'%20d='M364.9%20143.2v98.9l-1.4%201.2-165.6%20140.4%2035.7%2029.7v-34.2l1.4-1.2%20123.4-105.4%206.5-5.6v37l-1.3%201.2-99.7%2085.8v84.7l-6.6-5.7-97-83.1-3.6-3%203.5-3%20174.4-150.6v-25.1L153.7%20360.5l-6.6%205.7v-98.9l1.4-1.2%20161.6-138-35.7-29.7v34.4L273%20134%20153.7%20236.7l-6.6%205.7v-37l1.3-1.2%2097-83.2V36.3L252%2042l97.1%2083.1%203.6%203-3.6%203-171.7%20147.7V304l106.2-90.6%2056.4-48.8%2018.3-15.7z'%20/%3e%3c/svg%3e)
Every suffix isone of four things
A finished CFD case is a directory of unfamiliar extensions. Sorted by what they hold, the list gets short quickly.
Boundary conditions, solver settings, physical properties. Small, text, and the thing worth archiving.
The discretised domain. Expensive to build, so keep it even when you throw the results away.
Values per cell, per timestep. This is what fills the disk, and it grows with every write interval.
Extracted lines, surfaces and images. Small, and the only part you usually need to take home.
The ones thatcross packages
Three formats are worth writing deliberately, because anything can read them in five years.
CGNS is the standards-body answer for grids and solutions together. VTU and its siblings are what ParaView reads natively and what most post-processing assumes. PLOT3D is ancient, still everywhere in aerospace, and splits geometry into a grid file and solution into a .q file.
What each solverleaves behind
These are the ones you will actually be handed, usually without explanation.
An empty marker beside the case directory, so ParaView has something to open. Create it yourself.
Case and data, always written as a pair. One without the other is not a usable result.
Everything in one file: geometry, mesh, setup and results. Convenient, large, and not portable.
Model and results together in a single project file, tied to the version that wrote it.
Generic mesh files. Gmsh writes .msh, which most open-source solvers can be persuaded to import.
The OpenFOAM one confuses everybody: a case is a directory, not a file, and case.foam is an empty file that exists purely so ParaView has something to open. Creating it with touch is a normal and correct thing to do.
What to keep,and what to delete
Results directories grow faster than anything else in CFD, and disk is billed from the first run until you delete it.
Keep the case setup, which is small text and is what actually lets you reproduce the run. Keep the mesh if it took hours to build. Keep the timesteps you will plot. Everything else — intermediate write intervals, reconstructed copies of decomposed data — can be regenerated from the first two, and it is usually the bulk of what is sitting on the disk you are paying for.
ParaView isalready built
Nothing here needs compiling. Pick a machine, and these two commands put it on there.
$ dxflow workflow create --identity paraview hub://paraview
$ dxflow workflow start paraview