[MichaelK]

Hello everyone,

i have some issues to calculate / simulate the pressure drop over a pin-fin heat sink.

Geometry description:
Heat sink width: 50 mm
Heat sink length: 75 mm
Pin fin diameter: 1 mm
Number of pins in flow direction: 10
Number of pins each row: 10
Pin fin distance (equally distributed) X-direction: 8.22 mm
Pin fin distance (equally distributed) Y-direction: 5.00 mm

To reduce the computation time its converted into a 2D model.
In addition, only one row is simulated (--> "new width = 5 mm")
See Fig. 1 for the Geometry setup

Boundary conditions:
- Velocity inlet: 10 m/s (for example)
- Pressure outlet: 0 Pa
- Two Symetry planes (red walls in Fig.1)

Physics conditions:
- Segregated Flow
- Constant Density
- All y+ Wall Treatment
- Steady
- Two Dimensional
- k-Omega Model
- Turbulent
- Gas

- Ref. Pressure = Initial Pressure = 0 Pa
- Initial Velocity = Inlet Velocity = 10 m/s
- Density and Dynamic Viscosity = constant = default values

Results:
Surface average report at inlet: 82.5 Pa (Fieldfunction: "Pressure")
Fig2 shows the result in a scalar scene.

Problem:
If i compare the CFD result (82.5 Pa) with my analytical result (VDI Heat Atlas - chapter L1.4 (year 2013)) i have a huge gap. See Fig3 and Fig 4

Analytical result:
247.36 Pa




Does anyone know where my mistake is?
Do I forget a setting like reference pressure or pressure coefficient?

Thans a lot!

[MichaelK]

noone ?

[me3840]

Is 2D really a good assumption here? A pin fin will have significant 3D dynamics that you're ignoring.

[MichaelK]

I calculated both, 2D and 3D, versions.
Both Simulations have nearly the same results...

The simulated results are still very different from the analytic ones

What do i wrong?

If you need more information, feel free to ask !
Thanks for your reply!

My fluid is a Gas: Air

[me3840]

I misread you're using symmetry planes before. Since this is 2D and you're only using a single line of cylinders, is that really comparable to what that book states? It has some sort of grid. Make sure that there isn't anything you're missing since you're doing a 2D analysis to compare to a correlation, which is likely developed from 3D. Are there any area values which you might be misrepresenting? I can't read the text.


I'm not really sure if symmetry planes with only a single cylinder row is a good way to run this. Having multiple cylinders will allow them to shed better if they do indeed do so.