Since 1980 numerous explosion experiments have been performed by Gexcon in collaboration with other groups to further understand explosion phenomena. Gexcon used these experiments for development and validation of the FLACS explosion models. The image below demonstrates the FLACS simulation compared to the actual explosion experiment.
FLACS is a powerful CFD tool for risk assessment due to its extensive and ongoing validation efforts. The following validation studies, in which FLACS has been compared to experiments, are listed below:
- 1993-1996 50 methane explosion experiments with varying congestion, confinement and ignition location (JIP report 1995)
- 1993-1994 Extensive grid dependency study for explosion modeling (JIP report 1994)
- 1994-1995 Blast propagation in the far field, 10 tests (JIP report 1995)
- 1993-1996 >50 experiments with water deluge Gexcon and BG experiments (JIP report 1996)
- 1994-1997 ~50 BFETS 2 & 3A large-scale explosion tests, variation in ignition point, congestion and more (JIP report 1998)
- 1994-1998 Ventilation studies vs wind measurements at various oil platforms (e.g. Oseberg-C, Beryl-B, Nelson)
- 1997-1999 MOGELEG, Gexcon labscale experiments on the effect of nitrogen and CO2 dilution (JIP report 1998)
- 1996-1999 SMEDIS, EU project on evaluation of dispersion modeling (JIP report 1999)
- 1998-2000 Gexcon 50m3 and BFETS 3B large scale dispersion and explosion tests (JIP report 2001)
- 2001-2004 100 lab-scale (Gexcon)+ 25 large-scale (Sandia) hydrogen experiments (JIP report + ICHS-paper 2005)
- 2002-2004 Validation against Kit Fox, MUST and Prairie Grass atmospheric dispersion tests (Hanna Atm Env 2004)
- 2005-2008 15 large scale LNG dispersion experiments Burro&Coyote, Maplin Sands (Hansen, LPS 2007)
- 2004-2009 Various HySafe hydrogen dispersion and explosion studies (Several papers by Hansen & Middha)
Below are plots from the early validation studies. Due to competition with other softwares in the 1990s, validation efforts in the early years were only made available to sponsors of the FLACS development. In recent years there has been a much more open policy on this, and a number of papers have been made available on validation in the field of atmospheric dispersion / LNG, dust explosions and hydrogen safety. Below is an example comparing simulation and full-scale explosion experiments in the Advantica Spadeadam rig. For the validation work more than 30 pressure sensors are available for each of the more than 50 fullscale experiments.
Below is a plot of early validation studies showing how simulated pressures compared to experimental pressures for a range of different test configurations.
In general, the use of a CFD-tool is no guarantee for accurate or valid predictions, as the quality of the output strongly depends on the input and the models implemented. Since FLACS is utilized for calculations where accuracy is very important, Gexcon considers the validation of FLACS to be a top priority. Considerable efforts have been expended to validate the various models of FLACS.
Many CFD models have calibration parameters, or the possibility to choose from many different turbulence models, which may strongly influence the predictions. Gexcon provides the user with the appropriate methodology to eliminate such inaccurate modifications. The reliability of FLACS simulations is based on the extensive validation efforts and necessary user guidelines.