Coalescence model gives too big drops
 

Hi,

has anybody noticed the same as I have: the droplet coalescense model gives too big droplets.

I am computing sprays in a constant volume bomb and I have noticed that the activation of the coalescence model produces droplets that can easily be three times their original size (when injected from the nozzle) ! I really think that in reality such droplets do not exist. There is also experimental evidence to support the view that such big droplets do not exist in high pressure sprays.

Re: Coalescence model gives too big drops
 

An additional remark is that I am using initial droplet size distribution which might make the situation worse. It seems to be that with a constant initial drop size, the drops produced by the coalescense model are more reasonably sized than with an initial size distribution.

Re: Coalescence model gives too big drops
 

Ossi,

In the paper "Modelling of Drop Interactions in Thick Sprays and a Comparison with Experiments", O'Rourke and Bracco argue that an SMR of 3um produced by the primary atomisation of a 0.3mm nozzle at a back pressure of 11bar increases to some 20um at a downstream location due to coalescence. Is this less of an increase than you have in your results?

Hiroyasu and Kadota also state in their SAE paper "Fuel Droplet Size Distribution in Diesel Combustion Chamber" that the downstream SMD increases with higher gas density due to droplet coalescence. Accordingly, I suppose sprays in high pressure environments might well feature large droplet sizes locally. What is the experimental evidence you are referring to?

Regs.

Lars Ola

Re: Coalescence model gives too big drops
 

I have no doubt that the SMD will increase in high pressure sprays towards the tip of the spray due to coalescence effects. But I doubt very much the existence of single very large drops (>three times the nozzle hole diameter) especially near the nozzle orifice in the very high velocity region. Such drops are produced by the current coalescence model especially if an initial size distribution is used. I apologize for not being more specific in the earlier message.

The proof that such big droplets do not exist comes from many studies. Rantanen, P. et al. "Fuel Injection of a Medium Speed Diesel Engine, Spray Characterization" have used optical system to photograph the spray. They found no single very big droplet in the spray. Additionally, they found also (other studies available too) that the tip of the spray is always very smooth. In CFD-simulations on the other hand, especially if there exists very big droplets, the big droplets are penetrating much further than the rest of the drops making the tip of the spray very different than the experimentally observed spray tip. That is, only a few drops (the big ones) are penetrating ahead of the rest of the spray. This kind of behavior is not found in experiments.

Summa summarum, because the coalescense model by O'Rourke and Bracco may produce too big drops and others have found out this too, e.g. AVL has implemented some limiters into FIRE code that the drops produced by the coalescence model cannot be bigger than say (in their case) half their original size. This kind of a limiter worked well in the simulations I did with FIRE.