Fire performance of sandwich panels in a modified ISO 13784-1 small room test: the influence of increased fire load for different insulation materials

Abstract

Four sandwich panel rooms were constructed as prescribed in the ISO 13784-1 test. However, the construction followed normal industry practice, and the panels were also subjected to the kinds of damage typically found in commercial premises, although such damage may not typically be concentrated in such a small room. The fire load was increased to simulate fires actually occurring in commercial premises by stepping up the propane burner output from the usual maximum of 300–600 kW, and by placing a substantial wooden crib in two of the rooms. The results showed significant differences in fire growth rate and burning behaviour between those panels filled with polyisocyanurate (PIR) and those filled with stone wool in both the experiments without and with the wooden crib. Most significantly, the PIR pyrolysis products caused earlier ignition (by radiation from above) of the wooden crib 11 min into the experiment (1 min after the burner was stepped up to 300 kW), whereas the crib ignited 22 min into the test (2 min after the burner had been stepped up to 600 kW, which is beyond the test standard both in time and heat input) for the stone wool panels. This interaction between building and contents is most often ignored in fire safety assessments. After a few minutes, the PIR pyrolysis products that escaped outside the room, from between the panels, ignited. The extra thermal exposure from the PIR-fuelled flames distorted the panels, which in turn exposed more PIR, resulting in large flames on both the inside and outside of the enclosure. From a fire safety perspective this is most important as it shows that with the large fire loads that are commonly found in commercial premises, steel-faced PIR filled panels are not capable of acting as fire barriers, and may support flame spread through compartment walls and ceilings. In addition, the PIR panelled rooms produced very large quantities of dense smoke and toxic effluents, whereas the stone wool panelled rooms produced small amounts of light smoke of lower toxicity. Furthermore, the experiments showed that modifications to the standard test can lead to extremely different outcomes for some of the products. As the modifications simulated real-life situations, it seems important to discuss whether the standard is robust enough for property safety scenarios encountered in industrial premises.