Concrete filled steel hollow structural (CFS) sections are increasingly used to support large compressive loads in buildings, with the concrete infill and the steel tube working together to yield several benefits both at ambient temperature and during a fire. These members are now widely applied in the design of highly optimized multi-storey and high rise buildings where fire resistance ratings of two or more hours may be required. Whilst the response and design of these sections at ambient temperatures is reasonably well understood, their response in fire, and thus their fire resistance design, is less well established. Structural fire resistance design guidance is available but has been developed based on tests of predominantly short, concentrically-loaded, small-diameter columns in braced frames using normal strength concrete. The current prescriptive guidance is limited and the design of CFS columns is thus often based on a detailed performance based approach, which can be time consuming and expensive and which is generally not well supported by a deep understanding of CFS columns’ behaviour in real fires. This thesis aims to understand the fundamental thermal and mechanical factors at play within these sections so as to provide guidance on how to improve their design for fire resistance when applied either as unprotected or protected sections. A meta-analysis of available furnace test data is used to demonstrate that current guidance fails to capture the relevant mechanics and thus poorly predicts fire resistance. It is also demonstrated that the predictive abilities of the available design standards vary with physical characteristics of the CFS section such as shape and size. A factor which has been observed in furnace tests on CFS sections but which is not accounted for in available guidance is the formation of an air gap between the steel tube and the concrete core due to differential expansion; this affects their structural response in fire. The insulating effect of air gap formation has not previously been addressed in literature and an experimental program is presented to systematically assess the effects of a gap on the heat transfer through the section; showing that the presence of even a 1 mm gap is important. To explicitly assess the heat transfer response within both unprotected and fire protected (i.e. insulated) CFS sections, 34 large scale standard furnace tests were performed in partnership with an industry sponsor. Fourteen tests on large scale unloaded unprotected CFS sections are presented to assess current capability to predict the thermal response and to assess the effects of different sectional and material parameters on heating. New best practice thermal modelling guidance is suggested based on comparison between the models and observed temperatures from the tests. Twenty CFS specimens of varying size and shape, protected with different types and thicknesses of intumescent paint fire insulation, were also tested unloaded in a furnace to understand the thermal evolution within protected CFS sections and to develop design guidance to support application of intumescent coatings in performance based fire resistance design of CFS sections. These tests demonstrate that the intumescent coatings were far more effective than expected when applied to CFS sections, and that current methods of designing the coatings’ thickness are overly conservative. The reason for this appears to be that the calculation of effective section factor which is used in the prescription of intumescent coating thicknesses is based on the thermal response of unprotected CFS sections which display fundamentally different heating characteristics from protected sections due to the development of a thermal gradient in the concrete core. It is also demonstrated (by calculation supported by the testing presented herein) that the steel failure temperature (i.e. limiting temperature) of an unprotected CFS column in fire is significantly higher than one which is protected; procedures to determine the limiting temperature of protected sections are suggested. Finally, the residual strength of fire-exposed CFS columns is examined through structural testing of 19 of the 34 fire tested columns along with unheated control specimens. The results provide insights into the residual response of unprotected and protected CFS section exposed to fire, and demonstrate a reasonable ability to calculate their residual structural capacity. The work presented in this thesis has shed light on the ability of available guidance to rationally predict the thermal and structural response to fire of CFS columns, has improved the understanding of the thermal evolution within protected and unprotected CFS sections in fire, has provided best-practice guidance and material input parameters for both thermal and structural modelling of CFS sections, and has improved understanding of the residual capacity of CFS sections after a fire.

No / This paper presents experimental results of temperature distribution in fire in four typical types of steel-concrete composite joint (web cleat, fin plate, flush endplate and flexible endplate) with different fire-protection schemes. The test specimens were unloaded and the steelwork of each joint assembly was exposed to a standard fire condition [ISO 834, 1975: Fire Resistance Tests, Elements of Building Construction, International Organization for Standardization, Geneva] in a furnace. In total, 14 tests were conducted, including 4 tests without any fire protection and 10 tests with different schemes of fire protection. The main objective of these tests was to investigate the effects of three practical fire-protection schemes as alternatives to full fire protection of the entire joint assembly. The three alternative methods of fire protection were: (1) protecting a segment, instead of the entire length, of the beams; (2) unprotected bolts and (3) protecting the columns only. The main results of these tests are: (1) if all the steel work (excluding the bolts) in the joint assembly was protected, whether or not protecting the bolts had very little effect on temperatures in the protected steelwork other than the bolts. The bolt temperatures were higher if they were not protected than if they were protected, but the unprotected bolt temperatures in a joint with fire protection to other steelwork were much lower than bolt temperatures in a totally unprotected joint; (2) as far as joint temperatures are concerned, protecting a segment of 400 mm of the beam was sufficient to achieve full protection and (3) if only the column was protected, only the joint components that were in the immediate vicinity of the column (such as welds) developed noticeably lower temperatures than if the joint assembly was unprotected, but due to heat conduction from the unprotected steel beams, these temperature values were much higher than if the joint assembly was protected. Furthermore, the column temperatures in the joint region were much higher than the protected column temperatures.

Epoxy intumescent coatings are fire protection systems for steel structural elements that are widely used in applications that protection from severe hydrocarbon fires is required, such as oil and gas facilities. These polymer coatings react upon heating and expand into a thick porous char layer that insulates the protected steel element. In the typical fire scenarios for these applications, the intumescent coatings must resist very high heat fluxes and highly erosive forces from ignited pressurised gases. Hence, continuous fibre reinforcement is embedded in the thick epoxy coating during installation, so as to ensure the integrity of the weak intumesced char during fire exposure. This reinforcement is typically in the form of a bidirectional carbon and/or glass fibre mesh, thus under normal service conditions a fibre-reinforced intumescent coating (FRIC) is essentially a lightly fibre-reinforced polymer (FRP) composite material. This thesis examines the impacts of embedded high strength fibres on the tensile behaviour of epoxy intumescent materials in their unreacted state prior to fire exposure, and the potential enhancements that arise in the structural performance of elements protected with FRICs. An experimental programme is presented comprising tensile coupon tests of unreacted intumescent epoxies, reinforced with different fibre meshes at various fibre volume fractions. It is demonstrated that the tensile properties of FRICs can be enhanced considerably by including increasing amounts of carbon fibre reinforcement aligned in the principal loading direction, which can be tailored in the desired orientation on the coated structural members to enhance their load carrying capacity and/or deformability. An experimental study is presented on coated intact and artificially damaged I-beams (simulating steel losses from corrosion) tested in bending, demonstrating that FRICs can enhance the flexural response of the beams after yielding of steel, until the tensile rupture of the coatings. An analytical procedure for predicting the flexural behaviour of the coated beams is discussed and validated against the obtained test results, whereas a parametric analysis is performed based on this analytical model to assess the effect of various parameters on the strengthening efficiency of FRICs. The results of this analysis demonstrate that it is feasible to increase the flexural load capacity of thin sections considerably utilising the flexural strength gains from FRICs. Finally, a novel application is proposed in this thesis for FRICs as a potential system for structural strengthening or retrofitting reinforced concrete and concrete-encased steel columns by lateral confinement. An experimental study is presented on the axial compressive behaviour of short, plain concrete and concrete-encased structural steel columns that are wrapped in the hoop direction with FRICs. The results clearly show that epoxy intumescent coatings reinforced with a carbon fibre mesh of suitable weight can provide lateral confinement to the concrete core resisting its lateral dilation, thus resulting in considerable enhancements of the axial strength and deformability of concrete. The observed strengthening performance of the composite protective coatings is found to be at least as good as that of FRP wraps consisting of the same fibre reinforcement mesh and a conventional, non-intumescent epoxy resin. The predictive ability of existing design-oriented FRP confinement models is compared against the experimental results, and is found to be reasonably precise in predicting the peak strength of the tested columns, hence existing models appear to be suitable for design and analysis of column strengthening schemes with the proposed novel FRIC system. The research presented herein shows clearly that FRICs have a strong potential as alternative systems for consideration in the field of structural strengthening and rehabilitation, since they can provide substantial enhancements in the load carrying capacity for both applications considered. At the same time FRICs can thermally protect the underlying structural elements in the event of a fire, by intumescing and charring, thus potentially eliminating the need for additional passive fire protection that is common with conventional fire-rated FRP wrapping systems. Although this thesis provides a proof-of-concept for use of the proposed novel FRICs as structural strengthening materials, considerable additional research is particularly required to study their fire protection performance when applied to concrete substrates, to make use of the proposed hybrid functionality with confidence.

The expansion ratio (defined as the ratio of the expanded thickness to the original thickness) of intumescent coatings is the most important quantity that determines their fire protection performance. This thesis explores two possible methods of predicting intumescent coating expansion: an analytical method, and a detailed numerical simulation method using Smoothed Particle Hydrodynamics (SPH).The analytical method is based on a cell-model and predicts bubble growth due to pressure increase in viscous liquid with constant viscosity. It has been extended to non-uniform temperature field and temperature-dependent viscosity of intumescent melt. Accuracy of this extended analytical method is assessed by comparison against the cone calorimeter and furnace fire tests on intumescent coating protected steel plates with different intumescent coating thicknesses, steel plate thicknesses, and heating conditions. The extended analytical method is then used to investigate how intumescent coating expansion and intumescent coating effective thermal conductivity are affected by changing the coating thickness, the steel thickness and the fire condition (including smouldering fire). The main conclusion is that the expansion ratio decreases as the rate of heating increases. Therefore, the intumescent coating properties obtained from the Standard fire exposure may be safely used for slower realistic fires, but would produce unsafe results for faster fires. The second method explores the potential of a meshless numerical simulation: Smoothed Particle Hydrodynamics (SPH). SPH modelling of intumescent coating expansion has been implemented using the SPHysics FORTRAN open-source code as a platform. To check the validity of this modelling method, the modelling results are compared against theoretical solutions for surface tension (Young-Laplace theorem), and available numerical and analytical solutions for bubble expansion. A new algorithm for representing the mass transfer of gas into the bubble using SPH particle insertion and particle shifting scheme is presented to simulate the bubble expansion process. Close agreement with an analytical solution for the initial bubble expansion rate computed by SPH is obtained. Whilst this research has demonstrated the potential of using SPH to numerically simulate intumescent coating expansion, it has also revealed significant challenges that should be overcome to make SPH a feasible method to simulate intumescent coating expansion. The main challenges include:• Simulating gas-polymer flows when expansion is occurring where there are vastly different properties of these two fluids with a density ratio of about 1000. This high density ratio may easily cause numerical pressure noise, especially at the liquid-gas interface.• Extremely high computational cost necessary to achieve sufficient accuracy by using a large number of particles (higher resolution), especially for the multi-phase SPH program, and very small time step for the lighter fluid (air). • The behaviour of intumescent coatings involves expansion ratios on the order of 10-100 with thousands of bubbles which grow, merge and burst. Based on the results of this exploratory research, future improvements are outlined to further develop the SPH simulation method.

Fire presents a great risk to humans and their possessions. Polymeric compounds are highly flammable and compounds are added to increase their fire resistance. These additives are referred to as flame retardants. Intumescent flame retardants are systems that form a carbon rich foam on the surface of the polymer during thermal degradation. A new field of metal catalysed intumescence is discussed in this document. The influence of both the metal ion and organic backbone of polyhydroxylcarboxylic acid metal salts was studied. The intumescence and foaming properties of selected salts were studied in more detail. Gluconic acid proved to be the best intumescent and showed the best combination of carbon char yield, foam structure and volume, stability and open flame characteristics. As a general rule of thumb, the char volume and structure improved with an increased number of hydroxyl groups in the compound. Carbon char yield increased with an increase in the number of hydroxyl groups for a constant number of carbons in the complex. The number of acid groups in the compound played a lesser role in intumescence. Most metal complexes catalyse intumescent decomposition of polyhydric compounds such as polyhydroxylcarboxylic acids and pentaerythritol. The reduction in the volatilisation losses implies that the metal cations catalyse carbonisation reactions in the polyhydroxyl compound. Unfortunately, the resultant carbon-foams are unstable: the metal residues also catalyse the further oxidative degradation of the char. The low thermal stability of the compositions tested, as well as the afterglow effect makes them unsuitable for use as flame retardant additives in plastics. From the studies of the different metal complexes with acetylacetone and gluconic acid, it was concluded that the calcium ion exhibited the most promise. When pyrolysed in air with an open gas flame or in a furnace, calcium gluconate monohydrate produces voluminous foam. The foam is of a closed cell structure, densely packed and has no mechanical strength. The cell walls are very thin (5 to 50 nm) and the cells are between 50 µm and 200 µm in size. If the sample is compressed before heating, the resultant foam produced during pyrolysis is less voluminous than that of the loose powder. The BET surface area of the calcium gluconate monohydrate foamed at 300°C for 5 min. is 16.0 m2/g. The foam produced after 5 min heating at 300°C has a thermal conductivity similar to standard polystyrene foam. Adding leached silica to the calcium gluconate monohydrate increases mechanical strength of the foam, but influences the foam volume negatively. The optimum silica level was calculated to be 1:1 gluconate to silica on a mole ratio (11.8% SiO2 by mass). The mechanical strength of the foam can also be increased with the addition of expandable graphite. The graphite has a lesser influence on the foam volume than the silica, but also reduces the foam volume. However, the addition of the expandable graphite gives the sample more .opening. force, as the foam volume of a compressed sample pyrolysed, is similar to that of the heated loose powder. Calcium gluconate monohydrate starts to degrade at 120°C, losing its crystal water and shows a mass loss of 4% at 180°C. The exothermic peak (shown in the DSC/TGA data) associated with the metal catalysed carbon oxidation (afterglow) is observed at 570°C in air. For the calcium gluconate monohydrate the transition from CaCO3 to CaO occurs above 650°C in air. It has been shown that the foaming of polyhydroxylcarboxylic acids is due to the formation of water vapour during degradation. For calcium gluconate monohydrate, foaming starts due to the loss of crystal water and is continued by the loss of hydroxyl groups as water. The bulk of the foaming is due to the second reaction. It has been shown that compounds with crystal waters produce a more voluminous and lower density foam. The foam is an amorphous carbon rich residue. The molecular mass of the carbon residue increases up to a heating temperature of 300°C. This implies that the carbon residue crosslinks during formation, forming a stretchable cell wall for the foam. Compounds with a .free. hydroxyl group at the end of the carbon chain produce a foam of larger volume and lower density. This supports the crosslinking theory. The foam produced when calcium gluconate monohydrate is heated for 5 min at 300°C in air is of very low density . 2.5 kg/m3 based on residual mass. The density of the calcium gluconate monohydrate pyrolysed at 1000°C for 5 min in air yields a CaO with a density of 20 kg/m3 . This implies that the high temperature pyrolysis of calcium gluconate can produce an inorganic oxide of low bulk density and possible high specific surface area. The BET surface area of CaCO3 from the gluconate pyrolysed at 600°C is ~ 12 m2 /g. An intumescent coating containing calcium gluconate monohydrate, leached silica and expandable graphite as a system was prepared and compared to commercial formulations. This gluconate based system was at least as efficient as the commercial formulations when painted on balsa wood planks or aluminium plates. On cardboard sheets it did not perform as well as the commercial systems. More work should be done to overcome the afterglow effect observed with metal- based intumescent systems. The crystal structure of calcium gluconate monohydrate should be determined as to understand the decomposition better. It is unclear whether the decomposition is catalysed inter- or intramolecularly. / Thesis (PhD(Chemical Engineering))--University of Pretoria, 2005. / Chemical Engineering / unrestricted

Intumescent coating fire protection on steel structures is becoming widely popular in the UK and Europe. The current assessment for the fire protection performance method using the standard fire resistance tests is not accurate, owing to the reactive behaviour of intumescent coating at elevated temperature. Moreover, the available intumescent coating temperature assessment method provided in the Eurocode for structural steel at elevated temperature does not incorporate the steel beam's behaviour and/or assessment for partial protection and/or damaged protection. The research work presented provides additional information. on the assessment of partial and/or damaged intumescent coating at elevated temperature. In the scope of the investigation on the thermal conductivity of intumescent coating, it was found that the computed average thermal conductivity was marginally sensitive to the density and emissivity at elevated temperature. However, the thermal conductivity was found to be reasonably sensitive to the differences in initial dft's (dry film thicknesses). In this research, a numerical model was developed using ABAQUS to mimic actual indicative test scenarios to predict and establish the temperature distribution and the structural fire resistance of partial and/or damaged intumescent coating at elevated temperatures. Intumescent coating actively shields when the charring process occurs when the surface temperature reaches approximately 250°C to 350°C. Maximum deflection and deflection failure times for each damage scenario were analyzed by applying specified loading conditions. It was also found that the structural fire resistance failure mode of intumescent coating on protected steel beams was particularly sensitive to the applied boundary conditions. Careful selection of nodes in the element was necessary to avoid numerical instability and unexpected numerical error during analysis. An assessment of various numerical models subjected to a-standard fire with partially protected 1 mm intumescent coating was analysed using ABAQUS. An available unprotected test result was used as a benchmark. The outcome suggests that the fire resistances of the beams were found to be sensitive to the location of the partial and/or damage protection. The overall fire resistance behaviour of intumescent coating at elevated temperature was summarized in a 'typical deflection regression' curve. An extensive parametric analysis was performed on localized intumescent coating damage with various intumescent coating thicknesses between 0.5mm to 2.0mm. It was found that the average deflection was linear for the first 30 mins of exposure for all the variables, damage locations and intumescent thicknesses. It was concluded that a thicker layered intumescent coating may not be a better insulator or be compared to a much less thick intumescent coating at elevated temperature. The use of passive fire protection, however, does enhance the overall fire resistance of the steel beam, in contrast to a naked steel structure. The research work investigated the intumescent coating behaviour with different aspects of protection and damage and the outcome of the assessment provided a robust guide and additional understanding of the performance of intumescent coating at elevated temperature.

In the pursuit of lower cost intumescent flame retardant (IFR) systems, the compound expandable graphite (EG) was identified. This compound delivers high flame retardant performance but provides non-uniform thermal shielding when exposed to open flame from below due to negative gravitational effects. It was theorised that this may be remedied either through ion exchange of the interstratified ions with low glass transition ions or through use in binary systems with other compounds. Two classes of commercial EG were identified, namely a low and a high expansion onset temperature EG compound. Extensive characterisation of each EG compound was undertaken to assess its composition, expansion mechanisms and onset temperatures in order to identify compatible compounds for binary use. The susceptibility of each compound to ion exchange was also assessed. An industrial IFR ethylenediamine phosphate (EDAP) and a novel flame retardant were synthesised for assessment in binary use with EG. Coupled with the above study, this project developed two novel fire testing techniques as low cost alternatives to well-established fire testing methods such as cone calorimetry. The first technique involved an open flame fire testing method which allowed vertical or horizontal testing. Digital and infrared (IR) video recording during operation facilitated comparison of multiple performance indicators further strengthening this method. The second technique allowed assessment of the mass loss resistance of each compound during laser pyrolysis. Characterisation of the EG compounds allowed development of structural models to describe each compound and explain the mechanisms of their expansion and gaseous release. Exhaustive ion exchange testing did not deliver favourable results, necessitating the pursuit of compounds for binary use with EG. A novel IFR was synthesised by neutralising 3,5-diaminobenzoic acid hydrochloride salt with ammonium dihydrogen phosphate. This compound, which melts at 257 °C, decomposes concurrently to release carbon dioxide gas which promotes intumescent charring. The flame retardant performance of this compound and EDAP as primary flame retardants and in combination with expandable graphite was evaluated. As a proof of concept, the novel compound was tested as a primary flame retardant using cone calorimetry after which its utility in binary systems with low temperature expandable graphite was tested. Substantial decreases in peak heat release rate (pHRR) and flame out time were achieved for all binary systems. This success led to testing of a number of combinations of low and high expansion onset EG and the other IFRs to identify the highest performing combination, which proved to be the 10-10 EDAP-EG system. Combinations of EG and the novel compound also showed excellent results. The novel fire testing techniques proved effective in identifying high performance combinations and showed comparable trends to those measured in cone calorimetry, at a greatly reduced cost and material requirement. IR analysis of open flame fire testing indicated increases in the temperatures required for ignition and burn through of the substrate. Observations, corroborated by optical video, showed that cohesive and uniform thermal shielding was achieved in all binary systems tested. This study illustrates that systems of 10% EG combined with either 10% DABAP or 10% EDAP are both the most economical binary systems tested but are extremely high performance systems as well. Both of these systems delivered excellent results while being more economic than the widely used industrial system with a 25-30% EDAP loading. It is recommended that these compounds be considered for industrial use. Furthermore, the effective fire testing techniques developed in this study may be utilised in future fire testing to identify high performance compounds at a lower cost prior to further assessment through methods such as cone calorimetry. / Thesis (PhD)--University of Pretoria, 2017. / Chemical Engineering / PhD / Unrestricted

No description available.

The ASTM E119 is a large-scale test used to qualify assemblies for fire resistance, including heat transmission and structural integrity. The test requires specialized furnaces and full-scale assemblies that are 3.0 m (10 ft) or more on each side, making it very expensive to perform. In this study, we investigated the feasibility of the scaling methodology for a reduced-scale fire resistance test on different types of wood-based structures, specifically commercially available intumescent coating applied onto wood and bolted lap joints in wood. We build upon a previously developed scaling methodology for wood and gypsum boards, which integrated geometric scaling, Fourier number time scaling, and furnace boundary condition matching. Intumescent coating presents a particular challenge in scaling in that it expands when exposed to fire conditions. To account for this expansion, we identified a relationship between initial dry film thickness and final expanded thickness through cone calorimeter tests and integrated it into a modified scaling methodology. This approach was then validated through fire exposure tests in furnace on wood samples painted with intumescent coating at full, half, and quarter scales. Finally, we demonstrated the scaling laws for joints under combined thermo-structural loading, by subjecting wood-based half-lap joint samples to combined bending and thermal loading at half and quarter scale. The samples were subjected to static three-point bending with the load scaled to achieve structural similitude, while simultaneously being exposed to a scaled fire exposure on the bottom surface. Our study provides insights into the practical application of scaling methodology for testing the fire resistance of joints and fire-resistant coated wood, paving the way for more cost-effective and quicker fire testing for the wood-based composite industry. / Master of Science / The ASTM E119 is a critical test standard that evaluates the fire resistance of various building materials, including wood-based structures. However, the standard tests are quite expensive due to the need for specialized equipment and large-scale samples. In this study, we explored the potential of using a scaled-down fire resistance test on different types of wood-based materials, including commercially available fire-resistant coated wood and joints. We built on existing scaling methods for wood and gypsum boards and adapted it to account for the unique properties of intumescent coating - a fire-resistant material that expands when exposed to high temperatures. By conducting a series of tests, we developed a modified scaling approach to accommodate the expansion of the coating. We then validated this new method by performing fire exposure tests at various scales on wood samples coated with intumescent coating. Finally, we adapted the scaling methods to account for wood based bolted joints. We tested the fire resistance of wood-based half-lap joints under combined heat and structural stress at smaller scales. Our study offers valuable insights into a more cost-effective and efficient method for testing fire resistance in wood-based structures. By providing a reliable scaling approach for fire-resistant coated wood and joints, our work has the potential to make fire testing more accessible for the wood composite industry, ultimately leading to safer and better-performing buildings.

No / Based on temperatures measured in steel joints with different extents of fire protection, this paper proposes a simple method to calculate temperatures in steel joints with partial intumescent coating fire protection. The method combines the simple temperature calculation methods in EN 1993-1-2 (Committee of European Normalisation CEN, Eurocode 3: design of steel structures—part 1-2: general rules—structural fire design, 2005) for unprotected and protected steel structures through the introduction of an exposure factor, which is the ratio of the unprotected surface area of the joint region to the total surface area of the joint area. Using the measured temperatures for fully protected steel joints, this paper first extracts the effective thermal conductivity of the intumescent coating used in the fire tests. Afterwards, this paper presents validation results based on fire test results on joints with partial fire protection. Finally, this paper presents methods to calculate the exposure factor for different types of partially fire protected steel joints.