subdomain and cell volumes

Hello,

For some post-processing analysis, I need the volumes of the subdomains or alternately the individual cell volumes and a mapping of cells to subdomain.  I don't see volume info in the featool structure.  Is it stored somewhere or are there tools to easily do the volume calculation from the grid definitions?  I could do it manually for the subdomains but would prefer to use the model's own calculations of it for consistency and simplicity.  Thanks in advance for any tips.

Best,
Peter

peteri wrote

I need the volumes of the subdomains or alternately the individual cell volumes and a mapping of cells to subdomain.  I don't see volume info in the featool structure.  Is it stored somewhere or are there tools to easily do the volume calculation from the grid definitions?  I could do it manually for the subdomains but would prefer to use the model's own calculations of it for consistency and simplicity.

You can maybe use the pre-defined reserved variable "h_grid" for the mean cell diameter and subdomain integration defined as (https://featool.com/doc/physics#phys_coef_notes)

    h_grid :=  cell_volume^(1/n_sdim)

If that is not sufficient you would have to calculate it manually from the mesh information defined in (https://featool.com/doc/grid#grid_ref_mtrl):

    fea.grid.p  -  grid points/vertices
    fea.grid.c  -  grid cell connectivity
    fea.grid.s  -  subdomain indices

To correct the previous reply, the subdomain volumes can simply be gotten from:

    n_subd = max(fea.grid.s);
    for i_subd=1:n_subd
      vol_subd(i_subd) = intsubd( '1', fea, i_subd );
    end

or alternatively via "h_grid" and the cell volumes:

    n_sdim = size(fea.grid.p,1);
    p_centroid = squeeze(mean(reshape(fea.grid.p(:,fea.grid.c), [n_sdim,size(fea.grid.c)]), 2));
    v_cell = evalexpr( 'h_grid', p_centroid, fea ) .^ n_sdim;

    n_subd = max(fea.grid.s);
    for i_subd=1:n_subd
        vol_subd(i_subd) = sum(v_cell(fea.grid.s == i_subd));
    end

Thank you for the follow-up on this - I'll give it a try!

Another question came up regarding the FEAtool variables but I'll start a new thread.

Peter