Learning Forum

[aitor.amatriain]

I am performing DPM simulations, and Schmehl model is the most accurate one for my application. However, I am noticing that the result depends on the number of breakup parcels (sizes decrease as the number of breakup parcels increases), and I do not fully understand the reasoning behind the results.

As far as I am concerned, If the number of breakup parcels is set to one, that means that breakup occurs once in each particle. I imagine that this number has to be high enough in order to capture all breakup processes, am I right? In that case, why not set always a big number (25 -50)? (I imagine that the number of breakup processes is an upper bound)

Thank you,

Aitor

[Amine Ben Hadj Ali]

Default behavior: Parent particle won't really break into child droplet as mentioned in the manual.

if n is larger than 1 and depending on the regime, the parent will lose mass might still be "alive" or will disappear. The diameter of the child droplets underlies a normal distribution.

[aitor.amatriain]

If the number of breakup parcels is 25, then the parent droplet breaks 25 times into child droplets? Or this is an upper bound?
From my simulations, I see that if the number of breakup parcels is increased, then the d_50 of the resulting particle size distribution is lower. However, there is a kind of asymptotic value of the number of breakup parcels (25 in my case) from which the increase of this number does not modify the results.
In view of the above, what is the recommended approach? In my case, the value of 25 is the adequate? If I set the number of breakup parcels equal to 1, then I see practically no breakup (and this is not correct, based on experimental results)

[Rob]

I'll leave the theory for If you plot the d_50 with number of break up parcels what are you seeing? We do see mesh and temporal resolution changing the continuous phase field and number of (injected) particles the DPM results so it may be another parameter that needs careful thought.

[aitor.amatriain]

I see that d_50 increases as the number of parcels increases, but the curve approaches an asymptotic value around the number of parcels equal to 25. My particular problem is "special" as I am including solidification, but I would say that this behaviour should be present in all cases.