Honeycomb Impact

[flyforever85]

Hi all,
I've been working for months on a model of honeycomb subject to an impact. The model is analyzed with ABAQUS explicit and it converges with no problem. The deformation though is double the amount we get in real experiment and I run out of ideas to make the model working. I don't have a pic but the model is not much dissimilar from this:

https://www.youtube.com/watch?v=19nBbI7WC9U

[ul]
[li]Here some more info: the speed of the bullet is 108 in/sec so low speed impact.[/li]
[li]I tried with and without mass scaling to speed up the analysis with no big differences[/li]
[li]I'm using an elastic perfectly plastic model for every material involved (steel and nickel) but no damage modelling[/li]
[li]I checked the kinetic energy and it goes to zero, i.e. all the energy from the bullet is transferred to the model which plastically deforms[/li]
[li]As I've seen in other modelling, the bottom layer of the sandwich has the least amount of strain energy while the top layer has the highest[/li]
[li]Double checked time of analysis, materials properties, size, geometries, units and everything is correct[/li]
[/ul]

If you can think of anything else I could check, any our history output I should request or any other test that could prove my analysis is robust please advise.

Thank you!

 

[FEA way]

You should check all units first (especially regarding material properties like Young’s modulus). Then make sure that the model represents the real life test as closely as possible (especially in terms of boundary conditions - the model might be overconstrained).

To give you more detailed advice I’d have to know more about the model’s parameters and see the results.

 

[flyforever85]

I think the units are correct, I double checked them so many times and used the same in other models where everything was fine.

I noticed something: the vertical displacement changes a lot by changing the mass scaling. If mass scaling is 1, U3 is 0.036, if mass scaling is 128, U3 is 0.66 and if mass scaling is 500, U3 is 1.13 in or so. I thought max scaling was just an artifact to speed up a process (I know it's important not to exaggerate with it though). Also: the bullet the impacts the honeycomb is actually a box that is dropped on it. The box will be 200 lb so I added a mass of 200 lb to the bullet and I added gravity. Is this a mistake?

 

[FEA way]

Mass scaling is a feature that can speed up the analysis significantly but also cause large discrepancies in results (after all it artificially increases the density of elements). Especially in impact analyses you should be very careful with the amount of mass scaling.

Is the impactor modeled as a rigid body ? If so, you can assign mass to its reference point and enable gravitational acceleration but usually this is done via initial velocity assigned to the impactor placed right above the specimen.

I would also try with hardening included.

 

[flyforever85]

Initially I modeled it as rigid but the bullet never "bounced back", i.e. never stopped transferring all the energy to the honeycomb. So I moved to a deformable body with a mass assigned.

 

[FEA way]

If it's modeled as deformable body then it's enough to assign density to its material. Proper mass will result from the volume of impactor. Unless you model only part of its geometry.

 

[flyforever85]

Here is the thing: the honeycomb sandwich is 6 x 6 in2 while the box that drops on it is 40x40x40. So when I modeled it, I decided to take only a little part of the box (basically the corner) to reduce the number of elements and add a mass of 200 lb. As soon as I can get the picture I'll post it. Also, since the tile is a square I used both symmetry planes to reduce the analysis time so now I have 1/4 of the tile.
While writing I just relized I'm using 200 lb box on a quarter of a tile which is not correct... I'll get back to you