Mesh Generation of Rock Fracture (To study flow in fracture)

  • Last Post 21 January 2019
gdimadis posted this 17 January 2019

Hi there to everyone

I am new in Workbench /cfd/ fluent

My problem goes as:

Imagine a clean saw cut in let say a rectangular sample of Marble, if we consider marble rock matrix as almost impermeable , we will have flow through the artificial fracture (in our case a rectangular of low height).

If we had to model this volume that fluid can flow through would be a rectangular of a low height (lets say less than a millimeter)  so our geometry would be this rectancular (could be made either design modeler or with a CAD (3d) software). And we would assign one of the faces as FLOW-INPUT the opossite as FLOW-OUTPUT, the rest would be WALL and by giving an entry velocity of eg LiquidWater in INPUT, and a PRESSURE in OUTPUT, fluent would solve it and the streamlines would be more or less parallel lines from input to output (assuming zero wall roughness and very low velocities ).

In reality i have a rock sample with fracture that was scanned with CT (micro computer tomography) and from stacked slices (images) i manged to segment a 3d structure, isosurface of which inserted in CAD software (rhino 5) after MEsh lab simplification (I had enormous surface mesh in the beggining so i reduced the amount of faces with MEshlab). Any way  the issue is that I managed to obtain a good surface mesh (no naked edges, many non-manifold edges) and with a plugin for Rhino from ITASCA - Griddle (UDEC 3DEC FLAC FLAC3d PFC etc) I managed to create volume mesh for ABAQUS (for ANSYS too *.CDB file but cant use it (seems there are only nodes)). and closed polysurface (nurbs).

After failing to use ABAQUS and ANSYS 3dMEshes, i am trying to use the polysurface geometry and use ANSYS for meshing.


Though the result says MESH failed ..and cant really understand why ... And if i can correct something in the geometry to finaly have a useful mesh.

MAIN OBJECTIVE in the whole task is to observe the streamlines in the fracture and find out how much of the available volume is used for the fluid to flow (there will be preferential flow paths with no dought and i want to make them vissible -appart from making some calculations on their percentage etc)..

I am afraid i made the whole thing a bit confusing but maybe the enclosed images help a bit...

Thanks in advance potential users of ANSYS that may propose solutions

With FEmodeler i insert the Abaqus mesh

Attached Files

Order By: Standard | Newest | Votes
gdimadis posted this 21 January 2019

HI rwoolou thanks for answering

I tryied to reply on Friday but there was something wrong with posting (or maybe i had too many captures/images on my reply)

rwoolhou posted this 17 January 2019

Looking at the model I suspect the mesh is failing because the gap is very narrow compared with the facet size. Can you increase the surface resolution?  Note, the Academic version has a 512k cell count cap so you may not be able to model the whole crack. 

gdimadis posted this 17 January 2019

how on earth multiple images/files  can be uploaded on this forum!!????

PArt of the crack


The whole sample


GEometry inserted from RHINO/cAD  as closed polysurface (NURBS)

MEsh FAILED with messages


side view of the model