To improve the reliability of safety assessments in domestic performance-based designs (PBDs), the problem of the input parameters being dependent on fire-simulation users was quantitatively analyzed. Thus, the results of statistical analyses of domestic PBD reports evaluated over the last 5 years were examined. It was determined that the uncertainties of the input parameters might have a relatively larger influence on the statistical deviations than the measurement uncertainties. Accordingly, a sensitivity analysis was performed by considering the statistical deviations of the input parameters that could greatly influence the prediction results of visibility, which are important for the available safe egress time. The main results were as follows: a large change in visibility was observed owing to deviations of the heat release rate and soot yield. Based on this study, it is expected that more accurate results can be obtained if the objectivity of input parameters determined by user dependence can be secured in domestic PBDs.

In an oxygen consumption calorimeter system for measuring heat release rate, the CO/CO₂ analyzer and the O₂ analyzer for chemical species analysis of the sampling gas are the main factors that increase the construction cost of the system. In this experimental study, a low-cost device with performance similar to that of the existing high-cost device was developed. To evaluate the contribution of the CO/CO₂ analyzer to the measurement of the heat release rate, the performance of unsteady and steady-state conditions depending on whether CO/CO₂ was measured for diffusion flames with various fire growth rates was examined. As a result, it was confirmed that when only the O₂ concentration was considered, the heat release rate could be measured accurately. Based on these results, the measurement performance of a low-cost O₂ analyzer manufactured using an oxygen sensor for healthcare was verified. As a result of the verification of the low-cost O₂ analyzer for diffusion flames with various fire growth rates, it was confirmed that the relative error was within 10% for unsteady and steady-state fire sources and can be used to measure the heat release rate.