Assignment 3

ASSIGNMENT 3

OBJECTIVE:

To compare and validate results of base simulation and simulation with improved shell element properties.

PROCEDURE:

CASE-1 (Default shell properties)

1. For the first case, use the .rad file of the last assignment. Make changes to the run time and animation steps in the engine file.
2. The animation steps is set to 45. (between 25 and 60)
3. Run time is changed to 55ms.
4. Save the file as FIRST_RUN.
5. Use the default shell properties and run the simulation.
6. After simulation check out the results from the output file, check whether the energy errors and mass errors are under acceptable limits.
7. The energy error is at -10.3% and mass error is 1.65*E-4. The acceptable limits for the energy error is up to -15% to 10% and mass errors up to 10%.Thus, they are under acceptable limits.
8. Open the .h3d file in hyperview and then plot the rigid wall forces, contact energy, total energy, hourglass energy, internal energy in hypergraph. Case-1 is plotted in blue.

CASE-2 (Recommended shell properties)

1. For the second case use the same rad file and import it into hypermesh.
2. Make changes to the shell properties by going into the model browser and open the properties folder. Click on the different components and check their card image and their shell properties following it. Change them to the recommended properties as per the question.

• Ishell=24
• Ismstr=2
• Ish3n=2
• N=5
• Ithick=1
• Iplas=1

3. Export the file as .rad and then save it in a separate folder.
4. Open the engine file and make changes to it.
5. Set the run time as 55ms and animation steps are set to 45.
6. Save the file as beam_elem.
7. Run the simulation on Radioss.
8. After simulation check out the results from the output file, check whether the energy errors and mass errors are under acceptable limits.
9. The energy error is around -0.1% and the mass error is 1.65*E-4.
10. Open the .h3d file in hyperview and then plot the rigid wall forces, contact energy, total energy, hourglass energy, internal energy in hypergraph. Case-2 is plotted in red.

RESULTS:

In the first model, the default shell element parameters (Ishell=1, Q4) being used induces the presence of hourglass energy owing to hourglass control. Since hourglass energy is not involved in the energy error calculations it results in higher negative energy error compared to the second case which has no hourglass energy, rather physical stabilization through artificial stiffening (QEPH Ishell-24).

Fig 1. Energy error calculations

This variation in hourglass energy is also seen in the variation of total energy over time for both models. The absence of hourglass energy in QEPH formulation (CASE-2), results in accurate computation of total energy which remains constant over time. Whereas in Q4 shell formulation (CASE-1) due to the presence of hourglass energy the total energy gets dissipated as hourglass energy increases.

The internal energy, kinetic energy and force calculations remain almost the same as the both the shell formulations follow reduced integration schemes and thus their results are expected to be almost same.

Fig 2. Comparison of various results. CASE-1-BLUE CASE-2 RED

Fig 3. Rigid wall forces CASE-1

Fig 4. Rigid wall forces CASE-2

However there is a variation in contact energy calculations. The contact energy is due to the presence of friction or sliding effects in the model. In the first model with default shell parameters, the onset of hourglass deformation modes results in greater contact surface area during deformation, greater energy losses due to friction. Whereas with the recommended shell parameters, hourglass control takes place through artificial stiffening and hourglass modes are suppressed initially leading to less contact area while deformation thus dissipating less contact energy.

Fig 5. Contact energy comparison CASE-1 BLUE (FIRST_RUN) CASE-2 RED (beam_elem)

Hourglass modes:

Hourglass modes are zero energy deformation modes which occur in under integrated elements. Excessive deformation of these elements in certain ways results in net zero strain calculation at the central gauss point resulting in unreliable results.

Fig 6. Hourglass deformation modes. (7-8-12 Translational and 9-10-11-12 Rotational)

Hourglass control:

There are two methods to control hourglass deformation modes.

Viscous method: In this method, large deformation in the element is initially detected and its rate of deformation is slowed down by adding a viscous force (artificial viscosity). The energy that gets used up in resisting this deformation is dissipative and is called Hourglass energy. This results in decrease of total energy (dissipative) and since hourglass energy is not involved in energy error calculations, hence outputs of large negative energy error occur.

Stiffness method: In this method, artificial stiffness is added at the material level and thus large deformations are resisted. Thus hourglass modes fail to occur and no hourglass energy gets developed. The elements are under-integrated and no hourglass energy results in a good combination of cost and accuracy. The QEPH (Ishell=24) shell formulation is based on this.

CONCLUSION:

Understood the significance of the shell element properties and formulations through a detailed comparison of the simulation results. Learnt the concept of hourglass deformation modes and its control by changing element properties without compromising on the results.