Research and analysis

Fire safety: Roof insulation fire testing (Summary and recommendations)

Published 22 December 2025

Applies to England

1. Appointment

OFR Consultants Ltd have been engaged by the Building Safety Regulator (BSR), who are part of the Health and Safety Executive (HSE), to deliver the “Real Fires” project in support of fire safety technical policy. The Technical Policy Division of the Department for Levelling Up, Housing and Communities (DLUHC, formerly the Ministry for Housing Communities and Local Government, MHCLG, and whose responsibilities are now held by the BSR), originally commissioned this project on the 22nd of October 2021. The duration of the contract still stands from its initial award by DLUHC, running from the original commissioning date for three years.

As part of this project, the contract makes allowance for ad-hoc research to be undertaken to support fire safety technical policy on matters that emerge through dialogue with industry or through observations of real fires. Through this mechanism, OFR have been engaged to undertake research on the fire performance of roofs on residential buildings afforded retrospective spray applied foam insulation.

2. Background

On the 8th of April 2022, OFR attended an on-line meeting with (then) DLUHC staff from both the fire technical policy and sustainability divisions. At this meeting, DLUHC relayed their observations that suppliers in industry were incentivising the use of retrofitted spray applied closed and open-cell polymeric insulation to the underside of roofs of dwelling properties, with the intent of improving thermal performance of the buildings. However, there is a concern that such a retrofit could constitute a worsening of the existing fire safety expectations owing to a change in the fire performance of the roof to which it is applied. To this end, OFR were instructed to undertake a programme of experiments to establish the extent to which such products present an additional fire hazard, with any outcomes having the potential to impact technical policy on their future use.

3. Method of investigation

OFR were instructed to investigate the implications of retrospectively applied spray foam roof insulation in terms of the implications for compliance with Regulation B4(2). Compliance with Regulation B4(2) is typically addressed through the application of guidance in Approved Document B (ADB) (HM Government, ‘The Building Regulations 2010, Approved Document B (Fire Safety) Volume 1: Dwellings (2019 edition incorporating 2020 and 2022 amendments)’, Dec. 2022) with the intention that: “Flame spread over the roof and/or fire penetration from external sources through the roof is restricted.” In ADB, roof coverings are expected to achieve classifications to address the prospect of external fire spread. These classifications are expressed in function of the proximity of the roof to the relevant boundary. The classifications are evaluated through roof testing procedures, with the focus principally on fire penetration of a roof from outside-to-in. In England, BS 476-3 (BSI, ‘PD 476-3:2012 Classification rules for the end-use application of test results arising from BS 476-3, “Classification and method of test for external fire exposure of roofs”’, BSI, London, 2012) is adopted for the purpose of classification testing of roofs, which is known as CEN/TS EN 1187-4:2012 (BSI, ‘DD CEN/TS 1187:2012 Test methods for external fire exposure to roofs’, BSI, London, 2012) test 4 (t4) within the BS EN 13501-5:2005 classification framework (BSI, ‘BS EN 13501-5:2005+A1:2009 Fire classification of construction products and building elements. Classification using data from external fire exposure to roofs tests’, BSI, London, 2006). The test method can be described as two-stage, incorporating the effects of burning brands, wind and supplementary radiant heat.

This research study makes use of an ad-hoc format of the BS 476-3 protocol to investigate the implications of retrospectively applied, sprayed foam insulation on the fire penetration performance of three different roof coverings. The tests were undertaken at the Building Research Establishment (BRE) between June 2023 and November 2023.

4. Scope of this report

This report has been prepared to summarise the findings of the three BS 476-3 roof tests undertaken and to set out OFR’s view on the implications for technical policy. It then makes recommendations in respect of potential future work / research.

5. Outstanding information

At the time of writing, OFR are yet to receive BRE’s final consolidated report. Therefore, this report is based on interim information received by way of data logger information, test classification reports (see Appendix A), images and OFR’s observations having attended the tests. It is expected that a revision will be required as certain details require confirmation (from BRE) at the time of writing.

6. Conclusions

Retrospectively applied spray foam insulation is marketed as a means to simply upgrade the thermal performance of dwellings. The open or closed cell foam solutions are often applied by suppliers under the instruction of housing owners without specific consideration for the implications of Building Regulations compliance.

If the application of spray foam insulation were considered building work, from a fire safety perspective, the implications could affect compliance with Regulations B3(3) and B4(2). In the case of the former, it is noted that the extra fuel associated with the insulation could bring about a more severe enclosure fire if ignited. This could affect the fire separating ability of compartmentation elements and, thus, could represent a worsening of an existing condition. A further detailed study in such regard would be warranted but was not the focus of this study.  This issue has been highlighted to the BSR as an area for additional work, who consider that this would be best taken forward at a future date as part of a broader piece of work. 

In terms of compliance with B4(2), it is again clear that the introduction of a semi-permanent fire load increases the likelihood of fire spread across a boundary if the roof covering is penetrated and, thus, is a worsening of an existing condition. However, ad-hoc classification testing presented herein identifies that even traditional roof coverings that exhibit a poor performance, such as very old slate and clay tiles, can achieve the highest levels of roof classification when considering penetration performance. This is despite the inclusion of combustible sprayed insulation. This means such sprayed configurations can satisfy the highest demands for roof covering classification as currently included in ADB guidance. 

No guidance currently exists in ADB that controls the reaction-to-fire classification of bounding surfaces in a roof space. Developing such guidance could be a means by which insulation could be encapsulated behind fire protective linings, as exposed insulation is likely to achieve poor fire growth rate indices and total heat release rates. In some product literature, for example, British Board of Agrement (BBA) certificates, recommendations are made to enclose the insulation behind a plasterboard lining. 

Finally, it is highlighted that the test 4 method in BS EN 13501-5 potentially presents challenges in terms of mitigating fire spread across a boundary or over compartment walls. This is because:

• The provisions in ADB focus on the “roof covering”, yet the test method applies to a broader system. It is therefore unclear how provisions for isolated roof coverings are meant to be addressed without recourse to EC Committee Decisions. Also, if systems achieving the classifications recommended in ADB can be used in locations where guidance for coverings is provided.

• The test provides no quantitative measure of penetration. By focussing on if flaming or glowing is visible on the underside of the sample, high classifications can be achieved through the inclusion of multi-layered roof systems that can include large amounts of combustible insulation. For example, Sample C would suggest that fire spread occurred (there was ignition of the insulation below the tile), yet penetration was not visible.

The contractor is therefore of a view that further research is warranted to consider: (a) the relationship between ADB provisions for “roof coverings” and tested systems; and (b) if refinement of the test 4 method is necessary to include quantitative indicators of penetration, particularly where thick layers of charring insulation may be used.