Frame analysis - Go-kart frame with dirt plate

[soha]

Hey All

I'm simulating a go-kart frame made of circular tubes. Making a frame analysis gives me one result, but it isn't taking the bottom plate into considurations. 

Making a "simple" stress analysis seems wring and doesnt give me a sutable result. 

How can I make the best possible simulation of this chassis ??

[RajSchmidt]

The frame analysis calculates only the frame members as simple beams and ignores all other parts. What you can do to simulate the plate is to add additional rigid links. (You may have to define additional nodes first where you can attach the links.)

[soha]

hey

We did try with the rigid link features. But since the link has infinite strength, it's not a good pointer, since the plate only has a thickness of 1.5mm. 

[RajSchmidt]

Good point. You could try to build in some dummy beams to approximate the plate. But I guess you will be faster and more accurate if you do a full FE analysis instead. Since the frame is not too complex it should work.

[soha]

We tried to make a stress analysis, but for some reason the results were almost non existing.. making a frame analysis without the plate, did give us a "idea" and a result but far from as low as the stress analysis. 

[RajSchmidt]

That’s strange. I have rather seen the opposite: When you do an FE analysis of a frame you will get stress peaks at the joints. Remember that you have to “convert” some of the loads since you have different types. With frames you can use point and line loads which you should avoid in FE. (Better use area loads.)

[soha]

In the frame analysis, not taking the steel plate into account, we can see deformation and stresses.
Making a stress analysis, the whole things stays blue, stresses below 10MPa and almost to 0 deformation...

[RajSchmidt]

You can only check the obvious points: Do you use the same materials? Any open movement in places where the parts should stay connected? Any stress peaks in the FE analysis? (Sometimes the whole structure appears blue but there is a tiny area with huge stress.) Are these standard profiles from the CC? If you defined them yourself, are the inertia values correct? That’s where I would look.

And maybe you can check the whole procedure with a more simple model, i.e. just three or four beams? So that you can easier compare the results?

Bye for today, gotta go.

[swalton]

Do a quick free body diagram on the frame and loads to calculate the reactions at the supports.  Compare the hand calculation results to the reactions at each support in the FEA.  Are they close?  

Make a quick cross-section sketch in the assembly at the narrowest portion and use the Region Properties command to calculate the area moment of inertia.  Do a simple beam calc using that moment of inertia and basic loads.  Set up a quick FEA using the same supports and loads and see if the stress and deflection are close to the simple beam results.

Both of these steps should give you a basic sanity check on the problem.  If the FEA and hand results are close, then you can have more confidence in the FEA.

[cadman777]

Here's a thought: If you can find the stresses along the frame that touch the plate, and know the weld lengths and pitch distances around the perimeter of the plate, then you can use/adapt the formulas in 'Roark's Formulas for Stress and Strain' (book) to hand calc the plate. Then work it backwards. Maybe that would help? And maybe some of the procedures in Lincoln's 'Design of Welded Structures' can help w/the joint design and integrity?

[Cris-Ideas]

Post your model with a FE analysis defied.

This way we can actually check what is the problem.

Your model is very simple so there should not be any problem at all with FE analysis for entire structure.

Cris.

[soha]

I just added the parts here:

https://365basyd-my.sharepoint.com/:u:/g/personal/soha_iba_dk/EXf8msiXYdNBjCsIhNvZAMsBmMvUVJxlaVr7sc...

[soha]

DId you have any time to take a look at the mode?

[Cris-Ideas]

No I did not. I did not even turn my computer on this week.

But i have it downloaded.

Will have look tomorrow evening, or night.

Cris.

[Cris-Ideas]

Ok,

Had a look.

You need to explain design intent for the analysis.

1) What is this force? What is it simulating?

2) why you selected boundary condition as fixed for the end tube?

Overall this seems a little strange, so please try to explain in your own words what is that you want to simulate, as you would do to someone that has no idea of what you are doing and what your intent is.

Cris.

[Cris-Ideas]

Structure is not symmetrical (and I do not mean connection plate). is that intentional?

[Cris-Ideas]

why there are gaps in tube joints?

and not in all of them.

This connections will not do good in FEA.

What is that you tried to achieve with this frame design? gaps for welding?

Cris.

[Cris-Ideas]

Why did you have to define manual contacts?

I tire to avoid this as much has possible. This goes bad if you update the geometry as contacts reference model elements by reference and once references point to different physical (imaginary) part of the model than there starts to be a mess in the FE model definition.

Cris.

[Cris-Ideas]

As for now I would say:

This model, in hand, is not good to use for FEA, as results you may get are hardly to be reliable due to many model problems, mostly geometry and contact.

If you explain what what is that you are after, and answer questions in previous posts I can guide you threw needed updates.

Cris.