Unable to Run 1D Simulation

[Frank Servin]

Hello,

I have run into an error with the 1D-solver Simulation. I am able to "Create Files for Simulation" but I am unable to "Run Simulation". When I try to run the simulation I am returned with an 'nkown error return code".

335570.18 ERROR: Simulation job error = 1
335570.19 ERROR: nknown error return code
335572.08 WARNING: Simulation job '1D_PP' has failed with non-zero exit code.
335572.08 WARNING: Exit code: -1073740791
335572.09 WARNING: Error:


Here is what I am using:
Windows 10
Simvascular 2019.06.24 (MITK 2018.4.2 VTK 8.1.1 ITK 4.13.2 Qt 5.11.3)

I was also wondering about the units for some of the 1D parameters since I could not find them on the documentation. Here are the parameters I have questions about:

Inlet and Outlet BCs
Rp, C, Rd
Resistance (for BC Type Resistance)

Thank you for your time.

[David Parker]

Hi Frank,

The Exit code: -1073740791 error seems to mean that the 1D solver crashed with a memory problem. Could you post here the solver.inp file?

The 1D solver parameters are in CGS units. See http://simvascular.github.io/docs1DSimulation.html.

Cheers,
Dave

[Frank Servin]

Thank you Dave,

Attached is the 'solver.in' file

[David Parker]

Hi Frank,

I tried to run the 1D Solver on your solver input file solver.txt file at got

Code: Select all

   iter: 30 normf: 0.0249757 norms: 1.38765e-05 time: 0.001683
    iter: 31 normf: 0.0145429 norms: 8.08008e-06 time: 0.001683
Error: Newton not converged, exceed max iterations
norm of Flow rate:0.0145429, norm of Area:8.08008e-06
  Time = 0.01, Mass = -143.448, Tot iters = 31
    iter: 0 normf: 502627 norms: 130.597 time: 0.001687
    iter: 1 normf: 110230 norms: 30.9673 time: 0.001683
    iter: 2 normf: 68977.3 norms: 20.5454 time: 0.001685
    iter: 3 normf: 68724.5 norms: 50.9124 time: 0.001683
    iter: 4 normf: 21399.1 norms: 6.53327 time: 0.001683
    iter: 5 normf: 1765.14 norms: 1.11604 time: 0.001692
    iter: 6 normf: 372.741 norms: 0.26683 time: 0.001692
    iter: 7 normf: 88.8919 norms: 0.067705 time: 0.001708
    iter: 8 normf: 22.0029 norms: 0.0153841 time: 0.001684
    iter: 9 normf: 4.93594 norms: 0.00342103 time: 0.001686
    iter: 10 normf: 1.13368 norms: 0.000846806 time: 0.001681
    iter: 11 normf: 0.283049 norms: 0.000212648 time: 0.0017
    iter: 12 normf: 0.0688813 norms: 4.80043e-05 time: 0.001683
    iter: 13 normf: 0.0153229 norms: 1.04598e-05 time: 0.001684
    iter: 14 normf: 0.00345055 norms: 2.55939e-06 time: 0.001817
    iter: 15 normf: 0.00086292 norms: 6.58652e-07 time: 0.001773
    iter: 16 normf: 0.000215062 norms: 1.52434e-07 time: 0.001728
    iter: 17 normf: 4.85173e-05 norms: 3.25985e-08 time: 0.001701
    iter: 18 normf: 1.06222e-05 norms: 7.69666e-09 time: 0.00169
  Time = 0.02, Mass = -261.665, Tot iters = 19
    iter: 0 normf: 282933 norms: 139.797 time: 0.00169
    iter: 1 normf: 280571 norms: 104.523 time: 0.0017
    iter: 2 normf: 954554 norms: 123.59 time: 0.001681
Calculated a NaN for the residual.
Terminated.

Looking at the solver.txt file

Code: Select all

SEGMENT Group0_Seg0 0 155.25408096512353 1553 0 1 539.5504441356408 6.507825787297944 0.0 MAT1 NONE 0.0 0 0 RCR RCR_0

I see the segment area going from 539 to 6.5. That is a huge reduction in area and is causing problems in the nonlinear solver. We typically see a few percent reduction in areas along a segment. I'm not sure why you want to model such a large area difference but if you really need to do this then you could try to divide your domain into several segments, not sure if that will work though.

Cheers,
Dave