Heat transfer analysis of sand-filled steel sections

Square concrete-filled steel sections (CFT) have become more popular in recent years due to their more favourable fire performance than hollow steel sections. Although the fire performance of circular sections is widely discussed in the literature, square sections, which are more commonly used, are not. According to BS EN 13381-6:2012, sand can be used as a conservative fill material to simulate concrete, enabling simpler and faster laboratory testing. However, sand has rarely been investigated in high-temperature fire tests. This paper investigates the heat transfer in square sand-filled protected sections using finite element (FE) modelling with temperature data from experimental sand-filled specimens.
The research methodology involved using average thermocouple readings from the experiments as inputs for the FE model heat envelope and validation curves within the fill material. The thermal properties of steel and sand, including density, thermal conductivity, and specific heat, were incorporated into the model. The thermal conductance between the materials was thoroughly investigated using the FE model, which was instrumental in defining the contact interface. This rigorous methodology ensures the reliability and accuracy of the research findings. The thermal properties of sand were found to be more complex than most existing references in the current literature. Therefore, this paper uses laboratory quantification techniques of linear regression to determine the sand properties with moderate success. The FE model showed that temperature-dependent values for thermal conductivity and specific heat are required to best capture the thermal behaviour of sandfill. Further research needs to be done to improve values for temperature-dependent specific heat. The FE model for the sand-filled section was validated and can be extended to predict the behaviour of other section sizes. A predicted model for a concrete-filled square section is proposed that can be validated against future experimental data.