Yade

Uniaxial compression calibration parameters are not applicable to triaxial compression

Asked by Ziyu Wang

Hello,

I simulated a deep granite sample with a peak stress of about 120MPa and a peak strain of about 0.015.
After a lot of debugging, I selected a set of the most appropriate parameters and obtained the uniaxial compression stress-strain curve consistent with the test.
However, when I use the same parameters for triaxial compression simulation, I find that the parameters are not applicable. The problems are as follows:
The triaxial compression simulation step is to apply 10MPa confining pressure to achieve equilibrium, and then apply a certain axial strain rate.The termination condition I set is that the axial strain reaches 0.03(I think is enough..).
However,I found that the stress is only about 20MPa and the stress-strain curve keeps a linear rise...

How should I adjust the parameters?I think the ideal result should be that the stress reaches more than 100MPa and the stress-strain curve decreases significantly when the strain reaches about 0.015.

Thanks for help!

My parameters:
O.materials.append(JCFpmMat(type=1,density=2640,young=20e9,poisson=0.1,tensileStrength=25e6,cohesion=50e6,frictionAngle=radians(60),label='sphere'))

Question information

Language:
English Edit question
Status:
Solved
For:
Yade Edit question
Assignee:
No assignee Edit question
Solved by:
Ziyu Wang
Solved:
Last query:
Last reply:
Revision history for this message
Jan Stránský (honzik) said : 		#1

Hello,

> I selected a set of the most appropriate parameters and obtained the uniaxial compression stress-strain curve consistent with the test.
> However, when I use the same parameters for triaxial compression simulation, I find that the parameters are not applicable.

This is expected (or at least not surprising).
If you set suitable parameters for one type of test, the model response for a different type of test might be (and usually is) not good.

> How should I adjust the parameters?

If you want your parameters to be good for both uniaxial and triaxial conditions, you should fit them simultaneously.
I.e. select a set of parameters, run both uniaxial and triaxial test, and evaluate both results together.
According to the results, difference from expected behavior etc. you can try a different set of parameters.
E.g. using a trial-and-error approach or some more sophisticated optimization approach, neural network, genetic algorithms.........

Fitting model parameters is whole separate science branch, as I see it.
Search/google "fitting model parameters", "model parameters selection" or similar.

Cheers
Jan

Revision history for this message
Jan Stránský (honzik) said : 		#2

Clicked "Add Information Request" by mistake instead of "Propose Answer"

Revision history for this message
Ziyu Wang (ziyuwang1) said : 		#3

Thanks Jan,that solved my problem.

Revision history for this message
Luc Scholtès (luc) said : 		#4

Hello Ziyu Wang,

If you use the JCFPM model, you could have a look at this reference [1] where a calibration procedure is presented (section 4). It should help you determine a good set of parameters to match a given set of emergent properties.

Luc

[1]: https://www.sciencedirect.com/science/article/pii/S0022509612002268?via%3Dihub

Revision history for this message
Ziyu Wang (ziyuwang1) said : 		#5

Thanks Luc for your suggestion!