Boundary condition for UDS

October 30, 2007, 07:15

Hi:

Does anyone know how to difine zero normal gradient boundary condition for UDS? The boundary condition for the fluid is pressure outlet. The buildin options for UDS boundary are specific value and specific flux which are different from my case.

Any suggestions or examples are very appreciated. Thank you very much.

pmechz · February 7, 2016, 15:28

would you please help me?
I'm using UDS for simulation of mixing of two fluids (with concentration of 1 at one inlet and concentration of 0 at another inlet) but I dont know how can I specify uds in pressure outlet boundary condition?

pakk · February 8, 2016, 09:02

Open the boundary condition and click 'UDS'. Where else?

pmechz · February 9, 2016, 16:43

Quote:

Originally Posted by pakk

Open the boundary condition and click 'UDS'. Where else?

i mean that which amount is appropriate for uds in pressure outlet ?

pakk · February 10, 2016, 04:09

How did you find out which pressure is appropriate for the pressure outlet?

(What I mean to say: use the condition that your problem requires... If your problem requires a value of 0, use a value of zero. If your problem requires a flux of 0, use a flux of 0. We can not say that...)

pmechz · February 11, 2016, 09:05

Quote:

Originally Posted by pakk

How did you find out which pressure is appropriate for the pressure outlet?

(What I mean to say: use the condition that your problem requires... If your problem requires a value of 0, use a value of zero. If your problem requires a flux of 0, use a flux of 0. We can not say that...)

my problem is the article that i am modeling just shows pressure outlet for outlet boundary condition. and said nothing about amount of user scalar.
i sent you the pic of my question location

pakk · February 11, 2016, 09:14

Quote:

Originally Posted by pmechz

my problem is the article that i am modeling just shows pressure outlet for outlet boundary condition. and said nothing about amount of user scalar.
i sent you the pic of my question location

Your pictures shows that you are looking at the correct location. But I can not guess which boundary condition was used in the article that you are using. There are two 'standard' conditions:

1. Assume there is no scalar at this boundary, as if the boundary is completely 'clean'. In this case, set value to zero.
2. Assume there is equilibrium at this boundary, so there is no net flow of scalar across the boundary. In this case, set the flux to zero. (change "specified value" into "specified flux".)

But it is just as physical to set the value to "1.3712e-3" or the flux to "4.293e+2" or whatever values. It is just representing a different problem.

pmechz · February 11, 2016, 09:37

Quote:

Originally Posted by pakk

Your pictures shows that you are looking at the correct location. But I can not guess which boundary condition was used in the article that you are using. There are two 'standard' conditions:

1. Assume there is no scalar at this boundary, as if the boundary is completely 'clean'. In this case, set value to zero.
2. Assume there is equilibrium at this boundary, so there is no net flow of scalar across the boundary. In this case, set the flux to zero. (change "specified value" into "specified flux".)

But it is just as physical to set the value to "1.3712e-3" or the flux to "4.293e+2" or whatever values. It is just representing a different problem.

sorry. may i have your private email to sent my article to you and help me more, if it's possible?
thanx

Tharanga · April 26, 2018, 23:05

Quote:

Originally Posted by pakk

Your pictures shows that you are looking at the correct location. But I can not guess which boundary condition was used in the article that you are using. There are two 'standard' conditions:

1. Assume there is no scalar at this boundary, as if the boundary is completely 'clean'. In this case, set value to zero.
2. Assume there is equilibrium at this boundary, so there is no net flow of scalar across the boundary. In this case, set the flux to zero. (change "specified value" into "specified flux".)

But it is just as physical to set the value to "1.3712e-3" or the flux to "4.293e+2" or whatever values. It is just representing a different problem.

Hi , Could you explain further more about the net flux zero condition. When you think about the UDS flux equation, it does not say anything about the net flux. Please help. Thanks

c_023 · January 23, 2020, 11:29

Does anyone have experience on what kind of BC to use for a UDS where the scalar is mainly transported with convection? In my case, the flow has a strong convection where (Diffusion << Convection) for the Scalar. Therefore, any value set at the Outlet as BC should not be affecting the UDS, otherwise this would be non-physical (similar to Setting the temperature at an outlet BC, which is only used when backflow occurs).
I think that the "value" at the outlet BC for the UDS should, if using an upwind scheme for the UDS, not be influencing the solution as long as no backflow occurs, because a true upwind scheme would only use the Information of upstream cells. However, I am not sure if my thoughts on this are correct.
Has anyone got experience with this or any thoughts on the topic?

Kind regards!

pakk · February 4, 2020, 06:33

You don't know anything about the value of the UDS, but you know something about the value of the gradient of the UDS (in the direction of the normal). If your outlet is far enough away from the UDS sources, you expect the value of the UDS not to change that much anymore near the outlet, so the gradient should be zero.

vinerm · February 4, 2020, 07:52

That's exactly the definition of a developed flow profile, no changes along the flow direction, implying gradient along the flow direction is 0. This is the condition applied at the outlet for energy equation as long as the flow really is outgoing.

It is different from what you set for energy equation. For energy equation, usually a reverse flow temperature is provided. This is used when the flow reversal takes place. It can affect the solution if a lot of flow reversal happens or is expected. In physical perspective, if outlet is connected to another domain, which is not being modeled, then you should provide the value that could exist in the adjacent domain, such as, hot water from the duct entering a vessel with the water being maintained at 350 K. Then 350 K as backflow temperature is a good idea. However, if the duct is open to the atmosphere, then 350 K or any other temperature is a bad number since you do not expect water to return; it's the air that will return. But to avoid modeling two-phase flow, extend the domain to such an extent that the boundary condition at the outlet has no effect on the domain of concern. It still has effects near the outlet but that is far away from region of concern.

Sometimes, you cannot do that because extension is unrealistic. Then you have to think the situation in physical form. Whatever this scalar represents, what happens to it at the exit is important. May be it diffuses into the atmosphere, then the value at the outlet will always be 0 because atmosphere acts like a sink. And that's realistic in many cases. If the scalar does not diffuse as it comes out of the domain and the outlet is far from the domain of concern, then zero gradient is a good boundary condition. But if neither there is diffusion nor is the outlet far from some important region, then, you have to extend the domain realistically, i.e., you have to model the domain after the outlet as well.

c_023 · February 4, 2020, 11:52

Thanks, this makes sense.
I assume that the "Specified Flux" option in the UDS tab of an outlet BC refers to the UDS Gradient normal to the outlet ? So if I have an outlet far away from my region of interest, and I do not know much about my outlet but I know it will be a developed flow, I set "Specified Flux = 0".
Thanks for the replies.

October 30, 2007, 07:15	Boundary condition for UDS	#1
Tong Wu Guest Posts: n/a	Hi: Does anyone know how to difine zero normal gradient boundary condition for UDS? The boundary condition for the fluid is pressure outlet. The buildin options for UDS boundary are specific value and specific flux which are different from my case. Any suggestions or examples are very appreciated. Thank you very much.

February 4, 2020, 06:33		#11
pakk Senior Member Join Date: Nov 2013 Posts: 1,965 Rep Power: 26	You don't know anything about the value of the UDS, but you know something about the value of the gradient of the UDS (in the direction of the normal). If your outlet is far enough away from the UDS sources, you expect the value of the UDS not to change that much anymore near the outlet, so the gradient should be zero. c_023 likes this.

February 4, 2020, 07:52	Developed Flow	#12
vinerm Senior Member Vinerm Join Date: Jun 2009 Location: Nederland Posts: 2,946 Blog Entries: 1 Rep Power: 35	That's exactly the definition of a developed flow profile, no changes along the flow direction, implying gradient along the flow direction is 0. This is the condition applied at the outlet for energy equation as long as the flow really is outgoing. It is different from what you set for energy equation. For energy equation, usually a reverse flow temperature is provided. This is used when the flow reversal takes place. It can affect the solution if a lot of flow reversal happens or is expected. In physical perspective, if outlet is connected to another domain, which is not being modeled, then you should provide the value that could exist in the adjacent domain, such as, hot water from the duct entering a vessel with the water being maintained at 350 K. Then 350 K as backflow temperature is a good idea. However, if the duct is open to the atmosphere, then 350 K or any other temperature is a bad number since you do not expect water to return; it's the air that will return. But to avoid modeling two-phase flow, extend the domain to such an extent that the boundary condition at the outlet has no effect on the domain of concern. It still has effects near the outlet but that is far away from region of concern. Sometimes, you cannot do that because extension is unrealistic. Then you have to think the situation in physical form. Whatever this scalar represents, what happens to it at the exit is important. May be it diffuses into the atmosphere, then the value at the outlet will always be 0 because atmosphere acts like a sink. And that's realistic in many cases. If the scalar does not diffuse as it comes out of the domain and the outlet is far from the domain of concern, then zero gradient is a good boundary condition. But if neither there is diffusion nor is the outlet far from some important region, then, you have to extend the domain realistically, i.e., you have to model the domain after the outlet as well. c_023 likes this. __________________ Regards, Vinerm PM to be used if and only if you do not want something to be shared publicly. PM is considered to be of the least priority.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Domain Imbalance	HMR	CFX	5	October 10, 2016 05:57
Can anyone give me some hint on how to make traction free boundary condition?	poplar	OpenFOAM	3	January 14, 2015 02:37
Boundary Conditions	Thomas P. Abraham	Main CFD Forum	20	July 7, 2013 05:05
Setting outlet Pressure boundary condition using CAFFA code	Mukund Pondkule	Main CFD Forum	0	March 16, 2011 03:23
How to set boundary condition in Fluent for the fo	Peiyong	FLUENT	1	November 10, 2006 11:44

February 7, 2016, 15:28		#2
pmechz New Member Join Date: Feb 2016 Posts: 26 Rep Power: 10	would you please help me? I'm using UDS for simulation of mixing of two fluids (with concentration of 1 at one inlet and concentration of 0 at another inlet) but I dont know how can I specify uds in pressure outlet boundary condition?

February 8, 2016, 09:02		#3
pakk Senior Member Join Date: Nov 2013 Posts: 1,965 Rep Power: 26	Open the boundary condition and click 'UDS'. Where else?

February 10, 2016, 04:09		#5
pakk Senior Member Join Date: Nov 2013 Posts: 1,965 Rep Power: 26	How did you find out which pressure is appropriate for the pressure outlet? (What I mean to say: use the condition that your problem requires... If your problem requires a value of 0, use a value of zero. If your problem requires a flux of 0, use a flux of 0. We can not say that...)

January 23, 2020, 11:29		#10
c_023 New Member Join Date: Dec 2019 Posts: 19 Rep Power: 6	Does anyone have experience on what kind of BC to use for a UDS where the scalar is mainly transported with convection? In my case, the flow has a strong convection where (Diffusion << Convection) for the Scalar. Therefore, any value set at the Outlet as BC should not be affecting the UDS, otherwise this would be non-physical (similar to Setting the temperature at an outlet BC, which is only used when backflow occurs). I think that the "value" at the outlet BC for the UDS should, if using an upwind scheme for the UDS, not be influencing the solution as long as no backflow occurs, because a true upwind scheme would only use the Information of upstream cells. However, I am not sure if my thoughts on this are correct. Has anyone got experience with this or any thoughts on the topic? Kind regards!

February 4, 2020, 11:52		#13
c_023 New Member Join Date: Dec 2019 Posts: 19 Rep Power: 6	Thanks, this makes sense. I assume that the "Specified Flux" option in the UDS tab of an outlet BC refers to the UDS Gradient normal to the outlet ? So if I have an outlet far away from my region of interest, and I do not know much about my outlet but I know it will be a developed flow, I set "Specified Flux = 0". Thanks for the replies.