Accessing velocity at inverse position

fusionfire21 · June 30, 2017, 08:13

Hi guys,

I am trying to implement the following formula into my code:
$\vec{v}_a = \frac{1}{2}(\vec{v}(\vec{r})+\vec{v}(-\vec{r}))$

That is, I aim at calculating this "antisymmetric" velocity for every time step.
How can I implement especially the $\vec{v}(-\vec{r})$ part, i.e. the access to the velocity at the inverse position?
Of course, $\vec{v}(\vec{r})$ is simply implemented as

Code:

volVectorField U

Thank you very much in advance!

Regards
fusionfire

usv001 · June 30, 2017, 23:41

Hi there,

What is your domain and mesh like? On the top of my head, I cannot think of a way to access velocity at the inverse location unless the mesh is a uniform structured Cartesian or cylindrical one.

USV

fusionfire21 · July 1, 2017, 03:11

Hi,
the domain is an ellipsoid with its center at the origin, generated with a rather complicated snappyHexMesh file. The coordinate system is cartesian though.
Regards

usv001 · July 1, 2017, 03:24

Hi,

If your mesh is symmetrical about the origin, then perhaps you could build a volScalarField whose internalField consists of the indices of the cells centered at the corresponding inverse locations. Then, you could simply loop over the internalField to get your $\vec{v_a}$ . If you do it carefully, you only need to loop over half the internalField since the values should be symmetric. But you must first build the volScalarField somehow.

Can you attach an image of your mesh? That could give others some ideas.

USV

usv001 · July 1, 2017, 03:28

Actually, if you mesh only the first octant and then mirror mesh to get the full mesh, you would get a symmetrical mesh about the origin. Then, the cell index at the inverse location could probably be obtained by adding a constant number to the cell index in the first four octants where z (or x or y) is positive.

USV

June 30, 2017, 08:13	Accessing velocity at inverse position	#1
fusionfire21 New Member Join Date: Jun 2017 Posts: 3 Rep Power: 8	Hi guys, I am trying to implement the following formula into my code: $\vec{v}_a = \frac{1}{2}(\vec{v}(\vec{r})+\vec{v}(-\vec{r}))$ That is, I aim at calculating this "antisymmetric" velocity for every time step. How can I implement especially the $\vec{v}(-\vec{r})$ part, i.e. the access to the velocity at the inverse position? Of course, $\vec{v}(\vec{r})$ is simply implemented as Code: volVectorField U Thank you very much in advance! Regards fusionfire

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Multiple floating objects	CKH	OpenFOAM Running, Solving & CFD	14	February 20, 2019 09:08
Results of CFX Simulations	Sof_70_Usthb	CFX	3	July 9, 2011 06:15
Accessing Velocity Components	anfho	OpenFOAM Programming & Development	1	June 30, 2011 19:21
accessing another velocity field in bc	daviderzen	OpenFOAM	0	April 20, 2011 06:24
Plot velocity profile at different position	AirS	Main CFD Forum	2	December 3, 2010 06:03

June 30, 2017, 23:41		#2
usv001 Senior Member Join Date: Sep 2015 Location: Singapore Posts: 102 Rep Power: 10	Hi there, What is your domain and mesh like? On the top of my head, I cannot think of a way to access velocity at the inverse location unless the mesh is a uniform structured Cartesian or cylindrical one. USV

July 1, 2017, 03:11		#3
fusionfire21 New Member Join Date: Jun 2017 Posts: 3 Rep Power: 8	Hi, the domain is an ellipsoid with its center at the origin, generated with a rather complicated snappyHexMesh file. The coordinate system is cartesian though. Regards

July 1, 2017, 03:24		#4
usv001 Senior Member Join Date: Sep 2015 Location: Singapore Posts: 102 Rep Power: 10	Hi, If your mesh is symmetrical about the origin, then perhaps you could build a volScalarField whose internalField consists of the indices of the cells centered at the corresponding inverse locations. Then, you could simply loop over the internalField to get your $\vec{v_a}$ . If you do it carefully, you only need to loop over half the internalField since the values should be symmetric. But you must first build the volScalarField somehow. Can you attach an image of your mesh? That could give others some ideas. USV

July 1, 2017, 03:28		#5
usv001 Senior Member Join Date: Sep 2015 Location: Singapore Posts: 102 Rep Power: 10	Actually, if you mesh only the first octant and then mirror mesh to get the full mesh, you would get a symmetrical mesh about the origin. Then, the cell index at the inverse location could probably be obtained by adding a constant number to the cell index in the first four octants where z (or x or y) is positive. USV