Grenfell Tower: the road to ambiguity (part 2)

[Continued from part 1]

In part 1, I said that I would look in turn at three remarkable features of the BCA’s Technical Guidance Note 18. The third of these is that it purports to introduce an alternative whole system route to compliance, based on a Desktop Study rather than a real life fire test.

3. the new desktop study route to compliance

At the same time as tightening the requirements for the ‘Linear’ individual components route to compliance, BCA TGN 18 apparently relaxed the conditions for the alternative whole system route. AD B2 12.5 requires that ‘full scale test data’ be used to demonstrate that a cladding system meets the BRE 135 performance criteria:

It seems to me that the natural interpretation of this requirement is that the system that is tested to BS 8414 should be the same as the system that is to be installed. BCA TGN 18 refers to such a test of the ‘complete proposed external cladding system’ as Option 2:

It may be noted in passing that the BCA seem to misrepresent the ‘mechanical performance’ criterion by stating that it has to be ‘met’. BRE 135 itself does not set a minimum requirement for mechanical performance but only requires that mechanical performance be reported and ‘considered’ as part of an overall risk assessment when specifying the system:

which ad b2 paragraph is being satisfied?

The BCA created confusion by presenting this option as a means of showing compliance with AD B2 12.7:

This is quite wrong. Testing to BS 8414/BR 135 is a means of demonstrating compliance with the alternative route offered in AD B2 12.5. Paragraphs 12.6 to 12.9 only come into play when the ‘Linear’ individual components’ route is taken.

option 3: a desktop study

Remarkably, the BCA advised that:

If no actual fire test data exists for a particular system

then the client could show compliance with AD B2 ‘12.7’ [should be 12.5 as argued above] by submitting a desktop study report from an accredited testing body rather than such actual fire test data:

In passing, it is rather odd that there is no definite article before ‘BR 135 criteria’. As written, Option 3 requires only that some of the criteria are met rather than necessarily all of them.

The UKAS accredited body ¹ is to give an opinion as to whether the proposed system would meet the BR 135 criteria if it were to be tested to BS 8414. The report is to be supported by ‘test data’ but, rather alarmingly, the BCA do not specify what kind of tests the data is for. The reference to ‘situations/arrangements’ probably indicates that these are system tests that the BCA has in mind. But then it is odd that the final sentence concerns tests carried out on ‘the product’ in the singular. System tests are not carried out on a product but on a system. So could:

tests which have been carried out on the product

be or include small-scale product tests?

does the approved document allow desktop studies?

The alternative whole system route to compliance is defined as follows in AD B2 12.5:

External walls should … meet the performance criteria given in … BR 135 for cladding systems using full scale test data from BS 8414…

Does this allow for Desktop Studies which make use of data from tests to BS 8414 of systems other than the system that is to be installed?

It could only do so, it seems to me, on the basis of a very surprising and forced interpretation. The BR 135 criteria are that the temperature increase at the Level 2 thermocouples should not exceed 600° C within 15 minutes of the start time. Here for example are the temperatures at the external Level 2 thermocouples during the 2nd of the current series of DCLG tests at the BRE (PE ACM + mineral wool insulation):

It can be seen that the temperature at thermocouple 2032 (in yellow) exceeded the + 600° C threshold (red horizontal line) for a short period. As it happens, the test was in this case terminated shortly afterwards for another reason, resulting in the sudden drops in temperatures visible just after 300 seconds. But the principle remains that the determination of pass or fail from the test data is direct and specific. Either one of the lines cross the horizontal line (for more than 30 seconds), and the system fails, or none of them do on any of the Level 2 thermocouples, and it passes.

I rather think that this is what the Approved Document is referring to when it speaks of meeting the performance criteria using BS 8414 test data. The data yields a definite pass or fail.

A Desktop Study cannot perforce yield a definite result of this type. It can only make a prediction based on the performance of other systems. The expert gives an opinion that the criteria would be met if the system were to be tested to BS 8414. Could he or she say that the proposed system ‘meet[s] the [BR 135] performance criteria’, as the Approved Document has it? Or would a more conditional mode of expression be required, such as that an assessment has been made that the system would meet the criteria if it were to be tested, with a higher or lower degree of certainty?

Misrepresentation of the approved document

According to the Introduction to BCA TGN 18, the Approved Document ‘recommends’, ²  for the alternative whole system route to compliance:

to submit [sic: read ‘the submission of’] evidence that the complete proposed external cladding system has been assessed according to the acceptance criteria in BR135

There is no mention here of the ‘full scale test data’ that the Approved Document actually speaks of. Instead there is an assessment according to the BR 135 criteria. If the Approved Document did speak of such an assessment, then Desktop Studies would certainly be allowable. But it doesn’t. It speaks of performance criteria being met on the basis of test data. That is a very different thing, in my opinion.

origin of the desktop study option

At a seminar on ‘Facades to Tall Buildings’ hosted by the NHBC (National House Building Council) last year, the above mentioned Steve Evans was reported by his colleague at NHBC Diane Marshall to have explained (p. 12) the reasons why the BCA published TGN 18:

it became clear to building control bodies … that the routes to compliance described in regulations were insufficient…

The building control bodies reportedly came to the conclusion that the building regulations were too restrictive in allowing only two routes to compliance. They therefore took it upon themselves to, in effect, relax the regulatory regime. This is a remarkable admission. These bodies have no authority to make regulations. Their role is to interpret and enforce the regulations, not to make them, and most especially not to relax them.

The reasons he gave for the addition of the Desktop route to compliance were:

a) the demand for higher levels of insulation;

b) a multiplicity of cladding systems.

I have explained before why high insulation targets almost demand the use of polymer insulation rather than mineral wool, especially in the case of refurbishment, where it may be difficult to accommodate the sheer thickness of mineral wool required.

Now clearly, if high insulation targets are not easily compatible with fire safety regulations, then it does not necessarily follow that it is fire safety that is to be compromised. One alternative is to lower the insulation targets. Another would be to pursue the more difficult route to compliance through large scale testing. Since this is expensive, a further consequence might be that only a limited number of systems could be used economically. So be it. Are people’s lives to be endangered on the altar of architectural vanity?

It could be of course that the BCA were correct in seeing a need for a third route to compliance. But then this would be a policy decision to be made by Parliament, not by the bodies responsible for enforcing the regulations. It is true that the Approved Documents do allow for alternative approaches to meeting the statutory requirements (see, for example, AD B2 0.21). Compliance to a British Standard, for example, is specifically mentioned as a possible approach (AD B2 0.22). But the BSI is a standards-setting institution. Although it also does testing and certification work, this is as I understand it carefully separated from the development of standards, which is undertaken primarily by BSI Standards Ltd. The BCA is not a standards-setting body.

a change in a certificate

Evans gave further information about the origins of BCA TGN 18 in slide 7 of the presentation referred to in part 1:

Change in BBA certificate in late 2013 for most common product seen on site restricted its use on buildings with a floor over 18m

The change in the BBA certificate that he was referring to was the revision of the Kingspan Kooltherm K15 certificate BBA 08/4582 in December 2013, 6 months before the publication of BCA TGN 18. It is a little hard to piece together the bullet points. Evans mentions PIR and PUR boards, and appears to say that they are ‘preferred by the industry’. He then describes Kooltherm K15 as the ‘most common product seen on site’. But Kooltherm K15 is phenolic. I think probably phenolic boards are to be included along with PUR and PIR as those preferred by industry.

To trace the history of the BBA’s certification of Kooltherm K15 requires a separate post.

[continued in part 3]