UCS Test using rigid blocks stress calculation

Hello everyone. I was trying to do an UCS Test using rigid blocks and voronoi. My main topic is to simulate the failure of slope models using voronoi model. I use rigid blocks for slope simulation. But to calibrate the kn, ks, c and friction value for the voronoi joints in slope model, I am trying to do an rigid block UCS test to get the joint values to incorporate in the slope model. Now for the UCS test, since I need to calculate the stress on the top plate and its a rigid model, I use subcontact history to get the stress curves and values. But it can be seen from the results that the stress suddenly occurred in the beginning steps and then it comes to zero, same for force curves. And due to this I cannot calculate the strain since it didn’t even deform (d/L). Can anyone tell me what should I do to get the correct stress values from the rigid blocks? I have given my code and pictures below. Thank you.

And the reason I am not using the elastic voronoi model UCS test example from the Itasca handbook is because, as I said my slope failure simulations are using rigid blocks, thus we think that using elastic blocks results from UCS test and then putting them in a rigid slope model will have different effect. Thus keeping the UCS test blocks rigid for similarity with slope model.




1.txt (626 Bytes)
2.txt (1.5 KB)
cyl100.zip (1.1 KB)

1 Like

your .DXF file is need to be able to run and see and comment,
‘cyl100.dxf’ :slight_smile:

The reason is quite simple actually: Upon loading your model, the cylinder contacts quickly fail due to low strength (tension is zero, cohesion also very low) and/or too high loading speed (plot “joint subcontact” to see the failure states). Therefore your sample disintegrates and subsequently the normal forces on the subcontacts drop to zero.

With very high strength values, the model runs without problems. Re-check your parameters and loading speed, and everything should be fine.

1 Like

Yes sorry I forgot to send the file. This is the dxf file. Thanks.
cyl100.zip (1.1 KB)

Hello Markus,

The thing is, I am trying to keep cohesion and friction angle small for this UCS Test since I need them small for the future slope simulation. I have already simulated the slope failure models and all of them are working fine, but need smaller voronoi joint friction angle and cohesion for faster and easier simulation and computation time. Thus, to verify those two parameters in the slope model, I am trying to use this UCS test with small c and phi values. I wanted to ask what other options can be done to get better normal forces (longer time until the specimen deforms) while keeping the cohesion and friction angle small. Thank you.

why do you define your zone properties only with density differences?

block property density 982 range group ‘material’
block property density 7850 range group ‘plate’
block property density 7850 range group ‘plate2’

This will define your stiffness, bulk and shear modulus as zero and therefore your stress and strain calculate is incorrect.

Hello Mohsen,

I only define densities because they are all rigid blocks. There are no zones.

I am (hopefully) misunderstanding this sentence. Your material parameters should reflect the properties/strength of the real material and are not something to play around with in order to save computation time.

Again, you appear to be asking “how can I make my material stronger without making it stronger?”. It just does not make sense. You might be able to get a slightly better result by using a smaller displacement velocity, so that the initial starting shock will trigger less tensile fractures, but that is about it.

Another option would be to add gravity to your model and initialize stresses according to the 30m large specimen. This way, you would not start out at zero contact stresses and will get much less tensile fracturing immediately after starting.

Other than these options, your model reflects what your parameters allow and your parametrization approach seems to be “interesting” to put it mildly. Also, it is bad practice to assign anything using the “range id” command, since that ID might change after the slightest change to your model setup, especially using voronoi blocks.

1 Like

Yes, my sentence meant different from what I wanted to say, sorry. Yes I meant like since many voronoi joints in a model makes the overall model more stronger than a normal single block model, I needed to reduce the cohesion and friction angle a bit, so that it reflects exactly as the same way/strength to the original material. (similar like intact rock and rock mass).

Thanks for the suggestions. I will try them now.