[CFXer]

Originally Posted by ghorrocks

Both of those correlations are for flow in tubes, not flat plates. You should be careful about whether they are applicable to flat plates.

Also, for the interface conditions between your fluid and solid domains in CFX, the heat transfer model options you have are conservative interface flux, thermal contact resistance or thin material. You could also put a source term on the interface. How are you proposing to use the Nusselt Number given by those correlations with the available CFX heat transfer interface models? There is no option where you can select Nusselt Number or heat transfer coefficient. (Hint, if you have not worked it out already: the thermal resistance of the interface is 1/h, so you could use the thermal resistance option)

I'm sorry for your misunderstanding again...

I made a plate model to use porous domain, but each plate has several semicircular channel, and some paper said Gnielinski or Dittus-Boelter equation well fits the experimental result.

I set up the heat transfer coefficient in porous domain setting, I attached a figure for your better understanding.


Resistance loss coefficient : friction factor / hydraulic diameter
Heat transfer coefficient : Nusselt number * conductivity / hydraulic diameter

[ghorrocks]

Shouldn't Nusselt number have a characteristic length defined in the flow direction along the plate, not the hydraulic diameter? See: https://en.wikipedia.org/wiki/Nusselt_number

[CFXer]

Originally Posted by ghorrocks

Shouldn't Nusselt number have a characteristic length defined in the flow direction along the plate, not the hydraulic diameter? See: https://en.wikipedia.org/wiki/Nusselt_number

But in pipe flow, it needs hydraulic diameter as you can see in 'Forced convection in fully developed laminar pipe flow' part...

[ghorrocks]

Good point. It looks like you have everything in place to model what you intend, you have answered all my questions Just be aware that you are using empirical correlations for both pressure drop and heat transfer, so you can expect significant errors in both of them.

So if we go back to post #1 and your original question - you asked about a problem with your heat transfer. Was the heat transfer results you showed using the porous flow results you describe and the Gnielinski or Dittus-Boelter correlation? Can you clarify what your question as I do not understand exactly what you are asking.

[CFXer]

Originally Posted by ghorrocks

Good point. It looks like you have everything in place to model what you intend, you have answered all my questions Just be aware that you are using empirical correlations for both pressure drop and heat transfer, so you can expect significant errors in both of them.

So if we go back to post #1 and your original question - you asked about a problem with your heat transfer. Was the heat transfer results you showed using the porous flow results you describe and the Gnielinski or Dittus-Boelter correlation? Can you clarify what your question as I do not understand exactly what you are asking.

It is very difficult to compare the handwritten calculation result because the working fluid is supercritical CO2, having very steep variation of thermal properties.

Instead, I used code calculation result that can calculate its heat transfer by nodalization method.

The result is in the figure, the temperature variation of 'T_CO2_CFD_scale' is too weird. It means too high heat transfer coefficient in the left section.

'T_CO2_CFD_plate' is a result of 'thin material' assumption.

My question is : why this too high heat transfer coefficient in porous-solid interface, when the 'free slip wall' condition is applied instead of 'no slip condition'.

[ghorrocks]

I have no idea what that chart is showing, I do not know what heat transfer correlation you are using or how you have applied it, I have no idea what the "thin material" assumption is or how that is relevant.

This thread is very frustrating because you are not explaining what you are doing. I now know that:
* You are using some form of heat transfer empirical correlation but have not said which one or how you applied it.
* The fluid is supercritical CO2
* You are comparing between a scale model and a full sized device

These are all vital bits of information which you have only revealed after dozens of posts on this thread. Why did you not say these things in the first post? What other vital information have you not revealed yet?

[CFXer]

Originally Posted by ghorrocks

I have no idea what that chart is showing, I do not know what heat transfer correlation you are using or how you have applied it, I have no idea what the "thin material" assumption is or how that is relevant.

This thread is very frustrating because you are not explaining what you are doing. I now know that:
* You are using some form of heat transfer empirical correlation but have not said which one or how you applied it.
* The fluid is supercritical CO2
* You are comparing between a scale model and a full sized device

These are all vital bits of information which you have only revealed after dozens of posts on this thread. Why did you not say these things in the first post? What other vital information have you not revealed yet?

I'm sorry but I did not intentionally hide some information.

I'm trying to explain this domain to you.


Model :
*Printed Circuit Heat Exchanger(PCHE) with 7 hot and 7 cold plates and 15 solid plate between each hot/cold plates,
*Each fluid plate has 7 straight semicircular channels with D=1.8mm
*Plate thickness is 1.2mm and semicircular depth is 0.6mm
*Hot fluid is CO2 with 8MPa and Tin = 76C
*Cold fluid is Water with 200kPa and Tin = 25C
*Inlet and outlet header(distributor) is semi-cylinder shape.
*I want to describe heat transfer of hot/cold fluid with solid plate, to compare flow distribution of some conditions.
*It has too many channels, so I cannot use fluid domain because of computing cost, so I want to use porous domain.
*In porous media correlation, there are two terms but only V square term is valid if Re>1
*So, Resistance loss coefficient is composed with (f/Dh)
*At heat transfer setting in porous domain, there are 'fluid solid area density' and 'fluid solid heat transfer coefficient'
*I used interfacial Area Density[1/m] -> (Surface Area / Total volume)
and Heat Transfer Coefficient[W/m^2 K] -> (Nu * k / Dh)
*The total model has error with heat transfer calculation, so I used thin material option in porous-solid interface setting


I guessed heat transfer coefficient term in my handwritten calculation, so maybe I think it could be wrong.

I really have no idea why this model shows very weird result (too high heat transfer coefficient at inlet section).

If you need more information to understand, please let me know.

I'm trying to explain everything I can but it is difficult to say it completely.

[ghorrocks]

You have used many simplifications, such as using empirical relations (pressure drop and heat transfer), have complex material models (supercritical CO2), are not modelling the detailed fluid flow but are using simplified correlations. You have to expect that this will result in your model having errors, and in some cases big errors. You also seem to say that your heat transfer coefficient calculation does not seem to work so you used a thin material model - how are you going to get an accurate result when you use the wrong physics?

All I can say is that if you want an accurate result you need to use an accurate modelling technique, which is to model the fluid as a fluid domain and accurately resolve the boundary layer.

[CFXer]

Originally Posted by ghorrocks

You have used many simplifications, such as using empirical relations (pressure drop and heat transfer), have complex material models (supercritical CO2), are not modelling the detailed fluid flow but are using simplified correlations. You have to expect that this will result in your model having errors, and in some cases big errors. You also seem to say that your heat transfer coefficient calculation does not seem to work so you used a thin material model - how are you going to get an accurate result when you use the wrong physics?

All I can say is that if you want an accurate result you need to use an accurate modelling technique, which is to model the fluid as a fluid domain and accurately resolve the boundary layer.

Or.. this result could be right?

On the off change the flow distribution could affect heat transfer coefficient of core section, it could be right?

I'm continuously trying to check the heat transfer coefficient, heat balance of inlet/outlet Temperature, thermal conductivity of solid.. but everything has no problem just for heat transfer quantity...

[ghorrocks]

You really should check a model like what you propose against experimental data. You need to know what level of accuracy to expect from your approach as it is very unusual. I suspect you will find it quite inaccurate.