Active Explosion Barrier: Refinement of CFD Model and Application of Real Development Panels

After review of the successful ACARP funded active barrier demonstration trials at CSIR Klopperbos in 2012, two new physical models were introduced into the CFD active barrier simulations - a coal dust wetting model and a spray breakup model. Both of these models were shown to have a significant effect. The coal dust wetting removed as much as half of the dust from the model in the region of the water sprays, and the spray breakup model predicted a typical droplet size that was smaller than the nozzle manufacturer's specifications by a factor of two. Further refinements were made to the CFD code in the turbulent dispersion algorithm, combustion partially stirred reactor model, coal dust size, and droplet drag model.

The penetration of the water sprays into the air flow was recognised as a point of significance. To suppress an explosion it was found to be vital that the water was well mixed with the coal dust in the combustion region.

With the modifications in place, excellent agreement between the CFD model and the 2012 test results was attained. The CFD model predicted the correct deflagration speed and duration for the base case seminar explosion scenario, and predicted marginal suppression with 60 litres of water delivered in 500 milliseconds.

The CFD model was subsequently used to investigate possible barrier configurations to effect suppression in a 5 m by 3 m roadway. Smaller than the 2012 Kloppersbos Active Barrier Device, the roadway barrier utilised 120 litres of water delivered in 250 milliseconds (scaling based on section areas). Four possible barrier configurations were considered. The most encouraging results were from configurations involving just one or two nozzles located on units on the side of the roadway, the intent being to utilise the larger momentum of fewer spray plumes to carry the water to the centre of the roadway. With these configurations strong suppression was predicted.

Overall the results of the work presented here are very encouraging. Not only has the model been successfully revalidated with the 2012 Kloppersbos test data, but preliminary investigations into a barrier for a 5 m by 3 m roadway show promising results, with an indication that an effective barrier may be constructed from a pair of 60 litre units located opposite each other on the sides of a roadway.