STAAD Triangular Plate Stresses

[francis_intano]

Hello!

Can someone help me clarify about triangular plate stresses in STAAD. i know that this guide exist (see below)

but how would I interpret this in a triangular elements? something like this:

Thanks a lot btw.

 

[dik]

If memory serves, forces and moments are designated by the 'face' the force is acting on and the direction. For example, (sigma)xy would be a force on the +ve x face acting in the positive y direction. This would be a shear force of some kind. It appears they don't follow the international convention... anybody's guess. Mxy should be Myy since it is a torsional load, I think... this is going back decades.

Rather than think climate change and the corona virus as science, think of it as the wrath of God. Feel any better?

-Dik

 

[WARose]

As far as I know, the force/stress output in STAAD means the same thing for triangular elements as it does for quads. No interpretation is required. (Unless you are converting a force to a stress.)

 

[JoshPlumSE]

The plate/ shell stresses and forces mean exactly the same thing for tris as they do for quads. It's based on local axes, not the shape of the element.

Let me state that a slightly different way. First, take a look at the image of the quad plate element that you gave in your original post.

Now take the triangular element and look at an infinitesimally small piece of it. But, make that piece and infinitesimally small square, with the local axes aligned as they were in your in your original quad post.

 

[francis_intano]

@JoshPlumSE and @WARose

While I do understand that such triangular plate does have an infinetesimally small piece of quad where the local stresses do apply. I'm actually trying to convert such local stresses to force. Out-of-plane shear stresses (SQX and SQY) and moments (Mx and My) conforms to local axis which yields a result of force/unit width/thickness and force-length/unit width respectively. This would be easier with a normal quad but it was my first time trying to deal with it in a triangle. I feel like I had to do something, taking the result at face value sounds ridiculous to me(even at a right triangle element with local axis coinciding with right angle).

Thanks for your previous reply btw.

 

[JoshPlumSE]

Plate moments (like Mx, My and Mxy) are always given as Moment per unit width. That's standard notation in plate theory. Though I'm not particularly familiar with STAAD, I imagine it has to be the same way.

I expect the same to be true about Sx, Sy, Sxy, SQx, and SQy.

 

[rb1957]

You have results for a triangle element ... Mxx and Myy (in element axes, but not important).

You have moment per unit width in two (orthogonal) directions.
so bending stress is 6M/t^2 in these two directions, and use as per quad.

I think it is misleading (confusing?) to say the moment is along the edge of the element.

Hopefully you're not using a 3 node triangle.

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

 

[WARose]

Hopefully you're not using a 3 node triangle.

Why would that be a worry in STAAD?

 

[JoshPlumSE]

WARose -

It depends on the software. In many software, the 3 node element using the same formulation as a quad. So, essentially it's just a "collapsed" or highly distorted quad. I know that's the case with RISA. I don't know if that's the case in STAAD or not. In most cases, I believe even a true triangular element is not usually as good of a formulation as a quad.

Caveat: I used to work for RISA (a STAAD competitor) and currently work for CSI (another STAAD competitor). So, I'm not exactly a impartial observer.... Even if I try to play it totally straight (as I've done in this thread), there could be some unconscious bias that creeps into my opinions.

 

[WARose]

It depends on the software. In many software, the 3 node element using the same formulation as a quad. So, essentially it's just a "collapsed" or highly distorted quad. I know that's the case with RISA. I don't know if that's the case in STAAD or not. In most cases, I believe even a true triangular element is not usually as good of a formulation as a quad.

That's kind of what I thought rb was getting at. And STAAD's triangular elements perform as well their quads in my experience. rb is (IIRC) kind of old school [no offense rb; I am too] so he may think of triangular plate elements as being like the old CST elements. That's just not the case anymore. STAAD's "brick" elements (8 node solids) use to be prohibitively stiff as well.....but they solved that with the introduction of rotational degrees of freedom at the nodes. (Among other things from what I've heard; STAAD has never let me view the shape functions for some of their elements. Guess it is a trade secret.)

 

[rb1957]

hey, eff off !! pot calling the kettle ...

yes, old TRIA3 elements are overly stiff. TRIA6 work very well.

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

 

[Sye123]

As has been suggested by multiple posters in this thread, the interpretation of element stresses does not change for triangular plates. Visualize a cross section of unit width and depth = thickness of the element, at the location where the stress is being reported and that would help in converting a stress into a force, if you have to. For example if you multiply the out of plane shear stress SQX or SQY by a unit width and the element thickness, you would arrive at the out of plane shear force over a unit width at that location.

 

[Settingsun]

Do the Staad secret herbs and spices mean it's ok to use a triangle of the shape posted? It looks about 5:1 shape ratio.

 

[WARose]

Do the Staad secret herbs and spices mean it's ok to use a triangle of the shape posted? It looks about 5:1 shape ratio.

Yeah, the ratio of (internal) angles to each other don't look particularly good.

 

[Sye123]

Yes aspect ratio 1:1 is ideal. We should try not to exceed 4:1.

 

[francis_intano]

Thanks a lot guys. I guess it was just a confusion in my part upon seeing something unfamiliar but I understand it now. I'd take a look again over the ratio which is indeed doubtful.