Dense Particle Entrainment

kmh218 · February 21, 2019, 19:25

Hi all,

I am relatively new to cfd and am trying to model a turbulent compressible particle laden flow through a horizontally oriented cylinder. I am using an Eulerian Eulerian multiphase model (EMP) with the continuous phase being nitrogen (N2) and the dispersed phase being copper (mean diameter of 30 micron). I am using a k-epsilon turbulence model.

The issue I am running into is that gravity is causing the copper to fall to the bottom of the pipe and making the numerics to diverge. The pipe is about a meter long (diameter of 1cm) and the accumulation of copper begins about 0.2m down the pipe. Originally this was expected since copper is about 450x more dense than N2. However after reviewing experimental data of what I am trying to model, it is clear that the copper should be entrained within the N2.

At the inlet to the pipe, the copper and the N2 are moving at the same velocity (well above the velocity required for saltation) and the Reynolds # is about 30000.

I have also ensured that my mesh falls within the requirements for a k-epsilon model where the y+ of all the cells are all greater than 30.

I have employed several different multiphase interaction models to model the interaction between the copper and the N2: drag, lift, turbulent dispersion, virtual mass,particle induced turbulence, etc. and none seem to make a difference in the entrainment of the copper.

I am now considering adding a momentum source to the copper to allow the copper to stay entrained in the flow (assuming that the model is just not able to capture the physics of this flow). However this route seems a bit artificial.

I was wondering if there are any additional physics I could be missing or better yet some overall advice on what usually has to be done to entrain dense particles in a flow?

Any help would be greatly appreciated! Thanks!

fluid23 · February 25, 2019, 13:51

Maybe a stupid question, but did you verify that the density of N2 is correct WRT to the test data?

kmh218 · February 25, 2019, 14:02

The density of N2 in the simulation ranges between 16 and 20 kg/m^3. Even if it is slightly off from the N2 density in the experiments it shouldn't make too much of a difference since density of copper is about 9000 kg/m^3

fluid23 · February 25, 2019, 15:49

Just wanted to make sure you weren't an order of magnitude off or something crazy. That would have a big effect on particle response.

Sorry I can't be more help on this.

kmh218 · February 26, 2019, 10:40

Absolutely, thanks anyway

February 21, 2019, 19:25	Dense Particle Entrainment	#1
kmh218 New Member kmh Join Date: Feb 2019 Posts: 4 Rep Power: 7	Hi all, I am relatively new to cfd and am trying to model a turbulent compressible particle laden flow through a horizontally oriented cylinder. I am using an Eulerian Eulerian multiphase model (EMP) with the continuous phase being nitrogen (N2) and the dispersed phase being copper (mean diameter of 30 micron). I am using a k-epsilon turbulence model. The issue I am running into is that gravity is causing the copper to fall to the bottom of the pipe and making the numerics to diverge. The pipe is about a meter long (diameter of 1cm) and the accumulation of copper begins about 0.2m down the pipe. Originally this was expected since copper is about 450x more dense than N2. However after reviewing experimental data of what I am trying to model, it is clear that the copper should be entrained within the N2. At the inlet to the pipe, the copper and the N2 are moving at the same velocity (well above the velocity required for saltation) and the Reynolds # is about 30000. I have also ensured that my mesh falls within the requirements for a k-epsilon model where the y+ of all the cells are all greater than 30. I have employed several different multiphase interaction models to model the interaction between the copper and the N2: drag, lift, turbulent dispersion, virtual mass,particle induced turbulence, etc. and none seem to make a difference in the entrainment of the copper. I am now considering adding a momentum source to the copper to allow the copper to stay entrained in the flow (assuming that the model is just not able to capture the physics of this flow). However this route seems a bit artificial. I was wondering if there are any additional physics I could be missing or better yet some overall advice on what usually has to be done to entrain dense particles in a flow? Any help would be greatly appreciated! Thanks!

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
UDF for particle interception with pt_termination fortran routine	abcdefgh	CFX	6	October 6, 2019 14:30
Particle tracking error	alchem	OpenFOAM Bugs	5	May 6, 2017 17:30
particles leave domain	Steffen595	CFX	9	March 7, 2016 17:19
injection problem	Mark New	FLUENT	0	August 4, 2013 02:30
DPM UDF particle position using the macro P_POS(p)[i]	dm2747	FLUENT	0	April 17, 2009 02:29

February 25, 2019, 13:51		#2
fluid23 Senior Member Matt Join Date: Aug 2014 Posts: 947 Rep Power: 18	Maybe a stupid question, but did you verify that the density of N2 is correct WRT to the test data?

February 25, 2019, 14:02		#3
kmh218 New Member kmh Join Date: Feb 2019 Posts: 4 Rep Power: 7	The density of N2 in the simulation ranges between 16 and 20 kg/m^3. Even if it is slightly off from the N2 density in the experiments it shouldn't make too much of a difference since density of copper is about 9000 kg/m^3

February 25, 2019, 15:49		#4
fluid23 Senior Member Matt Join Date: Aug 2014 Posts: 947 Rep Power: 18	Just wanted to make sure you weren't an order of magnitude off or something crazy. That would have a big effect on particle response. Sorry I can't be more help on this.

February 26, 2019, 10:40		#5
kmh218 New Member kmh Join Date: Feb 2019 Posts: 4 Rep Power: 7	Absolutely, thanks anyway