Full-Scale Fire Performance of Combustible and Non-combustible Curtain Wall Systems with Firestops

Facade systems used in modern buildings have received much attention in recent times due to their involvement in the propagation of fires in various incidents. These systems comprise of multiple components—cladding frame (typically of aluminum), façade panels (glass, aluminum composite panels, etc.), perimeter firestop (to seal the gap between floor and façade) and spandrel fire protection (typically provided from the inside of a compartment). While many standardized testing methods exist for quantifying their fire performance, most of them consider the behavior of individual components in isolation. The test conditions are also quite different from those encountered in real fire scenarios. The current study highlights these gaps and provides inferences from six full-scale real fire experiments conducted in a three-storey structure with different façade (curtain wall) and firestop configurations. Combustible façade panels of aluminum composite panels (ACP) and medium density fiberboard (MDF) and non-combustible glass panels of single glazed units and double-glazed units were utilized whereas two different methods of edge of the slab firestop and spandrel insulation were employed. The fire scenarios were developed at all floor levels (fire loads as well as initial ventilation conditions) as per realistic residential/office type dwellings. It was found that once the flames leap out of the fire compartment due to the failure of the façade panels, the spandrel area was subjected to 31 kW/m2 of additional heat flux, which indicated the need to consider fire protection of the spandrel area from the outside, especially for combustible façade systems. Further, one of the firestop installation methods was found to be more robust to address site tolerances and installation uncertainties arising due to workmanship as it allowed significantly less ingress of hot gases and toxic fumes to the upper floors. In terms of the two combustible panels, ACP and MDF, it was found that there were differences in their performance in a bench-scale experiment (cone calorimeter) and the full-scale experiments. MDF was found to perform better in the full-scale experiments. The method used for securement of the façade panels (pressure tape/silicone sealant and screws) was also found to have significant effect on the overall performance of the façade system; the failure mechanism of the façade panels was found to be different in both cases. It is expected that the findings presented in this study will provide insights to the real performance of different façade components as a system and will help in improving the codes and standards in the future.