3. B modeling#

3.1. Characteristics of modeling#

Modeling B performs the same hydraulic calculation as modeling A, but analyze the stability of the slope according to the Mohr-Coulomb criterion by the Morgenstern-Price method assuming that the fracture surface is non-circular.

In order to quickly test the improved fireworks algorithm (EFWA), we take advantage of the result of the preliminary calculation to configure the initial state that is already optimal.

So by setting the authorized number of iterations (« ITER_MAXI = 1``) and the number of fireworks (N = 2), ordinary sparks (M = 2) and Gaussian sparks (MG = 1): it becomes almost impossible for the optimization result to become different than the initial state.

Since the mesh is coarse, we configure « MARGE_PENTE = 0.8`` which is relatively large in order to avoid the failure to create the mesh groups of slices when the fracture surface generated is close to the slope profile.

3.2. Tested sizes and results#

In the object evol_noli at the output of the macro-command we test the value of the slippage indicator (DX from field DEPL_NOEU) on node N391 close to the breaking surface and located outside the slippery part.

The result is shown in Tableau 3.

The values of the safety factors are also tested during the refinement of the mesh. and the geometric parameters contained in the table at the output of the macro-command.

Since the optimization algorithm (EFWA) for the MP and Spencer methods is stochastic, You can never control the arbitrarily generated fracture surfaces if you want test features EFWA.

The input result from which the optimization is resumed is already the optimal solution. So the FS result is controllable, while the coordinates of the intermediate points on the surface are not (small deviation around the optimal surface). Therefore, we do not test the ordinates of the intermediate points.

The results are shown in Tableau 4 and Tableau 5.

Table 3: Reference value at node N391 of the cham_defo (Modeling B)

Identification

Reference type

Reference value

Tolerance

X

“NON_REGRESSION”

0, 0

0.001%

Table 4: Reference values from the FS table (Modeling B)

Nb Refinement

Identification

Reference Type

Reference Value

Tolerance

0

FS

“NON_REGRESSION”

1.598078371690653

0.0001%

1

FS

“NON_REGRESSION”

1.60025990801935

0.0001%

Table 5: Reference values from the table of geometric parameters of the fracture surface