Isolate Cells in Layers Based on Height (FLUENT UDF)

Boudon · August 8, 2024, 13:34

Good evening, everyone.

I'm currently working on a UDF in ANSYS Fluent to calculate the Longitudinal Length Scale of turbulence, $L_{11}$ . I'm using the direct integration of correlation functions, which is expressed as:

$\rho^x_u(\Delta x, y_j)=\frac{\overline{ u'_1(x; y_j) \cdot u'_1(x + \Delta x; y_j) }}{ \overline{u'_1(x; y_j)^2}}$

$L_{11}=\frac{\sum_{j=1}^{N_y}\rho^x_u(\Delta x, y_j)} {N_y}$

Here,

is the longitudinal fluctuating velocity, $\Delta x$ is the grid spacing along the x-axis, and

represents the y-coordinate level at a constant j for $j \in [y_{\text{min}}$ , $y_{\text{max}}]$ . In other word

corespond to every layer of cells in the y directions. With a number of layer

The direct integration of correlation functions involves integrating $u'_1(x) \cdot u'_1(x + \Delta x)$ for each layer of y-cells and I already have every value for

and $u'_1(x + \Delta x)$

So I need to isolate each cell based on its y-layer within a UDF. But am having trouble finding an efficient method to do this. Do you have any ideas on how I can achieve this effectively?

My mesh is structural with inflation layer.

Thanks for the help.

Boudon · August 9, 2024, 08:56

I am also considering integrating over a sphere of radius

, where

is the turbulent length scale.

$\rho^r_u(\Delta r, y_j) = \frac{\overline{ u'_1(r; y_j) \cdot u'_1(r + \Delta r; y_j) }}{ \overline{u'_1(r; y_j)^2}}$

$L_{11} = \frac{\sum_{j=1}^{N_y} \rho^r_u(\Delta r, y_j)}{N_y}$

Using this method, I would need to isolate all cells within a sphere of radius

and $r + \Delta r$ . Do you have any ideas on how I can efficiently do this?

I was thinking of using a Boolean operation where the volume of the sphere at the current centroid point intersects the cells in the volume

inside the `get_domain` function. However, I’m having trouble finding an efficient way to implement this. If you have any suggestions, I would really appreciate it.

tomsjerry · August 23, 2024, 03:56

I'm having a similar issue, could you please share your workaround?

August 8, 2024, 13:34	Isolate Cells in Layers Based on Height (FLUENT UDF)	#1
Boudon New Member Baptiste Oudon Join Date: Dec 2023 Posts: 2 Rep Power: 0	Good evening, everyone. I'm currently working on a UDF in ANSYS Fluent to calculate the Longitudinal Length Scale of turbulence, $L_{11}$ . I'm using the direct integration of correlation functions, which is expressed as: $\rho^x_u(\Delta x, y_j)=\frac{\overline{ u'_1(x; y_j) \cdot u'_1(x + \Delta x; y_j) }}{ \overline{u'_1(x; y_j)^2}}$ $L_{11}=\frac{\sum_{j=1}^{N_y}\rho^x_u(\Delta x, y_j)} {N_y}$ Here, is the longitudinal fluctuating velocity, $\Delta x$ is the grid spacing along the x-axis, and represents the y-coordinate level at a constant j for $j \in [y_{\text{min}}$ , $y_{\text{max}}]$ . In other word corespond to every layer of cells in the y directions. With a number of layer The direct integration of correlation functions involves integrating $u'_1(x) \cdot u'_1(x + \Delta x)$ for each layer of y-cells and I already have every value for and $u'_1(x + \Delta x)$ So I need to isolate each cell based on its y-layer within a UDF. But am having trouble finding an efficient method to do this. Do you have any ideas on how I can achieve this effectively? My mesh is structural with inflation layer. Thanks for the help.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
cellZone not taking all the cells inside	rahulksoni	OpenFOAM Running, Solving & CFD	6	January 25, 2019 01:11
cellZone not taking all the cells inside	rahulksoni	OpenFOAM	0	January 16, 2019 02:16
[blockMesh] Create internal faces as patch in blockMesh	m.delta68	OpenFOAM Meshing & Mesh Conversion	14	July 12, 2018 15:43
[snappyHexMesh] Problem airfoil snappyhexmesh add layer	bye bye my blue	OpenFOAM Meshing & Mesh Conversion	1	May 5, 2017 06:56
Problem with entire domain comprised of split hex cells	Arzed23	Main CFD Forum	1	November 8, 2016 00:01

August 9, 2024, 08:56		#2
Boudon New Member Baptiste Oudon Join Date: Dec 2023 Posts: 2 Rep Power: 0	I am also considering integrating over a sphere of radius , where is the turbulent length scale. $\rho^r_u(\Delta r, y_j) = \frac{\overline{ u'_1(r; y_j) \cdot u'_1(r + \Delta r; y_j) }}{ \overline{u'_1(r; y_j)^2}}$ $L_{11} = \frac{\sum_{j=1}^{N_y} \rho^r_u(\Delta r, y_j)}{N_y}$ Using this method, I would need to isolate all cells within a sphere of radius and $r + \Delta r$ . Do you have any ideas on how I can efficiently do this? I was thinking of using a Boolean operation where the volume of the sphere at the current centroid point intersects the cells in the volume inside the `get_domain` function. However, I’m having trouble finding an efficient way to implement this. If you have any suggestions, I would really appreciate it.

August 23, 2024, 03:56		#3
tomsjerry New Member Join Date: Jan 2024 Posts: 3 Rep Power: 2	I'm having a similar issue, could you please share your workaround?