Hello there,

I suggest you take a look at the following link.

https://www.rocscience.com/help/rs3/pdf_files/theory/Constitutive_models_Manual/6-Hoek-Brown_Model.pdf

in general, PLAXIS considers a fixed limit for the yield condition and it does not model the residual state of material. Depending on which behavior you want to model, you can fit your simulation either with the peak state or residual state by changing the parameters.

regards,

Emad

Hi Rachel

I believe that the issue is related to the constitutive model and the numerical framework that you are using to reproduce the mechanical behavior of a poly-crystalline rock (e.g. granite).

According to the literature, and my research experience, the DEM is more suitable for rock and rock-like materials (both in laboratory and field scale). In the case of brittle rock, for instance, a constitutive model such as parallel bond model or Flat-joint model is particularly promising. Using DEM, you will also be able to take into account the influence of micro-structural properties of rock (e.g. grain size, mineral content and so on) on macroscopic response, which will make your numerical simulations more realistic.

Please see the following papers:

Article A Bonded-Particle Model for Rock

Conference Paper A flat-jointed bonded-particle material for hard rock

Article A numerical approach to investigate the effects of rock text...

Best wishes

Mahdi

Dear Rachel,

I agree with the answers given by Mr.Mahdi that Hoek-Brown constitutive model is used to simulate the rock material in numerical model. But, you are looking for simulate the soil material, so, i feel you could have tried with either Mohr-Coulomb Model or Drucker - Prager model to simulate soil. I suggest you to use these model in alternate to Hoek-Brown constitutive model. Hopefully, you would get correct results by using other models..ok all the best in your research..

Hi Rachel,

I think this should be due to the fact that you have not well simulated the post-peak behaviour of the materials. Rock-like materials are normally showing some sort of strain hardening/softening behaviours. This has well been investigated in geo-mechanical engineering (for example for the simulation of the strength of coal pillars in bord and pillar mining method). Depending on the width to height ratio of your samples, they may show strain hardening/softening behaviours. If the results of your lab test show a drop in the strength, you shall use a strain softening model to replicate such behaviour. The modified MC-Softening/Hardening model is capable of simulating such behaviours. Generally speaking, the shear strength parameters of geo-materials (e.g., the internal friction angle and cohesion) (also sometimes their tensile strength) are linearly/or even non-linearly and gradually reduced as a function of plastic work (normally for the sake of simplicity plastic strain is used) to account for the gradual degradation of rock/soil parameters due to mechanical loadings.

The other important thing is that strain hardening/softening models are normally showing some mesh dependencies. Therefore, you may perform some sensitivity analyses to ensure the you have selected a proper mesh size and shape in your numerical modellings.

I hope you find these useful.

Best regards,

E.F. Salmi