How to model pressure drop without modeling porous media

Sanyo · May 10, 2012, 08:08

Dear all,

Greetings to everybody.

I have come across a very strange situation. I am trying to simulate a 10mm perforated plate having air nozzles of height 150mm. The pressure drop across the plate & nozzles is 250mmWC. The high pressure drop is due to flow has to pass through very small holes (25mm perforation of plate & 10mm hole of air nozzle). Also it has change its direction at 90 Deg while leaving the air nozzle.

I tried to simulate the pressure drop using porous media assumption and defining the loss coefficient. This shows the uniform distribution of the flow across the plane above air nozzles. However velocity test at site shows the central band having higher velocity than the side bands. (60% higher)

When I ran the model with very low value of loss coefficient, I got similar picture of velocity as observed on field. High resistance value ought to distribute the flow uniformly. I cant model all holes on perforated plate & air nozzles. It would make my model impossible to solve.

Does anyone have any trick to overcome this situation?

I would appreciate your help. Thanks in advance.

Regards,

--Sanyo

ghorrocks · May 11, 2012, 00:02

Was the overall pressure drop correct when you modelled it with the correct pressure loss coefficient?

Sanyo · May 11, 2012, 00:38

Dear Glenn,

Thank you very much for the reply.

Yes, the overall pressure drop was correct. I have pressure inlet & pressure outlet boundary condition. So only if correct pressure drop is achieved, I get correct mass flow rate. I had tried with mass flow inlet & pressure outlet, but we find that tweaking of loss coefficient is easier with Pressure inlet & outlet condition. And it also maintains pressure values at different locations downstream.

Thanks & regards,

--Sanyo

ghorrocks · May 11, 2012, 07:12

OK, so if you are using pressure boundaries, let me ask it the other way around - so for the defined pressure drop, is the mass flow correct?

Sanyo · May 12, 2012, 04:53

Hello Glenn,

Yes, the mass flow was correct.

Thanks & regards,

--Sanyo

ghorrocks · May 13, 2012, 06:57

Something does not add up. What you are looking at is the ratio of the resistance associated with the porous media versus other resistances in the flow. If this ratio is low then there is an opportunity for the flow in the porous media to be affected by the other resistances (for instance boundary layers and developed flow will be apparent). If the ratio is high then the porous media resistance dominates all others and the flow across the porous media will be even.

You are going to have to investigate from there, or provide further details for us to help you more.

Sanyo · May 14, 2012, 08:14

Hi Glenn,

It is not possible for me to estimate the ratio you are mentioning. I am glad to supply further details. What exactly would you need to know. I would detail you as much as possible.

Its a AFBC (Atmospheric Fluidized Bed Combusting) Boiler. FD Air flow enters into a large box called windbox. Above windbox has a perforated plate (bedplate) having air nozzles. These air nozzles have small holes to let the air come into furnace to fluidize the sand particles. Combustion occurs in fluidized bed & flue gases pass to superheater tubes, economizer, APH, ESP & then chimney. My domain limits are FD air duct to ESP inlet.

As I stated earlier, my guess is that pressure drop could be because of the flow passing through tiny holes & changing its direction while entering into the furnace.

The only method I find to model the pressure drop is using the sink term for momentum.

I would appreciate your help.

Thanks & Regards,

--Sanyo

ghorrocks · May 14, 2012, 18:37

Sure you can estimate resistances. Just look at the pressure drops through your system. That tells you what the relative resistances are through the system.

But it is starting to sound like you have missed an important piece of physics in your simulation if you are getting the wrong flow profile. Something is making the flow uneven. You are going to have to look into your system in detail and work out what it is and put it your model.

swati_mohanty · May 25, 2016, 07:19

Have enabled "physical velocity" instead of "superficial velocity" ?

Sanyo · May 27, 2016, 10:50

Hi swati,

Its a long time & I dont recall much. But enabling true velocity didnt help much. As soon as loss coefficient is given the flow becomes uniform which was not seen at site during testing. Somehow we were not able to capture the physics.

May 10, 2012, 08:08	How to model pressure drop without modeling porous media	#1
Sanyo Member Sanyo Join Date: Apr 2009 Location: India Posts: 62 Rep Power: 17	Dear all, Greetings to everybody. I have come across a very strange situation. I am trying to simulate a 10mm perforated plate having air nozzles of height 150mm. The pressure drop across the plate & nozzles is 250mmWC. The high pressure drop is due to flow has to pass through very small holes (25mm perforation of plate & 10mm hole of air nozzle). Also it has change its direction at 90 Deg while leaving the air nozzle. I tried to simulate the pressure drop using porous media assumption and defining the loss coefficient. This shows the uniform distribution of the flow across the plane above air nozzles. However velocity test at site shows the central band having higher velocity than the side bands. (60% higher) When I ran the model with very low value of loss coefficient, I got similar picture of velocity as observed on field. High resistance value ought to distribute the flow uniformly. I cant model all holes on perforated plate & air nozzles. It would make my model impossible to solve. Does anyone have any trick to overcome this situation? I would appreciate your help. Thanks in advance. Regards, --Sanyo

May 14, 2012, 08:14		#7
Sanyo Member Sanyo Join Date: Apr 2009 Location: India Posts: 62 Rep Power: 17	Hi Glenn, It is not possible for me to estimate the ratio you are mentioning. I am glad to supply further details. What exactly would you need to know. I would detail you as much as possible. Its a AFBC (Atmospheric Fluidized Bed Combusting) Boiler. FD Air flow enters into a large box called windbox. Above windbox has a perforated plate (bedplate) having air nozzles. These air nozzles have small holes to let the air come into furnace to fluidize the sand particles. Combustion occurs in fluidized bed & flue gases pass to superheater tubes, economizer, APH, ESP & then chimney. My domain limits are FD air duct to ESP inlet. As I stated earlier, my guess is that pressure drop could be because of the flow passing through tiny holes & changing its direction while entering into the furnace. The only method I find to model the pressure drop is using the sink term for momentum. I would appreciate your help. Thanks & Regards, --Sanyo Kamal Kosta likes this.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Porous media model in OpenFOAM 14	gac	OpenFOAM Running, Solving & CFD	21	August 5, 2013 17:49
CFX Model for dispersion of gas through a porous media filled with Oil	deepmetal	CFX	13	November 27, 2010 20:34
Discrete phase modeling on porous media	magnounibo	FLUENT	0	April 9, 2009 08:18
pressure drop: radiator or porous jump	luigi	FLUENT	0	August 19, 2008 14:37
Hydrostatic pressure in 2-phase flow modeling (long)	DS & HB	Main CFD Forum	0	January 8, 2000 15:00

May 11, 2012, 00:02		#2
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	Was the overall pressure drop correct when you modelled it with the correct pressure loss coefficient?

May 11, 2012, 00:38		#3
Sanyo Member Sanyo Join Date: Apr 2009 Location: India Posts: 62 Rep Power: 17	Dear Glenn, Thank you very much for the reply. Yes, the overall pressure drop was correct. I have pressure inlet & pressure outlet boundary condition. So only if correct pressure drop is achieved, I get correct mass flow rate. I had tried with mass flow inlet & pressure outlet, but we find that tweaking of loss coefficient is easier with Pressure inlet & outlet condition. And it also maintains pressure values at different locations downstream. Thanks & regards, --Sanyo

May 11, 2012, 07:12		#4
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	OK, so if you are using pressure boundaries, let me ask it the other way around - so for the defined pressure drop, is the mass flow correct?

May 12, 2012, 04:53		#5
Sanyo Member Sanyo Join Date: Apr 2009 Location: India Posts: 62 Rep Power: 17	Hello Glenn, Yes, the mass flow was correct. Thanks & regards, --Sanyo

May 13, 2012, 06:57		#6
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	Something does not add up. What you are looking at is the ratio of the resistance associated with the porous media versus other resistances in the flow. If this ratio is low then there is an opportunity for the flow in the porous media to be affected by the other resistances (for instance boundary layers and developed flow will be apparent). If the ratio is high then the porous media resistance dominates all others and the flow across the porous media will be even. You are going to have to investigate from there, or provide further details for us to help you more.

May 14, 2012, 18:37		#8
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	Sure you can estimate resistances. Just look at the pressure drops through your system. That tells you what the relative resistances are through the system. But it is starting to sound like you have missed an important piece of physics in your simulation if you are getting the wrong flow profile. Something is making the flow uneven. You are going to have to look into your system in detail and work out what it is and put it your model.

May 25, 2016, 07:19		#9
swati_mohanty New Member sm Join Date: Jun 2009 Posts: 15 Rep Power: 16	Have enabled "physical velocity" instead of "superficial velocity" ?

May 27, 2016, 10:50		#10
Sanyo Member Sanyo Join Date: Apr 2009 Location: India Posts: 62 Rep Power: 17	Hi swati, Its a long time & I dont recall much. But enabling true velocity didnt help much. As soon as loss coefficient is given the flow becomes uniform which was not seen at site during testing. Somehow we were not able to capture the physics.