When doing the particle trajectory calculation, the result varied whenever the length scale value was reduced. How could I know if the length scale value is small enough to generate the accurate trajectory?

hi.

any suggestions? Thanks.

Isn't a correct result the result that is independant of grid, length scale ect. ? (a kind of convergence criteria)

If yes, you would have to reduce the length until the result stop changing.

But, what result are you talking about ?

I am looking at the fate of the injected particles. I have set the fate of particles hitting different boundaries as escape, trap. I started the DPM after the flow field solution was converged. The result I am talking about is that the reported numbers of particles escaped or trapped. Whenever I changed the length scale value in the discrete phase modeling panel, the reported numbers of different fate varied.According to the FLUENT user's guide, small length scale will give more accurate result,which means that result is not independent of the length scale. Am I right? I have reduced the length scale to the extent that the calculation of particle trajectory couldn't be completed with the max. number of steps of 1e+09. Still i couldn't have the result stop changing. Hope my question is clear this time. Thanks!

When using CFD, it is easy to focus on numbers - like the fraction of particles trapped or escaped. Everyone likes to have straight forward criteria for making decisions.

However, the most important information you might be getting from your simimulation is why are the particles trapped or why do they escape? And what happens in your simulation when you change the length scale such that it changes the fate of the particle? Do you have a lot of back eddies and recirculation?

Based on my experience with large particles (i.e. typical diameters 0.001 to 0.005m) in large domains (10m by 10m by 30m)your max no. of steps, 1e9, is huge. I normally take about 2500 as max.

Do your results look reasonable? Can you use more particles such that each one represents a smaller fraction of the overall condensed phase?

I'm not sure there is any general parameter that sets the step length for different problems. What is the time step between particle calculations in your case? And don't trust that printed on the screen while tracking particle history - look in the code.

While we are at it - if you really want the "right" numerical result - have you considered whether the particle phase effects the gas phase and whether you need to recalculate the gas phase after a particle phase calculation. And, are the differences that you calculate with a different time step large enough that they are significant compared to other uncertainties in boundary conditions and solving techniques?

Hope this helps.

Dear Walsh,thank you very much for your detailed reply. I think I should follow your suggestions to find out what really happend in my problem. There are a lot of things for me to learn and understand using FLUENT.

[Angelina]

Dear guys,

I have the same problem in my simulations. Do you unterstand what happend in your simulation?

kind regards
Angelina