[Yanlu]

In the literature, the patient's MRI data is read and the velocity value of the entrance at each moment is obtained. How to assign the values of different grid nodes at various times to the corresponding grid nodes in cfx as the entrance boundary? Any ideas?

[1]Pirola, S., Cheng, Z., Jarral, O.A., O’Regan, D.P., Pepper, J.R., Athanasiou, T., Xu, X.Y., 2017. On the choice of outlet boundary conditions for patient-specific analysis of aortic flow using computational fluid dynamics. J Biomech 60, 15–21. https://doi.org/10.1016/j.jbiomech.2017.06.005

[ghorrocks]

I assume you are trying to apply a varying boundary condition, where the value of the boundary condition varies in space (so across the boundary condition) and time.

You will probably need a user fortran routine to do this. This will give you complete control over your boundary condition. It is not easy to do, so read the CFX documentation on user fortran carefully, and look at the examples it gives.

[Yanlu]

Originally Posted by ghorrocks

I assume you are trying to apply a varying boundary condition, where the value of the boundary condition varies in space (so across the boundary condition) and time.

You will probably need a user fortran routine to do this. This will give you complete control over your boundary condition. It is not easy to do, so read the CFX documentation on user fortran carefully, and look at the examples it gives.

Thank you very much for your reply! Can the user fortran routine assign a value to the speed of each grid node? In my opinion, I should set the ID number, xyz coordinates and speed values of each grid node at different times in Matlab and put them in a matrix. This matrix is then set to the entrance velocity through the user fortran routine. But how can I know the ID of each grid node? When the grid is very dense, can a unique grid node be accurately located only by xyz coordinates?

[ghorrocks]

In user fortran you can program it to do whatever you want - assuming you are a good enough programmer to work out how to do it. As I said, have a look at the CFX documentation on user fortran and see what it looks like to you.

[Yanlu]

Originally Posted by ghorrocks

In user fortran you can program it to do whatever you want - assuming you are a good enough programmer to work out how to do it. As I said, have a look at the CFX documentation on user fortran and see what it looks like to you.

Ok, thank you! I will try to learn the user fortran.

[Opaque]

A lot of people tend to confuse things when using discrete data, and you seem to be doing so as well.

Your discrete data already has its own spatial and time grid/mesh. To deal with it, you can use UserFortran to organize and manage your data. However, keep in mind the mesh for discretization is NOT the same mesh for the discrete data --> there is a mapping to be done.

Who is responsible for the mapping between "discrete data mesh" and "solution mesh" -> UserFortran programmer either with local tools or with existing API if the exist.

One thing you can do is to use the mapping tools to map your "discrete data" to the "solution mesh" for each time step manually, and let the User Fortran work on the "mapped data" instead.

You should read how to use Ansys CFX .csv files, how they are loaded into CFX-Pre, how they can be transformed/edited/mapped and how that can help you simplify your tasks.

[Yanlu]

Originally Posted by Opaque

A lot of people tend to confuse things when using discrete data, and you seem to be doing so as well.

Your discrete data already has its own spatial and time grid/mesh. To deal with it, you can use UserFortran to organize and manage your data. However, keep in mind the mesh for discretization is NOT the same mesh for the discrete data --> there is a mapping to be done.

Who is responsible for the mapping between "discrete data mesh" and "solution mesh" -> UserFortran programmer either with local tools or with existing API if the exist.

One thing you can do is to use the mapping tools to map your "discrete data" to the "solution mesh" for each time step manually, and let the User Fortran work on the "mapped data" instead.

You should read how to use Ansys CFX .csv files, how they are loaded into CFX-Pre, how they can be transformed/edited/mapped and how that can help you simplify your tasks.

Yes, you are right. I found a video on how to set a csv file of x,y,z, and temperature as the entrance boundary. It may solve my problem by setting a CSV file of x,y,z,t and velocity. If I succeed, I will reply to this post again.

https://www.youtube.com/watch?v=NjiO9_wizzo

[Yanlu]

Originally Posted by Yanlu

Yes, you are right. I found a video on how to set a csv file of x,y,z, and temperature as the entrance boundary. It may solve my problem by setting a CSV file of x,y,z,t and velocity. If I succeed, I will reply to this post again.

https://www.youtube.com/watch?v=NjiO9_wizzo

This video solved my problem by setting a CSV file of x,y,z, and velocity. However, his function seems only to be able to set spatial coordinates as input parameters, and I can't turn time t into an input parameter.