you are here:   / News & Insights / Engineering Advantage Blog / A Top Ten List for Explicit Dynamics Analysis – Part 1

Engineering Advantage

A Top Ten List for Explicit Dynamics Analysis – Part 1

November 1, 2016 By: Steven Hale

I’ve taught quite a few classes on explicit dynamics over my career and I often get the following question at the end of the training: “Do you have a checklist of steps and best practices we could use to make sure we haven’t missed an important step?”..  This is a great question because there are a number of modeling steps that are unique to explicit dynamics that aren’t obvious even to seasoned analysts with years of experience running implicit analyses. And while many of these steps are critical to a successful analysis, most preprocessors won’t automatically provide a warning or error if you forget them.

In this blog, I’ll provide the first part of my top ten list of best practice steps for setting up robust, fast, and accurate explicit dynamics models.

Best Practice Steps for Explicit Dynamics Analysis:

1. Create or import geometry.
a. Defeature geometry: remove geometry and features that are not required to achieve an accurate solution in the critical region of the model.

2. Select an appropriate element type.
a. Use mostly reduced-integration, lower-order element types.

  • Full integration can be used to improve accuracy and eliminate hourglassing but these elements are generally avoided because they are susceptible to shear locking and will increase solver time.
  • Some explicit dynamics codes allow for higher-order elements but they can take much longer to solve, primarily because the resulting time step is much smaller.

b. Avoid tetrahedral elements created from degenerate bricks.  These element types are very inaccurate.  Most explicit codes have more accurate tetrahedral element types such as ELFORM = 10 or 13 in LS-Dyna. A comparison between a tetrahedral mesh and an all-brick mesh for a Taylor Bar impact model is shown in Figure 1 (above). More information on this topic can be found in my blog titled “Should Tetrahedral Elements Be Avoided in Explicit Dynamics Analysis?”.

c. Select a shell element type that is both fast and accurate. In many cases, the default shell type selected by a preprocessor is not very accurate, especially for warped shells and shells that will experience torsional deformations. In LS-Dyna for example, the Belytschko-Wong-Chiang and the Belytschko-Leviathan types are more accurate under these conditions than the default Belytschko-Tsay type.

3. Assign material models and properties.

a. Use rigid materials to assign rigid bodies wherever you are not concerned about stresses/strains in that body.  Rigid bodies can significantly reduce run time.
b. Use consistent units.  Some preprocessors do not allow you to change unit systems.
c. For material plasticity, a piecewise linear plasticity model is typically the most general because it allows you to enter stress-strain points directly and, in some cases, to enter different curves at different strain-rates (to include strain-rate sensitivity).  More information about strain-rate sensitivity can be found in my blog titled “The Importance of Including Strain Rate Effects in Explicit Dynamics Material Models”.

4. Assign controls to suppress hourglass modes.

a. Avoid applying loads and constraints to isolated nodes.
b. Increase the mesh refinement in contact zones.
c. Add hourglass controls.  These typically do an excellent job of suppressing hourglass modes.  Commonly-used types include the Flanagan-Belytschko viscous or stiffness forms, or the Belytschko-Bindeman method.  Please refer to my earlier blog titled “Why Worry About Hourglassing in Explicit Dynamics?” for more information. Figure 2 shows a corner impact analysis both with and without applied hourglass controls.
 

Figure 2: Corner impact without hourglass control (left) and with hourglass control (right)



5. Assign damping as needed.

  1. Damping can be used to reduce or eliminate undesirable oscillations, especially in quasi-static analyses.
  2. In some explicit dynamics codes, such as LS-Dyna, Rayleigh damping (mass and stiffness-weighted damping) is the primary means of applying damping to a model.  Mass-weighted damping is more effective at lower frequencies and is commonly-used in quasi-static analyses.  Stiffness-weighted damping is more effective at higher frequencies and represents solid, or material damping.
     

6. Assign section properties for beams and shells.

Now we are ready to proceed with the remaining four steps required to prepare the model.  These include meshing, applying loads and constraints, creating contact interfaces, and applying solution settings.  These best practice steps will be the subject of my next blog (“A Top Ten List for Explicit Dynamics Analysis – Part 2”).  Please let me know if you think I may have missed a critical step or best practice so far, and please relay any specific insight you may have about any of these steps.