Premium
Repeatability assessment of large‐scale fire experiment involving water spray system in a forced ventilated compartment
Author(s) -
Pretrel Hugues,
Querre Philippe
Publication year - 2018
Publication title -
fire and materials
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.482
H-Index - 58
eISSN - 1099-1018
pISSN - 0308-0501
DOI - 10.1002/fam.2679
Subject(s) - repeatability , scale (ratio) , full scale , environmental science , metric (unit) , simulation , ventilation (architecture) , cabin pressurization , computer science , engineering , statistics , mathematics , structural engineering , mechanical engineering , physics , operations management , quantum mechanics
Summary Repeatability of large‐scale fire test remains a key issue for code validation process. Most of the large‐scale experimental studies are based on single experiment, and the influence of repeatability is barely considered in the test analysis process. Due to the substantial cost, reproducing several trials of a given large‐scale fire scenario is not often performed. In the framework of the OECD PRISME 2 project, this topic has been identified, and a specific large‐scale fire test has been reproduced twice in the final goal of assessing the level of repeatability. The scenario is an oil pool fire in an enclosure mechanically ventilated and during which a water spray system is activated. The analysis consists in identifying a set of variables on which metrics is applied in order to quantify the levels of discrepancy between the two tests. A set of 27 variables are selected such as they characterize the whole fire scenario (the fire source, the gas phase, walls, the ventilation network, and the water spray system). The analysis points out that the repeatability levels are different depending on the type of variable. The gas temperature or species concentrations are more repeatable than gas pressure or air flow rate. In addition, a new methodology is proposed in comparing, for each physical variable, the variations due to repeatability (ie, the precision) and the uncertainty. A new metric is proposed helping modelers in code validation process.