Chapter 10 in the book: The Dalmarnock Fire Tests: Experiments and Modelling, Edited by G. Rein, C. Abecassis Empis and R. Carvel, Published by the School of Engineering and Electronics, University of Edinburgh, 2007. ISBN 978-0-9557497-0-4

An international round-robin study of fire modelling was conducted prior to the Dalmarnock Fire Tests in order to assess the state-of-the-art of fire modelling in real scenarios. The philosophy behind the Dalmarnock Fire Tests was to provide instrumentation density suitable for comparison to field models and designed the scenario for maximum test reproducibility. Each participating team independently simulated a priori the test using a common detailed description of the compartment geometry, fuel packages, ignition source and ventilation conditions. The aim of the exercise was to forecast the test results as accurately as possible, and not to provide an engineering analysis with adequate conservative assumptions or safety factors. The modelling results and experimental measurements are compared among themselves, allowing for conclusions on the robustness, reliability and accuracy of current modelling practices. The results indicate large scatter and considerable disparity among predicted fires and also differing from the experimental data. The Dalmarnock Fire Test One was benchmarked against a second test to establish the potential experimental variability. The scatter of the simulations is much larger than the experimental error and the experimental variability. The study emphasises on the inherent difficulty of predicting fire dynamics and demonstrates that the main source of scatter is originated in the many degrees of freedom and the uncertainty in the input parameters. The conclusions from the study are made public to encourage debate and exchange of views on the topic of fire modelling.