Return to search

Examination of the forensic engineering techniques employed on fire-damaged concrete structures

Portland cement (PC) concrete has historically been the most commonly used construction material within the United Arab Emirates (UAE), however, as the demand increases to reduce CO2 emissions most of newly-built facilities make use of modern concrete formulations adopting various PC substitutes, and these ‘blended cements’ typically involve mineral admixtures such as ground granulated blast-furnace slag (GGBS), fly ash (FA) and silica fume (SF). The percentage of PC replaced in these concrete mixtures varies depending on the type of cement and design criteria as well as other related issues such as fire resistance. The use of PC replacements in ready-mixed concrete has been made obligatory in Dubai from 1st April 2015. This recent move towards using greener concretes has been implemented with little research on their heat resistant properties and as such an understanding of their behaviour on exposure to high temperatures in structural fires is limited. Furthermore, the applicability of forensic engineering techniques for the assessment of any deterioration in these concrete formulations is largely untested. For this reason, a range of analytical techniques have been investigated as part of this research in order to establish the chemical and physical changes taking place as well as the practical applicability of the techniques used. Three key areas were addressed as part of this research. Firstly, a review of urban fires in Dubai and a survey of fire investigation related issues within the region was undertaken. This formed a base from which the research questions could be refined. Secondly, nine concrete mixtures were assessed using 15 analytical techniques. The concrete mixtures were exposed to 4 temperatures (150°C, 300°C, 600°C, and 900°C) within a muffle furnace and the chemical, mineralogical, physical and mechanical changes were investigated using TGA, DSC, FTIR, XRD, SEM, EDX, BET surface area, residual compressive strength, density loss, carbonation depth, visual colour change, rebound hammer, UPV, portable 3D laser scanning and micro CT scanning. Finally, a set of test concrete mixtures most closely linked to those used in construction in the UAE were exposed to a real fire and were analysed post fire using a reduced set of the analytical techniques. The techniques were assessed as to how well they could define the temperature range to which the concrete had been exposed as well as ascertaining the degree of concrete degradation based on the confirmation of the chemical, mineralogical, physical and mechanical changes which had occurred. Findings indicate that the use of blended cement concrete improves the thermal resistance of the material when compared with PC concrete up to a certain temperature, usually below 600°C. Discolouration in heated concrete sections were visualised using simple digital photography. Changing the cement composition influenced the rate of carbon ingress into the concrete matrix, however this did not result in any significant colour change in heated mortar surfaces. By contrast, colour changes within the aggregates was observed at temperatures > 300°C and was strongly determined by the mineralogy of the material. The analytical data demonstrated that there were three temperature regions that provide measurable data and information to inform fire investigators of the thermal history experienced by the concrete matrix. Between 70°C-200°C the evaporation of non-chemically bound water and dissociation of ettringite, gypsum and gel-like calcium silicate hydrate (CSH) occur, and can be detected using a range of the techniques used. The heat flux required for these reactions to occur was greater in all mixtures containing GGBS. Observed chemical and physical transformations between 300°C and 500°C were mainly due to the oxidation of iron hydroxide and the dehydroxylation of portlandite. Further chemical changes at 650°C and above were identified as a direct result of the decarbonation of CaCO3. At elevated temperatures, the absence of certain minerals within the concrete formulations provided an indication of the temperature which would have been reached by the concrete matrix. It was also revealed that in some cases the minerals present rehydrated during cooling of the concrete and this was also detectable using a number of the analytical methods employed. For the test samples exposed to real fire conditions, the rebound hammer, UPV and compressive strength measurements all provided good indications of physical losses experienced by the concrete, however these methods were not good estimators of the exposure temperature. The results from TGA, DSC and FTIR in particular were more reliable but differed from the reference models in that water used in suppression and absorbed by the concrete affected some of the predicted features. XRD also revealed peaks which could be related to various phases of change within the concrete, which was helpful in revealing the thermal history of concrete. Discolouration of cross-sections of the concrete samples produced trends similar to the lab-heated specimens however this was hard to visualise on the surface of the concrete due to the soot layer resulting from the fire. The results characterised, for the first time, chemical and physical changes occurring within a range of concrete mixtures used in the UAE and linked these to specific temperature ranges to which the concrete were exposed. Furthermore, this work has demonstrated that a number of the analytical techniques used can be helpful in the determination of the thermal history of concrete which has been exposed to fire conditions.

Identiferoai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:685627
Date January 2016
CreatorsAlqassim, Mohammad Ali Mohammad Abdulla
ContributorsJones, Martyn
PublisherUniversity of Dundee
Source SetsEthos UK
Detected LanguageEnglish
TypeElectronic Thesis or Dissertation
Sourcehttps://discovery.dundee.ac.uk/en/studentTheses/d3b61df4-ee76-4180-965f-da4db8b3c043

Page generated in 0.0023 seconds