|dc.description.abstract||Intumescent coatings are commonly used to protect steel structural elements from fire. In the event of a fire, an intumescent coating undergoes multiple chemical reactions and physical process, causing the coating to swell and form a layer of char around the steel. This char has an insulating effect on the steel and thus increases the time before the steel reaches a critical failure temperature. The increase in time before failure, provides additional time for the evacuation of people, and fire fighting activities.
Although intumescent coatings are used extensively in high-value residential, commercial and industrial settings, the intumescence process is itself poorly understood. In this research project, the performance of two different formulations of intumescent coatings under non-standard heating conditions was investigated.
The first part of this thesis describes the investigation of a constant heat flux additive method. This method aims to predict the performance of intumescent coatings over a wide range of variable heat flux conditions based on a relatively small number of constant heat flux experiments. To investigate the validity of this method, an experimental programme using a radiant panel array was carried out. In this experimental work, an epoxy-based intumescent coating was tested to four constant heat flux exposures, and using the data obtained, and the additive method, the backface steel temperatures under three variable heat flux exposures was predicted. It was found that the additive method under-predicted the backface steel temperature for all three variable heat flux conditions. It is demonstrated that the additive method relies on the assumption that char produced at different heat fluxes has identical heat transfer properties. By investigating the char which was developed in the constant heat flux experiments, it was shown that this assumption is not valid, and therefore the additive method can not accurately predict the backface steel temperatures in variable heat flux exposures for the coating studied.
The second part of this project investigated the parameters which can impact on the development of the char during the intumescence process. Specifically, the effects of incident heat flux, sample orientation, atmospheric conditions, and sample size on intumescence were investigated. Whether or not the choice of testing apparatus can influence the development of the char during intumescence was also explored. To investigate these parameters, a water-based intumescent coating was tested using three apparatus; the cone calorimeter, the Fire Propagation Apparatus (FPA) and the Radiant Panel Array (RPA). It was found that the sample orientation affected the backface steel temperatures recorded, with samples tested in a vertical orientation showing higher backface steel temperatures compared to horizontal samples. The atmospheric conditions did not have a significant impact on the performance of the coating. The investigation into the influence of sample size on the performance of the coating was inconclusive due to the potentially different average heat flux exposures to the coating, caused by the variability of the heat flux from the RPA. Finally, it was found that when identical samples were tested using different equipment, the performance of the coating was not consistent. This shows that conclusions reached on the basis of testing in one apparatus might not necessarily be applicable in other scenarios.
The practical significance of the findings described in this thesis are as follows. (1) Current industry practice of using standard furnace tests to certify intumescent coatings could lead to results which are not representative of the behaviour of an intumescent coating in a real fire. (2) Whilst the atmospheric conditions did not appear to impact on the heat transfer properties of the char which was produced (for the coating tested), the gas analysis and TGA data did show that the atmospheric conditions do impact on the chemical reactions taking place. Therefore it is the author's belief that the effects of atmospheric conditions on the intumescence process have to be determined for intumescent coating formulations before they can be used. (3) The RPA (like all experimental methods) has its own potential issues, i.e. the sample does not receive a uniform incident heat flux over its surface. However, in the RPA the source of this experimental error can be measured, and therefore quantified. (4) Comparisons between experimental results obtained using different apparatus can not automatically be made, as the intumescence process has been shown to be affected the apparatus which is used to study it. Furthermore, due to the impact of the apparatus on the intumescence process, the validity of using any apparatus in order to draw conclusions on how an intumescent coating will perform in a real fire cannot automatically be made.||en