Building Regulations. Analysing what actually happened at Grenfell Tower is more challenging, even with the information from the test, but a study of close-up post fire images can also help.

Large areas of blackened and charred insulation are clearly visible, as are what appear to be instances of buckled and melted aluminium. There also seem to be areas where the insulation has disappeared, perhaps because the flats below contained more combustible material than the norm, so burnt hotter for longer.

That the Grenfell Tower overcladding system failed the BRE test is hardly surprising. The ACM used for rainscreen cladding on the Grenfell Tower was 4mm thick Reynobond with a PE core. Manufacturer Arconic, in its pre-Grenfell publication Fire safety in tall buildings, recommends only Reynobond with either a fire retardant or non-combustible core for rainscreen cladding.     

Early ACMs often had polystyrene or polyurethane cores. PE-cored ACM is a fairly recent development. Arconic gives no information on what variety of PE it uses for its cores, but evidence from other sources suggests that low-density polyethylene is the most likely.

So, the BRE report can be summarised thus: when the flame from the crib impinges on the face of the ACM, the internal temperature of the core will rise rapidly. Aluminium is an excellent thermal conductor.

Melting core leaves cavity

The core melts from the bottom first, and starts to run out of the ACM. As it becomes exposed to the external flame and has access to oxygen, it ignites. By the time the aluminium begins to melt, the core has probably disappeared, leaving a 3mm wide cavity behind it.

At this point one key question must be asked. Are the melting/ignition of the core and the melting of the aluminium self-sustaining? In other words, once the melted core has ignited at the lower level, does it burn hot enough and long enough to melt and ignite more core from the next run of ACM above, and melt aluminium along the way?    

In terms of what happened at Grenfell Tower, the second key question is: could the burning of 3mm of PE, of slow burning PE, on the external façade of the tower produce enough thermal energy to set flats alight from the outside?    

Images in the BRE report might help answer the first question. The pattern of damage to the cladding seems to show it only occurred where the flames from the crib would have been at their hottest, ie at the centre. There is little lateral spread and no obvious evidence of self-sustaining damage to the aluminium.      

Another significant detail is that the burning PE at the foot of the rig could only manage a 300mm high flame. Given that, according to a BRE publication, flames breaking out from a burning flat will reach at least 2m above the top of the window, it would hard to conclude that flames from the PE would make a significant contribution to the spread of flame up the facade.    

So a simple question has to be answered. Did the burning cladding set light to the flats – or did the burning flats set light to the cladding?

The BRE tests and examination of video and still images are not enough to give a definitive answer. A computational fluid analysis should help resolve the issue. In the meantime, the test programme at BRE could yield useful information on the performance of other overcladding systems.


Readers' comments (1)

  • If you go to you will see a video of a cladding test carried out over three years ago in Spain. Four different combinations were tested simultaneously:-
    (panels are numbered from right to left)

    1. Rock wool with decorative panels
    2. PIR with metal plates and decorative panels
    3. Rock wool with a continuous decorative panel
    4. Coated XPS (extruded polystyrene) with firewall bands

    The video does not detail exactly what the decorative or metal plates are comprised of.

    The first thing you will notice about the testing, is that it is done in the open air. Oxygen supply is not limited, as I suspect might be the case in the indoor tests carried out by BRE.

    From my amateur's eye, it appears that number 2 (PIR insulation) becomes a self sustaining inferno within a few minutes. Now, I don't know exactly what exactly the outer panelling is, but could it be any worse than ACM with a PE core?

    I think it is highly plausible that PIR insulation combined with a flammable ACM with an air gap, is more than capable of spreading a fire vertically. In fact looking at the video, the test rigs include return corners, which No. 2 clearly shows the fire spreading across to.

    I would also suggest that the aluminium panels disintegrated from severe oxidisation caused by the very hot gases in the air void and released the PE filling in larger quantities than just a dribble from the bottom edge.

    Also, do not forget the heat generated from the cladding on the floors below and the upward rush of the hot gases.

    I believe Dave Parker's idea that it is hard to conclude that fire can be spread via the cladding is not justified, given the reality of the speed of fire spread at Grenfell and the results of the Spanish testing.

    More testing and analysis is required, but let's do it without drawing any preliminary conclusions which seem to lean towards absolving the cladding of blame.

    Unsuitable or offensive? Report this comment

Have your say

You must sign in to make a comment

Please remember that the submission of any material is governed by our Terms and Conditions and by submitting material you confirm your agreement to these Terms and Conditions. Please note comments made online may also be published in the print edition of New Civil Engineer. Links may be included in your comments but HTML is not permitted.