 |
 |
|
|
|
|
Forums /
NodeVolumeI'm using the semi-implicit bidomain (TstepSB?) to simulate activation from point stimulation. I'm using SCIRun? to build the matrices with the finite element method. The intra and extra conductivity tensors are defined at each element in the mesh. The domain is a regular mesh, so I'm using the 'DomainSPBfile?' module, then loading the two matrices into cardiowave. When calculating how much area a node owns in a regular triangular 2D mesh, I just calculate the area by the dx and dy spacing of the nodes, similar to example8. I then use this value to scale the area, with the 'scale_area' parameter, and it seems to work; being consistent with the finite differences method when using the domain123spb module. When I try 3D bidomain simulations, I take the amount of volume owned by a node by calculating it by the dx, dy, and dz node spacings (dx*dy*dz) in the mesh, then assign this value to 'scale_volume'; but this relationship doesn't seem to work. The simulated activating wave is too small. I've tried saving the node volumes using 'savevec.m', in matlab, but it had the same effect. I've got a 2x2x1 cm^3 regular grid of hexahedrals, with 100x100x50 nodes; giving dx=dy=dz=0.02 cm. So the volume of the hexahedral is 8.0e-6 cm^3. It seems like it needs to be smaller, by a factor of around 2 or 3. How would I calculate the node volume in this case? |