you are here:   / News & Insights / Engineering Advantage Blog / A Top Ten List for Explicit Dynamics Analysis – Part 2

Engineering Advantage

A Top Ten List for Explicit Dynamics Analysis – Part 2

November 8, 2016 By: Steven Hale

In my previous blog, I laid out the first six best practice steps for setting up robust, fast, and accurate explicit dynamics models. In this blog, I will describe the remaining four steps. Please note that these are the general steps that apply to almost all explicit dynamics models. In special cases, additional steps will likely be required. An example of this would be for a blast analysis where an Eulerian domain could be included to model the explosion and a coupling method would be needed to simulate the interaction between detonation gases and the solids.

Best Practice Steps for Explicit Dynamics Analysis:

7. Mesh the geometry

a. Create a mesh with a relatively uniform distribution of element sizes. Locations of a model with a very fine mesh will drive down the time step which can cause very long run times. Figure 1 (shown above) compares preferred meshes for implicit and explicit models. Peak stresses can occur almost anywhere in a high energy dynamic analysis due to stress wave motion and interaction, so having a fine mesh in the fillet is not as critical in the explicit model.

Some meshing tools, such virtual topology and mesh-based defeaturing in ANSYS Workbench/LS-Dyna, provide options for meshing through geometric features. This means that the mesh does not have to conform to the boundaries of the surface geometry. An example of this is shown below in Figure 2 where the default mesh contains very small elements at the sliver faces seen in the geometry on the left. The relatively uniform mesh on the lower right is vastly preferable because it allows for much larger time steps.

Figure 2: Default and Preferred Meshes – Both Created from Geometry Containing Sliver Faces

b. Brick mesh as much as possible. Tetrahedral elements will not only add significantly to the model size but will usually drive down the time step significantly.

8. Apply initial conditions, loads, and constraints

a. Assign initial conditions such as initial translational and rotational velocities.

b. Smooth load curves (e.g. sine curves) will help prevent shocks.

c. List and/or plot loads for verification.

d. For quasi-static problems, use a faster velocity and a reduced transient to speed up the run time.  More information on quasi-static analyses can be found in my blog titled “How Can Explicit Solvers Help with Stubborn Nonlinear Statics Models?”.

9. Create and plot contact interfaces

a. For shell models, contact types that can detect contact on both sides of the shell are preferred. These include the automatic contact types in LS-Dyna.

b. Make sure to include all surfaces in the contact definition that have the potential to be involved in contact.

c. Use a contact type that can handle self contact, if needed. This would be necessary when a material folds over on itself, which often occurs during vehicle impact or metal-forming analyses. An example of this is shown in Figure 3 for a compressed C-channel.

d. Avoid initial penetrations if possible. They can generate spurious stresses, energy errors, and undesired alterations to surface geometry.


Figure 3: Compressed C-Channel Section Showing Self-Contact

10. Apply solution settings and solve

a. Specify the transient time.

b. Specify the output frequency. Enough output points will be needed to capture results at critical time points, but too many output points can generate excessive data.

  • If possible, write output points at a high frequency for just a small subset of nodes and elements in the critical region of interest.

c. Request output for critical results such as global and part energies, reaction forces, and contact forces. Energy output is critically important for error checking. For example, if the hourglass energy is too high, it suggests that too much energy is required to control hourglassing or hourglass modes are significant. Also, the initial energy plus the external work should always be nearly equal to the total system energy at all times.

d. Use mass scaling judiciously. Mass scaling is a tried and proven method for reducing run times in quasi-static analyses where the velocity is low and the kinetic energy is very small relative to the internal energy. If applied carefully under certain circumstances, it can also be used in truly dynamic analyses. For more details, please refer to my blog titled “What is Mass Scaling and When is it Appropriate in Explicit Dynamics Analysis?”.

So now you have a comprehensive list of steps needed to create robust, fast, and accurate explicit dynamics models. I’ve tried to provide extra detail for steps that are unique to explicit dynamics. I hope you find these steps useful as you either begin or continue to develop your expertise in this type of modeling. Again, please let me know if you think I may have neglected a critical step or if you have any experience or insight into any of these steps.