Modelling a sandwitch structure in Abaqus

[spyros_ar]

Dear all,

I am working with a tapered cylindrical shell featuring the following layup: [45, -45, Aramid honeycomb, -45, 45]. The face sheets are made of CFRC. For the honeycomb core, I have the values for E1, E2, E3, ν12, ν13, G12, G23, G13, as well as the compression and shear strengths in both the W and L directions.

I aim to investigate the residual strength of the cylinder after an impact. My initial approach is to model the sandwich structure using a composite layup, treating the core as another ply to reduce computational effort.

While I understand that the Hashin damage model can be applied to the face sheets, I am unsure if it is suitable for the honeycomb core. Given that I prefer not to model the core in 3D and going to a detailed anaylis, is there an appropriate method to represent core damage in this context?

Thank you for your insights!

 

[SWComposites]

Hashin model definitely does not apply to h/c core. Its even debatable whether it is a good model for CFRP materials.
What values do you have for the core? Typically only E3, G13, G23 are available and important; the rest are small numbers which don't have any effect on the structural response.

Are you going to try to model the impact event? very very difficult.
Or are you just going to assume a damage state and try to predict the residual strength?
How is the cylinder loaded? If the sandwich is loaded in-plane, with little or no shear stress in the core, then ignore the core damage for the residual strength calc.

And you will need some test data to calibrate your analysis.

 

[spyros_ar]

Hashin model definitely does not apply to h/c core. Its even debatable whether it is a good model for CFRP materials.
What values do you have for the core? Typically only E3, G13, G23 are available and important; the rest are small numbers which don't have any effect on the structural response.

Are you going to try to model the impact event? very very difficult.
Or are you just going to assume a damage state and try to predict the residual strength?
How is the cylinder loaded? If the sandwich is loaded in-plane, with little or no shear stress in the core, then ignore the core damage for the residual strength calc.

And you will need some test data to calibrate your analysis.

For the core, I have values for E3, G13, and G23, with small values for the remaining parameters assumed based on literature. My focus is not on the detailed impact event, such as delamination or the core's exact failure mode. Instead, I aim to assume a post-impact damage state and predict the residual strength. The primary interest is in the global behavior of the model. Regarding loading conditions, a bending force is applied at one end of the cylinder.

 

[FEA way]

I prefer not to model the core in 3D and going to a detailed anaylis

Is the core's structure that complex ? Usually, it's better to model it with shells in such cases. You can just tie it to the face sheets if you don't want to account for delamination and such. I would look for some research papers though.

 

[spyros_ar]

Is the core's structure that complex ? Usually, it's better to model it with shells in such cases. You can just tie it to the face sheets if you don't want to account for delamination and such. I would look for some research papers though.

It's a honeycomb with hexagonal cells. The thing is that i don't have any other parameters besides compression and shear moduli as well as compression and shear strength. Also, I am not sure what damage model I should use for the core. That's why I thought to simplify the problem

 

[SWComposites]

as I said above, ignore the core for the damage modelling.

 

[FEA way]

As I’ve mentioned, I would look for some research papers. For example, "On Impact Damage and Repair of Composite Honeycomb Sandwich Structures" by H. Zhang et al. They also used aramid core and modeled it in Abaqus. Another one is "Study on the Cutting Damage Mechanism of Aramid Honeycomb Based on the Progressive Damage Model" by Y. Yang. And there’s definitely more on this topic.