Prediction of heat release in the single burning item test

Abstract

Abstract This paper presents a modification of an existing mathematical model that uses cone calorimeter test results to predict heat release rate in the single burning item (SBI) test. A method for classification of cases based on multivariate statistical analysis is incorporated in the modified model. This makes it possible to determine the development of the heat release curve in the SBI test with better precision than the original version was able to. The model has been applied to 100 cases of cone calorimeter test results from 33 different products. For most of the products the predicted shape and level of the heat release curves are close to measurements in the SBI test. Using the predicted heat release results as input to calculation of FIGRA 0.2MJ and THR 600s within the new European system for reaction‐to‐fire classification, we were able to predict membership of the correct class in 90% of the analysed cases. Copyright © 2002 John Wiley & Sons, Ltd.