5. C modeling#

5.1. Characteristics of modeling#

Subsequently, we will only work with plane stress modeling.

Young’s moduli and Poisson’s coefficients remain the same. The value of the pressure on the edge of the outer ring is set to (\(p\mathrm{=}1.0E7\)) and a rotation of the inner ring is imposed in accordance with equation 1.6 for a \(N=10\) number of time steps.

The outer ring defines the master surface.

5.2. Characteristics of the mesh#

Idem modeling A.

5.3. Tested sizes and results#

The contact pressure (LAGS_C) is calculated for the node \(A\) with coordinates \((0.6\mathrm{,0}.0)\), the node that at the initial moment is located farthest to the right of the interface between the two rings. The calculated values are compared to the value obtained according to equation 1.5 for an external pressure of \(p\mathrm{=}1.0E7\). The rotation of the inner ring is applied. We look at the variations in external pressure during this rotation. It is tested when the meshes are again facing each other.

Identification

Reference

Aster

tolerance

LAGS_C at node \(A\) Integration diagram AUTO

\(\lambda =9.26E6\)

Analytics

\(\mathrm{0,6}\text{\%}\)

LAGS_C at node \(A\) Integration diagram GAUSS

\(\lambda \mathrm{=}9.26E6\)

Analytics

\(\mathrm{0,8}\text{\%}\)

5.4. Comments#

In addition to calculating the contact pressure, we are also interested in its variation over time. The incompatibility of master and slave meshes induces fluctuations in the value of this pressure. Figure 5.4-1 shows (in a somewhat rough way) this effect. One way to mitigate it is to use either a finer mesh or higher-order elements.

_images/Object_33.png

Figure 5.4-1: Fluctuation in contact pressure due to mesh incompatibility

To estimate the contact pressure as correctly as possible, during rotation, when the master and slave contact surfaces are no longer compatible, it is necessary to use integration schemes of the highest possible order with a substantial refinement of the mesh compared to the situation with compatible meshes in models A and B. The result is very much improved by using quadratic elements (see modeling G).