AutoDyn Bullet Impact Analysis Help 1

[Albedo]

I am trying to analysis a bullet vs armor situation in AutoDyn and it comes out mostly correct although the bullet doesn't seem to be realistically simulated and I am concerned about it. Attached is the .MP4 file of the animation generated from the results. Notice how the entire bullet breaks down into particles, rather than holding together? The armor does fine and fragments at the end, which is realistic, but I am afraid the bullet itself isn't actually showing me what will happen in real life. All of my hand equations say it can penetrate this armor thickness, so I doubt I'll ever get a result showing it cannot.

The material for the armor is AR500 steel, the bullet has a copper jacket and a Molybdenum core. I had to use the same damage Johnson cook failure criteria (D1-D5) as tungsten carbide for it since there doesn't appear to be any damage models out there for Molybdenum and they are similar materials, in the same column of the periodic table.

Is this normal behavior? What exactly do all the red particles from the projectile mean? Is there a way I can get a more accurate simulation where the core doesn't just break into a node sized element? I was expecting something along the lines of this simulation:

https://www.youtube.com/watch?v=XYuKWuyKXII.

Granted, theirs are are against an angled armor, but everything correctly fragments in that simulation.

Any help on this is greatly appreciated!

 

[FEA way]

I don't use AutoDyn but I perform similar analyses in other software. It seems that the bullet is converted to SPH particles based on damage criterion. You should check the settings of material damage (maybe some adjustment of damage criteria will help) and behavior of failed elements. Maybe you could switch from SPH conversion to simple deletion of the elements (and decide at which point it will be done). Also, make sure that the software accounts for contact with interior faces (those that become exposed during the penetration). Apart from that, the simulation looks fine.

Can you say more about those hand calculations ? What was your approach in this regard ?

 

[Albedo]

Thanks for the reply! But is it accurate for the bullet to convert to SPH particles in this situation? It is expected the bullet core will erode and create a path forward for the remainder of the bullet, but the two concerns I have about these particles are: 1) they are influencing the penetration by acting as mass with velocity and continuing erosion of the plate 2) if I turn them off, isn't that just masking the problem? This video shows what I mean by erosion of the projectile and plate without having all these particles:

https://www.youtube.com/watch?v=Cye07MmEwvE.

See how the penetrator mushrooms a cavity inside the plate and there are no extraneous particles that failed, other than fragmentation. How do I adjust it to account for contact with interior faces? I don't remember seeing any options like that.

Attached is my document used to compute penetration depths. They mostly vary from one another, but even the ones that take into account more details say it can penetrate the depth I am looking at. The only formula I didn't use was Thompsons Formula (original, not F-Formula)because I couldn't find the values for which AR500 and Molybdenum under pressure begin behaving like a liquid rather than a solid.

 

[FEA way]

Conversion to SPH particles is usually used when a body undergoes huge deformations and is literally obliterated on impact or blown up. For example, such an approach is used in simulations of birds hitting jet engines. But in this case, I would avoid the conversion and stay with regular element deletion so that the bullet's fragmentation can be modeled properly. I don't know where this particular setting can be found in AutoDyn but if you search the documentation for SPH, you should find some settings including the criterion for that conversion and an option to deactivate it regardless of the strain level.

 

[Albedo]

I don't think there is a way to do that in AutoDyn, I looked everywhere, including the worthless manual. The only things I have found is:

The simulation (4.5 hours)I posted had the density cutoff option as "density limited" instead of "delete node". Upon changing to delete node, estimated times to solve for the full cycle (bullet exit after successful penetration) has steadily been climbing. I cut it at 135 hours estimated and 11% complete. Originally it predicted 22.6 hours. I tried messing with minimum and maximum density factors too but that didn't change anything. Guess I'm just gonna be stuck with having all these extra particles, because no way am I going to run a 135+ hour simulation without knowing if it will even work as intended. maybe there is another way to limit it that won't nuke the time to solve?

 

[FEA way]

These options refer to already existing SPH particles. There should be an option somewhere to not generate the SPH particles at all and use only traditional Lagrangian elements.

To reduce the time required to complete the analysis you could use a proper level of mass scaling (but only for a small set of elements - those with the lowest stable time increment since it’s not a quasi-static simulation).