Page 1 of 10

IIT Madras Student Satellite Project Indian Institute of Technology Madras

Interaction module experiments in Abaqus Standard

by Rakesh Ramakrishnan, Ajayraj, Lokaditya, Ananth Saran iitmsat-H-CRR,RA, RL, ASY-2-Feb-13 3 Feb 2013

1 Background: Any 3-D model in real life contains various individual surfaces in contact with each
other during its assembly. The property displayed by such a model depends on how the various surfaces interact with each other. For example, in a rotating fan: the blades are rigidly fixed (tied) to the main rotating body, The rotor is fixed rigidly to the motor but also has low friction in the tangential direction to avoid high damping. Abaqus/Standard provides us with a plethora of interaction properties which can be used to model various kinds of interactions. In addition to this, different methods of enforcing such interactions are present which yield different results each time. In this we use a dummy model to model and understand the effect of various constraints and hence understand which to use 2 Model used:

These 2 plates are given different interaction properties. The sides are " clenched" so as to form a realistic version which can even be experimented with. Just having two blocks introduces possibility of them flying apart which makes the simulation exit with an error. The red and blue dots indicate locations of boundary conditions which are the regions which are rigidly fixed throughout the simulation. The square region on top is where the load is applied and the direction of load is shown by the arrows. In the above shown case, the pressure load is downward-into.

Page 2 of 10 3 Experiments: Experiment 1: Interaction Interaction Allow separation after contact? Yes - ticked Load Pressure Direction Pressure Overclosure Hard Contact Master Surface Top Result

IIT Madras Student Satellite Project Indian Institute of Technology Madras

Constraint Enforcement Method Penalty - Linear Slave surface Bottom Inference / Additional Remarks

Stiffness (if any) 10 Relative meshing Slave finer mesh Should we use it for our model? No

Downward - Into

Penetration into Slave, Umax = 1.123x10-3m

Distinct Penetration observed even with slave having a finer mesh

Figure 1

Experiment 2: Interaction Interaction Allow separation after contact? Yes - ticked Load Pressure Direction Pressure Overclosure Hard Contact Master Surface Top Result Constraint Enforcement Method Penalty - Linear Slave surface Bottom Inference / Additional Remarks Stiffness (if any) 10 Relative meshing Equal meshing Should we use it for our model? No

Downward - Into

Penetration into Slave, Umax = 1.121x10-3m

Slightly lesser penetration than expt 1 contradicting previous knowledge

Page 3 of 10

IIT Madras Student Satellite Project Indian Institute of Technology Madras

Figure 2 Experiment 3: Interaction Interaction Allow separation after contact? Yes - ticked Load Pressure Direction Pressure Overclosure Hard Contact Master Surface Top Result Constraint Enforcement Method Penalty - Linear Slave surface Bottom Inference / Additional Remarks Stiffness (if any) 10 Relative meshing

Master finer mesh - ratio = 20

Should we use it for our model? No

Downward - Into

Penetration into Slave, Umax = 1.805x10-3m

More penetration than expt 1 - Follows previous knowledge

Figure 3

Page 4 of 10 Experiment 4: Interaction Interaction Allow separation after contact? Pressure Overclosure Hard Contact Master Surface Top Result

IIT Madras Student Satellite Project Indian Institute of Technology Madras

Constraint Enforcement Method Penalty - Linear Slave surface Bottom Inference / Additional Remarks

Stiffness (if any)

1 x 1010
Relative meshing

No - Unticked
Load Pressure Direction

Equal Meshing
Should we use it for our model?

Up - Outward

No Upward displacement where no load is applied, sides warped; Umax = 1.2x10-6m

Does strict enforcement of no separation mean more stiffness?

Yes

Figure 4

Experiment 5: Interaction Interaction Allow separation after contact? Pressure Overclosure Hard Contact Master Surface Top Result Constraint Enforcement Method Penalty - Linear Slave surface Bottom Inference / Additional Remarks Stiffness (if any)

10
Relative meshing

No - Unticked
Load Pressure Direction

Equal Meshing
Should we use it for our model?

Up - Outward

Upward displacement like a an arch, Umax = 1.121x10-3m

Extent to which separation allowed dependent on k??

Maybe dependence on k unexpected

Page 5 of 10

IIT Madras Student Satellite Project Indian Institute of Technology Madras

Figure 5

Experiment 6:
Interaction Interaction Allow separation after contact? Pressure Overclosure Hard Contact Master Surface Top Result Constraint Enforcement Method Penalty - Linear Slave surface Bottom Inference / Additional Remarks Stiffness (if any)

1 x 1010
Relative meshing

Yes - Ticked
Load Pressure Direction

Equal Meshing
Should we use it for our model?

Up - Outward

Upward displacement like a an arch, Umax = 1.121x10-3m

When there is separation, there is NO FORCE indicating non-linearity in system

No

Figure 6

Page 6 of 10

IIT Madras Student Satellite Project Indian Institute of Technology Madras

Experiment 7:
Interaction Interaction Allow separation after contact? Pressure Overclosure Hard Contact Master Surface Top Result Constraint Enforcement Method Stiffness (if any)

Default
Slave surface Bottom Inference / Additional Remarks

N/A
Relative meshing

No - Unticked
Load Pressure Direction

Equal Meshing
Should we use it for our model?

Up - Outward

No Upward displacement where no load is applied, sides warped; Umax = 5.2x10-6m

Umax higher than expt 4 expected since linear enforcement more strict than default enforcement

Yes - 4 is better

Output similar to figure 4 Experiment 8:

Interaction Interaction Allow separation after contact? Pressure Overclosure Hard Contact Master Surface Top Result Constraint Enforcement Method Stiffness (if any)

Default
Slave surface Bottom Inference / Additional Remarks

N/A
Relative meshing

Yes - Ticked
Load Pressure Direction

Equal Meshing
Should we use it for our model?

Up - Outward Experiment 9:
Interaction Interaction Allow separation after contact?

Upward displacement like a Similar to expt 6 - expected an arch, Umax = 1.120x10-3m Output similar to figure 6

No

Pressure Overclosure Hard Contact Master Surface Top Result

Constraint Enforcement Method

Stiffness (if any)

Default
Slave surface Bottom Inference / Additional Remarks

N/A
Relative meshing

No - Unticked
Load Pressure Direction

Equal Meshing
Should we use it for our model?

Downward - Into

Very minimal Penetration. Extra force transmitted to sides which warped outward

Exact result we were looking for - Confirms that not allowing separation removes penetration issues

Yes

Page 7 of 10

IIT Madras Student Satellite Project Indian Institute of Technology Madras

Figure 9

Experiment 10:
Interaction Interaction Allow separation after contact? Pressure Overclosure Constraint Enforcement Method Stiffness (if any)

Linear
Master Surface Top Result

Default
Slave surface Bottom Inference / Additional Remarks

1 x 1010
Relative meshing

N/A
Load Pressure Direction

Equal Meshing
Should we use it for our model?

Up - Outward

Upward displacement like a an arch, Umax = 1.121x10-3m

Same as expt 6 - Introduces non linearity

No - Never use linear

Output same as figure 6

Experiment 11:
Interaction Interaction Allow separation after contact? Pressure Overclosure Constraint Enforcement Method Stiffness (if any)

Linear
Master Surface Top Result

Default
Slave surface Bottom Inference / Additional Remarks

1 x 1010
Relative meshing

N/A
Load Pressure Direction

Equal Meshing
Should we use it for our model?

Downward - Into

Penetration observed in the load bearing regions, Umax = 5.9x10-6. Sideswarped upwards!

Different result! Unexplained behaviour

No - Never use linear

Page 8 of 10

IIT Madras Student Satellite Project Indian Institute of Technology Madras

Figure 11 Experiment 12:

Interaction Pressure Overclosure Constraint Enforcement Method Stiffness (if any)

Axial Connectors, Linear Uncoupled, F1 ticked

Allow separation after contact? Master Surface Top Result Slave surface Bottom Inference / Additional Remarks

1 x 1010
Relative meshing

N/A
Load Pressure Direction

Equal Meshing
Should we use it for our model?

Downward - Into

Penetration present only in circular region bounded by connectors, Umax = 8.949x10-4m

Connectors are working Absence of interaction caused penetration

Yes, try with interaction

Figure 12

Page 9 of 10

IIT Madras Student Satellite Project Indian Institute of Technology Madras

Experiment 13:
Interaction Pressure Overclosure Constraint Enforcement Method Stiffness (if any)

Axial Connectors, Linear Uncoupled, F1 ticked

Allow separation after contact? Master Surface Top Result Slave surface Bottom Inference / Additional Remarks

10
Relative meshing

N/A
Load Pressure Direction

Equal Meshing
Should we use it for our model?

Downward - Into

Penetration present only in circular region bounded by connectors, Umax = 1.149x10-3m

More penetration than expt 12 - expected

Yes, try with interaction

Figure 13 Experiment 14: Used a non constrained primitive model comprising only of 2 plates
Interaction Interaction Allow separation after contact? Pressure Overclosure Constraint Enforcement Method Stiffness (if any)

Hard Contact
Master Surface Top Result

Penalty - Linear
Slave surface Bottom Inference / Additional Remarks

1 x 1010
Relative meshing

No - Unticked
Load Pressure Direction

Equal Meshing
Should we use it for our model?

Downward - Into

Each plate rotates

Unexplainable result

No

Page 10 of 10

IIT Madras Student Satellite Project Indian Institute of Technology Madras

Figure 14

4 Conclusions: The linear type of interaction is never to be used. We now use a hard type of pressure
overclosure with penalty type of enforcement method and use a stiffness of 10 N/m for simulating hard 6 2 contacts and a stiffness of 10 N/m for simulating rubber like soft normal interactions. Also, the finer mesh surface is assigned the slave surface to minimise any possible penetration as penetration is not at all avoidable due to the approximate nature of the solutions.
10 2