[Weiqiang Liu]

Hi all,

I have a question about how fluent set adiabatic wall condition.

I have a straight channel without solid part. I want to set the wall adiabatic to the environment. Namely, there is no heat loss from the computation domain to the environment.

In the thermal condition panel of the wall, there are some options like : heat flux, constant temperature, convection, radiation, mix.

my question is if I set heat flux or convection coefficient to 0, how does the heat produced on the wall transfer to the inner gas phase. As far as I am concerned, zero heat flux on the wall means there is no heat transfer on the nodes of the wall. Therefore there is no way, the heat can be transported from the wall to the inner gas mixture.

Am I misunderstanding something? Can anybody give me some hints?

thanks very much

weiqiang

[LuckyTran]

You can use either a heat flux of 0 or a convection coefficient of 0 (but a heat flux of 0 would be slightly better for numerical reasons).


And yes no heat flux means there's no heat flux which means there's no heat transfer because there's no heat flux. There is no wall, so there is no heat produced on the wall. It is meaningless to talk about what heat is being generated in another universe. But in your simulation, you have perfect control over the heat flux. It matters not where this heat comes from, you can set it to any mythical value you like.

[Weiqiang Liu]

Originally Posted by LuckyTran

You can use either a heat flux of 0 or a convection coefficient of 0 (but a heat flux of 0 would be slightly better for numerical reasons).


And yes no heat flux means there's no heat flux which means there's no heat transfer because there's no heat flux. There is no wall, so there is no heat produced on the wall. It is meaningless to talk about what heat is being generated in another universe. But in your simulation, you have perfect control over the heat flux. It matters not where this heat comes from, you can set it to any mythical value you like.

do you mean my model is wrong? To simulate wall surface reaction, I have to draw the solid mesh?

[MKuhn]

heat flux of 0 means there is no heat flux between your wall and your environment (where is no mesh), not that there is no heat transfer between your fluid and your solid wall. If you are interested in the heat transfer between fluid and wall, the best way is to model the solid mesh as well. Another possiblity is to use shell conduction for the wall instead of a solid mesh, but I'm not sure about the accuracy. Maybe LuckyTran knows more :-)

[LuckyTran]

You still can have surface reactions. You don't need a solid mesh for this.

[Weiqiang Liu]

Originally Posted by MKuhn

heat flux of 0 means there is no heat flux between your wall and your environment (where is no mesh), not that there is no heat transfer between your fluid and your solid wall. If you are interested in the heat transfer between fluid and wall, the best way is to model the solid mesh as well. Another possiblity is to use shell conduction for the wall instead of a solid mesh, but I'm not sure about the accuracy. Maybe LuckyTran knows more :-)

so I just draw a fluid domain and set the upper boundary as adiabatic wall. I can have surface reactions on the wall while it's not enough for me to model the heat transfer between the wall and fluid?

[Weiqiang Liu]

Originally Posted by LuckyTran

You still can have surface reactions. You don't need a solid mesh for this.

so I can model both surface reactions and heat transfer between wall and fluid without any solid mesh?

[LuckyTran]

Originally Posted by Weiqiang Liu

so I can model both surface reactions and heat transfer between wall and fluid without any solid mesh?

You can always choose not to have a solid mesh. There can be an arbitrary heat flux (any number from negative infinity, to positive infinity, or 0) from the wall to the fluid, independent of whatever happens in the fluid domain. You can have surface reactions, nuclear reactions, quantum reactions, unicorns, whatever. The fluid also does not care the origin of this heat. It can be magically pumped in/out from another universe. Whatever. You can always terminate your domain somewhere when it's not important to your model. For example, you don't need a solid mesh of my bedframe in your simulation. I usually also don't (actually I never) mesh the andromeda galaxy in any of my simulations.

You need a solid mesh when you want to model the heat conduction process inside the wall. Or you have heat generation inside the wall from some other magical process.

Surface reactions, don't happen inside the solid. They happen on the surface which can be modeled by.... surface reactions! Voila! These have nothing to do with the interior of the solid.


I can't tell from your comments whether or not you need to do a solid mesh or not for your entire problem. But based on surface reactions alone.. no you don't need it (yet). But who knows... Maybe you do want to study the interaction your domain with the andromeda galaxy... You haven't told us.

[Weiqiang Liu]

Originally Posted by LuckyTran

You can always choose not to have a solid mesh. There can be an arbitrary heat flux (any number from negative infinity, to positive infinity, or 0) from the wall to the fluid, independent of whatever happens in the fluid domain. You can have surface reactions, nuclear reactions, quantum reactions, unicorns, whatever. The fluid also does not care the origin of this heat. It can be magically pumped in/out from another universe. Whatever. You can always terminate your domain somewhere when it's not important to your model. For example, you don't need a solid mesh of my bedframe in your simulation. I usually also don't (actually I never) mesh the andromeda galaxy in any of my simulations.

You need a solid mesh when you want to model the heat conduction process inside the wall. Or you have heat generation inside the wall from some other magical process.

Surface reactions, don't happen inside the solid. They happen on the surface which can be modeled by.... surface reactions! Voila! These have nothing to do with the interior of the solid.


I can't tell from your comments whether or not you need to do a solid mesh or not for your entire problem. But based on surface reactions alone.. no you don't need it (yet). But who knows... Maybe you do want to study the interaction your domain with the andromeda galaxy... You haven't told us.

I don't need to know heat conduction in solid domain. So I don't need solid mesh. In the thread I posted before, I though I got the same results with literature. However, the reaction zone moves very slowly out of the channel with iterations going on. Namely, ignition fails. The only difference between the ignited and unable-to-ignited case is thermodynamic and transport. one is from fluent database and the other is from GRI 3.0.

you might ask why I just use default value in fluent database. the reason is fluent database valued don't yield very reasonable results. like the temperature is too high or the surface coverage of species is not consistent with literature.

after a lot of efforts and failures, I started to think is it because my boundary condition is not right. I don't have solid domain in my case and the wall is set as adiabatic. I am wondering if the heat flux is 0. How can the heat generated on the surface be transported to inner gas phase?

[MKuhn]

How is the heat generated in your wall? In Fluent you have the possiblity to define a Heat Generation Rate in W/m³ for the wall with a defined Wall Thickness. This is also possible with an adiabatic wall where the heat flux is 0 to the andromeda galaxy.

[Weiqiang Liu]

Originally Posted by MKuhn

How is the heat generated in your wall? In Fluent you have the possiblity to define a Heat Generation Rate in W/m³ for the wall with a defined Wall Thickness. This is also possible with an adiabatic wall where the heat flux is 0 to the andromeda galaxy.

yes, but I don't have a wall thickness in my case.

[MKuhn]

But in reality every channel has walls with a certain wall thickness. And this is only one click in the boundary conditions in Flunet, you can switch it on and off, you do not have to setup a new mesh.

[Weiqiang Liu]

Originally Posted by MKuhn

But in reality every channel has walls with a certain wall thickness. And this is only one click in the boundary conditions in Flunet, you can switch it on and off, you do not have to setup a new mesh.

do you mean I need to set a wall thickness in fluent even though I don't have wall mesh in my model?

[MKuhn]

yes try it if you want to model the heat exchange betwenn the wall and the fluid. For this you have to define a solid material in fluent, even if you don't have solid mesh. Switch shell conduction on, if you are also interested in the heat conduction along the tube (planar direction) and not only in the normal direction of you boundary.

Wall thickness and wall mesh are different things. You can model a tube wall using solid mesh or via defining a wall thickness without solid mesh. You can also combine boths e.g. having a wall mesh and defined wall thickness to consider the coating of the wall. For more details and limitations please refer to the Fluent User's Guide.

[Weiqiang Liu]

Originally Posted by MKuhn

yes try it if you want to model the heat exchange betwenn the wall and the fluid. For this you have to define a solid material in fluent, even if you don't have solid mesh. Switch shell conduction on, if you are also interested in the heat conduction along the tube (planar direction) and not only in the normal direction of you boundary.

Wall thickness and wall mesh are different things. You can model a tube wall using solid mesh or via defining a wall thickness without solid mesh. You can also combine boths e.g. having a wall mesh and defined wall thickness to consider the coating of the wall. For more details and limitations please refer to the Fluent User's Guide.

thanks very much for your suggestion. I'll try this and report the results in this thread.