Sediment transportation

mdallali · September 27, 2013, 12:02

Hi every one, If I model a flow field with non.cohesive sediment and consider the effect of sediment on flow just as a density change in gravity term of N-S equation, the MODEL IS TWO -PHASE?

ghorrocks · September 30, 2013, 18:29

Sediment obviously is a two phase flow but replacing it with a density change in the NS equations sounds like a single phase approach to me. But I do not understand how modelling sediment with a density change is going to accurately model anything.

MD276 · November 7, 2013, 04:51

Hi every one, I used a sediment flux at bottom boundary for A diffusion /convective Eulerian aproach to sediment transport in free surface channel , but as the channel is closed in inflow and out flow, the concentration is increasing continuously, and pass the value 1 at bottom! physically the concentration should be a value between 0-1 , but as we impose the flux continuously we have a incremental value , so the idea is to fix the value somehow, any idea?

ghorrocks · November 7, 2013, 05:06

How are you modelling this? Are you using an additional variable or a multiphase model?

Volume fractions do sometimes go outside of 0 and 1, this is a sign of poor boundedness in the simulation. It can be caused by unstable numerical schemes, insufficient convergence or incorrect simulation setup.

If you have a flux at the bottom then you have to make sure the flux is acting correctly, as it is easy for the flux to simply add more and more stuff until the VF exceeds 1. You will need to make the flux act on a mass or volume basis, so it asymptotically approaches 1.

MD276 · November 7, 2013, 05:13

it considers N-s equation for the flow field and then i solved the convection-diffusion as for sediment as scalar. the flux is just specified the dc/dy at bottom which is a function of instantaneous wall shear stress

ghorrocks · November 7, 2013, 05:23

Well, your flux appears to have nothing to enforce bounding (ie limiting volume fraction to 0 and 1) so that is your problem. You need to make sure your flux results in a bounded volume fraction.

Why didn't you use CFX's built in multiphase models for this?

MD276 · November 7, 2013, 06:47

i m working on my own code, in fortran, the bottom flux is modeled as a relation for shield Parameter, how could i bound it?

ghorrocks · November 7, 2013, 16:17

The first question to answer is why can't you use the existing multiphase models in CFX for this?

MD276 · November 11, 2013, 04:53

I m not familar with CFX model , is it in Fluent?

ghorrocks · November 11, 2013, 05:04

I have no idea (but it almost certainly is). This is the CFX forum. Try that question on the Fluent forum.

But if you define an unbounded boundary condition in Fluent then you will get the same result.

sharif88 · February 26, 2015, 16:54

Hi Glen

Can you please explain why bed erosion cannot be modeled in CFX or FLUENT?
I am working on a problem where basically I want to simulate bed erosion using water, I do not understand why thus is different than the case when particles are in the flow?
I mean the important things are drag and other forces which anyways are solved for,

Thanks

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Sediment Scour model	mshotm2000	FLOW-3D	1	February 6, 2014 20:30
Sediment Scour model	mshotm2000	FLOW-3D	3	March 16, 2011 07:06
Sediment Scour model	mshotm2000	FLOW-3D	2	February 4, 2011 03:54
How to calculate suspended sediment concentration ?	fang	FLOW-3D	4	January 20, 2010 11:28
Re: Can I simulate sediment transport with FLUENT	li li	FLUENT	0	October 21, 2003 20:42

September 27, 2013, 12:02		#21
mdallali New Member marjan Join Date: Aug 2013 Posts: 11 Rep Power: 12	Hi every one, If I model a flow field with non.cohesive sediment and consider the effect of sediment on flow just as a density change in gravity term of N-S equation, the MODEL IS TWO -PHASE?

September 30, 2013, 18:29		#22
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	Sediment obviously is a two phase flow but replacing it with a density change in the NS equations sounds like a single phase approach to me. But I do not understand how modelling sediment with a density change is going to accurately model anything.

November 7, 2013, 04:51		#23
MD276 New Member ronak Join Date: Jul 2011 Posts: 6 Rep Power: 14	Hi every one, I used a sediment flux at bottom boundary for A diffusion /convective Eulerian aproach to sediment transport in free surface channel , but as the channel is closed in inflow and out flow, the concentration is increasing continuously, and pass the value 1 at bottom! physically the concentration should be a value between 0-1 , but as we impose the flux continuously we have a incremental value , so the idea is to fix the value somehow, any idea?

November 7, 2013, 05:06		#24
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	How are you modelling this? Are you using an additional variable or a multiphase model? Volume fractions do sometimes go outside of 0 and 1, this is a sign of poor boundedness in the simulation. It can be caused by unstable numerical schemes, insufficient convergence or incorrect simulation setup. If you have a flux at the bottom then you have to make sure the flux is acting correctly, as it is easy for the flux to simply add more and more stuff until the VF exceeds 1. You will need to make the flux act on a mass or volume basis, so it asymptotically approaches 1.

November 7, 2013, 05:13		#25
MD276 New Member ronak Join Date: Jul 2011 Posts: 6 Rep Power: 14	it considers N-s equation for the flow field and then i solved the convection-diffusion as for sediment as scalar. the flux is just specified the dc/dy at bottom which is a function of instantaneous wall shear stress

November 7, 2013, 05:23		#26
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	Well, your flux appears to have nothing to enforce bounding (ie limiting volume fraction to 0 and 1) so that is your problem. You need to make sure your flux results in a bounded volume fraction. Why didn't you use CFX's built in multiphase models for this?

November 7, 2013, 06:47		#27
MD276 New Member ronak Join Date: Jul 2011 Posts: 6 Rep Power: 14	i m working on my own code, in fortran, the bottom flux is modeled as a relation for shield Parameter, how could i bound it?

November 7, 2013, 16:17		#28
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	The first question to answer is why can't you use the existing multiphase models in CFX for this?

November 11, 2013, 04:53		#29
MD276 New Member ronak Join Date: Jul 2011 Posts: 6 Rep Power: 14	I m not familar with CFX model , is it in Fluent?

November 11, 2013, 05:04		#30
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	I have no idea (but it almost certainly is). This is the CFX forum. Try that question on the Fluent forum. But if you define an unbounded boundary condition in Fluent then you will get the same result.

February 26, 2015, 16:54		#31
sharif88 New Member Majid Join Date: Jan 2014 Location: Canada Posts: 23 Rep Power: 12	Hi Glen Can you please explain why bed erosion cannot be modeled in CFX or FLUENT? I am working on a problem where basically I want to simulate bed erosion using water, I do not understand why thus is different than the case when particles are in the flow? I mean the important things are drag and other forces which anyways are solved for, Thanks