Numerical Modeling Approaches

Lagrangian and Eulerian Codes

Historically this difference in the approaches is due to the different approaches used for solid mechanics
and fluid mechanics. In solid mechanics, we always follow a particle tracking or lagrangian approach, that
is the Finite element mesh keeps on deforming with the structure. As Belytschko (1) puts it “Every
material point is compulsorily also a finite element node. ”Here the mesh continuously follows the
deformed structure and hence this approach has always some limitation with respect to the very large
deformation problems. In Fluid Mechanics, the attention is put on a control volume bound by control
surfaces and then one monitors the mass momentum and energy exchange across it. Essentially we solve
the same equation as that in a solid mechanics, but we put them in a solution suitable form also called
as conservation form. The conservation form has lot of practical applications in problems with shock
waves and if used in the integral form is the basis of Finite Volume Method. This approach is also called
the Eulerian or Field approach. Thus, the mesh remains fixed all the time in space and you don’t have
any severe or stringent limitations on the mesh quality.

The physics thus dictates the following requirements:

• Lagrangian meshes for slight to moderate deformation.
• Eulerian meshes for large deformation.

It is also possible to use another type of formulation called Arbitrary Lagrangian Eulerian (ALE)
formulation in which the nodes can be programmed to move arbitrarily. The nodes on the boundaries
are moved on the boundaries itself while the interior nodes are moved to minimize the mesh
deformation.

The difference can be clearly observed in the following figure. In Lagrangian meshes, element boundaries
remain coincident with boundaries and material interfaces. In Eulerian, they do not and then you will
have to use other approximate or tracking methods for the treatment of moving boundaries.
Mesh rezoning is also another approach which can be used with the projection of variables between the
meshes (similar to sub-modeling technique used in static analysis). But this is also a time-consuming
procedure and introduces spurious jumps in the state variable histories as it violates the conservation
laws. Zukas (1992) gives a detailed account of the applications of various Lagrangian and Eulerian codes
available for impact dynamics along with their detailed capabilities.

Smooth Particle Hydrodynamics (SPH)

The principal idea of SPH is to treat hydrodynamics in a completely mesh-free fashion, in terms of a set
of sampling particles. This is a mesh free method and there is no physical contact between the particles.
The behavior of SPH particles after an impact is shown in the image below. We can see that the object is
splashing as soon as it contacts with the floor. The typical examples of SPH formulations are bird strike
analysis, fuel sloshing etc. Smoothed particle hydrodynamics is a remarkably versatile and simple
approach for numerical fluid dynamics.
Some Points to Note While Running a Dynamic Simulation

• Always mention the system of units which you have used for simulation.
• You must have a fairly good idea of what would be the total file size of the output since a wrong input
could generate a large output of ascii and binary files, and you will run out of space very soon.
• A thumb rule to remember is: If T is the termination time, then your animation output file generation
time step should be T/10 (so that you have minimum 10 frames available) and the time history file
size should be a further ten times lesser as you would require an accurate representation of energy,
contact forces and stresses data.
• There should be a zero penetration in the model so that there are no initial contact related energies
in the model. It is always a good practice to run a 0 msec run for the model making you understand
if there are gross modeling errors in the model.
• It is always better to set the global element length to 5 mm which corresponds to a time step of 1 e-
06 seconds for steel material. Try to make your initial runs with coarse mesh so that you know how
your model behaves.
• There are lot of element formulations and different material models, and a beginner could naturally
be very confused with understanding the software overall, it is hence a good idea to understand the
element formulations that perform best and why. This comes through a lot of verification runs,
reading the software manuals carefully and experimental validation of the material models.
• Just defining the contacts and leaving to the software to do the rest of the things is not enough. You
must also check if it is giving you a physically compatible behavior or result.
• Usually a crash simulation software has different cards suitable for rigid bodies and deformable
bodies. Carefully study these options so that you get a full idea of the capabilities of the software.
• Try to use restarts to the maximum extent as this is a powerful utility associated with explicit
dynamics code.
• Crawl, walk, run approach: start with the simplest model and progressively add complexity. An FE
model should be as simple as possible, but as complex as necessary. In other words, numerical
models should capture the required physical behavior with the minimum amount of input
parameters.
• Always check the CAE results with some hand calculations and make a note if it makes some sense.
Always check the energy balance. If hourglass energy is large compared to internal energy (strain
energy) of the system, then it can mask the physical failure and can give a totally erroneous result. A
thumb rule is that the hourglass energy should be a maximum of 10 % of the internal energy.
• Note: By using reduced integration you save a lot of computational time which is always necessary
in transient dynamics. Any philosophy which makes your run fast should be used e.g. lumped mass
matrix instead of consistent mass matrix. The risk of using reduced integration is that it generates
non –physical zero strain energy modes which are to be compulsorily eliminated by using artificial
stiffness / viscosity thus generating additional strain energy which we call as hourglass energy.
Another strategy to make a fast run is to use mass scaling which increases the time step, but you
should set reasonable limitations on the maximum mass increase.
• There can be many errors and mistakes by a beginner. Always remember that the cause of these
errors is very much present in the deck itself. Common are wrongly formatted input, initial contact
penetration, improper load curve definitions and massless nodes due to temporary nodes created
not deleted. Floating point exceptions can be caused by parts with zero density or thickness, over
 constrained nodes. Always take a closer look whether you have inputted the correct information and
whether the mesh is of a good quality. Normal run doesn’t necessarily guarantee that you have
solved the problem in a correct way
as what is more important is always checking the CAE simulation result with the test data or
physical behavior of the system.
• If the time step is too small due to the presence of a small element, it is always better to coarsen the
mesh thus allowing an considerable increase in the time step.
• Total energy of the system is always constant:
Total Energy = Internal Energy + Kinetic energy + External Work
• For most of the impact problems, there will be initial kinetic energy and no external work at time
t=0.0. As the simulation advances Kinetic energy will decrease, Internal energy will increase, and the
total energy of the system should always remain Constant. If the total energy plot shows a very big
jump then the model has an error and you must check the definition of contacts. The momentum
printout also shows whether the bodies after hitting are going in the right direction. Always ensure
that parts are going in the right direction with right velocities at the right time.
• Effect of Process and Residual Stress on Crash Analysis is important for correlation between physical
crash test and FE. The work hardening effect and formation of residual stresses during metal forming
causes change in subsequent yield stress and deformation behavior of formed components. Thus,
the mechanical properties of formed components can significantly different from those of the blank.
Consideration of these material property changes in crash simulation is vital for achieving good
simulation accuracy. One step forming simulation technique from HyperForm is used improve
accuracy of crash simulation result by mapping residual stress-strain data predicted by HyperForm
onto crash simulation model.