Digital image processing for the advanced characterization and simulation of experimental fire tests
Por:
Schiaroli, Alice, Mata-Miquel C, Scarponi, Giordano Emrys, Habib, Abdel Karim, Kluge, Martin, Ustolin, Federico and Cozzani, Valerio
Publicada:
1 sep 2025
Ahead of Print:
1 jul 2025
Resumen:
The fire engulfment of storage tanks of hazardous materials is among the
most critical scenarios in hazard assessment of industrial value chains.
Laboratory and full-scale experimental trials are used to test the tank
performance and integrity in such scenarios. However, strong
uncertainties usually affect the actual fire load experienced by the
tank, in particular when large-scale experiments are carried out in open
test fields. Such uncertainties arise from several factors difficult to
control during experimental tests, such as the atmospheric conditions as
there is the influence of wind drifts influencing the actual fire
engulfment, the flame temperature and the flame dynamic distribution
around the target. Consequently, verifying the concordance of an
experimental test with standard test criteria and defining accurate
boundary conditions in correlated model simulations is challenging. In
this study, the development of a novel method for the analysis of fire
conditions based on image processing is presented. The approach allows
identifying the flame coverage on the target surface during the test and
provides an accurate map of the flame distribution on the equipment over
time. The approach is tested using experimental data from a full-scale
fire test campaign carried out on liquid hydrogen cryogenic tanks. The
results prove to be accurate in replicating the experimental
temperatures measured on the outer tank shell during the test. The
proposed methodology can be used to better understand the results of
experimental fire tests and to characterize realistic fire scenarios,
also supporting the definition of fire test requirements. Moreover, the
approach produces results that can be implemented as advanced
space-time-varying boundary conditions in simulation models, improving
their accuracy in reproducing real cases.
Filiaciones:
Univ Bologna, Dept Civil Chem Mat & Environm Engn, LISES Lab Ind Safety
& Environm Sustainabil, Via Terracini 28, I-40131 Bologna, Italy
Norwegian Univ Sci & Technol NTNU, Dept Mech & Ind Engn, N-7491
Trondheim, Norway
Univ Politecn Cataluna, Res Ctr Biochem Engn CREB, Barcelona 08028,
Spain
Bundesanstalt Mat Forsch & Prufung, Eichen 87, D-12205 Berlin, Germany
Open Access
|