Gusset plate calculation/approximation

[orionz06]

I am trying to avoid modeling gusset plates in a beam/frame model and am stumped as to how to work up a hand calculation to prove the plates are not overstressed or buckling.

The frame is 80"x160"x80" deep with 15" triangular plates. The customer does not want anything other than beam elements to be used. The best approximation I have thus far is treating the plate as a beam and coming up with approximate section properties, but I cannot figure out the means to prove them to be adequate with a hand calculation.

Any thoughts?

 

[rb1957]

free body diagram ?

i'm guessing you're looking at in-plane loads ??

why the customer insists on modelling a plate with beam elements is beyond me ?? maybe model the plate as a series of beam, each representing a width (1', 1", ?) of the gusset, with transverse beams representing the shear between the modelled loadpaths ...

but then i guess you're modelling 160" deep beams as beam elements ... so a 15" gusset is pretty localised.

a free body of the gusset should help a lot in understanding the loads on the gusset and maybe some "waffly" story for the gusset stresses ... the loads are small so that 2" effective width carries the load required ... ???

 

[orionz06]

I think I have the problem solved. I set up a few simpler models of the same size and was able to show that the gussets are not effective to begin with so they are negated from the larger model.

 

[corus]

You wouldn't use beam elements to model a triangular plate.

Bear in mind the purpose of the gusset is to reduce the bending stress at the bottom of the frame. It also reduces the effective length of a beam so by omitting the gusset you'll have a worst case scenario, and hence err on the safe side.

Tara

 

[rb1957]

i'm thinking you mean that they're so localised that you don't need to model them expressly, and they "just" provide the loadpath for the nodal forces to move from one member to another.

 

[Qshake]

I can understand modeling the frame with beam elements provided that is the correct element for the loading. For example if there are no out-of-plane loads and only in-plane loads, beam elements in the frame are ok.

If you have out-of-plane loads you'll need to make sure your beam elements have the necessary DOFs to handle the out of plane plane loading.

Once the member forces and or joint forces are known from the frame analysis it is commone to design the gusset plates by hand. These plates are not typically modeled explicitly unless a smaller "substructure" model is warranted.

If a smaller yet more robust model is required for the joint itself then you'll have to use plate/shell elements for the gusset plate and plate or shell elements for the connecting beams. The beams should be modeled as plates/shells for a distance of at least 2 or 3 x the depth of the member. After that it can be modeled as a beam element.

A little more info would help us understand better what you're trying to do and thus provide better input.

Regards,
Qshake

Eng-Tips Forums:Real Solutions for Real Problems Really Quick.