Spalling of concrete specimens exposed to elevated temperatures using gas-fired radiant panel arrays

Abstract

Portland cement concrete is one of the most used construction materials the modern world; its use spans singular structural elements to having mega structures built almost entirely from it. Ever since its initial development (in its current form), concrete has been the subject of ongoing innovation and technological advances aimed at making it more efficient, more durable, and easier to produce. However, given its very significant carbon footprint, Portland cement concrete technology has more recently been the subject of intense research focusing on making it more environmentally sustainable and reducing its carbon intensiveness. One such advance has been pioneered by researchers at the Swiss Federal Laboratories for Materials Science and Technology (Empa), where controlled expansive concrete has been used to achieve moderate levels of prestressing (i.e., self-prestressing), when used in conjunction with ultra-high modulus carbon fibre reinforced polymer (UHM CFRP) bars/tendons. Given the corrosion-free nature of CFRP bars, the cover to reinforcement could be reduced dramatically, without risking damage to them from weathering. The reduced cover is big environmental advantage given it leads to less resources being wasted. This type of novel concrete element generates potential hazards that have hitherto not been explored with respect to its behaviour at elevated temperatures, or in fire. Therefore, the project presented in this thesis focused on investigating the behaviour of small-scale samples fabricated from such self-prestressing concrete mixes under exposure to steep thermal gradients, such as those that are likely to be experienced in a building fire scenario.

A large number of test samples were cast, varying a rang eof relevant parameters related to both the concrete mix design and the internal reinforcement used, and their curing was monitored to characterise the levels of prestressing that had been achieved prior to elevated temperature testing. Elevated temperature testing was conducted in the Structures Test Hall at the University of Edinburgh using a mobile gas-fired radiant panel array (RPA). In order to properly characterise the thermal boundary conditions created by the RPA, an experimental campaign was performed to determine the effects of the presence of samples with heated surfaces (such as concrete) on the calibrated heat-flux exposure conditions. Experiments included measurements of calibrated heat fluxes (HFs) with no samples present, and also with samples made from concrete, or vermiculite, or a water-cooled metal plate (thus altering thermal feedback to the RPA). The results showed that the surface temperature of the RPA rose significantly when there was a heated sample present, with consequent increases (of up to 78%) in the value of the calibrated imposed HF to the sample. Using this increase in the surface temperature, a correction method was suggested to allow for the necessary correction of the results. Failure to include this correction results in a significant misrepresentation of the thermal insult to the test specimen when using this type of testing methodology.

Results from the high temperature tests on self-prestressed concrete samples showed that these were prone to explosive heat induced spalling. It was also observed that the addition of 2 kg/m3 of specific polypropylene (PP) microfibres to the fresh concrete led to the effective elimination of heat-induced spalling in such samples, thus corroborating the effectiveness of this methodology of spalling mitigation also for self-expanding concrete mixes such as those tested herein. Furthermore, it was shown that samples with no/reduced prestress were less likely to spall, thus corroborating the theory that imposed compressive stress can exacerbate heat-induced explosive concrete spalling. For pre-dried (i.e. conditioned) samples, no spalling was recorded in samples that were otherwise very susceptible to spalling, thus corroborating the theory that pore moisture plays a central role in heat-induced explosive spalling of concrete, including for self-prestressed mixes such as those tested herein.

Results from high temperature tests on mechanically prestressed and non-prestressed samples, made using more conventional (i.e. non self-expanding) high performance self-compacting concrete (SCC) were non-conclusive as regards their heat induced spalling hazards. Regardless of the severity of the imposed HF, the sample orientation, the presence or lack of prestress, none of these samples spalled during exposure to elevated temperature. Every sample with CFRP internal reinforcement/prestressing was observed to develop longitudinal cracks along the CFRP (due to mismatch between the coefficient of thermal expansion for the CFRP and the surrounding cementitious materials), which allowed for pore moisture to exit the samples’ rear face when exposed to heating. The lack of spalling in these sample corroborates the complexity of factors influencing heat induced spalling in high-performance concrete samples. All the different testing arrangements, including possible reasons why spalling was or was not observed (in samples that were expected to spall) are presented and discussed in detail.

This thesis shows the complexity of heat induced spalling in high-performance concrete, and the vast range of factors that can affect the outcome of experimental investigating it. It also highlights common assumption made during experimental work, that have the potential to severely skew the results of experimental work. Based on the results, further research into specific areas is recommended to further the boundaries of knowledge into the explored topics.