Run a ballistic impact

One part fired at another, hard enough that the target may open: a fragment against a protective plate, debris against a housing. The projectile and the target are both meshed, a contact between them carries the load, and elements that reach a stated failure strain are deleted as the shot goes through — so the hole is the mesh’s own.

Solved by OpenRadioss.

Before you start: read this one honestly

This build has no penetration data to check itself against. The case is verified as an integration test — that the contact and the erosion work together, that a plate opens at one speed and holds at a lower one — and not against measurements. Two consequences:

Do not read a ballistic limit off it. Whether a stated shot goes through is a real prediction of the model, and it is not a validated one.
Both bodies are the same material. The case carries one material for the whole model, so a hard penetrator against a soft target cannot be stated. A projectile that should stay intact will deform along with what it hits.

If your question is “does this armour stop this round”, this is not the tool yet. If it is “where does this plate tear and how far does the damage spread”, it will show you something worth looking at.

Build the geometry as one assembly

The projectile and the target must be two solids in one geometry file, with a small gap between them — a tenth of a millimetre is plenty. Each solid becomes a part, and parts are numbered in the order they appear in the mesh.

Import a STEP with both bodies in it, or build both with the geometry primitives. A single solid cannot be a ballistic case, and the pane says so.

Mesh it

First-order tetrahedra, as everywhere in this family. Two things to get right:

The target needs elements through its thickness. A hole is made by deleting elements, so a plate one element thick either has a hole or has not. Three or more through the thickness is a floor, not a target.
Mesh the impact area finely. The failure strain is reached element by element; coarse elements tear coarsely.

Set the material and the failure strain

The elastoplastic Johnson-Cook material the rest of the family uses; see Run a drop test for what each number means.

Then, on the same Materials step under FAILURE, set the failure strain — the plastic strain at which an element is deleted. Zero means nothing ever fails, and the shot dents rather than perforates. No library material carries one, so it is the case’s own number and stays when you pick a different material.

The strain to use is a property of your material and how it is loaded, not a setting to tune until the picture looks right. If you do not have one, the honest reading of any result here is qualitative.

Say which part is fired

Also under PARTS:

Parts in contact — leave this on. Off means the two bodies ignore each other and pass straight through, which conserves momentum and energy perfectly and answers nothing.
Contact friction — zero is frictionless, and is the usual starting point.
Part fired — the number of the projectile. This one matters more than it looks. Left at 0 the velocity is given to every part, so the target sets off alongside the projectile and nothing ever strikes anything. The result still runs, and the two bodies fly along together.

Then set the velocity under ANALYSIS, pointing from the projectile toward the target.

Hold the target, if you can

An unrestrained target is knocked away as much as it is pierced — and at moderate speeds, mostly knocked away. In one measured case a free plate left at 33.6 m/s while the projectile that hit it trailed at 15.6: a shove, not a perforation, and one that looks like a perforation if you only check where the projectile ended up.

Under RESTRAINT, fix a face of the target if the real thing is held. This build fixes one named face, so a plate clamped all round cannot yet be stated exactly; pick the face that carries the mounting.

Run it, and read it

Press RUN. Then read the result the way the case was set up:

Did it go through, or push it away? Compare the projectile’s final speed with the target’s. If the target is moving away as fast as the projectile, the shot did not perforate — whatever the picture looks like.
Plastic strain shows where the material was worked; deleted elements are the hole.
The state shown is the end of the event, not its worst moment.

What this build does not do

Not possible	Why
A ballistic limit or a V50	No penetration data here to check against
A hard penetrator against a soft target	One material for the whole model
Thin sheet as a shell	A thin plate here is a thin solid, and needs elements through its thickness
A fragmenting or exploding projectile	The projectile is one deformable body
A target clamped all the way round	One named face is fixed

How far to trust it

The two pieces this case is built from are each checked against a closed form: the contact against the exchange of velocities and the 2L/c contact duration of two identical bars, and the failure strain against ln(1 + v t / L) for a bar pulled until its elements are deleted. What is not checked against anything measured is the perforation itself.

See the verification page, which says the same thing at greater length and shows the runs.

Run a drop test — one part against a rigid floor.
Run a blast analysis — a pressure load instead of an impact.
Explicit analysis — every control on the pane.