Physical interpretation of temperature data measured in the SBI fire test: Nordtest technical report 416: Nordtest project No. 1381-98

In the SBI fire test, there are temperature sensors in the exhaust duct allowing measurement of the temperature rise of the exhaust gases (DT measurement). In principle these sensors provide means to monitor the production rate of thermal energy. However, while traversing to the exhaust duct the heated gases lose energy to their surroundings which in a rigorous determination of the heat release rate of the specimen must be taken into account. There is also another factor hampering the conversion of the DT values to the thermal energy readings, namely the fact that the temperature readings do not directly show the temperature of the gases, but rather reflect it through a heat transfer process involving also the duct wall temperatures. These complications make the experimentally simple approach to use DT for rate-of-heat-release evaluation a complex task with inherently high uncertainty and proneness to systematic errors. Thus the method is not suitable for routine testing and classification purposes. However, for some other applications, such as quality control and product development purposes, the DT method can in some cases provide a well-suited option for RHR assessment. The method can also be used in fire research, e.g., to analyse the convective portion of the total heat release. In this report we present a method to interpret the DT values in terms of physically relevant factors describing generation and loss of heat. The analysis yields straightforwardly the convective part of RHR of the specimen. However, to evaluate the total RHR of the specimen, external sources of information must be exploited to establish the radiative contribution of the total RHR. The analysis method is based on a simplified model of heat transfer in the SBI system and its mathematical formulation. The solution of the problem entails handling of an inverse heat-transfer problem. The novelty of the presented solution lies in the use of the regularised output least-squares method to tackle with this problem. The dextrous numerical solution of developed enables execution of the whole analysis by a spreadsheet program (EXCEL). As majority of the operations can be executed automatically by EXCEL Macros, integration of the program into, e.g., the SBI data-analysis software is straightforward. Experiments were carried out to establish and verify the heat transfer model and to check the operation of the calculation codes. Besides the use in connection with the SBI apparatus, the data analysis method developed in this study can in principle be applied also in other fire tests.