Gordon’s Ramblings

The strategy laid out a few posts ago did not work. So I am diong this incrementally. Generate ls200000 on a coarse resolution (32×64x64) with zero levels. Input ls200000 into Avizo. Use a ROI to determine the shift in newxblob (which is a translation factor) in sci…f90 code. I have now set minima correctly.

xblob(1)=0.0
xblob(2)=0.0
xblob(3)=0.0
newxblob(1)=0.5
newxblob(2)=0.5
newxblob(3)=1.0
radblob=12.5

! try this ?

xblob(1)=0.0 ! will become y
xblob(2)=-1.0 ! will become -x
invertfactor(2)=-1.0
xblob(3)=0.0 ! will become z
newxblob(1)=0.5
newxblob(2)=1.0
newxblob(3)=0.5

!invertfactor(3)=-1.0
newxblob(1)=0.4
newxblob(2)=0.9
newxblob(3)=0.4

newxblob(1)=0.3
newxblob(2)=0.7
newxblob(3)=0.3
—–
Now I will setup the maximum sizes, which I would like to be: (1.3, 0.6, 0.6)
Do this by changing geometry_problem_hi in inputs.whale. This worked.
Now choose the ROI for the mouth: (.25 .45 .17 .43 .17 .43). Change this in inputs.whale, parameter ncoarseblocks, which states that anything outside this region has a tag=0 (no adaptivity). Anything inside is adapted based only on the level set.. set amr.max_level=2 (2 levels of refinement in the mouth only). Set amr.n_cell=(32,64,64) for now to check the code out.

It is already 8 min and still computing to generate the first level set file ls200000.
I stopped the code.

I am now experimenting with the plot format to avoid translating data to the finest level, which is quite impractical. Even if the output is restricted to the head of the whale, the refinement occurs over the entire grid.

test

Currently, the open mouth whale domain has bounding box B1 of (0 2 0 1 0 1). A smaller simulation box would have dimensions (..4 1.6 .2 .8 .2 .8). The whale itself has dimensions (0.54 1.4 0.345 0.66 0.374 0.59).
These dimensions were calculated by run3d/sci_clsvof.f90 and are not the original dimensions of the whale defined in the input file. To reduce the size of the box. Sussman recommends keeping the bounding box with minimum values of (0,0,0) and only changing maximum values (these are the values of geometry.prob_lo and geometry.prob_hi in inputs.whale. I should therefore shift the whale forward by a distance of (0.4,0.2,0.2) and set geometry.prog_hi=(1.2, .6, .6).

The mouth area is located (relative to B1) at (0.5 0.75 0.35 0.65 0.35 0.65).