I have a 3D model that experiences

[Mohammad_Structural]

I have a 3D model that experiences different boundary conditions under different conditions (for example when it has +y acceleration the right wheels are constrained; when it has +z acceleration four wheels are constrained). There are six parameters: x, y, z linear acceleration and x, y, z rotational accelerations (6 degrees of freedom in total). Would it be possible to scale the unit acceleration cases and superimpose them on each other to obtain the stresses and deformations that the vehicle undergoes?

I know that the principal of superposition tells us we can do this when we have a structure with a defined set of boundary conditions. We can then break up the loads in 2D, analyze separately, and add the results. “The total displacement or internal loadings (stress) at a point in a structure subjected to several external loadings can be determined by adding together the displacements or internal loadings (stress) caused by each of the external loads acting separately.” But does this also apply to the above case?

 

[rb1957]

different boundary conditions implies different models, at least different constraint sets.

You can run the different models (different constraint sets) within the same project, which makes sense (the same physical structure is subject to different loads and reactions). But combining them together sounds "odd" ... for a common constraint set you can run Fz and Fx separately (and superimpose) or together

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

 

[rb1957]

thinking more about it ...
maybe "all" you're saying is that for y_accel I don't want to have Fx constrained 'cause there should be no reaction there (and not because the structure moves in such a way at it is physically impossible to have this constraint).

if it is the former (ie no reaction 'cause you as the intelligent designer don't think there should be any) then maybe it is ok to consider the different constraint sets together, and superimpose.

if the latter, then I think you've more of a problem ... how will the structure react to a combined load ?

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

 

[BUGGAR]

Are you talking of combining modes as is done in seismic?

I'm not sure of your design approach but don't you have to look at each condition by itself since some situations may cause overstress in torsion, some in bending?

 

[Mohammad_Structural]

A) rb1957, I think it's the "former" I'm talking. I'm looking at a vehicle that can undergo simultaneous loading. I've created a simple 2D analogy in the attachment just so there in no confusion here. You will see that in case one, I have a moment around Z. And in case two, a force in X acting on the center of the vehicle. Now you see because of the loading types, the boundary conditions on vehicle structure change. Now my question is would it be valid to combine the stress results from 1 and 2 to obtain the stress results in 3?

In other words, is 1+2=3 as in the figures?
I know for deformation this is definitely not the case but I don't know about stress.

B) After giving it some thought, I realize that I would never have a scenario such as in 3. I should realistically have the constraint sets as 1 or 2 depending on the magnitudes of the loads. e.g. Either the bending moment or the lateral force will govern under combined loading. In any case, I'm still interested in answering the hypothetical scenario in A.


C) Buggar, I dont' believe this is related to seismic analysis. But your second point is spot on and that's after some time I came to the same realization in B.

 

[BUGGAR]

And in addition to all that, are you considering that your circle of traction may be an ellipse? The dimensions of the ellipse change with the loading conditions and directions. That further complicates things if you don't take them singularly.

 

[GregLocock]

There is no physical difference between a boundary condition and a loadcase. That is, you can replace a set of boundary conditions by a further set of loads for each explicit loadcase. Therefore you can use superposition in a linear model (tautological I know). What you are proposing is commonly done for initial strength estimates for cars, that is, imposing large static forces at the contact patches to represent things like 3g bump and 2g pothole strike.

Cheers

Greg Locock


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

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