simtk.org

hello, everyone!
I am following the guide of SimVascular. However, i still not very clear about how to determine the value of the time step size and the number of the timesteps. i know there is a relationship between the time step and the mesh size through CFL number, but what is the procedure, firstly i should determine which number and how to determine. (when i increase the time step size, there needs more time to convergence ),
Also how to check if the result get stable. sometimes when the pressure is stable, but the velocity is still not,is it normal?
Thanks so much please.

Generally, before running a simulation, your mesh size and characteristic velocity (say mean velocity) are known. Then you can find a dt such that CFL<1. The number of time steps, ts, will be T/dt, where T usually is the time of a cardiac cycle. If you want to simulate n cycles, your total time steps will be ts*n. To check your results, usually you need to monitor the residual values printed in the histor.dat file. You can also plot pressure and flow at inlets and outlets and see if you have periodic outputs.

Thanks so much for your reply.
Please allow me to ask one more question further, For the mesh size,you say before a simulation we should know that, but how to get it. just take the vessel simulation for example, how could you determine the mesh size firstly.
Thanks for your patience to answer my question.

Hi Jianfei,

Sorry for the late response. Your mesh size is what you specify in the "Meshing" tab in SimVascular. We define mesh size to be the average length of an edge of a single tetrahedral element.

Generally, you want to choose your mesh size such that your simulation can adequately capture the pheneomena you are interested in. This is a vague statement, as it heavily depends on the anatomy you are simulating and the simulation results you are interested in. For example, our lab group is often interested in the wall shear stress, which is computed from the gradients of velocity. The accuracy of the computed velocity gradients depend heavily on the mesh spacing near the wall, so we often make sure we choose a sufficiently small mesh size to capture the gradients well.

On the other hand, accurate calculations of pressure often do not require a very fine mesh. So in cases where we primarily are only interested in pressure we select a larger mesh size to reduce the computational cost of running a simulation.

Hope that helps! Feel free to reach out if you have additional questions.