Yade

Manipulating blocks created using jointedCohesiveFrictionalPM

Asked by velimier

Hi,
I want to create model that represent highly jointed rock mass using jointedCohesiveFrictionalPM. At first I want to create rock block and then to add joints into that block. Since I have dense fracture network that I want to add I need to use spheres with small radii in order to be able to create even smallest subblocks that are created by joint intersections. But before i proceed to simulation I would like to rationalize whole model and to reduce overall number of spheres but to keep all subblocks. And here comes my question: Is it possible to determine what spheres belong to each subblock? Or to transform subblocks into clumps and then replace spheres with small radii with ones with larger radii?
I'd appreciate any suggestions.

Veljko

Question information

Language:
English Edit question
Status:
Answered
For:
Yade Edit question
Assignee:
No assignee Edit question
Last query:
Last reply:
Revision history for this message
Jérôme Duriez (jduriez) said : 		#1

Hello Veljko,

The example script examples/jointedCohesiveFrictionalPM/testingJoint.py may be related to your question, though in a much simplified manner. This script is a model of two blocks separated by one rock joint.

Each block includes obviously a lot of particles that are clumped together, turning both blocks into rigid bodies. Here, the spheres are identified as belonging to one or another block directly from their position... If zSphere > zJoint then sphere belongs to block 1 else sphere belongs to block 2...
Unfortunately there is no general method that comes to my mind to identify subblocks in a general manner. You would have to work on the DFN surfaces to investigate what closed volumes they define, I guess. I never did it..

Your 2d question seems to me deeply related to the 1st one. In order to transform subblocks into clumps, you need first to identify what spheres belong to which subblock...

Jerome

Revision history for this message
velimier (velimier) said : 		#2

Jerome,
thank you for your answer. I forgot to mention that I already saw example in ~/testingJoint.py. Here I have a lot more complicated situation since I have thousands of intersecting joints. Obviously I'll have to work on this.
Thank you once again!

Revision history for this message
Luc Scholtès (luc) said : 		#3

Hi Veljko,

After the first identification stage (DFN over sphere packing with lots of particles), you could try to identify the cohesive blocks (the ones in between the discontinuities) according to the status of the particles' interaction with their respective neighbours (cohesive or not). This can work only if you define the jointed interactions as non cohesive of course.

For instance, looping over all bodies (for o in O.bodies), you loop then over all their neighbours (using O.interactions.withBody(o) ) and find those with which the interaction is cohesive. If the interaction is cohesive, they belong to the same subblock. If not, they don't. Doing so, you should be able to identify the subblocks.

This is what I did to obtain the last figure of this paper: http://www.sciencedirect.com/science/article/pii/S1631072114002034

Afterwards, I did not get what you want to do precisely. I guess that replacing the particles contained in the subblocks by clumps is quite straightforward (using what Jerome proposed). You would then end up with a model that would be similar to what 3DEC proposes, with blocks presenting rough surfaces. Now, replacing these clumps by sphere assemblies... I am not sure it would be of interest... But, again, I did not get precisely your objectives.

Cheers

Luc

Revision history for this message
velimier (velimier) said : 		#4

Luc,
thank you for your answer. I'll try to use your suggestion to solve thi problem.
Regarding my objective, all I want to do is to reduce number of particles in some larger subblocks in order to optimize simulation.

Once again, than you for answering!

Veljko

Can you help with this problem?

Provide an answer of your own, or ask velimier for more information if necessary.

To post a message you must log in.