How to add outputs for instantaneous variables in output.f90 file

[chenjiaweiDr]

I'm trying to output some instantaneous variables, such as the velocity gradient. To achieve this, I modified the output.f90 file and noticed the function:
subroutine out1d_single_point_chan(fname,ng,lo,hi,idir,l,dl,dzc_g,dzf_g,zc_g,zf_g,u,v,w,p,visct)
I see that there are some space-averaging operations, such as:
buf16 = buf16 + tmp_x
buf17 = buf17 + tmp_y
buf18 = buf18 + tmp_z
If I want to output instantaneous variables, I assume these space-averaging operations should be removed. Is that correct?

[soaringxmc]

Your should be right if you want to first compute instantaneous quantities (such as gradients) for each point, then do spatial average in the uniform directions, and then output the statistics. However, if you want to just output the instantaneous quantities (such as gradients) for the whole field, and then use it for what you want. The postprocessing code
utils/read_binary_data/python/read_single_field_binary.py
may be helpful to you.

[chenjiaweiDr]

Thank you for your great work! The postprocessing script utils/read_binary_data/python/read_single_field_binary.py is very useful. I also found the single-point-stats-cavity.py file to be a helpful example. However, I have two questions regarding this script:

**### #
# u along z direction
#
data, xp, yp, zp, xu, yv, zw = read_single_field_binary('vex_fld_0005000.bin',iskip)
fname = resultsdir + 'stats-single-point-cavi-vertical-' + casename + fname_ext
np.savetxt(fname, np.column_stack((zp, 0.5*(data[63,63,:]+data[63,64,:])/2.0)), fmt='%16.6e', delimiter='')**

（1）Is there any documentation or explanation about the contents of vex_fld_0005000.bin, particularly regarding the meaning of the stored data?

（2）Why does 0.5*(data[63,63,:] + data[63,64,:])/2.0 represent the instantaneous velocity?

I realize that these two questions might be closely related, but I would appreciate any clarification.

Thanks in advance!

[soaringxmc]

Happy it is helpful to you! The case is a cavity with 128 cells in each direction. The u velocity along z at the center of the cavity should be calculated as 0.5*(data[63,63,:] + data[63,64,:])/2.0, since a staggered grid is used. I would suggest you to read the output code for the details, and it is simply the x velocity for (i, j, k)=(1:64, 1:64, 1:64). The useful postprocessing code read_single_field_binary.py is actually provided by @p-costa.

[soaringxmc]

It is to plot a figure similar to figure 2 in pedro's paper:
https://www.sciencedirect.com/science/article/pii/S089812211830405X

If you want to further compute the gradients after the instantaneous field is obtained, the grid spacing info can be accessed in grid.out. Only z direction may be useful, since it is uniform in both x and y.

[soaringxmc]

To be more clear, the following figure may help undetstand how we compute the u velocity along the cavity centerline

[chenjiaweiDr]

Thanks for the clarification. It seems that the CalES use staggered grid and is in forward, not backward. If we want to get velocity u at centroid [64,64,64], why not use 0.5*(data[63,64,:] + data[64,64,:]) ?

[soaringxmc]

Because we do not have u locations there.