Sensitivity analysis on the microscale combustion calorimeter for polyurethane foam using a full factorial design methodology

Abstract

A sensitivity analysis on test outcomes from the microscale combustion calorimeter was performed using flexible polyurethane foam as the sample material. Three input factors were investigated: sample preparation, heating rate and pyrolysis chamber atmosphere. The experiments were setup according to the principles of design of experiments and a full factorial design was implemented. In total, 39 experiments were performed as part of this study: 27 unique tests and 12 repeat experiments. Results from the microscale combustion calorimeter showed excellent repeatability in most cases. Sample inhomogeneity was suggested as a possible explanation for the variation observed. A statistical analysis was performed to quantitatively assess how each input factor affected the test results. Analysis showed that all three factors had some effect on the test outcomes, with heating rates and pyrolysis chamber atmosphere proving to be the most influential in regard to the analysed output quantities. Increasing the heating rate from 15 to 180°C/min illustrated the effects of thermal lag, increasing the temperatures at which the sample reactions took place. Interaction effects between input parameters were also observed. The pyrolysis chamber atmosphere, either inert or oxidative, was observed to have the largest effect on the tested output parameters. The energy release was observed to decrease with the introduction of 20% oxygen. Increasing the oxygen levels also produced a drop in the peak reaction temperatures; this was consistent for all tested atmospheres (from 0% to 10% and 10% to 20% oxygen). This study illustrated that outcomes from this apparatus, for materials such as flexible polyurethane foam, can be dependent on sample preparation as well as heating rate and testing atmosphere. In addition, it highlights the importance of apparatus characterisation for less homogeneous materials to understand how input factors choices can influence the outcomes of the tests performed.