7. F modeling#

7.1. Characteristics of modeling#

We use a AXIS model. Four calculations are performed with different matching options or contact algorithms.

7.2. Characteristics of the mesh#

_images/100000000000036A00000321F96CDE73DCFEB23F.png

Knots: 376 knots.

Meshes: 30 TRIA3 and 324 QUAD4.

7.3. Tested sizes and results#

First calculation (nodal pairing, master-slave normal and algorithm “PENALISATION”)

Identification

Reference type

Reference value

Tolerance

_images/Object_345.svg

mesh \(\mathit{M31}\) knot \(\mathit{N291}(G)\)

“ANALYTIQUE”

-2798.3 \(N\)

7.0%

\(\mathrm{DX}\) knot \(\mathit{N291}(G)\)

“ANALYTIQUE”

0 \(\mathit{mm}\)

1.0E-10

\(\mathrm{DX}\) knot \(\mathit{N287}(H)\)

“NON_REGRESSION”

-0,110211 \(\mathit{mm}\)

\(\mathrm{DY}\) knot \(\mathit{N287}(H)\)

“NON_REGRESSION”

-0,162912 \(\mathit{mm}\)

\(\mathrm{DX}\) knot \(\mathit{N285}(I)\)

“NON_REGRESSION”

-0,165946 \(\mathit{mm}\)

\(\mathrm{DY}\) knot \(\mathit{N285}(I)\)

“NON_REGRESSION”

-0.629667 \(\mathit{mm}\)

Second calculation (normal master-slave, algorithm “PENALISATION”)

Identification

Reference type

Reference value

Tolerance

_images/Object_368.svg

mesh \(\mathit{M31}\) knot \(\mathit{N291}(G)\)

“ANALYTIQUE”

-2798.3 \(N\)

7.0%

\(\mathrm{DX}\) knot \(\mathit{N291}(G)\)

“ANALYTIQUE”

0 \(\mathit{mm}\)

1.0E-10

\(\mathrm{DX}\) knot \(\mathit{N287}(H)\)

“NON_REGRESSION”

-0,110678 \(\mathit{mm}\)

\(\mathrm{DY}\) knot \(\mathit{N287}(H)\)

“NON_REGRESSION”

-0,162865 \(\mathit{mm}\)

\(\mathrm{DX}\) knot \(\mathit{N285}(I)\)

“NON_REGRESSION”

-0,167194 \(\mathit{mm}\)

\(\mathrm{DY}\) knot \(\mathit{N285}(I)\)

“NON_REGRESSION”

-0.628961 \(\mathit{mm}\)

Third calculation (normal master-slave, algorithm “CONTRAINTE”)

Identification

Reference type

Reference value

Tolerance

_images/Object_391.svg

mesh \(\mathit{M31}\) knot \(\mathit{N291}(G)\)

“ANALYTIQUE”

-2798.3 \(N\)

7.0%

\(\mathrm{DX}\) knot \(\mathit{N291}(G)\)

“ANALYTIQUE”

0 \(\mathit{mm}\)

1.0E-10

\(\mathrm{DX}\) knot \(\mathit{N287}(H)\)

“NON_REGRESSION”

-0,110678 \(\mathit{mm}\)

\(\mathrm{DY}\) knot \(\mathit{N287}(H)\)

“NON_REGRESSION”

-0,162901 \(\mathit{mm}\)

\(\mathrm{DX}\) knot \(\mathit{N285}(I)\)

“NON_REGRESSION”

-0,167194 \(\mathit{mm}\)

\(\mathrm{DY}\) knot \(\mathit{N285}(I)\)

“NON_REGRESSION”

-0.628947 \(\mathit{mm}\)

Fourth calculation (smoothing, master-slave normal and algorithm “CONTRAINTE”)

Identification

Reference type

Reference value

Tolerance

_images/Object_414.svg

mesh \(\mathit{M31}\) knot \(\mathit{N291}(G)\)

“ANALYTIQUE”

-2798.3 \(N\)

7.0%

\(\mathrm{DX}\) knot \(\mathit{N291}(G)\)

“ANALYTIQUE”

0 \(\mathit{mm}\)

1.0E-10

\(\mathrm{DX}\) knot \(\mathit{N287}(H)\)

“NON_REGRESSION”

-0,110211 \(\mathit{mm}\)

\(\mathrm{DY}\) knot \(\mathit{N287}(H)\)

“NON_REGRESSION”

-0,162911 \(\mathit{mm}\)

\(\mathrm{DX}\) knot \(\mathit{N285}(I)\)

“NON_REGRESSION”

-0,165946 \(\mathit{mm}\)

\(\mathrm{DY}\) knot \(\mathit{N285}(I)\)

“NON_REGRESSION”

-0.629666 \(\mathit{mm}\)

7.4. notes#

On this modeling, we note that smoothing makes it possible to regain the symmetry of the problem (nodes perfectly facing each other once contact has been established), the last calculation in fact obtaining the same results as in nodal matching.

The values obtained are in good general agreement with the analytical solutions.