Grenfell Tower Inquiry: Phase 2 Report, Volume 1 · 2024
Chapter 7: The government and the Building Research Establishment
Chapter 7: The government and the Building Research Establishment
Introduction
¶7.1 In this chapter, we describe the role played by the government in investigating cladding fires and regulating the use of combustible materials in the external walls of high-rise buildings. The name of the department responsible for those matters has undergone several changes during the period with which we are concerned, from the Department of the Environment, Transport and the Regions (from 1997), to the Department for Transport, Local Government and the Regions (from 2001), to the Office of the Deputy Prime Minister (from May 2002), the Department for Communities and Local Government (from May 2006), the Ministry of Housing, Communities and Local Government (from January 2018), the Department for Levelling-up, Housing and Communities (from September 2021 to July 2024) and the Ministry of Housing, Communities and Local Government (since July 2024). In this chapter we refer to it simply as "the department". We also examine the part played in those matters by the Building Research Establishment (BRE). The period covered by this chapter runs from the fire at Knowsley Heights in 1991, widely regarded as the first cladding fire of relevance to what later happened at Grenfell Tower, to the amendment of Approved Document B published in April 2007, the relevant provisions of which remained current at the time of the Grenfell Tower fire, despite later amendments.
¶7.2 Our findings are based on the documents and the evidence of four principal witnesses, Dr Deborah Smith and Dr Sarah Colwell, both employees of BRE, and Anthony Burd and Brian Martin, both officials in the department. Mr Martin was employed by BRE for nine years, during which he was seconded part-time to the department.
Knowsley Heights: 1991
¶7.3 On 5 April 1991 a fire was started in an external refuse area at the base of Knowsley Heights, an eleven-storey block of flats in Huyton, Merseyside. By the time the firefighters arrived, the entire external wall on one side of the building was alight and flames were issuing from windows on all floors and from the roof.288 {BRE00035385/3} paragraph 3.1; {BRE00035385/5} Figure 3.1.1 "External view of the damaged block". Fortunately, there were no fatalities; the fire had spread vertically up an unoccupied area of the building that did not interfere with the escape route289 {BRE00035385/4} third paragraph. and all the residents were safely evacuated.290 {BRE00035385/3} paragraph 3.1.
¶7.4 In 1988 Knowsley Heights had been clad in what was later described by BRE as "Class 0 GRP rainscreen cladding". GRP (Glass Reinforced Plastic) is a combustible polymer material.291 For further details of the GRP panels used at Knowsley Heights, see Bisby, Phase 2 Report {LBYP20000001/94-95} paragraph 465. Behind those panels, a layer of mineral wool insulation had been fixed to the building with a gap between the insulation and the cladding to allow the movement of air.292 {BRE00035385/3} paragraph 3.1. The refurbishment at Knowsley Heights was carried out as part of a pilot scheme run and funded by the government's Housing Management Estates Action programme to improve the appearance of high-rise blocks, to reduce energy consumption and to prevent the damp from which many such buildings suffered at the time.293 {BRE00035385/3} paragraph 3.1.
82¶7.5 The fire was investigated by BRE, then a government-owned laboratory,294 BRE (previously FRS) was Government-owned until it was privatised in March 1997. Field {BRE00043710/6} page 6, paragraph 28; Smith {MET00081237/7} page 7, paragraph 18; Shipp {BRE00047594/6} page 6, paragraph 38; Bisby, Phase 2 Report {LBYP20000001/113} paragraph 592. which produced a report for the department in November 1992 under the Investigation of Real Fires contract.295 {BRE00035385}; the report's authors were Penny Morgan, Derek Jones and Sharon Clinch {BRE00035385/2}. In a section headed "Implications for Building Regulations" BRE drew attention to the absence of cavity barriers in the cladding system and concluded that there was no reason to suggest that cladding in general posed a risk to life unless there were cavities large enough to allow the vertical spread of fire.296 {BRE00035385/5} first paragraph. It also suggested that where the use of a Class 0 cladding material was recommended the material should be Class 0 on both sides exposed to air.297 {BRE00035385/4} fourth paragraph. It also advised that there were implications for the protection of window reveals in circumstances where refurbishment had involved the use of certain types of combustible materials in close proximity.298 {BRE00035385/5} first paragraph.
¶7.6 Notably absent from the report was any reference to the combustibility of the GRP panels themselves or the contribution they had or might have made to the spread of fire. When noting that cavity barriers had not been fitted as part of the system, the report said that the guidance in force at the time allowed them to be omitted if the cladding system was not combustible.299 {BRE00035385/3} penultimate paragraph. Morgan {BRE00043866/18} page 18, paragraph 96; Shipp {BRE00047594/48} page 48, paragraph 237. In the same vein, it noted that the fire had called into question the departmental guidance on cavity barriers because it had spread vertically up the entire height of the building.300 {BRE00035385/4} fourth paragraph.
¶7.7 It is not clear what that guidance was.301 Asked what "DOE/BRE guidance" she had been referring to in the report, one of the authors, Penny Morgan, could not recall; Morgan {BRE00043866/18} page 18, paragraph 97. At the time Knowsley Heights was refurbished, Approved Document B recommended that the external walls of buildings over 15 metres in height should be constructed of materials of limited combustibility.302 {CLG10002325/15} paragraph 2.7; Bisby, Phase 2 Report {LBYP20000001/87} paragraph 424. Cladding used on walls over 15 metres in height could be combustible if it was not relied on to contribute to the fire resistance of the wall and if it was certified Class 0.303 {CLG10002325/15} paragraphs 2.9 and 2.13, Table 2.2. Although Approved Document B did not specifically address cavity barriers in ventilated rainscreen cladding systems, it did recommend that they be fitted in any cavity abutted by a wall, ceiling, roof or floor.304 {CLG10002325/65}; Bisby, Phase 2 Report {LBYP20000001/96} paragraphs 473 - 475. It seems likely, therefore,305 Bisby, Phase 2 Report {LBYP20000001/96} paragraph 480. This view is also supported by the evidence of Martin Shipp and Dr Raymond Connolly, Shipp {BRE00047594/48} page 48, paragraph 238; Connolly {BRE00047667/4} page 4, paragraph 8(b). that BRE was referring there to the first edition of BR 135, which had been produced in the same year as Knowsley Heights had been refurbished (1988) and stated that in ventilated cladding systems constructed of non-combustible materials fire barriers were not essential.306 {BRE00001077/9} paragraph 1, second bullet point. That remained the advice at the time of the fire and could fairly be described as "departmental",307 {BRE00035385/4} fourth paragraph. given that at the time BRE was a governmental organisation. We have not been able to find any other guidance to which that passage could have been referring.
83¶7.8 However, the cladding at Knowsley Heights was not non-combustible and it is therefore difficult to see why BRE said that cavity barriers could be omitted if the cladding was non-combustible. Anyone reading that passage might well understand that the omission of cavity barriers from the system installed at Knowsley Heights was justifiable precisely because it was a non-combustible system, which of course was not the case.
¶7.9 Given the passage of time, very few of those who had been involved in the investigation into the fire or in drafting the report were able to assist us by providing background information, details of the investigation or the thinking behind the report. The only two who could were Penny Morgan, one of BRE's investigators,308 Morgan {BRE00043866/1-16} pages 1, 2 and 16, paragraphs 6, 12 and 83-85. and her line manager, Martin Shipp, who in November 1992 was the Head of BRE's Fire and Security section with responsibility for fire investigations.309 Shipp {BRE00047594/3} page 3, paragraph 14. He was aware of what was going on in relation to the fire at Knowsley Heights, since both Penny Morgan and Derek Jones, both of whom visited the scene of the fire on 10 April 1991,310 Morgan {BRE00043866/31} page 31, paragraphs 174 and 178. reported to him,311 Shipp {BRE00047594/43-45} pages 43-45, paragraphs 205 and 218. but he did not personally take part in the investigation into the fire.312 Shipp {BRE00047594/45} page 45, paragraph 220. However, it is very likely that he reviewed the report both for technical and editorial purposes.313 Shipp {BRE00047594/47} page 47, paragraph 229.
¶7.10 Mr Shipp said that, as he read the report, the authors had been aware that the cladding was combustible, since they had recorded that it had been destroyed.314 {BRE00035385/4} second paragraph. They must therefore have understood that the panels had burnt and therefore were combustible.315 Shipp {BRE00047594/49} page 49, paragraph 245. However, although that is a reasonable analysis of the report, we do not think that it necessarily reflects Penny Morgan's understanding. The tenor of the report suggests that she did not realise that the system was combustible, which would explain the reference to the omission of cavity barriers and the calling into question of the current guidance that cavity barriers were not required in non-combustible systems. If the authors of the report had understood that the fire had been caused by a failure to comply with a requirement for cavity barriers in what was recognised to be a combustible cladding system, there would be no reason for the guidance to be called into question. Instead, we should have expected BRE to have identified and questioned the use of combustible GRP panels.
¶7.11 Penny Morgan told us that she had considered the vertical fire spread to be significant and unusual because a fire that had started at ground level had spread over what was purportedly non-combustible cladding.316 Morgan {BRE00043866/22-23} pages 22-23, paragraphs 125 and 127. It appears that she thought that the GRP rainscreen panels used at Knowsley Heights were non-combustible because they had been certified Class 0.317 Morgan {BRE00043866/21} page 21, paragraph 115; Morgan {BRE00043866/19} page 19, paragraph 105; Morgan {BRE00043866/20} page 29, paragraph 113; Morgan {BRE00043866/24} page 24, paragraph 140. That is supported by a note dated 5 April 1991 from Alison Curtis of the Housing Management Estates Action Division, who had spoken with Penny Morgan by telephone on the day of the fire,318 Morgan {BRE00043866/29} page 29, paragraph 162. in which she said that the cladding system had no vertical fire stopping "in line with BRE recent advice in cases where all materials used in the over cladding and insulation are non-combustible".319 {INQ00014995} paragraph 4.
84¶7.12 Although Penny Morgan's recollection of events so long ago was understandably limited,320 Morgan {BRE00043866/15} page 15, paragraph 80. it seems likely to us that she and the other authors of the report had equated Class 0 with non-combustibility and that others, both in BRE and the department, repeated that error. That might explain why the report produced by BRE in November 1992 contained no analysis of the contribution of the cladding panels themselves to the fire. It might also explain why the report was silent on the combustibility of the panels, their contribution to the vertical spread of fire, their compliance with current guidance or the adequacy of Class 0 as a classification for external walls of buildings over (at the time) 15 metres in height.
¶7.13 On the day of the fire, Penny Morgan was alerted to its outbreak by a telephone call from Alison Curtis, followed by a note which said that the fire was of particular interest to the department because the building had been overclad using techniques that were relatively new to public sector housing and were currently being used on other blocks.321 {INQ00014752} paragraph 1. Ms Curtis also referred in the note to concerns about the innovative and high profile nature of the improvements to the block.322 {INQ00014752} paragraph 4. A second, handwritten, note dated 11 April 1991 from "Lyn" to "Mr Sage" contains the following comment,
"We have received, via HMEA,323 Housing Management Estates Action Division. a request from M. St Press Office [the department's press office]324 Morgan {BRE00043866/32} page 32, paragraph 187. Bisby, Phase 2 Report {LBYP20000001/105} paragraph 543. to play down the issue of the fire. Our briefing for [the Secretary of State] is purely factual and as far as I am aware Knowsley will not be making an issue of the fire."325 {INQ00014755}.
¶7.14 In the light of the fact that other high-rise residential buildings had been, and were still being, overclad using the same or similar materials under the department's programme,326 {INQ00014752} paragraph 1; {INQ00014995} paragraphs 1-12. we find it difficult to understand why the press office thought it necessary to play down the fire at Knowsley Heights.
¶7.15 The failure by the BRE's investigators to understand the distinction between combustibility and Class 0 was a basic error. In our view, the fire represented a significant opportunity for the department, and perhaps industry more widely, to recognise and respond to the fact that the use of combustible cladding panels on high-rise buildings, even where the surface of those panels was certified as Class 0, might lead to the uncontrolled spread of fire to the full height of the building if fire were to take hold in the external wall. If that and the tendency of many in the industry to treat Approved Document B as containing a definitive statement of what was required had been more clearly recognised, it might, and indeed should, have prompted consideration whether Class 0 was a suitable standard to include in Approved Document B for cladding panels used on high-rise buildings.
85¶7.16 The fire at Knowsley Heights was frequently referred to in BRE reports and presentations327 {BRE00043751/5} "The most significant of the historic fires is that of the 1991 fire in Knowsley Heights."; {CLG00019445/3} second paragraph, "One of the most significant of the historical fires is the 1991 fire in Knowsley Heights.". and by witnesses328 Crowder {Day230/158:21-25} "…one of, you know, the major fires in terms of understanding external fire spread"; Colwell {MET00080530/8} page 8 "...those fires were our anchor points. They were the reference points that we were using in terms of moving guidance and understanding"; Martin {Day250/76:17-23}; Smith {Day234/129:5-12}. as an event of major significance in the understanding of the spread of fire across the external walls of high-rise buildings. However, when asked why it was of such significance, Dr Crowder (who in 2014 was appointed to lead the fire investigation business group at BRE)329 Crowder {BRE00043716/50} page 50, paragraphs 164-165. reiterated that the key factor was the absence of cavity barriers.330 Crowder {Day230/158:25}-{Day230/160:15}; {Day230/182:2-10}. He went so far as to say that, if the cladding system at Knowsley Heights had included cavity barriers, the fire would probably not have spread to anything like the same extent.331 Crowder {Day230/182:8-10}; Martin {Day250/77:20-22} "I'm pretty sure the focus was on undivided cavities". That suggests to us that, even with the benefit of hindsight, some in BRE had not recognised that the combustibility of the cladding panels themselves had been an important factor in the development of the fire.
¶7.17 Dr Sarah Colwell, who became involved in BRE's large-scale testing programme in 1996,332 Colwell {BRE00047571/2} page 2, paragraph 9. In 1991, Dr Colwell was a Higher Scientific Officer in the BRE's Explosion Protection section: Colwell {BRE00047571/2} page 2, paragraph 6. By 2016, Dr Colwell was the Director of the Fire Suppression Team, Colwell {BRE00047571/3} page 3, paragraph 14. told us that one of the lessons from the fire was that the testing of individual components of a cladding system could not tell one much about its overall performance in fire,333 Colwell {Day231/127:1-14}. and it appears that following the fire at Knowsley Heights both the department and BRE accepted334 Smith {Day234/130:15-22}. that a large-scale fire test method was needed to provide a better means of assessing and controlling the potential fire hazards associated with external cladding systems.335 Smith {BRE00005624/8} page 8, paragraph 24. The department therefore funded further research by BRE into large-scale testing,336 Smith {BRE00005624/8} page 8 paragraph 25. Dr Sarah Colwell agreed that it was the fire at Knowsley Heights that led to funding for further research on large-scale testing, Colwell {Day231/128:24} -{Day231/129:2}. a subject to which we return below.
¶7.18 However, even if the department had recognised that small-scale testing (such as the methods used to determine Class 0) was not adequate to assess the danger of the spread of fire over the walls of high-rise buildings, it does not appear that at that stage it gave any consideration to restricting the use of combustible cladding on high-rise buildings. Other than including in paragraph 12.7 a general warning of the risks involved in the use of combustible materials,337 {BLA00005482/74} paragraph 12.7, "The use of combustible materials for cladding framework, or of combustible thermal insulation as an overcladding or in ventilated cavities, may present such a risk in tall buildings, even though the provisions for external surfaces in Diagram 36 may have been satisfied". no action was taken in that regard in the revised version of Approved Document B, published in 1992.338 {BLA00005482}. Instead, the only practical action taken was to strengthen the guidance on the use of cavity barriers339 {BLA00005482/2} Section B3, paragraph w; {BLA00005482/64}; Crowder {Day230/159:9-14} and {RCO00000001/7} second paragraph. and to provide that the panels forming the external walls of buildings over 20 metres in height should have Class 0 surfaces on both sides.340 {BLA00005482/74} paragraphs 12.5- 12.6; {BLA00005482/75} Diagram 36; Martin {Day250/73:17-22}. Those amendments directly reflected the contents of BRE's report on the fire at Knowsley Heights.
861994: Dr Raymond Connolly's ten large-scale tests
¶7.19 In 1993, the department commissioned BRE to examine the hazards associated with external cladding systems and to assess the influence of a range of fire barriers on both the fire hazard and on the movement of air within the cavity in relation to the control of damp.341 {RCO00000001/3} penultimate paragraph; {RCO00000001/7} points (i)-(iii). The research was carried out by Dr Raymond Connolly, a Senior Scientific Officer in BRE's Structural Fire Protection section,342 Connolly {BRE00047667/1} page 1, paragraph 2(b). and was published on behalf of the department in 1994.343 {RCO00000001}. He designed and carried out full-scale tests on ten different external cladding systems344 {RCO00000001/12} section 3. chosen as representative of systems in use at the time.345 Connolly {BRE00047667/9} page 9, paragraph 21(d). The four-storey (nine metre high)346 {RCO00000001/46} section 4.2.3 final paragraph. test rig had been designed specifically for the purpose of the tests.347 {RCO00000001/8} section 2.1.
¶7.20 In his statement Dr Connolly said that the purpose of the research had been to develop a means of examining all the relevant variables in a holistic manner by full-scale testing, rather than by relying on tests of individual components.348 Connolly {BRE00047667/9} page 9, paragraph 21(g)(ii). In the opening paragraph of his report, however, he had said that the purpose of the work had been to assess the effectiveness of fire barriers349 {RCO00000001/3} first paragraph. and in that context to develop the most appropriate heat source for large-scale testing.350 Connolly {BRE00047667/9} page 9, paragraph 21(a). That suggests that at that stage the department was still more interested in learning how to limit the spread of flame by the use of cavity barriers than in assessing the combustibility of different products. That is significant because successive versions of the large-scale test method that became BS 8414 were all derived directly from Dr Connolly's original work.351 The theoretical basis of Fire Note 3 "is as set out in Dr Connolly's 1994 report", Colwell {BRE00047571/11} paragraph 63; {Day231/132:15-22}; {Day231/152:14-17}. Fire Note 9 simply incorporated the method in Fire Note 3 and contained no changes to the method or the assessment criteria, Colwell {BRE00047571/19} paragraph 127; {BRE00047571/21} paragraph 138; {Day231/200:8}-{Day231/201:16}. The BS 8414 test series represented the adoption by the British Standards Institution of the test method set out in Fire Note 9, Colwell {BRE00047571/26} paragraph 168; {Day232/135:6-15}. The test method did not change between Fire Note 9 and BS 8414, though the classification element of Fire Note 9 was removed and published separately in the second edition of BR 135, Colwell {BRE00047571/29} page 29, paragraph 184 and {Day232/139:21}-{Day232/140:1}.
¶7.21 The second of the ten tests was carried out on a system comprising 6mm polyester glass reinforced sheeting with a Class 0 certification, mineral wool insulation and no cavity barriers. The object was to examine the need for fire barrier protection. Dr Connolly expected the fire hazard to be limited by the fire properties of the cladding sheet material (i.e. its Class 0 rating).352 {RCO00000001/16}. After the test, however, he reported unlimited spread of the fire over the full height of the test facility353 {RCO00000001/18} final paragraph. and that flames had reached the level of the roof 15 minutes after ignition of the crib.354 {RCO00000001/17} Plate 6. He also reported more generally that the polyester- bound sheet Class 0 cladding, which had been used in eight of the ten tests,355 {RCO00000001/12} section 3. had suffered from extensive surface spread of flame in nearly all configurations,356 {RCO00000001/46} section 4.2.3 second paragraph. often spreading to the top of the test facility,357 {RCO00000001/46} section 4.2.3 final paragraph. and that fire barriers had not been not completely effective in any test.358 {RCO00000001/46} section 4.2.3 second paragraph.
87¶7.22 Dr Connolly considered it clear that BS 476 Parts 6 and 7 (i.e. the tests relevant to a Class 0 classification) did not accurately reflect the fire hazards that might be associated with cladding systems.359 {RCO00000001/46} section 4.2.3 fourth paragraph. Furthermore, in his overall conclusions he recorded that the way in which cladding materials reacted to fire in small-scale tests did not reflect the fire hazard associated with a full-scale system360 {RCO00000001/48} paragraph 8. and were an inadequate measure of the hazard if used in isolation.361 Connolly {BRE00047667/13} page 13, paragraph 24(c)(i).
¶7.23 Dr Connolly told us that that was neither new or surprising; it simply confirmed what had already been suspected and what he understood already to have been publicly recognised by the department,362 Connolly {BRE00047667/12} page 12, paragraph 23(g). albeit in what he described as an "open-ended"363 Connolly {BRE00047667/12} page 12, paragraph 23(g). general warning in Approved Document B that Class 0 might not provide adequate assurance of safety. It is surprising, therefore, that the department took no action, either then or for many years after, to amend the guidance in Approved Document B to take account of what was known to be the unsuitable nature of Class 0 as a standard for assessing the spread of flame over an external wall. That is particularly so when the cladding systems used by Dr Connolly in his tests had been chosen precisely because they were then in current use on high-rise buildings.
¶7.24 There is no evidence that anyone in the department gave any consideration to introducing a recommendation that only non-combustible materials be used in the construction of external walls of high-rise buildings. Dr Connolly did not think that was necessary364 Connolly {BRE00047667/16} page 16, paragraph 25(f)(i). and neither he, nor it seems anyone else, thought that there was a need to take action as a matter of urgency.365 Connolly {BRE00047667/13-14} pages 13-14, paragraph 24(c)(ii). He considered that the results of the tests that BRE had carried out confirmed the need for a full-scale test to enable evaluation of a complete system366 Connolly {BRE00047667/12} page 12, paragraph 23(f). and that BRE should develop such a test. He envisaged that Approved Document B would then be amended to require external wall systems used on high-rise buildings to pass that test.367 Connolly {BRE00047667/12-13} pages 12-13, paragraphs 23(g)(i) and 24(c)(ii).
1996–1998: Fire Note 3
¶7.25 In the years that followed, staff at BRE (including Dr Connolly368 Connolly {BRE00047667/17} page 17, paragraphs 26(c) and 27(b). and Dr Colwell369 Colwell {Day231/155:6-10}.) worked on the development of a large-scale test method for the fire performance of external cladding systems. That method was initially known as Fire Note 3, which was submitted to the department in 1996370 Colwell {BRE00047571/18} page 18, paragraph 120; Colwell {Day231/179:25} - {Day231/181:4}. and published in 1998.371 {BRE00005868}. It built on the work Dr Connolly had done in 1994372 Colwell {Day231/150:3-5}; {Day231/152:14-17}. and adopted the theoretical basis of his experimental research.373 Colwell {BRE00047571/11} page 11, paragraphs 63 and 67. Following various additional tests on various types of rainscreen cladding and insulated renders (no records of which have been provided to us)374 Colwell {Day231/150:3-5}; Colwell {BRE00047571/11} page 11, paragraphs 61, 62 and 67; Colwell {Day231/161:24} -{Day231/162:22}. and discussions with the department and other interested bodies,375 Colwell {Day231/173:6-11}. three performance criteria were eventually selected against which the performance of a system was to be evaluated: mechanical performance, external fire spread and internal fire spread.376 {BRE00005868/9} paragraph 10.1; Colwell {BRE00047571/12} page 12, paragraph 68.
88¶7.26 No pass or fail criteria were set for mechanical performance.377 {BRE00005868/10} paragraph 10.5; Colwell {BRE00047571/12} page 12, paragraph 68. Fire Note 3 simply stated that observation should be made of any collapse or partial collapse judged to be hazardous.378 {BRE00005868/10} paragraph 10.5. Dr Connolly, who said that he had no specific recollection of the evolution of, or the reasoning underlying, the test method,379 Connolly {BRE00047667/17-18} pages 17-18, paragraphs 27(b) and 27(e). thought that that might have been the result of a comparison with glazing systems, which were not required to exhibit any mechanical resistance when exposed to fire.380 Connolly {BRE00047667/19} page 19, paragraph 27(k). By contrast, Dr Colwell told us that it had proved too difficult to set criteria for mechanical performance because that depended very much on the particular system.381 Colwell {Day231/168:9-15}; {Day231/177:15}-{Day231/178:1}. It was therefore left to the end-user to assess the significance of whatever signs of collapse had been observed.382 Colwell {Day231/178:1-9}.
¶7.27 In relation to both the external and internal spread of fire, Fire Note 3 provided that failure would occur if the temperature of the relevant thermocouples at certain levels on the test rig exceeded 600 degrees for a period of at least 30 seconds within 15 minutes of the start of the test.383 {BRE00005868/9} paragraphs 10.3 and 10.4. Although Dr Colwell was asked to explain the basis on which each of those criteria had been chosen,384 Colwell {Day231/168:25}-{Day231/169:18}; {Day231/170:21}-{Day231/171:12}; {Day231/173:12}-{Day231/174:7}. she was unable to give us any precise information.385 Colwell {Day231/174:11}-{Day231/175:24}. Ultimately, she agreed that she and the other authors of Fire Note 3 had worked collaboratively and had sought to reach a consensus on its contents. The criteria they had adopted were generally accepted as suitable by those involved in testing of that kind,386 Colwell {Day231/176:19-23}. by which we understood her to mean that the department considered them to reflect an acceptable degree of risk. Dr Connolly had a rather vague recollection that he and his manager, Tony Morris,387 Tony Morris was Dr Raymond Connolly's immediate line manager: Connolly {BRE00047667/2} page 2, paragraph 3(c). He supervised the work on and was one of the authors of Fire Note 3: Smith {Day234/159:3}-{Day231/160:1}. had calculated that flames spreading vertically at the rate of 2.5 metres in 15 minutes from a level of 2.5 metres immediately above the source of the fire388 2.5 metres was chosen as representative of the minimum height between storeys, Connolly {BRE00047667/18-19} pages 18-19, paragraph 27(j). did not represent undue fire spread, but he could not remember how or on what basis that calculation had been made.389 Connolly {BRE00047667/18-19} pages 18-19, paragraph 27(j). We have been left with the impression that the criteria emerged from discussions among those involved in the process without being supported by any specific data or calculations.
891999: the Garnock Court fire
¶7.28 On 11 June 1999, a fire broke out in a flat on the fifth floor390 {BRE00035375/4}. of Garnock Court, a fourteen- storey391 {BRE00035375/3} second paragraph. block of flats in Irvine, North Ayrshire, Scotland. By the time firefighters arrived the fire had spread externally up the face of the building to the eighth floor. It continued to spread rapidly thereafter, eventually enveloping a vertical section of the external face from the fifth floor to the roof.392 {BRE00035375/4} third paragraph; {BRE00035375/11} Figure 3.1.3.13 and Bisby, Phase 2 Report {LBYP20000001/126} Figure 15. Three residents were rescued from the seventh floor of the block, several were evacuated safely and one, who had been in the flat where the fire had started, died.393 {BRE00035377/5} section 1 first paragraph; {BRE00035377/8} section 3.2.3.
¶7.29 Garnock Court had been built in 1968394 {BRE00035375/3} second paragraph. and was one of a group of five blocks that had been refurbished in 1991 in an attempt to solve problems of damp.395 {BRE00035375/3} fourth paragraph. As part of that work, glass reinforced polymer (GRP) spandrel panels had been used to clad certain parts of the external walls396 Bisby, Phase 2 Report {LBYP20000001/123} paragraphs 634-635. together with new windows, each of which had been enclosed in a GRP "pod".397 {BRE00035375/3} penultimate paragraph.
¶7.30 BRE, which had become a commercial organisation in 1997, was asked by North Ayrshire Council to investigate the fire.398 {BRE00035375/3} first paragraph; {BRE00035377/5} section 1 first paragraph. It produced two confidential reports dated 8 September 1999399 Report 79902 {BRE00035377/2-13}. and 5 May 2000.400 Report 81310 {BRE00035377/22-56}. The investigation was undertaken in phases and was carried out by Penny Morgan, Brian Martin (both of whom reported at the time to Martin Shipp)401 Shipp {BRE00047594/8-70} pages 8 and 70, paragraphs 47 and 380. and Tony Morris.402 {BRE00035377/2}; {BRE00035377/5} section 2 first paragraph; {BRE00035377/22}. The reports were approved by Nigel Smithies, then BRE's Fire Safety business group manager.403 Shipp {BRE00047594/7} page 7, paragraph 45. This group had responsibility for fire investigations and for the Investigation of Real Fires contract with the department, Shipp {BRE00047594/23} page 23, paragraph 116.
¶7.31 After some years working in construction, first as a carpenter, then as a site manager and finally for nine years as a building control officer,404 Martin {Day250/6:12}-{Day250/7:16}. Brian Martin joined BRE in 1999405 Martin {Day250/15:2-5}. as an expert on the Building Regulations specialising in fire.406 Field {BRE00043710/8} page 8, paragraph 39; Martin {Day250/18:8-14}. It is unclear to us how he was chosen for that position or what experience he had in fire safety matters. He had no fire engineering qualifications and could tell us only that his experience in fire safety came from his nine years working as a building control officer.407 Martin {Day250/18:12-19}. Very shortly after he joined BRE Mr Martin was seconded part-time to the department's Building Regulations Division as the lead consultant providing technical support on fire safety, principally working and advising on Part B of Schedule 1 to the Building Regulations and Approved Document B.408 Martin {CLG00019469/2} page 2, paragraph 6; Martin {Day250/25:11-23}. He remained in that dual role, dividing his time about equally between BRE and the department,409 Martin {Day250/22:20-24}. for nine years until September 2008, when he joined the department full-time as the Principal Construction Professional in the Building Regulations Division.410 Martin {Day250/29:23}-{Day250/30:3}. He worked in that role until November 2017.411 Martin {Day250/39:2-5}. During that period Mr Martin played a central role in reviewing, drafting, implementing and shaping government guidance and policy on the Building Regulations as they related to fire safety. During the period in which BRE was working on the fire at Garnock Court, Brian Martin therefore had one foot in the department and one in BRE. It is not clear whether any of those involved recognised the risk that Mr Martin's involvement might result in the department's interests influencing the advice it obtained from BRE, or if so, how that risk should be managed.
90¶7.32 Separately from its reports into the fire at Garnock Court for North Ayrshire Council, BRE produced a report for the department under the Investigation of Real Fires contract.412 We explain that project in more detail in Chapter 8. The report, dated August 2000,413 {BRE00035375}. was written by Penny Morgan414 {BRE00035375/2}. and submitted to Anthony Burd, then Principal Fire Safety Professional at the department.415 Burd {CLG00019461/1-2} pages 1-2, paragraph 4. As we explain below, the contents of that third report replicated the first report to North Ayrshire in September 1999 almost verbatim, save in two key respects.
¶7.33 In contrast to the report into the fire six years earlier at Knowsley Heights, the report produced by BRE into the fire at Garnock Court416 8 September 1999 {BRE00035377}. recorded in clear terms the contribution of the GRP cladding, which it identified as the main material involved in the fire.417 {BRE00035377/10} section 4, first paragraph. It said that the GRP panels had been ignited by the fire plume spilling from the living room of the fifth floor flat in which the fire had started418 {BRE00035377/4} paragraphs 1 and 2. and that the GRP had then generated a self- propagating fire.419 {BRE00035377/10} section 4, first paragraph. It recorded rapid external fire spread to the full height of the building, with full involvement of the GRP cladding within 15 minutes, and noted that the even burning of the GRP cladding and the production of flames and dense black smoke indicated the involvement of the GRP alone.420 {BRE00035377/7} first paragraph. Investigators were able to make these observations from video footage of the fire captured on security cameras, {BRE00035377/5} Section 2, second paragraph; {BRE00035379/12} paragraphs 3-5. In a section dealing with the remedial measures planned for the five blocks in Irvine, BRE suggested that non-combustible materials should be chosen wherever possible.421 {BRE00035377/10} section 4, paragraph 6.
¶7.34 BRE noted that, in order to comply with the provisions of Approved Document B at the time, the cladding panels should have been certified Class 0422 {BRE00035377/10} section 5, paragraph 4. and expressed doubt about whether in their aged state at the time of the fire the panels met that requirement.423 {BRE00035377/10} section 4, second paragraph and section 5, paragraph 5. For that reason, BRE recommended a series of tests on undamaged samples of the panels to assess (amongst other things) their surface spread of flame and fire propagation characteristics.424 {BRE00035377/10} section 5, paragraph 6.
91¶7.35 To that end, on 16 February 2000 BRE carried out tests in accordance with BS 476 Part 6 on three specimens425 {BRE00035377/38} paragraph 4.1. from the panels.426 {BRE00035377/37}. They reacted so badly427 The sample (tested as three specimens) achieved a fire propagation index of 20.8 {BRE00035377/42} paragraph 8. that they put the integrity of the test equipment itself at risk. Following that test, BRE decided to abandon any further testing of the GRP panels in accordance with BS 476 Parts 6 or 7. It was clear from that test alone that the GRP panel would not obtain a Class 0 classification.428 {BRE00035377/26} "Results of Test 2". Shipp {BRE00047594/74-75} pages 74-75, paragraph 416. As a result, BRE concluded not only that it was highly unlikely that the panels used at Garnock Court had been Class 0 at the time of the fire but also that it was highly unlikely that the panels would ever have obtained a Class 0 classification.429 {BRE00035377/29} paragraphs 1 and 2. In other words, the fire at Garnock Court was primarily caused by a failure to follow the statutory guidance then in force.430 As BRE noted in an internal outline document for the report for North Ayrshire, "Can we say that we think they were sold a pup?": {BRE00035380/18}; Martin {Day250/135:7-20}.
¶7.36 Despite its obvious importance, neither that finding nor any mention of the test on the GRP panels in accordance with BS 476 Part 6 was included in the report BRE sent to the department in August 2000.431 {BRE00035375}. In fact, that later report contained no reference at all to Class 0. Having examined the two reports carefully, we think that in at least one case432 {BRE00035377/6} section 3.1.1 as compared with {BRE00035375/4} first paragraph, where the words "either... or Class 0" from the former has clearly been removed in the latter. (and probably in all the others) the reference to Class 0 must have been deliberately removed from the text of the later report, which otherwise reproduces the earlier one.433 Bisby, Phase 2 Report {LBYP20000001/134-136} paragraphs 704-710. We have been unable to determine why or by whom the changes were made, not least because none of the witnesses was able to shed any light on the matter.
¶7.37 Penny Morgan, one of the authors of the reports provided to North Ayrshire Council and the sole author of the report provided to the department, could not remember how the references to Class 0 had been omitted from the report she had sent to the department in August 2000, although she said that if there had been a conscious decision to remove them, she would have remembered it.434 Morgan {BRE00043866/51} page 51, paragraphs 302 and 303. Brian Martin, the other author of the report provided to North Ayrshire Council in September 1999, had no recollection of reading either the report given to the department or the second report given to North Ayrshire Council.435 Martin {Day 250/112:12-21}; {Day250/115:6-13}. He told us that he knew nothing about the removal of any reference to Class 0 from the report to the department, had not been aware of it at the time or since436 Martin {Day250/118:23}-{Day250/119:17}; {Day250/137:16-24}. and could offer no explanation for the discrepancy between the reports.437 Martin {Day250/118:13-22}; {Day250/157:11-16}. Their line manager, Martin Shipp, could not recall whether he had read or reviewed the reports to North Ayrshire Council at the time438 Shipp {BRE00047594/70} page 70, paragraph 383. and did not know whether a decision to remove references to Class 0 had been made.439 Shipp {BRE00047594/78} page 78, paragraph 432.
¶7.38 Martin Shipp's evidence that the report given to the department contained no more than a factual description of the findings of the investigation440 Shipp {BRE00047594/78} page 78, paragraph 430. does not provide an answer to the mystery. The fact that the panels did not meet the requirements of Class 0 was the key finding, as Mr Martin agreed.441 Martin {Day250/135:7-20}. Even if Mr Shipp's suggestion that that report (although dated August 2000) could have been written before the reports for North Ayrshire Council442 Shipp {BRE00047594/78} page 78, paragraph 431. were correct, it would not explain why BRE did not amend the text to include the important finding that the GRP panels did not meet the requirements of Class 0 before sending it to the department in August 2000. Mr Martin accepted that that the absence of any reference to Class 0 in the report sent to the department was strange,443 Martin {Day 250/137:20}; Martin {Day250/157:15} "It does seem odd". but denied that the references had been removed on the instructions of either the department or BRE.444 Martin {Day250/118:18-22}.
92¶7.39 Equally surprising, if true, is the fact that, as Brian Martin told us, he did not discuss BRE's findings about the cladding panels with Mr Burd, to whom he reported at the time445 Martin {Day250/24:14-16}. and with whom he worked closely throughout the period.446 Martin {Day250/127:18-23}; {Day250/128:14-22}; {Day250/136:21}-{Day250/137:8}. Anthony Burd told us that he had not previously seen BRE's second report to North Ayrshire Council dated 5 May 2000, which contained the results of its tests.447 {BRE00035377/22}; Burd {Day238/136:20-25}. Mr Burd was concerned by the absence of any reference to Class 0 in the report to the department.448 Burd {Day238/142:2-7}. He said that it was imperative that the department should receive the fullest information about fires investigated for it by BRE.449 Burd {Day238/143:8-11}. Following the fire neither the department nor (so far as we are aware) any other authority carried out any investigation into the use of cladding panels that did not comply with existing guidance,450 Martin {Day250/137:25}-{Day250/140:17}. but without some kind of investigation the department could not have known whether the problem was limited to the refurbishment of the blocks in Irvine or was more widespread.451 Martin {Day250/140:6-17}. Viewing the evidence in the round, we think it much more likely than not that the omission of any reference to Class 0 in the report to the department was deliberate, but we are unable to identify the person responsible for it or the reason why the information was suppressed.
¶7.40 There was a second important omission from the report to the department, namely, the suggestion that non-combustible materials should be used wherever possible, which had been included in the second report to North Ayrshire.452 Compare {BRE00035377/10} Section 4, third paragraph "The remedial measures planned for the high-rise blocks in Irvine should address the problems identified i.e. damp penetration and the avoidance of an external route for fire spread. We suggest that non-combustible materials are chosen wherever possible" with {BRE00035375/7} fifth paragraph "The remedial measures planned for the high-rise blocks in Irvine should address the problems identified i.e. damp penetration and the avoidance of an external route for fire spread". Again, none of the witnesses could explain that.453 Martin {Day250/136:21}-{Day250/137:8}. We accept that the suggestion was probably directed to the remedial work at the five blocks in Irvine,454 Martin {Day250/125:12-19}. but we are nonetheless surprised that BRE did not advise the department that only non-combustible materials should be used in the construction of the external walls of high-rise buildings and that the department itself did not consider the implications of using combustible materials in the context of the Building Regulations and guidance.455 Martin {Day250/143:15-21}; {Day250/125:12-19}.
¶7.41 We find those omissions surprising when considered against the background of the fire at Knowsley Heights, which was not a case of a failure to comply with the recommendation in the Approved Document for the use of Class 0 materials, and the apparently clear understanding that the panels used at Garnock Court had been highly combustible.456 Martin {Day250/101:2-5}; {Day250/131:15-19}; Burd {Day239/135:23-25}; Morgan {BRE00043866/44} page 44, paragraphs 264 and 266. They were particularly surprising, given that Mr Martin himself told us that one of the most important lessons drawn from the fire at Garnock Court had been not to use combustible cladding.457 Martin {Day250/102:10-14}. In contrast to his colleagues Martin Shipp,458 Shipp {BRE00047594/66} page 66, paragraphs 354 and 356. Dr Smith, Dr Colwell459 Colwell {Day231/130:13-17}; {Day231/131:15}-{Day231/132:2}. and Dr Connolly,460 Martin {Day250/93:10-17}. Brian Martin did not share the view (which on their evidence had become widely accepted following the fire at Knowsley Heights) that Class 0 was an inadequate standard for assessing or controlling the danger of fire spreading across the external walls of high-rise buildings. He had regarded Class 0 as a reasonable form of control461 Martin {Day250/124:6-14}; {Day250/126:15}-{Day250/127:5}. and told us that he had never seen the report produced by Dr Connolly in 1994.462 Martin {Day250/93:5-11}.
93¶7.42 The section of BRE's report to the department entitled "Implications for the Building Regulations" stated only that the matter had been thoroughly explored by a Select Committee.463 {BRE00035375/7} final paragraph. In fact, by August 2000, a Parliamentary Select Committee had recommended that the full-scale test and performance criteria developed by BRE that were subsequently adopted by the British Standards Institution as BS 8414 should be substituted in Approved Document B for the previous requirements relating to the fire safety of external cladding systems.464 {CLG00019478/9-10} paragraph 20. However, as BRE was well aware by August 2000, that recommendation had not been implemented by the department.465 {CLG10000347/2-3} paragraphs 8-11.
Fire Note 9
¶7.43 In 1999, BRE published Fire Note 9,466 {CTAR00000019}. a revised version of the test method that had been set out in Fire Note 3.467 {BRE00005868}. Fire Note 9 was written by Dr Colwell and David Smit, who at the time was a laboratory technician and therefore junior to her.468 Colwell {Day231/198:13-16}. As Dr Colwell explained,469 Colwell {BRE00047571/19-21} pages 19-21, paragraphs 127 and 134-135; Colwell {Day231/200:8}-{Day231/202:7}. Fire Note 9 contained no technical or substantive changes to the test method or assessment criteria set out in Fire Note 3. Similarly, the drafting of Fire Note 9 did not involve any reconsideration of the theoretical basis for the test method. The only revisions to Fire Note 3 were the inclusion of a definitions section,470 {CTAR00000019/4} section 3. a sample graph (for showing the determination of test start time and temperature)471 {CTAR00000019/10}. and a maximum thickness for test samples to ensure that samples would fit on to the test facility.472 {CTAR00000019/4} clause 1.2.
1999: recommendations of the Parliamentary Select Committee
¶7.44 Following the fire at Garnock Court, the Environment, Transport and Regional Affairs Select Committee,473 {CLG00019478/2} for a list of members of the Committee. acting through its Environment sub-committee, established an inquiry to examine the potential risk of the spread of fire through external cladding systems.474 {CLG10000349/1}.
94¶The Select Committee heard evidence on 20 July 1999 from a variety of witnesses,475 {CLG10000349/4}. including witnesses from BRE476 {CLG10000349/19-22}. and the department.477 {CLG10000349/32-37}. It also received memoranda from the department478 {CLG10000349/29-32}. and several other organisations.479 The memoranda are listed at {CLG10000349/3} and included either as appendices to the minutes of evidence {CLG10000349/39-51} or within the evidence {CLG10000349/5-11}; {CLG10000349/17-18}; {CLG10000349/23}; {CLG10000349/29-32}.
¶7.45 The Select Committee reported on 14 December 1999.480 {CLG00019478}. It concluded that the evidence it had heard did not suggest that the majority of external cladding systems currently in use in the UK posed a serious threat to life or property in the event of fire481 {CLG00019478/9} paragraph 18. and it made a number of recommendations. Notwithstanding that conclusion, however, it added that it did not believe that it should take a serious fire in which many people were killed before all reasonable steps were taken towards minimising the risks and it went on to note that the evidence it had heard strongly suggested that the small-scale tests then in use to determine the fire safety of external cladding systems were not fully effective in assessing the performance of such systems in a fire.482 {CLG00019478/9} paragraph 19. (Those small-scale tests included the BS 476-6 and BS 476-7 tests which supported the Class 0 classification.) The Select Committee noted that the department had said that Fire Note 9 would be referred to in Approved Document B but pointed out that that would represent only one of the ways of ensuring that the system complied with the Building Regulations and would not amount to a requirement that cladding systems pass the test.483 {CLG00019478/9} paragraph 17. The Select Committee therefore expressed the view that all external cladding systems should be required either to be entirely non-combustible or be proved through full-scale testing not to pose an unacceptable level of risk through the spread of fire484 {CLG00019478/9-10} paragraph 20. and recommended that compliance with Fire Note 9 be substituted in Approved Document B for the existing requirements relating to the fire safety of external cladding systems.485 {CLG00019478/9-10} paragraph 20 and {CLG00019478/7-8} paragraphs 9-10 and 12. The Select Committee's recommendation was therefore clear: that only non-combustible materials should be used, unless the entire construction had been tested as a complete system and met the performance criteria in Fire Note 9.
2000: The department's response
¶7.46 However, the Select Committee's recommendation was not implemented. Instead, in the revised edition of Approved Document B that was published in January 2000 and came into force on 1 July 2000486 {INQ00014107}. the guidance in the 1992 edition that the external surfaces of walls of high-rise buildings should be Class 0487 {BLA00005482/74-75} paragraphs 12.5, 12.6 and Diagram 36. remained unchanged,488 {INQ00014107/89-90} paragraphs 13.5, 13.6 and Diagram 40. although Fire Note 9 was referred to as an alternative to meeting the provisions of Diagram 40, which contained the Class 0 guidance.489 {INQ00014107/89} Note to paragraph 13.5.
95¶7.47 In his response to the Select Committee on 6 April 2000,490 {CLG10000347}. the Rt Hon Nick Raynsford MP, Parliamentary Under Secretary of State and Minister for Construction at the time,491 Raynsford {CLG00035627/3-4} pages 3-4, paragraphs 6-7. who had given evidence before the committee, explained that decision on behalf of the department.492 {CLG10000347/3} paragraph 9. He said that during its review of the 1992 edition of the Approved Document, which had been completed by December 1999,493 {CLG10000347/2} paragraph 5. there had been no suggestion that the existing guidance was insufficient or that, if followed, it would tend to create an unsafe situation.494 {CLG10000347/2} paragraph 6. Mr Raynsford also said that when Fire Note 9 had been adopted as a British Standard and Approved Document B had been amended to reflect the new status of the test method, the department would consider whether the reference to that method of demonstrating compliance should be strengthened.495 {CLG10000347/3} paragraph 11.
¶7.48 In the event, there is no evidence that the department ever considered strengthening the reference to full-scale testing in 2002 when the European supplement to the Approved Document was published and references to European test methods were incorporated into Diagram 40.496 {CLG10000740/90-91} paragraphs 13.5, 13.6 and Diagram 40; Martin {Day250/166:2-15}. In fact, for the purposes of that review it was agreed by the members of the working party that consideration of any amendments to Approved Document B for the 2002 revision would be limited to facilitating the process of harmonisation and would not involve reconsideration of any of the general technical guidance.497 {CLG10000508/2} paragraph 7; Martin {Day250/166:13-15}. Nor have we seen evidence that any serious consideration was given to adopting the Select Committee's recommendation in the reviews, consultations or discussions leading to the publication of the 2006,498 {CLG10000007/95-97} paragraphs 12.5, 12.6 and Diagram 40. 2007,499 {CLG10000005/98-100} paragraphs 12.5, 12.6 and Diagram 40. 2010500 {CLG00000161}. or 2013501 {CLG00000224/95-97} paragraphs 12.5, 12.6 and Diagram 40. editions of Approved Document B. In all those editions the guidance continued to refer to a Class 0 surface as an appropriate standard for external cladding on high-rise buildings. Meeting the criteria in BR 135 following full-scale testing in accordance with BS 8414 was included as an alternative.502 Burd {Day238/177:18-25}; Martin {Day250/167:6-16}.
¶7.49 Largely due to the passage of time, few witnesses were able to give us any insight into the reasoning behind the department's response to the Select Committee's recommendations. Paul Everall, then Deputy Director of the Building Regulations Division503 Everall {CLG00019482/1} page 1, paragraphs 3-5. could not recall what had led to the formulation of the department's response.504 Everall {CLG00019482/10-13} pages 10-13, paragraphs 44 and 57. Anthony Burd thought505 Burd {Day238/174:12-20}. that the advice to ministers not to adopt the Select Committee's recommendation had been given by Tony Edwards506 Principal Fire Safety Professional in the Department's Building Regulations Division at the time, Burd {Day238/136:15-18}. and did not know why or how it had been formulated.507 Burd {Day238/175:2-5}.
¶7.50 Mr Raynsford told us that he had no recollection of the Select Committee's report but that he was confident that he had read it and had taken advice on it from officials at the time.508 Raynsford {CLG00035627/21} page 21, paragraph 52. He recalled that the department had had concerns about the application of small-scale fire tests in the context of large-scale cladding systems,509 Raynsford {CLG00035627/18} page 18, paragraph 40. but did not recall any concerns about reliance on Class 0 as a classification for materials that could be used on the external walls of high-rise buildings.510 Raynsford {CLG00035627/18} page 18, paragraph 41. He did not recall having given any consideration to the creation of guidance recommending that all components of an external wall system should be of limited combustibility and did not recall the proposal having been brought to his attention by any official.511 Raynsford {CLG00035627/20} page 20, paragraph 48. He told us that his role in the department's response had been to consider the advice of officials and give directions at a high level.512 Raynsford {CLG00035627/22} page 22, paragraph 54.
96¶7.51 There was only one ministerial submission which touched on the subject. It was dated 9 March 2000 and was drafted by Caroline Cousin, the Building Regulations Division's Head of Technical Policy. The submission was sent to Deputy Director Paul Everall, who approved it, and then sent to the minister513 {CLG00019457/1}. It stated that none of the Select Committee's recommendations was substantial in nature514 {CLG00019457/1} paragraph 4. and explained that the draft response the minister was being asked to approve had assumed that he would wish to accept the spirit of the recommendations.515 {CLG00019457/2} paragraph 9. It referred to the recommendation that the BRE large scale test for the fire performance of external cladding systems should be substituted in Approved Document B as a requirement for the safety of cladding systems but it did not tell the minister that the Select Committee believed that all external cladding systems should be entirely non-combustible unless proved through full-scale testing not to pose an unacceptable level of risk. It therefore failed to provide the minister with enough information to question the rationale behind the department's response or to decide to adopt a different course. None of the witnesses was able to assist us with any aspect of the drafting of the submission. Although Mr Burd suggested that more comprehensive advice and options might have been given to the minister in separate discussions,516 Burd {Day238/182:21}-{Day238/184:10}. we saw no evidence of that.
¶7.52 Brian Martin said that in the period after the publication of the 2000 edition of Approved Document B he had not asked his colleague Anthony Burd why the Select Committee's view had not been followed.517 Martin {Day250/170:5-10}. It seems that within the department the matter was considered closed.518 Martin {Day250/165:20-22}.
2000–2001: review of BR 135
¶7.53 In response to the Select Committee's inquiry and the pending adoption of Fire Note 9 as a British Standard test, in January 2000 the department entered into a contract with BRE519 Contract CC1924. to carry out research leading to the proposed revision of BR 135,520 {CLG10000347/2} paragraphs 6-7; {BRE00041836/3} final paragraph; Colwell {BRE00047571/29-30} pages 29-30, paragraphs 185-187; Colwell {Day232/12:15-25}; Smith {Day234/195:23}-{Day234/196:8}; Martin {Day250/206:13-18}. the first edition of which had been published in 1988.521 {BRE00001077}.
97¶7.54 Among other things, the department's objectives were to review the guidance in Approved Document B on the external walls of multi-storey buildings, to revise the Building Regulations and guidance based on a series of experimental studies on new and existing cladding systems522 {BRE00041836/3} section 2. and to use the large-scale test to determine the most appropriate method for specifying the fire performance requirements of cladding systems.523 {BRE00041836/4} section 2, second bullet point; {BRE00041836/6} task 6. The work fell into five main parts: a literature review, a survey of building owners, analysis of the survey results, an experimental testing programme and the drafting of a revised edition of BR 135,524 Colwell {Day232/13:1-14}. each of which we consider below.
¶7.55 Dr Colwell was the project leader.525 {BRE00041836/1}; {BRE00041836/9} "Project Team"; {BRE00041836/11} sixth paragraph. Brian Martin was responsible for the survey and the drafting of BR 135.526 {BRE00041836/9} "Project Team"; Colwell {Day232/14:17}-{Day232/15:13}; Martin {Day250/207:18}-{Day250/208:8}. Both Dr Colwell and Mr Martin were involved in the project from beginning to end527 Colwell {Day232/13:18-20}; {Day232/15:14-16}; Martin {Day250/208:10-14}. and all the work on the project, including in relation to technical detail, was supervised and monitored by Dr Smith.528 {BRE00041836/11} second and sixth paragraphs; {BRE00041836/12} "Project Responsibilities"; Martin {Day250/211:19}-{Day250/212:1}; Colwell {Day232/17:19-21}. Dr Colwell and Dr Smith met frequently during the course of the project.529 Colwell {Day232/17:19}-{Day232/18:6}. Dr Smith was responsible for supervising the work and exercising quality control530 {BRE00041836/12} "Project Responsibilities". and Dr Colwell consulted her whenever necessary.531 Smith {Day234/198:5-14}. All project reports were submitted to Anthony Burd at the department.532 Burd {Day239/99:11-14}.
The literature review: 2000
¶7.56 The first piece of work produced under the project was a report entitled Fire Spread in External Cladding – a Literature Review dated 30 March 2000.533 {BRE00001353/4-29}. It was written by Dr Colwell, Brian Martin and J Foster534 Mr Foster was a junior member of the Reaction to Fire team: Colwell {Day232/29:25}-{Day232/30:4}; Smith {Day234/205:6-8}. and was approved by Dr Smith.535 {BRE00001353/4}. The executive summary stated that the report identified and summarised the types of external cladding systems then in use, the existing requirements and guidance contained in the 2000 edition of Approved Document B and the research previously undertaken on the spread of fire across the external walls of high-rise buildings.536 {BRE00001353/6} Executive Summary. Dr Colwell told us that the purpose of the review was to provide a point of reference based on previously published information that would provide a platform for future work.537 Colwell {BRE00047571/24} page 24, paragraph 155; Colwell {Day232/32:22}-{Day232/33:3}.
¶7.57 Among other matters, the review resulted in a number of findings, two of which are important for present purposes. First, that a large-scale test method was necessary to assess the performance in fire of a complete external cladding system. Secondly, that the revised (2000) edition of Approved Document B had gone some way to addressing "the issues of fire performance of external cladding systems" and that the review of BR 135 would help to "clarify any remaining issues".538 {BRE00001353/27} paragraphs 2-3.
98¶7.58 The word "issue" used in that way is notoriously imprecise but usually connotes something unsuitable or unsatisfactory. It is not clear from the paper itself what "the issues of fire performance of external cladding systems" were and neither Dr Colwell nor Mr Martin could tell us. Although there were references to Approved Document B, none of them suggested that it was defective in any way. Nor was either of them able to explain the basis of the finding that the 2000 edition of Approved Document B had gone some way towards addressing them. The most that either of them could tell us was that they had been referring to the addition of the new reference to the large-scale test method539 Colwell {BRE00047571/25} page 25, paragraph 159. and the guidance that both surfaces of cladding panels used in the construction of the external walls of high-rise buildings should be classified Class 0.540 Martin {Day251/19:8-21}. Neither of them explained what "issues" they had thought would be clarified by BRE's work on BR 135.541 Colwell {Day232/46:13}-{Day232:50:2}; Martin {Day251/19:8}-{Day251/21:12}. Mr Burd was equally unclear about those matters, saying only that he could not recall precisely what the issues were but that there was a need to look at full-scale testing and the guidance provided by BR 135.542 Burd {Day239/4:14}-{Day239/6:1}.
¶7.59 The literature review was, in our view, a poor piece of work, as the uncertainty surrounding its conclusions demonstrates. It was superficial at best, even as a reference point, and contained at least one significant inaccuracy. The report appeared to misrepresent what Class 0 actually signified and conflated it with non-combustibility, thereby repeating the error in the report of BRE's investigation into the fire at Knowsley Heights. It was an error that was repeated many times in the years to come, particularly in BRE's summaries of the relevant provisions of Approved Document B.543 {BRE00041986/15}; {BRE00005881/18}; {CLG00019202/2} paragraph 7; {CLG00019445/3}. Most significantly, in a section summarising the guidance given in the 2000 edition of Approved Document B, the report stated that "Diagram 40…restricts the combustibility of external walls of high buildings".544 {BRE00001353/14} first paragraph. That was not correct. Diagram 40545 {INQ00014107/90}. did not restrict the combustibility of materials used in the external walls of buildings over 18 metres in height; it described an appropriate classification for the surfaces of external wall products that was primarily concerned with the spread of flame.
¶7.60 Dr Colwell said that Mr Martin had written the paragraphs summarising the provisions of the guidance,546 Colwell {Day232/36:15-18}. that Approved Document B was outside her area of expertise547 Colwell {Day232/36:11-14}; {Day232/37:11-13}. and that she had not noticed the error at the time.548 Colwell {Day232/37:14-18}. Dr Smith, who had approved the report and was responsible for the technical accuracy of the project as a whole,549 Smith {Day234/198:8-11}; {Day235/85:11-14}. told us that her review had been essentially editorial and typographical.550 Smith {Day234/206:15}-{Day234/207:6}. She accepted that there appeared to be a degree of confusion in the mind of the author about the effect of Diagram 40551 Smith {Day235/5:6-17}. and said that she would not have described it in that way.552 Smith {Day235/3:23-24}. All in all, the review does not reflect the work of an institution wedded to rigorous thinking, careful analysis or precise expression.
99¶7.61 Although he said that he could not remember having done so, Mr Martin probably did write the sections of the literature review dealing with the 2000 edition of Approved Document B.553 Martin {Day251/5:8-12}. He accepted that the statement that Diagram 40 restricted combustibility was not accurate but did not go so far as to agree554 Martin {Day251/8:19}-{Day251/9:6}. with Mr Burd that the use of the word "combustibility" had been a basic mistake.555 Burd {Day238/193:20-23}. He attempted to justify his use of it by saying that he had used it in a lay sense.
¶7.62 Just as we are critical of the way in which BRE expressed itself in the literature review, we are equally critical of the loose way in which Mr Martin used expressions such as "combustibility" and "Class 0" when giving his evidence. We have already referred to his use of the word "combustibility". He also said that he considered Class 0 to form part of a sliding scale on which there were degrees of combustibility.556 Martin {Day251/7:2-10}; {Day251/10:6-7}. We found this aspect of his evidence unsatisfactory because we should have expected someone in his position to use language more carefully when dealing with matters on which he was credited with some expertise. That would not have been "pedantic", as he put it, merely careful and precise. However, his casual use of language had led us to question whether his grasp of the concepts with which he was dealing was as sure as it should have been. In our view Mr Burd was correct to describe the use of the word "combustibility" as a basic mistake. Despite those misgivings, we think that Mr Martin was in fact well aware of the distinctions between Class 0, limited combustibility and non-combustibility and of their importance in the context of Approved Document B and that his unsatisfactory evidence reflected an attempt to justify his poor choice of language.
¶7.63 The BRE's literature review contained three tables setting out the typical costs of using different cladding systems based on information obtained from the February 1998 edition of the Architects Journal.557 {BRE00001353/12-13} Tables 1-3. In the third table, relating to "In-fill panels", figures were given for a composite panel made of 0.5mm aluminium skins with a 3mm polyethylene core,558 {BRE00001353/13} row 4. i.e. ACM panels with polyethylene cores.559 Colwell {Day232/42:21-23}; Burd {Day238/194:23}-{Day238/195:6}; Smith {Day234/13:19-20}; Martin {Day251/12:14-16}. Dr Colwell, Dr Smith, Mr Martin and Mr Burd each told us that, at the time of the report, they had not been aware that material of that kind was being used in the external walls of high-rise buildings in England and Wales.560 Martin {Day251/12:18-22}; Colwell {Day232/42:24}-{Day232/43:2}; Smith {Day234/14:23}-{Day234/15:4}; Burd {Day238/195:12-19}.
¶7.64 Mr Martin told us that he had not noticed at the time the reference to aluminium composite panels with polyethylene cores.561 Martin {Day251/17:9-11}. Anthony Burd said that he had not noticed it either.562 Burd {Day238/195:2-9}. Dr Colwell said that she had no experience of the product and that although she would have been aware that it was combustible (as were Mr Martin and Dr Smith),563 Martin {Day251/14:9-14}; Smith {Day235/17:20-24}. she was not particularly struck by its inclusion and could remember no discussions about it either within BRE or with the department.564 Colwell {Day232/43:7}-{Day232/45:6}. Nor could Dr Smith or Mr Martin.565 Smith {Day235/19:21-24}; Martin {Day251/16:15-17}. None of them paid sufficient attention to the fact that the product, or something very similar to it, had been included in the experimental phase of the project because it was of interest to the industry566 {BRE00041882}. and none of them appears to have realised that it would have been difficult to obtain costs for aluminium composite panels from the Architects Journal if they had not been in use in the United Kingdom at the time.567 Colwell {Day232/43:16-19}.
100The survey
¶7.65 One of the Select Committee's recommendations had been that the department should instruct local authorities to undertake a review of existing buildings in order to ascertain how many of them had external cladding systems and how many of those might not comply with the Building Regulations. The Select Committee had recommended that in the light of the responses fire safety assessors should be called on to advise on the work required to ensure that none of them posed any undue risk.568 {CLG00019478/10} paragraph 22. As a result,569 {BRE00041887/5} first and second paragraphs; Colwell {Day232/51:3-20}. one of the tasks set by the department as part of the project was to carry out a survey of high-rise buildings in Great Britain in order to determine the composition and design of existing cladding systems and the changing nature of the materials currently in use.570 {BRE00041836/3} section 2, fourth bullet point.
¶7.66 To that end Dr Colwell designed a questionnaire571 Colwell {Day232/52:7-12}. directed to local authorities and companies in the construction industry seeking information about the number of buildings over 18 metres in height with which they had been involved, their age, the proportion that had external cladding, whether it had been installed as part of a refurbishment and the nature of the cladding systems and the materials used.572 {BRE00041885}. The questionnaire also asked respondents whether they had experienced the spread of fire due to external cladding systems and if so, to provide details.573 {BRE00041885/5} question 11.
¶7.67 In the event, only 45 questionnaires were sent out.574 {BRE00041886} "Circulation List"; {BRE00041887/5} "Distribution and Selection". Responses were received from 17 recipients, four of which contained no relevant information575 {BRE00041886} "Introduction". and only eight of which came from local authorities.576 {BRE00041887/7} Figure 1. The decision to send out only 45 questionnaires and to proceed with only 13 completed returns was taken by Mr Burd,577 Colwell {Day232/53:25}-{Day232/54:5}; {Day232/57:24}-{Day232/58:23}. who told us that it had been taken on the advice of BRE and the Industry Advisory Group,578 Burd {Day239/9:2-8}; {Day239/9:18}-{Day239/10:24}. a group of 27 representatives drawn from building owners and the construction industry579 {BRE00001392/5}; {BRE00001392/9} Table 1. to provide advice and support during the project.580 Colwell {Day232/22:2}-{Day232/23:4}; Smith {Day235/201:2-13}; Burd {Day238/190:13-20}; Smith {MET00081237/12-13} pages 12-13, paragraph 44. One of its functions was to make sure that the products and systems tested during the experimental phase of the project were representative of materials and designs in use at the time.581 Colwell {Day232/23:7}-{Day232/24:10}.
¶7.68 It is clear that the data from which BRE worked had not been derived from the comprehensive review of the country's building stock that the Select Committee had originally envisaged.582 Colwell {Day232/61:1-9}; Burd {Day239/13:3-6}. Nonetheless, Dr Colwell thought it contained enough information to provide a reliable view of the types of external cladding systems currently being used in the public housing sector583 {BRE00041887/3} first bullet point. and the department apparently considered it to be representative.584 Burd {Day239/14:10-22}.
101¶7.69 Mr Martin collated the responses to the survey,585 Colwell {Day232/54:16-24}. which were also provided to the department in tabulated form.586 Colwell {Day232/56:25}-{Day232/57:2}. It is clear that no immediate action was taken as a result. The report to Mr Burd informed the department that the use of cavity barriers appeared to be very sporadic,587 {BRE00041887/10} "Fire Breaks". a finding that piqued his interest and eventually led to a further piece of research on the use of cavity barriers some years later for the guidance in the 2006 edition of Approved Document B.588 Burd {Day239/15:21}-{Day239/16:11}.
¶7.70 One of the companies that responded to the survey, WS Atkins, said in response to the question about the spread of fire caused by external cladding systems "Spread of flames generally rapid due to loss of integrity of composite aluminium panels using combustible cores".589 {BRE00041886} row 54, column C. Neither Dr Colwell nor Dr Smith could remember receiving that response or knew whether the company had been asked for further information.590 Colwell {Day232/55:22}-{Day232/56:6}; Smith {Day235/52:7-16}. Nor is there any evidence that anyone approached WS Atkins on behalf of the department to seek further information. That is surprising, because it ought to have alerted BRE and the department to a potentially widespread problem relating to ACM panels.
Experimental testing programme
¶7.71 The information produced by the survey was used as the basis for the programme of tests that subsequently took place, including the choice of the types of system to be tested.591 {BRE00041887/10-14}; Colwell {Day232/69:22-24}. In 2001 and 2002, BRE carried out a series of full-scale tests on cladding systems and small- and intermediate-scale tests on various products. They included 14 full-scale tests (using the test method in Fire Note 9),592 {BRE00041882/23} Table 8; {BRE00041913}. national reaction to fire tests in accordance with BS 476 Parts 6 and 7 on 11 products593 {BRE00041882/12} Table 2. and several tests relating to the relevant European classifications (BS EN ISO 11925 Part 2,594 {BRE00041882/13} Table 3. BS EN 13823595 {BRE00041882/15} Table 4. and ISO 9705596 {BRE00041882/17} Table 5.). The testing was done as part of the project that was itself carried out under a framework contract with the department bearing the number cc1924. As a result, the testing series became known as "cc1924".597 {BRE00001353/1}.
¶7.72 The 11 products and systems selected for testing were chosen in collaboration between BRE, the Industry Advisory Group and the department, with the department having the final say.598 Colwell {Day232/22:8-15}; {Day232/24:2-4}; {Day232/69:25}-{Day232/70:10}; {Day232/75:22}-{Day232/76:24}; Smith {Day235/92:22-25}; {Day235/94:9-22}. Although Mr Burd approved the testing programme, it was devised by BRE and the Industry Advisory Group with little contribution from the department.599 Burd {Day239/23:8-11}; {Day239/25:17-24}. Materials and products were chosen for testing largely on the advice of the Industry Advisory Group about what was most relevant and pertinent to the industry.600 Colwell {Day232/23:1-15}. The aim was to test materials that were widely used so that the tests would be of maximum value both to the department and the industry.601 Colwell {Day232/70:5-10}.
102¶7.73 BRE's report on the cc1924 tests, written by Dr Colwell and approved by Dr Smith, was dated 19 September 2002 but the results of those tests were not published until after this Inquiry had begun its investigations.602 Burd {Day239/104:22}-{Day239/105:9}. They revealed two matters of great significance, neither of which appears to have been met with any appropriate response from either BRE or the department, let alone any action on the part of either of them.
Products sold as Class 0 did not achieve Class 0
¶7.74 All 11 of the products selected by BRE for testing in accordance with BS 476, Parts 6 and 7 had been chosen on the basis that they were marketed as being classified Class 0.603 Colwell {Day232/112:15-24}. However, only four actually achieved that classification when tested.604 {BRE00041882/11} first paragraph; {BRE00041882/19} first paragraph; {BRE00041882/12} Table 2, final column. BRE's report simply recorded that the results were not as expected605 {BRE00041882/11} first paragraph. and said that there did not appear to be any one reason for that lack of performance.606 {BRE00041882/19} first paragraph. Brian Martin said that he could not remember having seen any of the small-scale test data,607 Martin {Day251/65:22-24}. and doubted that he had seen it until he was preparing to give evidence before us.608 Martin {Day251/66:22-23}. He said that he had no recollection of any reaction to or discussions about those findings at the time609 Martin {Day251/65:22}-{Day251/67:16}. but that they might have been discussed in his absence.610 Martin {Day251/67:23-25}.
¶7.75 There are two possible explanations for those results: either the manufacturers were misrepresenting the characteristics of their products (deliberately or otherwise) or there were problems reproducing the results of the tests used to achieve the original classification. Either way, the results were a matter of serious concern to BRE.611 Colwell {Day232/113:25}-{Day232/114:2}. Dr Smith did not think that the explanation lay in the reproducibility of the tests612 Smith {Day235/133:11-16}. and she and Dr Colwell agreed that, since it was not possible to know whether the same situation applied more widely across the market, the matter needed to be considered further.613 Colwell {Day232/115:1}-{Day232/116:1}; Smith {Day235/136:24}-{Day235/137:3}.
¶7.76 Nonetheless, neither BRE nor the department contacted the manufacturers of the seven products that had failed to achieve Class 0 in the tests to tell them of the results614 Colwell {Day232/116:2-5}. and neither of them took steps to report the manufacturers to National Trading Standards or any other authority.615 Colwell {Day232/116:6-9}; Smith {Day235/134:9-20}. Despite its concerns, BRE did nothing other than present the results to the department. Dr Colwell and Dr Smith considered that it was for the department to review the position and take any appropriate action.616 Colwell {Day232/116:11-19}; Smith {Day235/134:18}. As far as Dr Colwell could recall, there had been no reaction from the department when the test results were presented to it.617 Colwell {Day232/114:19-21}; Smith {Day235/136:12-19} who could not recall the reaction of Anthony Burd. Mr Burd's recollection was very limited. He thought that misrepresentation by the manufacturers must have been a consideration618 Burd {Day239/73:21-25}. and that he might have had discussions with Dr Smith about it,619 Burd {Day239/73:3-20}. but he could not remember his or the department's reaction, beyond saying that it would have been one of surprise620 Burd {Day239/75:9-13}. and would have given rise to concern.621 Burd {Day239/74:5-7}. However, he could not recall any contact having been made with the manufacturers, National Trading Standards or any other authority and confirmed that ministers had not been informed of the results.622 Burd {Day239/88:15}-{Day239/89:12}.
103¶7.77 Accordingly, less than three years after the Select Committee had made recommendations directly connected to the fatal fire at Garnock Court623 {CLG00019478}. the discovery that the majority of a selection of typical cladding products marketed as Class 0 did not in fact meet that standard was simply ignored by the department. It did not alert the industry to the potential danger and took no steps to review the statutory guidance. We do not understand its failure to act in relation to a matter of such importance.
Aluminium composite panels with polyethylene cores
¶7.78 One of the four products that did achieve Class 0 in BRE's tests was a panel described in its report as an "aluminium sheet",624 {BRE00041882/12} Table 2 "Rainscreen Systems", row 3; Colwell {Day232/98:16-18}; Smith {Day235/107:11-14}. which was also tested under the three European test methods625 {BRE00041882/13} Table 3 "Rainscreen Panels", row 5; {BRE00041882/15} Table 4 "Rainscreen System", row 5; {BRE00041882/17} Table 5 "Rainscreen System", row 5. and was incorporated into one of the 14 systems subjected to a full-scale test.626 {BRE00041882/23} Table 8 "Rainscreen System" row 5. Despite the generic description of the product in almost all the BRE reports,627 {BRE00041912}; {BRE00041895/9}; {BRE00042045/5}. Dr Colwell told us that it was the practice of BRE not to refer to specific products or manufacturers in its reports in deference to the wishes of the department: Colwell {Day232/28:23}-{Day232/29:8}; Colwell {Day232/84:5-11}. Burd {Day239/44:15-22} and {Day239/45:24}-{Day239/46:7}. it is clear from other documents that it was in fact a composite aluminium panel with a polyethylene core.628 {BRE00041882/10} Table 1 "Rainscreen", row 5; {BRE00041909} "Rainscreen", final row.
¶7.79 Aluminium composite panels with polyethylene cores were selected for testing because they represented a product that was available on the market and was of interest to the construction industry.629 Colwell {Day232/95:14}-{Day232/96:11}; Smith {Day235/104:2-23}. When subjected to a full-scale test, a system incorporating panels of that kind reached the failure criteria in Fire Note 9 at a very early stage,630 {BRE00041912} row 15, column R; {BRE00041882/23} Table 8, row 5. exceeding the external temperature limits within three minutes of the start631 {BRE00041882/23} Table 8, row 5, column "Ext". and the internal limits after 4.34 minutes.632 {BRE00041882/23} Table 8 "Rainscreen System", row 5, column "Insulation". Staff intervened to terminate the test, manually extinguishing the crib fire at 5.45 minutes633 {BRE00041911/3}. and ending the test at 5.75 minutes.634 Colwell {Day232/97:9-20}; {BRE00041882/23} Table 8, row 5, column "Test Time (min)". Detailed observations recorded during the test included the fact that at 3.05 minutes molten aluminium was dropping from the front face of the system, that at 4.20 minutes flames had reached the top of the test rig and that 5.00 minutes from the start flames had reached 20 metres in height, approximately twice the height of the test facility.635 {BRE00041911/3}. On any view, that was a catastrophic escalation.636 For further analysis of the test data Bisby, Phase 2 Report {LBYP20000001/148-151} paragraphs 807-816.
104¶7.80 Dr Colwell was present at the test and was shocked both by the rapid rate at which the fire grew and the extent of its growth.637 Colwell {Day232/96:23-24}; {Day232/97:5-8}; {Day232/99:25}-{Day232/100:4}. By contrast, Dr Smith did not recall the test as being particularly notable638 Smith {Day235/101:19}-{Day235/102:2}. and did not remember anyone at BRE being particularly surprised at the way in which the system failed.639 Smith {Day235/106:10-17}; {Day235/109:17}-{Day235/110:7}. She simply noted that the test had failed at an early stage.640 Smith {Day235/107:3-10}.
¶7.81 Neither Dr Smith nor Mr Martin recalled having seen the test data or detailed observations from that test at the time.641 Smith {Day235/107:24}-{Day235/108/8}; Martin {Day251/51:5-10}; {Day251/52:6-7}. Mr Martin did recall that Dr Colwell had come into his office at BRE to tell him about the test and to show him a sample of the aluminium cladding panel,642 Martin {Day251/47:14}-{Day251/48:15}. but did not remember her being shocked or alarmed.643 Martin {Day251/48:20-22}; {Day251/49:16-20}. As far as he could recall, they had not discussed the performance of the system in any detail.644 Martin {Day251/49:21}-{Day251/51:10}.
¶7.82 In the light of the test it was clear to Dr Colwell645 Colwell {Day232/100:11-25}. and Dr Smith646 Smith {Day235/111:7-13}. that ACM panels with unmodified polyethylene cores were not suitable for use in cladding systems on high‑rise buildings. Nonetheless, in its report dated 16 September 2002647 {BRE00041895}. BRE said only that the aluminium sheet product had satisfied Class 0 but had proved to be one of the worst performing products in the intermediate scale and Single Burning Item tests. It concluded by saying that, as the current guidance in Approved Document B asked for Class 0 performance in Diagram 40, those matters might require further consideration.648 {BRE00041895/13} paragraph 5. That was as far as the report went.
¶7.83 BRE expected the department to decide what it wanted to do in response.649 Smith {Day235/152:3-16}. Apparently, no report it produced for the department spoke in overt terms about an immediate risk to life, because its function was simply to present the evidence, including the technical facts and the data.650 Smith {Day235/152:24-25}. It was for the department to decide what to make of it.651 Colwell {Day232/127:24}-{Day232/128:25}. It is clear that BRE did not share with the department, at least in any formal way, its opinion that aluminium composite panels with unmodified polyethylene cores should never be used on tall buildings.652 Smith {Day235/113:8-18}.
¶7.84 We are not convinced that it was as clear to the department as apparently it was to Dr Smith and Dr Colwell what the results of the tests incorporating the ACM PE panels meant in practical terms for the degree or immediacy of the danger their use presented or why there was a need to make changes to the guidance. Whatever may have been the accepted way of reporting the results of experimental work carried out for the department, the disastrous performance of the panels in the full scale-test should have prompted BRE to draw the department's attention to the dangers inherent in the use of ACM PE rainscreen panels.
105¶7.85 Dr Smith eventually agreed that the results of the tests had led her to have serious misgivings about the existence of statutory guidance that treated Class 0 as an appropriate standard for external panels on high-rise buildings.653 Smith {Day235/141:8-19}. Dr Colwell went so far as to agree that the results indicated that there was a serious problem with the relevant provisions of Approved Document B.654 Colwell {Day232/111:14-24}. In those circumstances, given the performance of the product both in full-scale tests and in small-scale tests using the European methods,655 {BRE00041882/14} fifth paragraph "The aluminium system generated high rates of fire growth and in both cases was extinguished early due to excessive temperatures and fire growth. This is reflected in the indicative classification of D-s2, d0." Smith {Day235/141:20-23}; Colwell {Day232/118:12-20}. BRE's comments on the performance of the ACM panels fell far short of what was required.656 {BRE00041895/13} paragraph 5.
¶7.86 One object of the research project was to find a test method that ensured that products that performed badly in response to fire would not be able to pass it. Dr Smith appears to have thought that the broader reaction of those at BRE to the tests had been communicated to the department in the course of discussions about the criteria to be included in BR 135.657 Smith {Day235/111:14-24}. However, that fails to take account of the fact that full-scale testing was not the only way of demonstrating compliance with functional requirement B4(1) recommended by Approved Document B658 Smith {Day235/29:6-13}; Colwell {Day232/111:14-21}. and apart from anything else, the guidance continued to recognise the use of ACM panels with a Class 0 surface.
¶7.87 BRE itself was not asked for its opinion on the retention of Class 0 as an appropriate standard for panels to be used on the external walls of high-rise buildings and did not give any consideration to that question.659 Smith {Day235/114:23}-{Day235/115:6}; {Day235/143:4-13}. In its view, that was a matter for the government.660 Smith {Day235/143:23}-{Day235/144:3}.
¶7.88 Mr Burd was aware that the full-scale test on the system incorporating ACM panels had produced a very fierce, fast fire.661 Burd {Day239/51:25}. That made him think that it was imperative to ensure that the work being carried out on BR 135 would produce a test that would exclude the use of products of that kind.662 Burd {Day239/51:13-21}; {Day239/54:13-22}; {Day239/55:23}-{Day239/56:1}; {Day239/59:2-7}; {Day239/64:6-11}. He appears to have thought that it was the responsibility of BRE to explain the significance of the results to him,663 Burd {Day239/53:19-21}. but could not recall that BRE had communicated to him any sense of shock at the test results.664 Burd {Day239/52:7-10}.
¶7.89 The department discussed the report with the Industry Advisory Group but Mr Burd could not recall whether any consideration had been given to intervening in the use of ACM on high-rise buildings.665 Burd {Day239/58:18}-{Day239/59:7}. Nor could he recall whether the department had taken any steps to alert others to the danger posed by the product666 Burd {Day239/57:19}-{Day239/58:17}. or to find out how many high-rise buildings might already have been clad in ACM products.667 Burd {Day239/60:6-17}; {Day239/61:10-21}. There is no evidence of any discussions within the department about withdrawing Diagram 40 (and thus Class 0) as a relevant standard for the fire performance of the external surfaces of high-rise buildings.668 Burd {Day239/75:14-17}; {Day239/76:17}-{Day239/77:4}.
106¶7.90 One might reasonably question whether Mr Burd or anyone else in the department, including Mr Martin, studied the BRE report in any detail or really gave any thought to the implications of the results of the full-scale test using the ACM product. If they had done so, they would surely have realised that although it had a Class 0 surface, it was highly combustible and entirely unsuitable for use on the external walls of buildings, particularly high-rise buildings. Mr Burd's view at the time was that the use of ACM panels on buildings over 18 metres in height did not comply with functional requirement B4(1) of the Building Regulations.669 Burd {Day239/55:5-22}; {Day239/61:15-18}; {Day239/64:14-23}; {Day239/65:8-12}; {Day239/94:12-19}; {Day239/100:22}-{Day239/101:13}; {Day239/167:17}-{Day239/168:8}. In his view, the general warning in paragraph 13.7 of Approved Document B against the use of combustible materials670 {INQ00014107/89} paragraph 13.7 and the existence of functional requirement B4(1) were sufficient to prevent the use of such products on high-rise buildings.671 Burd {Day239/59:5-7}; {Day239/70:8-11}. That view is difficult to reconcile, however, with the selection of ACM panels for testing on the grounds that they were a form of material then being widely used.
¶7.91 Brian Martin also thought that, although Diagram 40 allowed the use of ACM panels with a Class 0 surface to be used on buildings over 18 metres in height,672 Martin {Day251/71:16-21}. their use would contravene functional requirement B4(1) of the Building Regulations,673 Martin {Day251/71:25}-{Day251:72:1}. but he could not explain why Approved Document B contained guidance that appeared to allow the use of a product that would result in a breach of the Building Regulations.674 Martin {Day251/72:22-25}; {Day250/87:4-21}. On any view, that created a deeply unsatisfactory position for at least two reasons: first, because following the guidance is evidence of compliance with the requirements of the Building Regulations and, secondly, because although the contents of Approved Document B are no more than guidance, their detailed nature encourages people in the construction industry to treat them as prescriptive and not to pay sufficient attention to the functional requirements themselves. If the department thought that the use of some cladding panels with Class 0 surfaces did not comply with the Building Regulations, it should have given urgent thought to withdrawing Class 0 as a standard or at least to including in Approved Document B an explicit warning about the use of composite products, substantial parts of which were combustible. As it is, there is no evidence that any discussions of that kind took place.
¶7.92 We have been unable to establish how or why such important information was ignored. There are three possibilities: the first is that those in the department who were responsible for the research project, principally Mr Burd and Mr Martin, did not read it carefully or appreciate its implications; the second is that they did read it but did not understand the full seriousness of its implications for the use of ACM and put it aside without taking appropriate action; the third is that they understood the seriousness of the implications for the use of ACM panels and decided to suppress it.
¶7.93 There is no evidence to support the conclusion that the department deliberately decided to refrain from acting in response to what it knew constituted a serious danger to people working and living in high-rise buildings. The fact that the results of the research were not published at the time suggests that their significance was not fully appreciated and is consistent with the failure of Dr Smith and Mr Martin to refer to the cc1924 tests when they took part in a meeting in June 2017 to advise the government whether the cladding used on Grenfell Tower had been fit for its purpose.675 Smith {Day238/109:7-13}; {Day238/109:18}-{Day238/110:4}; {Day238/111:1-9}; Martin {Day257/208:7-14}; {Day257/215:12}-{Day257/216:9}; {CLG00005247/5}. BRE's failure to draw the attention of the department to the performance of the ACM panels in the full-scale test in suitably forceful terms bears some responsibility for that.
107¶7.94 No one was able to explain why the results of the tests had never been made public,676 Burd {Day239/104:22}-{Day239/106:4}; Martin {Day251/85:16}-{Day251/86:20}; Smith {Day235/155:21} -{Day235/156:25}; {Day235/159:12-24}; Colwell {Day232/129:24}-{Day232/130:10}. even in the wake of the fire at Grenfell Tower, when they would have been highly relevant to the department's investigations into the fire performance of similar products. We think that the explanation lies in the fact that having once been shelved they had subsequently been entirely forgotten.
Harmonisation and RADAR 2
¶7.95 In February 2000 the European Commission directed that the European reaction to fire classification system, EN 13501, which we have described in a separate chapter of this report,677 Chapter 5. be adopted across all European member states and that national testing standards should be harmonised.678 2000/147/EC implementing Council Directive 89/106/EEC. There was no harmonisation of building regulations, however. Although it remained open to member states to determine how and in what circumstances particular test methods and classifications would be required or recommended, each member state was obliged to incorporate the new European classification standards into its own regulations and associated guidance, if necessary alongside existing national standards.679 {CLG00007308/2} paragraph 3.2; Burd {Day238/166:10-19}; {Day239/109:18}-{Day239/110:1}; Martin {Day250/65:17-23}; Smith {Day235/160:12-18}; Bisby, Phase 2 Report {LBYP20000001/115} paragraph 600. In the UK the approach taken by the department was to recognise both the European and national testing methods and standards.680 Burd {Day239/157:21-23}; {CLG00007308/3} paragraph 3.7.
The RADAR research
¶7.96 In preparation for the introduction of the European classification system the department commissioned research from Warringtonfire Research Centre Ltd681 Now Warringtonfire Testing and Certification Ltd. to help it decide how the European test methods and classifications should be incorporated into UK regulations and guidance and to understand the effect of their adoption.682 {CLG00001464/4} paragraph 17; Smith {Day235/160:19-25}; Burd {Day239/109:23}-{Day239/110:14}; Martin {Day250/175:15-25}. Warringtonfire produced a series of reports under the title "RADAR" (Research on Approved Document B and Revision).683 {CLG00001068/1} paragraph 4; Burd {Day239/109:15-17}; RADAR 1 addressed fire resistance and RADAR 3 dealt with tests for roofing materials. A two-part RADAR 2 report delivered in May 2000 specifically addressed the reaction to fire test methods under BS 476 Parts 6 and 7 and related national classifications.
¶7.97 The RADAR 2 report entitled "Correlation of UK Reaction to fire classes for building products with Euroclasses and Guidance on Revision of Approved Document B"684 {CLG00000950/2}; {CLG00000951/1}. was intended to determine whether satisfactory correlations could be established between the different methods of testing and classification.685 {CLG00000951/3} paragraph 1. The focus of the work, therefore, was on comparing the classifications of various products under the Euroclass and national systems.686 {CLG00000950/4} section 3. For that purpose, 64 different products687 {CLG00000950/5} paragraph 4.1. The seven industry sectors referred to were: Wood, Mineral Wool, Paints, Cellular Plastics, Wallcoverings, Board and Sheet and Plastics {CLG00000950/5-6} paragraphs 6.2-6.8. were chosen for testing.688 {CLG00000951/3} section 1. The RADAR research was commissioned under the Partners in Innovation Scheme, meaning that it was funded half by the department and half by industry, including manufacturers.689 Martin {Day250/175:25}-{Day250/176:1}; Smith {Day235/162:14-20}; Burd {Day239/111:23}-{Day239/112:3}. Mr Burd told us that the products selected for testing were chosen following discussions with those industry partners.690 Burd {Day239/113:2-12}.
108¶7.98 Following the testing, the results for 60691 {CLG00000951/5} section 3, first paragraph. of the 64 products were plotted onto a matrix. The second RADAR 2 report stated that where a high density (in other words a large number) of products occurred at any given transposition point on the matrix, that could be interpreted as giving confidence that a "representative correlation" existed and that a "reliable transposition" could be made.692 {CLG00000951/3} section 2, first paragraph.
¶7.99 The report contained a table that suggested possible options for the transposition of classes for performance in reaction to fire. It suggested a transposition between national Class 0 and Euroclass B (or better).693 {CLG00000951/6} Table 6. Of the products tested, 35 satisfied the requirements of Class 0 under the national system. When those 35 products were tested under the Euroclass systems, 21 achieved Class B, 10 achieved Class A1 or A2 and, at the other end of the scale, one achieved Class C, two achieved Class D and one achieved only Class E.694 {CLG00000951/4} Table 2. The Class 0 products which achieved Euroclass C, D or E were not identified in the report.
¶7.100 The possible options for transposition were discussed by members of the Part B Working Party of the Building Regulations Advisory Committee (known as "BRAC")695 The Building Regulations Advisory Committee was a statutory advisory public body sponsored by the department ordinarily made up of twelve members. The Secretary of State was required to consult the Committee on any changes to the Building Regulations. When changes were being considered, a technical working party (such as the Part B Working Party) was normally set up, Ledsome {CLG00019465/30} page 30, paragraph 116; {CLG00007308/1} for a list of members of the Part B Working Party. at a series of meetings between 2000 and 2002,696 See for example meetings on 28 February 2001 {CLG00001051}; 10 May 2001 {CLG00007308}; 15 April 2002 {CLG00000720}; 2 May 2002 {CLG00001462}. as well as with the Industry Advisory Group that had been set up to steer the project.697 {CLG00000950/3-4} section 2. Mr Burd attended a number of those meetings, as did Dr Smith,698 {CLG00007308/1} and {CLG00001462/1}. who told us that she had been called in to advise because of her expertise in fire testing and European harmonisation.699 {CLG00007308/1} paragraph 1.3; Smith {Day235/174:19-21}; {Day235/194:10}-{Day235/195:1}. Her role was to act as a co- ordinator of BRAC's discussions.700 {CLG00000950/3} section 2, second row, third column; Smith {Day235/164:2-4}.
109Transposition
¶7.101 The department concluded that the most appropriate transposition point for Class 0 was Euroclass B and so the proposal made by Warringtonfire was accepted.701 Martin {Day250/181:9-13}; Burd {Day239/156:11-15}. In the amended (2002) edition of Approved Document B, Diagram 40 therefore referred both to Class 0 and to Euroclass B-s3, d2 (or better) as the recommended standard for the external surfaces of walls over 18 metres in height.702 {CLG10000740/91} Diagram 40.
¶7.102 The second RADAR 2 report had suggested that a decision to treat Class 0 as equivalent to Euroclass A2 would severely restrict the choice of materials available to designers and clients. Euroclass A2 is a more stringent standard than Euroclass B and there may have been a view that very few Class 0 products were capable of achieving Euroclass A2. Accordingly, placing the transposition point at Euroclass A2 rather than at Euroclass B might not only have restricted market choice but would potentially have met resistance from, and caused losses to, the UK construction industry.703 Smith {Day235/192:7}-{Day235/193:3}. However, if the transposition point for Class 0 had been Euroclass A2, at the end of any transition period buildings would be clad in materials that performed better in a fire.704 Smith {Day235/193:4-9}. The decision to use Class B as the transposition point to Class 0 was a compromise between the need to maintain fire safety standards and the need to avoid distortion of the market.705 Smith {Day235/193:15-22}.
¶7.103 The effect on manufacturers of the harmonisation scheme rested largely in the hands of the department, since it determined the transposition points, any notional equivalence between classification under the two systems, and any period of co-existence.706 Burd {Day239/134:7-15}. However, Mr Burd said that the likelihood of resistance from the construction industry to a correlation between Class 0 and Euroclass A2 had not been a factor in his consideration of the possible transposition points707 Burd {Day239/128:15-21}. and that he had not considered whether that transposition point might restrict choice at the end of any period of co-existence.708 Burd {Day239/128:5-8}. He did concede, however, that it might have been too demanding. Indeed, one of the stated objectives of the RADAR research was to ensure that the introduction of the new European test methods did not significantly change the existing regulatory position.709 {CLG00000950/3} section 1, third paragraph. The department appears to have made no attempt to find out whether the four products that had achieved Class 0 but only Euroclass C, D or E were in widespread use,710 Burd {Day239/118:5-10}. and if so, in what circumstances.
The retention of Class 0
¶7.104 When considering what amendments to make to Approved Document B in the light of the RADAR report, the department gave no thought to the removal of Class 0 as the appropriate standard for the fire performance of materials used for the external surfaces of walls of buildings over 18 metres in height, despite the catastrophic performance of a system incorporating Class 0 aluminium composite panels with polyethylene cores in a full-scale test under contract cc 1924.711 Burd {Day239/209:12}-{Day239/210:1}; Smith {Day236/37:17}-{Day236/39:7}. One reason for that was the fact that Class 0 represented the established method of determining the suitability of products for use in the external walls of high-rise buildings712 Burd {Day239/207:10-21}. and the purpose of the exercise then being conducted was seen as being to harmonise national and European classification and testing standards, not to review the technical requirements of Approved Document B generally.713 Burd {Day239/121:19-20}; {Day239/123:4-5}.
110The transitional period
¶7.105 It was originally intended that the national classes (including Class 0) would be retained in use alongside the Euroclass system for a limited transitional period, after which they would no longer be relied on as the relevant classification standards.714 {CLG00000720/4} paragraphs 6.11 and 10 "The UK currently utilises our own British Standards – mainly the BS 476 suite of documents, however these will eventually be withdrawn"; {CLG00001464/3-4} paragraphs 15-16; {CLG00001462/3} paragraph 6.7 "The transpositions proposed for AD B are for Class B to be substituted for Class 0 in time". In the event, the transitional period continued until December 2018, after the fire at Grenfell Tower.715 Martin {Day250/167:12-16}.
¶7.106 A Regulatory Impact Assessment prepared by the department in May 2002 suggested that the transitional period would begin in 2002 and would last for three years,716 {CLG00001464/4} paragraph 16. but no date for the end of the period of co-existence was fixed at the time.717 Burd {Day239/164:3-14}; {Day239/175:21}-{Day239/176:19}.
¶7.107 It was difficult to discern what other factors led to the retention of Class 0 for so many years. As far as we can see, it did not form part of any of the discussions or consultations leading to the amendments of Approved Document B in 2010718 Burd {Day240/12:8-22}. or 2013.719 Burd {Day240/13:20}-{Day240/14:3}; {Day240/16:5-14}.
¶7.108 Reliance on the national classes in Approved Document B could have been discontinued at any time,720 Burd {Day239/188:12-16}; {Day239/191:4-25}. provided the proper processes for amendment had been followed.721 Burd {Day240/13:2-9}; {Day240/14:9-25}. Similarly, Diagram 40 could have been amended at any time to recommend the use of materials of limited combustibility for the external surfaces of buildings over 18 metres in height.722 Burd {Day240/18:15-25}.
Equivalence and the note in Diagram 40
¶7.109 As everyone agrees, there was no technical correlation or equivalence between Class 0 and Euroclass B because the two classification regimes were completely different, testing for different properties according to different criteria and in different circumstances.723 Smith {Day235/184:17-19}; {Day235/185:1-13}; {Day235/188:24}-{Day235/189:1}; {Day235/197:8-23}; {Day236/17:14-15}; Burd {Day239/144:19}-{Day239/145:8}; {Day240/6:17-19}; Martin {Day250/190:11-16}; {CLG00007308/7} paragraph 5.8; {CLG00001051/1-2} paragraph 6. Nonetheless, a footnote to Diagram 40 was added in the 2002 version of Approved Document B stating:
"National classifications do not automatically equate with the equivalent European classifications, therefore products cannot typically assume a European class unless they have been tested accordingly".724 {CLG10000740/91} Diagram 40.
¶7.110 The note was intended to warn readers that Class 0 could not be achieved by testing under the European system and obtaining a Class B (or vice versa). Only tests under the national system (i.e. BS 476-6 and 476-7) could provide a Class 0 classification and only tests under the European system could provide a Class B classification.725 Burd {Day239/202:13-17}; Smith {Day236/30:14-18}; {Day236/33:22}-{Day236/34:1}; Martin {Day250/188:9-18}. Although we agree that that is one possible reading of the footnote, we think that it could equally be understood to mean that there could be an equivalence of some kind between the two classifications, albeit not an automatic one. Indeed, the note uses the word "equivalent" in that very context.
111¶7.111 In our view the footnote was certainly confusing. The use of the word "automatically" suggested that there could be an equivalence in some cases, though not all, and the expression "equivalent European classifications" suggests that there is an equivalence between national classes and European classes. Although some readers might have understood that testing under each system was necessary in order for a product to obtain the related classification, they might have understood that Class 0 and Euroclass B were technically equivalent, which was not the case. They were equivalent only for the limited purposes of complying with the guidance in Diagram 40, but they were not technically equivalent and did not confirm the same characteristics.
¶7.112 Although Brian Martin, Anthony Burd and Dr Smith all told us that they had not been aware of any manufacturers claiming that Class 0 was equivalent to Euroclass B,726 Martin {Day250/204:9-18}; {Day250/205:7-10}; Burd {Day240/8:6-17}; Smith {Day236/35:19-21}; {Day236/60:12-21}. they all saw an email to the department from the fire safety authorities in Singapore sent in December 2009727 {CLG10004229}. asking specifically whether the two were equivalent. However, that inquiry did not lead any of them to think that the note in Diagram 40 might not be clear or that consideration ought to be given to amending it.728 Martin {Day250/196:14-22}; {Day250/197:14-22}; {Day250/200:18-21}; {Day250/203:16-24}; Burd {Day240/7:6}-{Day240/8:4}; Smith {Day236/59:12-19}. Mr Martin's view, which seems to have been shared by others involved, was that people understood that the two systems operated in parallel and were not interchangeable.729 Martin {Day250/204:15-18}. The department appears to have assumed that the inquiry from Singapore was an isolated instance, but that may have been wishful thinking. There was evidence before us that some manufacturers and others in the construction industry thought that Class 0 and Euroclass B were technically equivalent and that others did not know whether they were equivalent or not, whether technically or for the purposes of Diagram 40.730 {KIN00000060/3} left column, fourth paragraph, "Class 0 correlates with Euroclass B in the new regime"; Rochefort {Day80/60:13}-{Day80/61:1} "...if I remember correctly, you can read across from a C-s1 d0 and that was deemed to be equivalent to a BS 476 Class 0"; French {Day87/84:21}-{Day87/85:9}; {Day87/86:15-25}; Gregorian {Day105/19:24}-{Day105/20:14}.
The development of the BS 8414 test series and the second edition of BR 135
BS 8414 Part 1: 2002
¶7.113 While the work on harmonisation of testing standards had been going on and the work towards the revision of BR 135 under contract cc 1924 was continuing, Fire Note 9 had been submitted to the British Standards Institution for adoption as a British Standard. Following editorial amendments and approval,731 Colwell {BRE00047571/27} page 27, paragraph 172. BS 8414 Part 1 was published in December 2002.732 {CEL00001205}.
112¶7.114 BS 8414-1 set out the test method for full-scale testing of external wall systems. The performance criteria against which the test data were to be assessed were moved from Fire Note 9 into the second edition of BR 135, which Dr Colwell and Brian Martin drafted as part of their work under contract cc1924.733 Colwell {BRE00047571/30} page 30, paragraph 190; Colwell {Day232/150:1-4}. The second edition was published in 2003.734 {BRE00005554/22} Annex A. Separating the test method from the performance criteria reflected the approach then being taken in relation to European standards, in which test methods and classification criteria were contained in separate documents.735 Colwell {BRE00047571/29} page 29, paragraph 184; Colwell {Day232/140:2}-{Day232/141:16}.
BR 135, second edition: 2003
¶7.115 Dr Colwell and Mr Martin, together with the department, decided that the first edition of BR 135736 {BRE00001077}. provided a suitable basis for the second.737 Colwell {BRE00047571/30} page 30, paragraph 192; Colwell {Day232/150:16-23}. They reviewed the text of the first edition and revised the contents to take account of developments in the testing of external cladding systems and advances in knowledge about the mechanism of fire spread, including those that had arisen from the work undertaken for the department under contract cc1924.738 Colwell {BRE00047571/31-32} pages 31-32, paragraph 196. The classification performance criteria, which had themselves been derived from Fire Note 3, were taken from Fire Note 9.739 Colwell {Day232/152:5-11}; Smith {MET00081237/12} page 12, paragraph 41; Colwell {BRE00047571/32} page 32, paragraph 200; Colwell {Day232/165:21}-{Day232/166:4}.
¶7.116 The second edition of BR 135 placed heavy emphasis on the potential difference that cavity barriers could make to the performance of external cladding systems.740 Colwell {Day232/165:13-17}; {BRE00005554/14} "Fire Barriers". However, it took no account of the fact that all five of the ventilated rainscreen systems subjected to a full-scale test under contract cc1924, including the system with cavity barriers fitted, had failed to meet the criteria in Fire Note 9.741 Colwell {Day232/162:18}-{Day232/163:10}. It also took no account of the fact that Dr Connolly had concluded in 1994 that a number of the fire barriers used in his testing had proved to be ineffective.742 {RCO00000001/48-49} paragraphs 9-11. It is not clear why the second edition of BR 135 placed so much emphasis on cavity barriers.743 Colwell {Day232/163:3}-{Day232/164:16}. There seems to have been a widely held view that properly designed cavity barriers could potentially improve performance,744 Smith {Day236/102:2-13}. but that does not appear to have been supported by any empirical evidence and in our view BRE should not have allowed its conclusions to be influenced by unsubstantiated opinion in that way.
¶7.117 The first (1988) edition of BR 135 had contained an explicit warning that BS 476 Parts 6 and 7 could provide information only about external surfaces.745 {BRE00001077/4} under the heading "Regulatory aspects". That warning was removed from the second edition, but it is not clear why. Neither Dr Colwell nor her co-author, Mr Martin, was able to provide a satisfactory explanation.746 Martin {Day251/113:9-15}. He expressly rejected the suggestion that the passage in the first edition had been deliberately omitted to avoid upsetting the industry.747 Martin {Day251/111:9-16}; {Day251/111:24}-{Day251/112:2}; {Day251/112:19}-{Day251/113:2}.
113¶7.118 We find it difficult to accept that the omission was an oversight. On the contrary, the removal of a passage pointing out the limitations of Class 0 was wholly consistent with the attitude that the department had taken to Class 0 since the Knowsley Heights fire in 1991, namely, that it must be retained, despite the evidence that it was unreliable and much misunderstood and despite a Parliamentary Select Committee expressly recommending its abrogation. We think it likely that commercial considerations played a part in the decision to delete the warning from BR 135.
BS 8414 Part 2: 2005
¶7.119 In October 2005, BS 8414 Part 2, setting out the large-scale test method for cladding systems on steel-framed structures, was published.748 {BSI00000097} BS8414-1 applied only to masonry structures. By that time, Dr Colwell had become chair of the British Standards Institution's joint committee working on the test standard.749 Colwell {BRE00047571/52} page 52, paragraph 330. She produced the initial draft750 Colwell {Day232/186:4-9}. and was responsible for obtaining a consensus on the draft text before it went out to consultation.751 Colwell {BRE00047571/52} page 52, paragraph 330. The test method in BS 8414-2 was based on the same principles as BS 8414-1. The need for a test method for steel-framed structures had become apparent during BRE's work on large-scale cladding tests under contract cc1924752 Colwell {Day232/70:13}-{Day232/71:25}; Colwell {BRE00047571/53} page 53, paragraph 333. and during the development of BS 8414-1,753 Colwell {BRE00047571/58} page 58, paragraph 367. as a result of which it was realised that there was insufficient knowledge about the fire performance of external cladding systems where no masonry wall was present.754 {BRE00041887/13} Option 2; {BRE00041895/9-10} tasks 6 and 7. Following its publication, BRE produced a document containing the performance criteria and classification method for BS 8414-2.755 Smith {MET00081237/22-23} pages 22-23, paragraph 89. They were attached to BR 135 as Annex B.756 {BRE00005552}.
BR 135, third edition: 2013
¶7.120 In 2013, a revised version of BR 135 was published, written by Dr Tony Baker757 Dr Tony Baker joined the BRE in 2004. He was the Certification Scheme Manager for the Passive Fire Protection Group and from 2013 onwards the fire resistance Test Laboratory Manager: Baker {BRE00005774/1-2} pages 1-2, paragraphs 4-7. and Dr Sarah Colwell.758 {CEL00000584}. The performance and classification criteria remained unchanged.759 Smith {MET00081237/23} page 23, paragraph 91; Colwell {BRE00047571/59} page 59, paragraph 378; Colwell {Day233/46:10-14}. The guidance in the third edition contained new wording making it clear that a classification in accordance with BR 135 applied only to the system tested and not to any other.760 {CEL00000584/9} left-hand column, final paragraph; {CEL00000584/27} left-hand column under bullet points; {CEL00000584/33} right-hand column under bullet points. Dr Baker told us that he had thought that the second edition of BR 135 had not been quite as clear as it could have been in that respect. New clauses had been added in order to clarify the position761 Baker {Day100/122:3-13}. because he was concerned that the message was not getting through to the industry.762 Baker {Day100/122:15-24}. For the same reason, BRE decided to insert the same wording into all classification reports.763 Baker {Day100/122:12-13}.
114Approved Document B 2006
The background to the amendment
¶7.121 In March 2004, the department commissioned BRE to undertake a review of Part B of Schedule 1 to the Building Regulations and Approved Document B.764 {BRE00001953}; {CLG00001508}. It did so because it felt that the technical aspects of the requirements of Part B and the guidance in Approved Document B were due for revision.765 {BRE00001953/4} section 1, third paragraph. The work was led and managed by Brian Martin in his capacity as an employee of BRE766 {BRE00001953/10} section 3; Martin {CLG00019469/2} page 2, paragraph 6. (although he continued to be seconded part-time to the department). It was split into a number of tasks, one of which was a survey of interested parties intended to identify matters they might think it important to include in the review.767 {BRE00001953/5} section 2 Task 1. In the light of the responses to those enquiries Mr Martin and Mr Burd prepared a draft revision of Approved Document B for public consultation.768 Martin {Day251/195:21-24}; {Day251/181:16-24}.
¶7.122 The topics considered as part of a "Forwards Look" exercise and on which recipients were consulted769 {CLG00002410/10-11} and {CLG00002410/22-23} in relation to B4 External Fire Spread, the topics consulted on were space separation and roof coverings. did not include the retention of Class 0, any amendments to Diagram 40 or the guidance relating to the construction of the external walls of buildings over 18 metres in height.770 Martin {Day251/145:19}-{Day251/146:4}. For whatever reason, none of those involved in the survey raised the question whether Class 0 should be retained as one of the standards by which the suitability of materials used in the external walls of buildings over 18 metres in height was to be judged.771 Martin {Day251/145:19}-{Day251/146:4}. Mr Martin did not suggest that topic for discussion because the object of the exercise was to find out what other interested parties thought the department should be considering.772 Martin {Day251/142:20}-{Day251/143:4}; Burd {Day240/25:11-24}. At that time he himself had not given any thought to the suitability of Class 0 for that purpose773 Martin {Day251/171:6-11}. and thought that it would have been odd to identify that particular subject, given the wide range of matters covered by Approved Document B.774 Martin {Day250/171:12-22}.
The 2002 edition
¶7.123 The 2002 edition of Approved Document B provided in paragraph 13.7 as follows:775 {CLG10000740/90}.
115"External wall construction The external envelope of a building should not provide a medium for fire spread if it is likely to be a risk to health or safety. The use of combustible materials for cladding framework, or of combustible thermal insulation as an overcladding or in ventilated cavities, may present such a risk in tall buildings, even though the provisions for external surfaces in Diagram 40 may have been satisfied. In a building with a storey 18m or more above ground level, insulation material used in ventilated cavities in the external wall construction should be of limited combustibility (see Appendix A). This restriction does not apply to masonry cavity wall construction which complies with Diagram 32 in Section 10."
¶7.124 Three things should be noted. First, the paragraph contained a warning that the fact that the surface of an external wall might satisfy Diagram 40 by having either a Class 0 or Euroclass B classification did not eliminate the risk of the spread fire across it. In other words, it drew attention to the fact that satisfaction of the requirements of Diagram 40 would not necessarily result in compliance with functional requirement B4(1). The person carrying out the work, therefore, always had to make a judgement about whether the wall as a whole (including the proposed cladding system) met the functional requirement, even if the surface of the outer material was classified Class 0.776 Burd {Day239/81:14}-{Day239/84:14}. That was how the department understood that provision and it does not appear to have crossed the mind of anyone within the department that it might be understood in any other way.777 Burd {Day 239/64:14}-{Day239/67:17}.
¶7.125 Secondly, notwithstanding the reference to external surfaces and Diagram 40 in the first part of paragraph 13.7, external surfaces were dealt with in paragraphs 13.5 and 13.6. Moreover, Diagram 40 had nothing to do with insulation.
¶7.126 Thirdly, the guidance suggested that only insulation used in a ventilated cavity in an external wall of a building over 18 metres in height need be of limited combustibility.
The 2006 edition
¶7.127 In the 2006 version of Approved Document B the former paragraphs 13.5, 13.6 and 13.7 were reorganised and re-worded as paragraphs 12.6 – 12.9. The first part of paragraph 12.5, which now bore the heading "External wall construction", read:
"The external envelope of a building should not provide a medium for fire spread if it is likely to be a risk to health or safety. The use of combustible materials in the cladding system . . . may present such a risk in tall buildings."
¶External surfaces were then covered by paragraphs 12.6, and insulation by paragraph 12.7.
¶7.128 On the face of it, the reorganisation did to some extent make the section easier to use, but in the attempt to achieve greater simplicity the warning that had existed in the first part of paragraph 13.7 of the 2002 edition was removed, so that there was now no overt statement that the use of a material or product that was Class 0 did not itself mean that the external wall would comply with functional requirement B4(1). Neither Mr Burd nor Mr Martin could offer any satisfactory explanation for the change.778 Martin {Day252/28:7}-{Day252/31:5}; Burd {Day240/79:25}-{Day240/82:21}.
¶7.129 Paragraph 12.7, which was the successor to paragraph 13.7 of the 2002 edition, now read as follows:
"Insulation Materials/Products In a building with a storey 18m or more above ground level any insulation product, filler material (not including gaskets, sealants and similar) etc. used in the external wall construction should be of limited combustibility (see Appendix A). This restriction does not apply to masonry cavity wall construction which complies with Diagram 34 in Section 9."
¶7.130 There were two other amendments of linked significance: first, the title of the paragraph was new, and secondly, the paragraph itself now contained the phrase "filler material (not including gaskets, sealants and similar) etc…". The significance of those changes lies in the fact that the department's contention, at least after the fire at Grenfell Tower, was that the words required the core of an external wall cladding panel to be of limited combustibility. Acceptable materials therefore, it argued, did not include ACM panels with unmodified polyethylene cores. However, we do not think that an ordinary reader would understand those words in that way and it is clear that in the years that followed many in the construction industry did not do so.
116The origin and evolution of the amendments
¶7.131 In the light of subsequent events we think it is useful to consider the origin of those amendments and the way in which they were introduced.
¶7.132 On 6 January 2005 a fire broke out at The Edge, a partly-occupied 19-storey block of flats in Salford, Greater Manchester.779 {BRE00035368/1} under heading "Type of building". Within ten minutes it had spread from the point of origin on a second floor balcony to the top of the building.780 {BRE00035368/1} final paragraph. BRE investigated the fire for the department under the Investigation of Real Fires contract781 Martin {Day251/148:5-8}. and Mr Martin discussed it with BRE colleagues, including Dr Colwell.782 Martin {Day251/146:25}-{Day251/147:11}; {Day251/149:5-9}. BRE found that the fire had spread externally over the building's aluminium composite cladding panels,783 {BRE00035368}. which had delaminated, exposing the combustible polystyrene cores, which had then become involved in the fire.784 {BRE00035368/1} final paragraph; {CLG00019455/76} paragraph 4.1.148. In its report it also pointed out that what had then been paragraph 13.7 of Approved Document B785 {CLG10000740/90} paragraph 13.7. had limited the combustibility of insulation alone and that, although polystyrene might ordinarily be used as an insulation material, in that case it had been used to stiffen the decorative panels rather than for its insulating properties.786 {BRE00035368/2}; Martin {Day251/154:6-13}. It referred to the polystyrene core of the panels as "a low cost filler".787 {BRE00035368/2}.
¶7.133 It is clear that neither BRE nor the department thought at the time that those responsible for the choice of the panels used at The Edge had failed to follow the existing guidance in Approved Document B; on the contrary, they thought that there was something wrong with the guidance that needed to be corrected. It does not seem to have occurred to anyone in the department that many in the industry were treating the guidance as prescriptive and that the warning in section 13.7 was insufficiently clear.788 Burd {Day240/57:21}-{Day240/58:17}.
¶7.134 At the time of the fire at The Edge Approved Document B did not warn against the use of aluminium composite panels with combustible cores on the walls of high-rise buildings; on the contrary, Diagram 40 tended to support the use of such panels if they had Class 0 surfaces.789 Martin {Day251/161:1-17}. BRE's second report into the fire dated 31 March 2005 warned that even if the construction had complied with current guidance, that would not have restricted the speed at which the fire had spread through it. It also reported that the fire and rescue services had grave concerns that a fire of that nature could quickly get out of control and put the lives of occupants at risk.790 {CLG00019455/76-79} paragraph 4.1.148 (Investigation of Real Fires Report April 2004 to March 2005, dated 31 March 2005). To that extent the guidance in Approved Document B was ineffective.
117¶7.135 On 18 January 2005, Brian Martin sent Anthony Burd some suggested amendments to section B4 of the working draft of Approved Document B then being prepared for consultation.791 {CLG00018832}. The reference to ventilated cavities in what was proposed as paragraph 11.8 was removed and the words "or any other material" were added, so that it read:
"In a building with a storey 18m or more above ground level, insulation or any other material used in ventilated cavities in the external wall construction should be of limited combustibility."792 {CLG00018833}.
¶7.136 If that form of words had been adopted, the guidance would have been that all materials used in the external wall construction of a building over 18 metres in height should be of limited combustibility.793 Martin {Day251/197:9-15}; Burd {Day240/45:19-23}. However, they thought that a blanket provision of that kind would go too far, because it would prevent the use of many materials that might otherwise have been considered acceptable in various circumstances.794 {CLG00018832}.
¶7.137 Accordingly, paragraph 11.7 of the draft sent out for consultation in July 2005795 {CLG00000022}. was amended to provide that:
"In a building with a storey 18m or more above ground level, insulation used in the external wall construction should be of limited combustibility."796 {CLG00000022/147}.
¶7.138 The intention at that stage was to ensure that what eventually became paragraph 12.7 covered materials which, although normally used as insulation, were being used for purposes other than their thermal performance.797 Burd {Day240/53:13-21}. Apparently, the department was not willing to delay the consultation to allow time to improve the wording but was willing to reconsider it after the consultation had been completed.798 Martin {Day251/209:15}-{Day251/210:8}.
The response of NHBC to the consultation
¶7.139 The origin of the word "filler" lies either in the first BRE report into the fire at The Edge or in NHBC's response to the consultation,799 Burd {CLG00019461/8-13} pages 8-13, paragraphs 23-31; Martin {Day252/16:25}-{Day252/17:11}. in which it sought clarification whether sandwich panels that used polystyrene as what it described as a "filler" could be used on buildings over 18 metres in height. "Filler" was not a word that had any settled technical or customary meaning. Mr Martin did not have a clear understanding of what it meant800 Martin {Day251/155:4-19}. and Mr Burd described it as "something that fills a void".801 Burd {Day240/101:20-21}, {Day240/102:24}. Both said that they had intended it to include the core of a sandwich panel, regardless of whether it was a material that could in other circumstances be used for the purposes of insulation,802 Martin {Day252/19:14-17}; Burd {Day240/72:22}-{Day240/73:9}; {Day240/100:8}-{Day240/101:6}. and thus to control the combustibility of the material below the surface.803 Martin {Day251/171:18}-{Day251/172:21}. Mr Martin may have thought that if the combustibility of the core was controlled, the classification of the surface would become less of a problem,804 Martin {Day251/191:7-16}. but that seems unlikely, because at the time no one in the department thought there was any problem with the use of Class 0.
118¶7.140 We accept that the use of the broad expression "filler material" was also intended to capture the use of materials in ways or circumstances that Mr Martin and Mr Burd had not yet come across,805 Martin {Day252/16:2-11}; {Day252/23:15}-{Day252:24:1}; {Day252/51:8-13}. such as the use of sheet materials to provide additional stiffness to a steel frame.806 Martin {Day252/21:24}-{Day252/22:16}. Mr Martin807 Martin {Day252/13:13-25}; {Day252/20:12-23}; {Day252/47:3-8}. and Mr Burd808 Burd {Day240/73:10-13}. both said that they had wanted people to think about it and work out for themselves what did or did not fall within its scope.809 Martin {Day252/24:2-7}. That approach was wholly inappropriate, however, since the purpose of Approved Document B was to provide clear guidance on how to comply with the regulations.
¶7.141 A new heading to paragraph 12.7, "Insulation Materials/Products", was introduced when the revised edition of Approved Document B was published. Mr Martin said that its purpose had been to prompt people to think more broadly about other combustible material they might be thinking of using, but he was unable to explain how it might achieve that.810 Martin {Day252/35:2-20}. It is much more likely that it was included simply to reflect the scope of the paragraph, which on a natural reading is directed to insulation materials and products.
¶7.142 Later confusion over the meaning of paragraph 12.7 may have reflected two competing aims we have mentioned: preventing the use of combustible materials in the cores of composite panels and in other, as yet unforeseen, ways811 Martin {Day252/16:4-11}. and allowing some combustible materials to be used in the construction of the external walls of high-rise buildings.
The course of the amendment
¶7.143 The reference to "filler material" was added at a very late stage in the revision process. The expression made its appearance for the first time, together with the new heading, in a single-page document drafted by Mr Martin on 22 November 2006.812 {CLG10002070}. The final revision came even later with the insertion of the words "(not including gaskets, sealants, and similar) etc.", which are not found in any drafts and appear to have been added to the version approved for publication by Anthony Burd on 18 December 2006 as the final proofs went to press.813 {CLG10002200/96}. Neither Mr Martin814 Martin {Day252/42:5-24}. nor Mr Burd815 Burd {Day240/95:11-23}. was able to recall any discussion about the addition of the reference to gaskets and sealants or the reason for it.
¶7.144 It was highly unusual to make changes to Approved Document B without going through a number of established procedures,816 Burd {Day240/14:9-25}; Martin {CLG00019469/6} page 6, paragraph 18. which included public consultation and scrutiny by the Building Regulations Advisory Committee, an independent body whose function was to ensure that any changes were appropriate.817 Harral {CLG00019487/38} page 38, paragraph 149. In this case, however, none of the established procedures was followed in relation to the additional wording and we have been left with the clear impression that the changes were rushed through by the back door without proper consideration.
119¶7.145 No statements appear to have been made by the department about the meaning of paragraph 12.7 after the revised edition of Approved Document B had been published. The reference to filler material was not listed as one of the main changes818 {CLG10000007/2}. and there was no reference to it in a circular issued by the department to building control organisations in December 2006819 {CLG10000038}. alerting them to the main changes. Mr Martin accepted that the circular should have contained a reference to it and could not explain why it had not done so.820 Martin {Day252/65:11-25}.
¶7.146 The term "filler material" was later addressed in a series of821 Burd {Day240/124:1-2}. presentations given by the department to professional organisations and building control bodies following the publication of the revised version of Approved Document B,822 Burd {CLG00019461/13} page 13, paragraph 32; Burd {Day240/117:10}-{Day240/119:11}. but the slides used in those presentations simply set out the text of paragraph 12.7 and gave no indication that the expression applied to the core of a composite panel or to any material other than insulation.823 {CLG00019451/55}; Burd {Day240/120:6-22}. No effort was made to draw attention to the change; on the contrary, as Mr Martin accepted, it was made quietly and without publicity in order to avoid a fuss.824 Martin {Day252/39:21}-{Day252/40:3}; {Day252/36:13-19}.
¶7.147 Although from time to time during the four years or so leading up to the Grenfell Tower fire Mr Martin was asked about the meaning of the phrase "filler material", it was not until he attended the meeting organised by the Centre for Windows and Cladding Technology in March 2016 that he said that it was intended to cover the core of a composite panel. Even when he did offer that explanation, it was restricted to a relatively small group of industry participants.825 See Chapter 11. Why he was so reticent is not entirely clear. There appears to have been some concern on his part that being more specific about the products that fell within the guidance might lead to legal challenges by manufacturers.826 Martin {Day252/39:11-20}. The fact that the changes had been slipped in without consultation was also one of his concerns827 Martin {Day252/39:21}-{Day252/40:3}. and he may also have thought that it was not for him to say what the statutory guidance issued in the name of the Secretary of State was intended to mean. However, he could have drawn the attention of more senior officials to the confusion in the industry about the meaning of paragraph 12.7, albeit at the risk of having to explain how the amendment had found its way into Approved Document B in the first place. At all events, no document or statement emanating from the department until after the fire at Grenfell Tower explained that the expression "filler material" was intended to include the cores of cladding panels.
121