How to Calculate Resistance Values with Limited Patient Data

[Jake Sanchez]

Hello,

I am an undergraduate student who is enjoying the learning process of this amazing program, but have hit a roadblock.

I am having trouble producing a valid simulation with SV using a model of the ascending to descending aorta, with only the subclavian and common carotid arteries for simplification. I have watched your Youtube tutorials and read through most of your documentation but am unsure what the correct process is to calculate resistance values without PC-MRI data.

My simulations quickly diverge, reaching 10,000 mmHg within half a second. I believe the problem lies in the resistance values which I pulled directly from an article. It used the same section of the aorta so I thought the estimation would be ok, but I quickly learned this is not the case. From reading articles, I also saw success using Ohm's law for fluidics, however, I haven't found an in depth guide on how to accomplish this when the only given data is stroke volume and heart rate.

Are you able to refer me to such an article or give me some advice? It would be much appreciated.

The URL below contains my latest project with a short simulation to lower computational cost.

https://drive.google.com/drive/folders/ ... sp=sharing

Thank you for your time,
Jake

[David Parker]

Hello Jake,

Looking at the 12-procs_case/histor.dat file I see that the nonlinear residuals are large (should be < 1e-3)

Code: Select all

     7 3.000E+02 3.138E+02  (  45)   2.815E-01   1.324E-01  <  1967- 7|  17> [   0 -   0]
     7 3.400E+02 2.914E+02  (  44)   1.679E-01   1.289E-01  <  1967- 7|  17> [   0 -   0]
     8 3.640E+02 3.943E+02  (  46)   9.024E-01   3.540E-01  < 28233- 7|  17> [   0 -   0]
     8 4.240E+02 4.047E+02  (  46)   1.318E-01   1.369E-01  < 28233- 7|  16> [   0 -   0]
     9 4.710E+02 5.795E+02  (  47)   2.152E-01   5.318E-01  < 26831- 7|  15> [   0 -   0]
     9 5.010E+02 6.219E+02  (  48)   7.684E-01   1.062E+00  < 11184- 7|  16> [   0 -   0]
    10 5.340E+02 1.037E+03  (  50)   1.045E+00   1.761E+00  < 19898- 7|  15> [   0 -   0]
    10 5.750E+02 1.157E+03  (  50)   5.000E-01   9.491E-01  < 24488- 7|  15> [   0 -   0]

This is likely caused by the time step being too large (0.05), should be much smaller and satisfy the CFL condition.

There are lots of details that you need to get right to ensure that the simulation results you obtain are meaningful: good geometric model (no small features), mesh size suitable to capture the flow features you are interested in, well-defined BCs, proper solver parameters and visualize/interpret the results. A solid background in CFD is essential.

It is a good idea when doing a numerical study to start with as simple a simulation as possible (steady, 0 resistance BCs, etc), make sure that converges to a reasonable solution and then increase the complexity of the BCs. As a

Cheers,
Dave