Some functions (such as: Coordinate Initialization Method = String "box random") are still broken in windows build of Elmer8.2, but not in the nightly builds. I was using elmerfem-nightly_Windows-AMD64 downloaded on to run those tests. Is there some parameters that are not properly set in my test case or what I observed is expected in the current development stage of the module? In summary, I would like to know if there is something I could do to correctly model the interaction of particle with sidewalls. For example, the first particle on the left arrives straight down on the bottom wall, but reflects at 45o angle… Note that I removed the initial velocity in x for this test. However, particles are bouncing from their center (i.e., particle size is ignored). When activated, the particles start bouncing from all surfaces, including the circular wall. I found this BC keyword in the source code (it is not in the documentation…). Changing size of time steps does not appear to improve things. Contrary to what I observed for “Box Particle Contact”, there seems to be no spring value that provides physically plausible wall interaction. >0.5), the particles often “explode” (i.e., bounce with higher velocity than before the interaction). 0.1), the edge of bouncing particles cross significantly the wall boundary (as visible here). By changing the wall or particle spring value, it is possible to get one or more different particles to bounce (but not all of them). This “special” particle continues to bounce every time it hits a wall or the external box. When this keyword is active both in BC and in solver, I found that only one particle bounces off the sidewall. Note that, if “Wall Particle Collision” and “Wall Particle Contact” are activated together, the solver often crashes with segmentation fault when a particle reaches an external boundary or the circular wall. The keyword has apparently no effect as all particles stop at sidewalls. I could not find how to make this keyword working (despite activating it both in BC and Solver). At very high spring force (>1000), particles can “explode” (i.e., they are gaining energy at each interaction). Is this the expected behavior? At high spring force (>1), the particles seem to bounce from all the edges of the box correctly. Note however that, for high particle speeds, the wall interaction force is not strong enough to prevent the center of the particle from touching the wall and the particle is then stuck (as visible here for the 4 particles that reached the bottom of the domain). the interaction take longer and the center of the particle comes closer to the wall). 0.1 in this geometry), interaction is similar to a soft particle deforming during collision (i.e. Particles again have interaction only with the external box. This keyword also appears to be working mostly as expected. Note however that the “Wall Particle Bounciness” or “Particle Bounciness“ keywords do not seem to be taken into account here. Particle size is also correctly taken into account during the collisions. I assume that the absence of bouncing from the internal circular wall is expected here. When activated, particles are bouncing correctly from all sides of the external box but not from the circular wall. This keyword appears to be working mostly as expected. In this case all particles stop as soon as their center hits a sidewall (i.e., particle size is ignored). I'm using 2.73.Test summary.jpg (157.14 KiB) Viewed 3595 times Test 1: No collision or contact The object that is being duplicated for the particle system does not have it's renderable turned off, neither do any of it's parents. I will try to provide as much detail as I can. It just doesn't appear to render but works in the 3D view. I am having problems with my particle system.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |