Lagrangian Particle Transport in the Gas Phase

[Alex V]

Hello!

My question pertains to the subject of Lagrangian particle transport. I want to ensure my understanding of the FDS calculations is accurate.

I conducted a study where particles were introduced using a horizontal dot array. I placed 2500 dots within a 1 m2 area. The particles (5 µm spherical shape with a density of 2000 kg/m3) were simply released into the air without any initial velocity.
In general, the settling velocity of particles (which is defined as the vertically travelled distance over time, roughly 1 m / 900 s) is not too far (see below) from the analytical values obtained from the Stokes Law. My simulations explore scenarios where the number of particles (N_PARTICLES) from each dot falls within the range of 1 to 1000 (in increments of 10).
Depending on the cell size and the number of particles, we observe variations in settling velocity. For instance, with a coarse mesh (0.4 or 0.5 m), the difference between simulated settling velocities falls within the 2-3% range across all particle number variations. Conversely, with fine meshes (0.025 m), the difference in velocities exceeds 16% when comparing models with 1 and 1000 particles per dot (the velocity for the high particle number case is greater). Therefore, with fine meshes, introducing more particles results in higher settling velocities.
For a clearer understanding, here's a chart displaying the simulated data (positive relative error values indicate that the modelled settling velocity is higher than the analytically obtained value):

Am I correct in understanding that this variance in velocity values between coarse and fine meshes, as well as between scenarios with low and high numbers of particles, is due to the cell volume considered in the calculation of the "bulk" force transferred from particles to the gas and the total bulk mass of particles itself?

This implies:
1) Fewer cells result in fewer calculation steps needed to cover the distance travelled by particles, leading to more generalized results.
2) Introducing more particles increases the total bulk mass dragging the gas, subsequently increasing the particle acceleration.

Please, kindly correct any inaccuracies in my interpretation.

BR

[dr_jfloyd]

The 0.025 m results are 2).  At 2,500,000 milllion particles in 1 m2 you have ~8 % of the grid cell area covered by the projected area of the particles. If all you did in your inputs was change the MESH, then some of 0.4 and 0.5 may be error due to timestep. All else being equal an 0.5 m grid will have a much larger timestep then 0.025 m. If you did not set NOISE=F on MISC, then some of your error may also be due to the initial noise which scales with grid size.