Use and uncertainty of an inverse heat conduction code for estimating heat flux to a large pipe calorimeter in a jet fuel fire

Heat flux to a large pipe calorimeter is determined by use of an inverse heat conduction code for the purpose of benchmarking large scale fire simulators. Benchmarked results from Container Analysis Fire Environment (CAFE) simulations are used to calibrate an inverse heat conduction code by optimizing number of future times (NFT) at 11 and identifying a linear correlation and uncertainty range for the unknown heat flux. Time-varying heat flux is estimated for three fire tests at 58 locations along the calorimeter occupied by thermocouples in the original experiments. A Curie transition in the pipe calorimeter steel causes high levels of instability in the calculation, and as a result any data obtained from the inverse heat conduction code during this transition are discarded. Results for a single thermocouple location are presented and are representative of other thermocouples in the experiments. The calibration is then used to adjust the heat flux prediction for the calorimeter during the three experiments. Estimated heat fluxes are approximately an 11 second window average of the actual heat fluxes. The maximum heat flux occurred at the beginning of the first experimental fire and was 195 ± 37 kW/m 2 at a 95% confidence level.