Questions about prestress simulations and assesing simulation performance
Posted: Thu Jan 18, 2024 1:46 pm
Hello,
I am following the workflow outlined in this example: https://github.com/SimVascular/svFSI-Te ... _prestress to conduct svFSI prestress simulations. I am conducting a rigid, prestress, and then final FSI simulation according to the workflow. For the final overall simulation, I want to run it for 5 cardiac cycles, or 5000 timesteps. However, this is a rather lengthy simulation to run, so i was curious: if I want the final fsi ALE simulation to run for 5000 timesteps, must the preceding simulations (rigid and prestress) also run for 5000 timesteps? In the linked example, I noticed that the rigid and prestress simulations ran for much less timesteps as compared to the final svFSI ALE simulation which ran for 3000 timesteps, but I was curious as to whether this creates incongruity between the prestress results and then final simulation which runs for more cardiac cycles.
Next, in the literature I noticed that they run simulations for 5-6 cardiac cycles to allow for stabilization, and then use the results from the final cardiac cycle in their post-processing analysis on paraview. By what mechanism can I convert/extract results for only the final cardiac cycle? For reference, I am running the simulation through the command line and not through the GUI.
Lastly, I was curious about how I can validate the results of my simulations/ensure that my simulations are converging correctly and generating reasonable predictions. Is there anything I can look for in my histor.dat file, or in my simulation results (velocities, pressures, etc), to validate the outputs? Thanks so much.
Joanne
I am following the workflow outlined in this example: https://github.com/SimVascular/svFSI-Te ... _prestress to conduct svFSI prestress simulations. I am conducting a rigid, prestress, and then final FSI simulation according to the workflow. For the final overall simulation, I want to run it for 5 cardiac cycles, or 5000 timesteps. However, this is a rather lengthy simulation to run, so i was curious: if I want the final fsi ALE simulation to run for 5000 timesteps, must the preceding simulations (rigid and prestress) also run for 5000 timesteps? In the linked example, I noticed that the rigid and prestress simulations ran for much less timesteps as compared to the final svFSI ALE simulation which ran for 3000 timesteps, but I was curious as to whether this creates incongruity between the prestress results and then final simulation which runs for more cardiac cycles.
Next, in the literature I noticed that they run simulations for 5-6 cardiac cycles to allow for stabilization, and then use the results from the final cardiac cycle in their post-processing analysis on paraview. By what mechanism can I convert/extract results for only the final cardiac cycle? For reference, I am running the simulation through the command line and not through the GUI.
Lastly, I was curious about how I can validate the results of my simulations/ensure that my simulations are converging correctly and generating reasonable predictions. Is there anything I can look for in my histor.dat file, or in my simulation results (velocities, pressures, etc), to validate the outputs? Thanks so much.
Joanne