Extracting flux data from faces

[hisl6802]

Hello,

I am simulating bone-cartilage fluid transport during cyclic compression. I am hoping to extract the fluid flux across the faces shared by elements with model parameters for bone and cartilage. Currently after simulation I zoom into the wanted surface, select the faces of interest, and extract the data. After which


1) Is there a way I can extract the face z-fluid flux data into the log file? Or do I need to extract element or nodal z-fluid flux values and solve for flux across the face?

If I do need to calculate the face flux using nodal or elemental fluid flux values. Can you clarify how FEBio does this for hexahedron elements and penta elements?

I am trying to to analyze the faces shared by part1 (Bone) and part 2 (DZ1)

Thank you,

Brady

[maas]

Hi Brady,

In FEBio, fluid flux is an element variable, so there is no direct way of outputting fluid flux on a face via the logfile. In FEBio Studio, when displaying a datafield, it will always project the data onto the nodes, and from that, face values can be extracted.

In order to do the same with the logfile, you'd probably have to write some code. First, you'll have to get the fluid flux data out of FEBio. You can do that using this logfile tag:

<element_data data="wx;wy;wz"/>

(The variables for exporting fluid flux to the log file are wx, wy, and wz. See here and scroll down to the table for biphasic analysis.)

Note that this is an element average of the fluid flux (it is the average value of the integration point values for each element).
Once you have this data, you can extract the face values. You can either find the element that a face belongs to and use that value, or if you want to do something more similar to what happens in FEBio Studio, you can first project the element values on to the nodes (for each node, the nodal value is the average of the surrounding element values), and then to get a face value, you would average the nodal values of the face. Note that this may introduce smoothing of the data as a result of the two averaging procedures.
Hope this helps!

Best,

Steve

[hisl6802]

Hi Steve,

Thank you for replying. I would like to know if you can clarify a discrepancy I am finding.

I have been double-checking the calculations I am getting for z-fluid flux across a face and the ones FEBio Studio is calculating. If I follow the procedure you mentioned above, where I take the elements surrounding a node of interest and calculate the average z-fluid flux, my results match those of FEBio Studio for that node. However, I am finding that for the faces of interest, the averaging of the nodal values of a face in FEBio Studio doesn't match the average of the nodal values I am getting. In fact, I am getting a 44.6% - 80.4% error depending on whether I consider the FEBio Studio calculations or my calculations as the baseline.

As you mentioned above, I only have access to the averages of the integration points (i.e., the element average of the fluid flux). Would this discrepancy arise from FEBio Studio projecting from the integration points to each node and then averaging the nodal values for fluid flux? Is FEBio Studio performing a different set of projection steps from those outlined above, leading to the differences?

Thanks,

Brady

[ateshian]

Hi Brady,

I think that the issue you are encountering is a boundary layer problem + the dominant flow is radial, not axial. So you should not expect tremendous accuracy for the negligibly small component of the axial flow. I remeshed your model to create finer boundary layer meshes near the cartilage-bone interface, then I converted the mesh to quadratic ( FEBioForumModel_BH_GAA_quadratic.feb ) and ran it. Here is a snapshot of the results, which support what I am saying:
FEBioForumModel_BH_GAA_quadratic.png
Best,

Gerard

p.s. If you want to test continuity of flux in the axial direction, perhaps you could try a uniaxial confined compression analysis, but please remember to use a highly biased mesh at the interface between the cartilage and bone.

[hisl6802]

Hi Gerard,

Thank you so much for helping me understand my question better.

Can you point me towards how you went about re-meshing my model? I would like to learn this for future models. I can see how you added the bias to the osteochondral interface, but I am curious how you did this for the merged parts. Previously, I relied on meshing individual parts and then merging before setting up boundary value problems and running. For example, from your image above, it seems possible to select the "yellow" part and add additional z-stacks. Is there an example of this that I have missed in forums?

Thanks for the help.

Brady

[ateshian]

Hi Brady,

I followed a tedious procedure because I encountered some bugs (which I have reported to Steve).

Basically the trick is to purge the selections of boundary conditions, detach the mesh of the bone from that of the cartilage, use the boundary layer tool to add a finer mesh to the bone (at the interface with the cartilage) and to the cartilage (DZ1, at the interface with the bone), then merge the two meshes back together. The bugs I encountered are (1) detaching the bone changes the material assignments in the cartilage section, and (2) merging the two parts back together causes FEBioStudio to crash. Once these are resolved, I am hoping that you'll be able to do this easily.

Best,

Gerard

[hisl6802]

Hi Gerard,

Thank you for providing me with more detail on this.

I am currently rebuilding the models completely, but look forward to future updates.

Best,

Brady

[ateshian]

Hi Brady,

Steve has fixed the bugs, once the next development version is released you should be able to try it.

Best,

Gerard

[hisl6802]

Hi Gerard,

Thank you very much!

Best,

Brady