RelaxationStepLevelSetCorrectionInner doesn't work for 3d multi-resolution

[WeiyiVirtonomy]

The relaxation using RelaxationStepLevelSetCorrectionInner produces weird particle distribution

Lattice:

After relaxation:

Gaps:

RelaxationStepInner doesn't have this problem:

There might be a bug in the 3d level-set shape of this beam

[Xiangyu-Hu]

@WeiyiVirtonomy yes. It seems like.
@Nafug22 could you have check on the implementation? Thanks.

[Xiangyu-Hu]

@WeiyiVirtonomy Could you make small test case for multi-resolution relaxation? Thanks,

[WeiyiVirtonomy]

@WeiyiVirtonomy Could you make small test case for multi-resolution relaxation? Thanks,

Sure. Would you like me to use this beam in the test, or would you prefer more complicated shapes?

We have a 3d teapot with surface refinement, and I wonder why that one works well. Could the problem lie in the shape generated by simtk?

[Xiangyu-Hu]

@WeiyiVirtonomy Just use this simple case.

[WeiyiVirtonomy]

The problem is now solved after fixing dW(h_ratio) of 3D and defining the level set refinement level correctly