Learning Forum

[G.Troulliotis]

Dear all,

I am trying to create a coupled cardiac electromechanical model using LS-DYNA. The electrophysiology solver requires very fine meshes (millions of elements) whilst the mechanical solver requires (approx. 50 thousand elements). This makes the cost of solving both physics on the same mesh prohibitivelly high. Is there an existing method in LS-DYNA to solve the two models in a co-simulation setting by mapping the fine electrical solution to the course mechanical mesh and vice-versa.

Thanks in advance   

[Jim Day]

To better understand your simulation, please state which LS-DYNA solvers are involved in the two coupled parts to the problem. Have you employed the implicit mechanical solver on the mechanical side? What is the time step and what controls that time step? I'm not familiar with the electrophysiology side, but I do notice there are multiple hits on dynalook.com when searching for electrophysiology. Perhaps your problem is addressed in one of those papers. Please check.

[G.Troulliotis]

Hi Jim,

I am currently using the Implicit EM_EP Solver and the Implicit Mechanical solver.

The systems of equations are coupled through MAT_295 which uses transmembrane potential (in my case) as  a switch to activate the active contraction module. As a result, I would assume that the mechanics solver would control the timestep. I looked through the articles on dynalook.com but none of them seem to mention anything about multigrid implementation of the electromechanical model (ie. solving the problem on two different meshes to avoid unnecessary computations on the mechanical side). I am asking if such approach is already implemented in LS-DYNA or whether such a method would need to be user-coded. Please let me know if you need any further clarification  

[Jim Day]

Comments from LS-DYNA developers in off-line discussion: 1. We are currently working on the capability of splitting a coarser mechanical mesh (used for the mechanics) given by the user into a finer EP mesh (used for EP) within the LS-DYNA simulation. The mapping mechanics-EP between the 2 meshes will be done automatically. 2. The subroutines required to do the mapping of the EP results to the [material] integration points are currently in development. Once they are finished and tested, we'll probably be ready to give you a development version to test. 3. Until developments mentioned above are ready to test, a compromise (coarser) EP mesh is suggested.