Reduced beam model for a complex assembly

[NaikD]

Hi All,

I need help with general guide lines to create a beam model for a complex assembly with steel and aluminum parts, that would match at least first few frequencies. Since customer wants a beam model, shell model is not an option.

For known cross sections (angle, channel etc.,) I can build the model. However, for complex assemblies with many sub assemblies, I am assuming the process could be of trial and error by representative cross sections and adjusting density/modulus of elasticity to match first couple of frequencies (say 5).

Any help is appreciated.

Thanks,
- Dipak

 

[rb1957]

Presumably you have a model of the complex assembly, and you're looking for a way to reduce the size/complexity of this model, presumably to add to some larger model ?

Then yes, model with simple beams, trail and error to match the frequencies. Or you could use the stiffness matrix of the complex model, like a "super-element"; but probably your customer won't let you ?

another day in paradise, or is paradise one day closer ?

 

[GregLocock]

Believe it or not I worked on this lousy concept in 1983. Basically how could we represent a complex shell model as a series of beams and compliant joints? It was the A pillar to windshield beam to doortop beam (can't remember the words for those). ie three pressed steel sections coming together in a complex curve.

It's a research project not a production method.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[FEA way]

Since you are looking for general guidelines, I think that you may find the NAFEMS book „Finite Element Dynamics Primer” (by D. Hitchings) quite interesting. There’s a whole chapter about problem reduction in dynamic analyses.

 

[NaikD]

Rb1957
Correct. Customer is looking to reduce the size and complexity of this model specifically to a beam model.

Greg Locock,
I am leaning toward, representing beam by sections that may approximately represent cross sections at various places. I can split the assembly in four segments and then play with density and modulus of elasticity to get correct weight and frequencies. I never had to do this so far hence the request for help.

FEA way,
I will check the book by Hitchings to find out more.


I wish to thank all of you for helping me.
With Best Regards,

- Dipak

 

[rb1957]

How well does a beam mirror the "complex structure" ? If there are more than 2 or 3 attachment nodes, it'll get very complicated very quickly.

You can use a stiffness matrix to represent the structure accurately, like a NASTRAN "superelement". Maybe accuracy is a selling point ?

What code are you using ?

another day in paradise, or is paradise one day closer ?

 

[NaikD]

rb1957

I am using ANSYS. I have used NASTRAN/ PATRAN ten years ago.

Customer is specifically asking for beam model that would be accurate enough for first couple of frequencies only. I did not enquire whether customer would be satisfied with the stiffness matrix in lieu of the beam model but I can find out. Getting this info from ANSYS is not that complicated.

Here is how I am planning to go about it but have never done it. Not sure if it is the correct way to do. If not please correct me.
The complex structure is a box structure that I can represent by hollow closed section (7075T7350, Al alloy). However it will not be a simple rectangle within a rectangle.

The above structure I can cut in five sections.

Section 1, 3 and 5: box section (Al alloy)

Section 2 and 4: The middle two sections contain sub assemblies and are complicated.
The support structure for these sub assemblies I can approximate by suitable cross section (Ph13-8, H1000, Steel alloy).
The middle sections then would be composite structure made of Al and Steel alloys.
I can run hand calculations to arrive at equivalent modulus of elasticity and density for the composite structure material as a first approximation for the middle beams.

Input these cross sections to ANSYS design modeler for five sections of the beam with respective material properties (Al for forward, middle and aft sections and equivalent for composite sections).
Fixed boundary conditions are in sections 2 and 4. Any frequencies coming from the middle sections would need adjustment in modulus of elasticity and density of the composite beam.

I have never tried to do this and do not know if this is correct or even if it will work. I am just trying to figure out if it is possible to arrive at a beam model. More I think about it, as you suggested, stiffness matrix would get me out all these complications.

Thanks,

With Best Regards,
- Dipak