Usually numerical models don't make distinction between damage directions associated to tensile strains and to compressive strains, assuming that both are aligned with the direct strain which induces the damage process.
For plane problems, I estimate the crack propagation direction in the Mohr/Coulomb plane, by joining the pole of Mohr with the crisis point (the point, or the points, in which the circle of Mohr is tangent to the failure domain). The failure criterion I use is the Leon criterion, which has a parabolic shape. I do not use the criterion of Coulomb, since it does not fit the experimental data well for tensile failures. Moreover, the parabolic shape of the limit domain allows you to obtain different crack propagation directions in traction and compression.
The parabola of Leon has a curvature that, in the vertex of the parabola, is always greater than the curvatures of the circle of Mohr in uniaxial traction and pure shear. This means that you will have just one tangent point in uniaxial traction, pure shear, and for a shear/traction failure. Consequently, you will have just one direction of propagation in pure traction pure shear, or shear/traction failure. Things work differently in compression and shear/compression, where you will have two possible direction of propagation. This latter time, you must evaluate whether or not the constraint degree is the same along the two directions: if it is, you will have two propagation direction, while, if it is not, just one direction activates, the one for which the constraint degree is lower.
The paper where I deepened this matter better is:
https://www.researchgate.net/publication/235998942_Masonry_walls_under_shear_test_a_CM_modeling?ev=prf_pub
you can also take a look to this second paper:
https://www.researchgate.net/publication/235999584_MODELING_OF_THE_PULLOUT_TEST_THROUGH_THE_CELL_METHOD?ev=prf_pub
Article Masonry walls under shear test: A CM modeling
Chapter MODELING OF THE PULLOUT TEST THROUGH THE CELL METHOD
- Badges
- Science method