Contact controls

This problem contains basic test cases for one or more Abaqus elements and features.

This page discusses:

ProductsAbaqus/Standard

Contact stabilization

Elements tested

C3D8 CPE4

Features tested

Contact stabilization

Problem description

Contact stabilization can be used to control rigid body motions that may exist in a model before contact is fully developed. The option adds viscous damping in both the normal and tangential directions. By default, the damping is calculated automatically, but it is possible to modify the damping coefficient, the variation of the damping coefficient over the step, the range over which the damping works, and the ratio between normal and tangential damping. The controls specified with this option remain in effect until they are either changed by another contact controls procedure or reset to their default values. Contact stabilization can be defined for a specific contact pair or for the entire model. Further description of the stabilization controls can be found in Adjusting Contact Controls in Abaqus/Standard. In these tests various combinations of stabilization controls are tested in multistep analyses with multiple contact pairs.

The first group of analyses consists of six pairs of blocks that are pushed together in Step 1, subjected to tangential sliding in Step 2, and pulled apart in Step 3. The blocks are elastic, and the motion of the blocks is controlled with boundary conditions. Contact stabilization parameters are specified for the whole model and are overridden by different parameters for several individual contact pairs. The stabilization parameters vary from step to step. A restart file is written, and some restarts are made to test the restart functionality.

The second group of analyses consists of three blocks that are pushed together in Step 1, subjected to tangential sliding in Step 2, and pulled apart in Step 3. The blocks are elastic; and the top and bottom blocks are controlled with boundary conditions, whereas the middle block is completely free and held in place by contact stabilization. Different contact stabilization parameters are used for each contact pair. In addition, frictional properties are prescribed for one contact pair. This group contains two-dimensional and three-dimensional static analyses as well as a dynamic analysis.

Results and discussion

The results show contact damping pressures CDPRESS as well as contact damping shear stresses CDSHEAR1 and CDSHEAR2 that are in agreement with expectations. In addition, in the second group of problems the rigid body motions of the middle block are controlled and no solver messages are observed.

Input files

controlsstab_3d.inp

Static analysis with six pairs of blocks and different control parameters.

controlsstab_restart1.inp

Restart from the results of the analysis with six pairs of blocks.

controlsstab_restart2.inp

Restart from the results of the first restart analysis.

controlsstab_free_2d.inp

Static analysis with two fixed and one free block in two dimensions.

controlsstab_free_3d.inp

Static analysis with two fixed and one free block in three dimensions.

controlsstab_dyn.inp

Dynamic analysis with two fixed and one free block in three dimensions.

Tangential contact controls

Elements tested

C3D20R C3D27R

Features tested

Modifying the tangential penalty stiffness for all contact pairs in a linear perturbation step.

Problem description

During linear perturbation steps, all points in contact (i.e., with a “closed” status) are assumed to be sticking if friction is present. However, stick conditions are not enforced for contact nodes for which a velocity differential is imposed by the motion of the reference frame or the transport velocity. Stick conditions are enforced with a penalty method by default, and the perturbation tangent scale factor can be used to scale the penalty stiffness. For example, setting this parameter to zero will result in zero penalty stiffness, such that the stick conditions are not enforced during the perturbation step. Setting this parameter to a value greater than unity results in a larger-than-default penalty stiffness and, thus, stricter enforcement of stick conditions during the perturbation step.

The model consists of two blocks of different sizes in contact, with a nonzero friction coefficient in effect. In the first and second general steps we establish contact and apply a tangential displacement boundary condition such that the small block slips along the larger block. Natural frequencies are computed in subsequent perturbation steps:

Step NamePerturbation Tangent Scale Factor
Frequency1 Not specified (default setting is 1.0)
Frequency2 Set to 1.0 (same as default)
Frequency3 Set to 0.0 (same as frictionless)
Frequency4 Set to 103

Material:

Young's modulus 2 × 107
Poisson's ratio 0.3
Friction coefficient 0.2

Results and discussion

Steps 3 and 4 (step names “Frequency1” and “Frequency2”) provide identical results, as expected. Step 5 (step name “Frequency3) has three zero-frequency eigenmodes corresponding to relative sliding between the two blocks, consistent with frictionless behavior. Strict enforcement of stick conditions is apparent in the eigenmodes for Step 6 (step name “Frequency4”).

Pressure-dependent contact controls

Elements tested

CPS4 C3D8

Features tested

Pressure-dependent contact constraint enforcement

Problem description

By default, during linear perturbation steps contact constraints are usually fully enforced for surfaces where contact is active independent of local normal pressure in the base state. This treatment of contact constraints may excessively stiffen the structure because certain areas of the model where the contact pressure is close to zero may actually come in and out of contact during vibrations. The nature of linear perturbation procedures does not allow such contact chattering. However, you can use contact pressure-dependent constraint enforcement to relax or even completely remove the constraints that have low pressure.

You can control the behavior through two user-specified contact pressure coefficients, p0 and p1; by default, p0 is equal to p1. For pressures greater than p1, the constraints are fully enforced, as in the default contact constraint treatment. For pressures between p0 and p1 (if p0 is less than p1), the constraints are gradually weakened. For pressures less than p0, the constraints are totally removed.

The model consists of a plane stress disk placed between two anvils. A sequence of linear perturbation frequency steps and general steps is applied. During the general steps the disk is compressed between the anvils, and the linear perturbation frequency steps compute eigenpairs with different parameter settings as shown in Table 1.

Table 1. Sequence of linear perturbation frequency and general steps.
Step Namep0p1
Frequency1 Not used Not used
General1 Not used Not used
Frequency2 Not used Not used
Frequency3 0. 5 × 104
Frequency4 5 × 104 5 × 104
Frequency5 0. 5 × 104
General2 Not used Not used
Frequency6 0. 0.
Frequency7 Not used Not used
Frequency8 1 × 1015 1 × 1015

Material:

Young's modulus 2 × 105
Poisson's ratio 0.3
Friction coefficient 0.1

Results and discussion

During the Frequency1 step minimum boundary conditions are imposed to compute the nonrigid body eigenpairs of the disk without any prestress. The General1 step compresses the disk between the anvils, so contact constraints are established with a maximum contact pressure value of more than 6 × 104. The Frequency2 step computes the eigenpairs of the stressed structure, and the Frequency2 spectrum increases compared to Frequency1. In the Frequency3 step p0=0 and p1=5 × 104, so some of the contact constraints are relaxed. Accordingly, the eigenvalues of Frequency3 decrease compared to Frequency2. In Frequency4 p0 = p1 = 5 × 104, so some constraints are removed completely, and the spectrum decreases further. Frequency5 has the same pressure-dependent constraint enforcement specification as Frequency3 but uses a local contact pair instead of a global specification, and the results of Frequency5 and Frequency3 are the same. The General2 step produces additional compression, so the eigenvalues of the next Frequency6 linear perturbation step increase compared to Frequency2. The Frequency6 p0=p1=0 results are coincident with the Frequency7 default constraint enforcement results. In Frequency8 both p0 and p1 are equal to a large number, so all contact constraints are removed. However, the Frequency8 results are not coincident with the Frequency1 results because of geometrically nonlinear effects in the loaded structure.

Two simple two-dimensional and three-dimensional two-element problems are also run using the same step sequence described above. The problems are geometrically linear, however, so the Frequency8 results are the same as the Frequency1 results.