DPM Particle Tracking Study
 

Hi, this is my first post on the forum so apologies if I give too much or too little information, I will try to be thorough and clear!

I am investigating the way in which Fluent's DPM model calculates a particle's path through the continuous phase. I have been reading a lot of information online and understand that Fluent calculates a particle's position as it enters and leaves a cell and also calculates its position a number of times through each cell according to the "Step Length Factor".
(http://www.cfd-online.com/Forums/flu...-tracking.html)

What I want to know is whether settings such as the Step Length Factor (SLF) and Length Scale affect the overall path of the particle or whether they act just to smooth out the trajectory? For example if I have a really coarse mesh and a SLF of 5, would Fluent calculate the same position for a particle entering and leaving a cell as it would if the SLF was increased to, say 20 or higher? My results for particle tracking do not change at all regardless of what I specify for the SLF, but that could just be due to my specific setup which is why I'm seeking help from the forum.

If this is the case should the mesh density have a significant impact upon an individual particle's trajectory? Obviously a more refined mesh will give a different solution for the continuous phase which will impact the discrete phase but is that the only way the mesh affects the particles being tracked? Or should the particle path change because of the fact there are more cells in the domain and thus more points at which the particle positions are calculated? (I attached a sketch of a basic mesh to try and illustrate what I mean if it isn't clear.)

My model is setup using one-way coupling for the DPM and the flow regime is laminar. The only forces acting upon the particles are gravity and fluid drag.

Thanks in advance for any help!

Ralph

Hi Ralph,

a late answer is better than none :) The way I see it, you already have the basic understanding of how the DPM works.

The only thing that you need to keep in mind is that the integration time step for the trajectory integration is chosen based on the specified step length factor (or length scale), which definitely affects the computational accuracy. In other words, the trajectory computation is resolved finer with higher step length factor, but if the mesh is dealing with the flow as it should there should no reason for very high factors, close to the default value usually does the job.

Your system does not seem to be very sensitive to the factor, which kind of makes sense for a laminar one-way coupled one, but this is not always the case. Especially for highly trubulent or strained flows, you would be able to see a clear difference when chaning the factor significantly.

But you are correct, the mesh size indirectly influences the DPM in this way, hence one thing to keep an eye on.

Best, Mo