NaN results in coronary artery simulation

[Rodrigo Romarowski]

Hi everyone,

I am currently simulating a few coronary geometries and I'm having the following result after mesh partitioning when running the solver in our cluster:

1 4.980E+02 NaN (****) NaN NaN < 2452- 6|****> [ 0 - 0]
1 9.950E+02 NaN (****) NaN NaN < 1664-14|****> [ 0 - 0]

This error occurs in the coronary geometry with the aberrant origin of the right coronary artery (see attached file, left) while the other one works just fine. Boundary conditions have all been taken from the website tutorial. I checked the quality of the meshes with Paraview and everything seems OK. I used radius-dependency in the coronaries and a fixed edge length in the root of the aorta. Different parameters for this meshing scheme derive in the same problem in the solver.

Any hints for the behaviour in this particular geometry?

Many thanks,

Rodrigo

[Justin Tran]

Hi Rodrigo,

Thank you for your question! That is strange that the left model diverges immediately while the other one seems fine. Since the meshes seem to be OK, my hunch is that there might be something funky with the boundary conditions. If you are following the tutorial online, probably you have a Dirichlet inlet condition with a pulsatile waveform, RCR boundary conditions at the aortic outlet, and coronary-type boundary conditions at the coronary outlets.

One method I use to debug simulation issues is to run the simplest simulation possible: steady inflow with resistance outlets. Can you try setting up another simulation job for the left geometry with these conditions? You can set a steady inflow BC by setting the BC Type to "Prescribed Velocities" and setting the value to be the cardiac output for your model (ex: -60 mL/s). For all of the outlets (both the aortic outlet and coronary outlets) you can set the BC Type to "Resistance" and set the resistance value to the Rd values from your previous simulation (or whichever of the resistances was the highest in the coronary outlets).

Give that a try and let us know how it goes!

[Rodrigo Romarowski]

Thanks for the advice! that certainly solved the issue.
By modifying one boundary at a time i found an error with one of the Coronary capacitances.