[Shomaz ul Haq]

Hello everyone. Hope all are well. I am changing a simple parameter heat flux in wall boundary condition in CFX-Pre. When I run the solver and open CFD-Post although the new Wall Heat Flux is output I get the same ANSYS computed Wall Heat Transfer Coefficient. I am perplexed. Why is that so. Would be grateful for help. Maybe somebody has faced a similar issue. Thanks.

[urosgrivc]

Are you doing a CHT simulation including solid and fluid, or where do you put your heat flux ?
Give a bit more info about what you are doing

[Shomaz ul Haq]

Thanks urosgrivc. There is no solid just a circular pipe with fluid in it. I think CHT is for solid regions for which heat transfer is involved. Right? Also I am inputting my heat flux at the wall. Hope to hear on this matter. Thanks.

[Opaque]

The Wall Heat Transfer Coefficient is not a function of the amount of heat flux at the wall. It may be affected by it via material properties as a function of temperature.

What are you trying to determine?

[Shomaz ul Haq]

Thanks Opaque but doesn't Wall HTC computation in ANSYS involve Wall Heat Flux as said in the FAQs? I just want to determine the effect on different values of Wall Heat Flux on Wall HTC? Does it mean if I use a solid wall only then Wall HTC would be affected by Wall Heat Flux. Would be grateful if you could shed some light on it. Thanks.

[Opaque]

Perhaps is better rewind back to what the Wall Heat Transfer Coefficient represents. Then, everything else will come into perspective.

If you take a simple flow configuration w/o complex models, say laminar flow over a flat plate or within a pipe, you can compute the wall heat flux from the fluid to the wall by using Fourier's law, i.e. K * Grad T . Normal Area. Done!!. If you know the applied wall heat flux, you can compute the wall temperature. Alternatively, if you know the wall temperature, you can compute the wall heat flux. That is: one helps to compute the other, no need for another quantity.

I used the simple example above to show the Wall Heat Transfer Coefficient does not exist, nor it is needed for simple cases. It is an engineering quantity defined to simplify complex flows into a single number and a simple formula for the computation of the wall heat flux:

Wall Heat Flux = Area * Delta Temperature * HTC

known as Newton's cooling formula. The HTC is a consequence of the flow around a wall based on flow conditions and material properties.

My advice is to read a good book on convective heat transfer keeping in mind where it comes from, what is used for, and how/where to use it.

Hope the above helps,

[Shomaz ul Haq]

Thanks a lot Opaque. I revised heat transfer especially convective heat transfer. I have realized that changing wall heat flux doesn't change heat transfer coefficient for forced convection. So average htc is independent of delta T i.e. the difference in temperature between the body and the fluid being contained (internal flows) by it or surrounding (external flows). I think it is due to this reason changing heat flux doesn't change htc. I believe now calculating htc is a fairly complicated task. Thanks for guiding me.

[ghorrocks]

There is an expert parameter which fixes the ambient reference temperature for HTC evaluation. This may be useful in your case.

[urosgrivc]

Yes take a look at na expert parameter (T-bulk) like glenn wrote. If you dont set this one to some reference temperature your HTC results will be dependant on your mesh thicknes near the wals and this is usualy not what users want and there is a lot about it here on the forum. it is posible ofcourse to compute the usefull HTC later in post with the help of an expression but is bit more complicated.

To obtain HTC walues I think it is more apropriate to just set walls to a fixed temperature and I always use this T-bulk parameter
I am mostly working with external flows so I set it to 298.15K But you have internal flow so your inlet temperature might be more apropriate.
HTC is mostly dependant on the temperature gradient near the wall which is dependant on velocity gradient near the wall so the higher the speed of the flow the higher the HTC. Temperature of the wall will have almost negligable efect as this only changes properties of the fluid and becouse of this flow properties a litle bit so if you can estimate what the temperature of the wall is that is a good robust BC. Results are best if you use fine mesh inflation layers y+~1 so those gradients near walls are computed better

[Shomaz ul Haq]

Thanks Glenn and Urosgrivc. I am already setting the expert parameter "tbulk for htc". In my case I had set it to the inlet temperature of 20C or 293.15 K so my h results are independent of grid thickness near the walls. Despite that I am not seeing a change in h when flux changes. One think I would like to ask though is that I am computing velocity at inlet for each concentration of nanoparticle as V=Reμ/ρD with thermophysical properties μ and ρ at that concentration (I have made expressions for that in CFX-Pre). Is it a good way to set initial inlet velocity? The velocities computed this way are very small (although they should be to get a laminar flow) of the order of around ~0.0177 m/s due to low values of μ of water and high values of effective nanofluid ρ and doesn't change much even at high concentration or Re number. Can this be the problem? Should I set higher order values of velocities? It is extremely tedious for me to compute velocities, input in pre, run the solver, and run post processor at each concentration. Take a lot of my time. Or there is no need to compute inlet velocity? Urosgrivc, you said increasing velocity changes velocity gradient which in turn changes temperature gradient increasing h. Would be grateful if you guys could guide me here. Thanks.

[Shomaz ul Haq]

Hello everyone. Hope all are well. I have battling with this problem for a week. I read in some articles on forums that changing heat flux doesn't change htc in single phase heat transfer. Why is that? Also my question remains about computing velocity at inlet for each concentration of nanoparticle as V=Reμ/ρD with thermophysical properties μ and ρ at that concentration (I have made expressions for that in CFX-Pre). Is it a good way to set initial inlet velocity? The velocities computed this way are very small (although they should be to get a laminar flow) of the order of around ~0.0177 m/s due to low values of μ of water and high values of effective nanofluid ρ and doesn't change much even at high concentration or Re number. Can this be the issue for Should I set higher order values of velocities? Would be grateful. Thanks.

[ghorrocks]

changing heat flux doesn't change htc in single phase heat transfer. Why is that?

Unless changing the heat flux changes something which affects the HTC then the HTC will not be affected (that is an obvious statement, really...). In other words, if you have an incompressible fluid with constant properties then changing the heat flux will change the HTC. For the HTC to change it needs some link to the heat flux - maybe a compressible fluid, temperature dependant properties or something like that.

I do not understand what you are saying after your first question.

[Shomaz ul Haq]

Thanks Glenn. As you said yourself "for an incompressible fluid with constant properties changing heat flux will change htc". The fluid is incompressible and has constant thermophysical properties. Why still the htc doesn't change with heat flux?

With regards to my second question what I want to know is related to the computation of velocity at the inlet boundary condition. I am setting a Re number by calculating the required velocity for getting such a number i.e. by V=Reμ/ρD. I am getting very low values for laminar regime and also the change from Re number like from 400 to 2000 doesn't change velocity a lot. Can this be the issue I am getting no effect of heat flux on htc and getting low htc numbers. Am confused. Thanks.

[ghorrocks]

Oops, I left out an important word: "In other words, if you have an incompressible fluid with constant properties then changing the heat flux will NOT change the HTC." Sorry about that.

The HTC is a function of the velocity field and fluid properties. If the heat transfer does not affect the velocity field (ie incompressible flow) or the material properties (ie constant properties) then the HTC will remain unchanged.

An exception to this would be natural convection. The stronger the heat flux the greater the natural convection flow and you will get a larger HTC.

If you are changing Re from 400 to 2000 and everything else is constant then the velocity should be increasing by a factor of 5.

[Shomaz ul Haq]

Thanks Glenn. Yeah I figured out that was a mistake on your part No worries. I perfectly understand now. Thanks a lot for clearing my understanding. Just perfect the way you described. So the variable properties should be a function of temperature?

About the velocity problem. Yes it should change by a factor of 5 if I change Re number from 400 to 2000. But calculating velocity from that equation, is it a good idea to use that as an initial boundary condition value? Also would using mass flow rate and changing it give a higher htc value and its variation than velocity? Thanks.

[urosgrivc]

Try simulating a wall with temperature of 100 and 150°C you will observe that HTC is almost the same if flow properties stay unchanged, as it holds [K] in the unit [W m^-2 K] so it tries to be independent of the wall temperature and it actualy stays exactly the same if the fluid has constant (temperature independant) properties (if properties of the fluid are efected by temperature thigs change but only a tiny usualy insignifacnt amount).
As I wrote before HTC is mostly dependant on the flow velocity (gradient) near wall, other things dont change htc much.

This value HTC was actualy developed becouse it includes a lot of flow properties, heat transfer properties and everithing in a simple value so people could calculate some wery dificult stuf simply by hand.

Changing mass flow will ofcourse change HTC walue but only becouse it changes velocity.
What I mean by that is if you multiply mass flow by 2 and increese the cross section area of your pipe by 2 than again HTC will not change (maybe a litle, but we can easily neglect that)
But if you change mass flow and leave the pipe the same velocity will go up and HTC with it.
it is calculated from velocity gradient near wall and if you think about it it is actualy quite reasonable
(on a stove, water will boil quicker if you stir it, so boundary layer is as thin as posible (hight velocity gradient near wall) the more you stir the thiner the boudary layer the higher the htc the more Joules per second in to the water and les seconds for it to boil).

but be cerfull when using mass flow if you will use gasses as again, if the preesure would be higher
mass flow change might not cause velocity to change (it might change gas properties like heat capacitance and conduction a bit and htc becouse of that but that is again far less efective than speed)

[Shomaz ul Haq]

Thanks Urosgrivc. I have seen in an article what you just said. Changing heat flux even with temperature dependent properties change htc by a very little amount even with temperature dependent properties. It depends on the flow gradients near the wall.

For the mass flow rate problem I think the equation is m_dot=ρVA. You said "multiplying mass flow by 2 and increasing the cross section area of your pipe by 2" that means m_dot=2*(ρV*2*A) that means increasing both mass flow rate as a whole by 2 and area by 2 i.e. by a factor of 4. Right? So does that mean even four times mass flow rates won't have an affect on htc? Is this what you are trying to tell me? Hope you can clarify this point.

That example of water being boiling and stirred in a cooking stove for explaining the boundary layer and gradients effects is illuminating. By the way I am using liquids. Hope to hear from you soon. Would be grateful. Thanks.

[Nourhan Bahgat]

Hi Guys,
Any one can help me to solve this problem.
why the contour of wall heat flux change in CFX post in bottom wall absorber tube although i assumed it's constant value in CFX pre ?

[ghorrocks]

No idea. If you don't tell us anything about what you are doing then we can't help.

Please post an image of your mesh and your output file.

[Opaque]

Just in case, how does the plot for "Heat Flux" variable looks like.

Note the missing "Wall" prefix when looking for it in the Variables list dialog