Transient impact analysis using IDEAS, and in general. 1

[jurs]

Dear Group,
I have strain data from an impact experiment using a cantilever beam type structure. Data was taken with a resist strain gauge at various points along the structure. In parallel I am modeling this structure with a couple different FEA programs to see how hard it is to accurately model an impact event. Needless to say, there are many things going on that can drastically effect the results(impulse time and shape, damping, impact location, etc...).

In my experimental data I can clearly see the transverse wave motion and reflection off free and clamped surfaces (I think). One thing that has been confusing for me is that I-DEAS does not seem to be able to capture the initial transient wave motion. My initial thought is that its method of solution (from theory manual) "mode-displacement" or "mode-acceleration" is not able to describe the transient phenomenon, but rather is a steady state solution...? But I am not really familiar with these methods so my assumption could be somewhat naive.

When I model the event in LSDyna which uses numerical integration techniques for solution, the transient wave and reflection is apparent (assuming my time step is small enough to capture it).

Does anyone out there have experience with these things? I am very excited to learn more about what I am observing.

Thanks in advance!
Craig

 

[tld23]

I am not sure about IDEAS, but it is probably an implicit code which is not good for capturing short duration events (it can do it but it will be very inefficient). LSDyna is a explicit code which should handle the impact event very well - that what explicit codes were designed for. This may explain the difference between your models.

If you want to capture the shock waves use LSDyna or another explicit code.

TERRY

 

[brad]

I second Terry's statement. IDEAS appears to be the wrong tool.
The "right" tools:
ABAQUS/Explicit
DYNA
DYTRAN
PAMCRASH
which (as Terry has noted above) are all explicit codes.
Brad

 

[jurs]

Thanks, there is no doubt that I-Deas IS the wrong tool (that was never the question). The question is why...The fact that it does not use an explicit/implicit intergration scheme still doesnt answer the true question.

What is it about mode-displacement solution method (IDEAS solution method) that will prevent you from approximating a true transient event?

I appreciate the responses I have received but I am still in the dark.

CJ

 

[brad]

By "mode-displacement" solution method, I am assuming that you must mean you are doing a frequency extraction and then using these modes as a basis for your transient solution.

There are two possible reasons why this is bad (and they both may apply):
1) this assumption is limited to linear behavior, and the solution has nonlinearity (even if it's only geometric nonlinearity);
2) The frequency content early on in the analysis is so high that it is not captured by the combination of modes which you are using (this is almost certainly true).

In all likelihood, both (1) and (2) are problems for you. After considering my earlier post, if there is not severe deformation in this analysis you may be able to use a nonlinear implicit code to do the analysis. If the model essentially behaves linearly (i.e. point 1 above is moot), then you could do a transient dynamics analysis with a linear code (using full degrees of freedom, not a modal solution). You should then arrive at the correct solution.

So in hindsight this may not necessarily be an issue of implicit vs. explicit methods; rather an issue of frequency-based dynamics and/or nonlinearity.

Brad