CHS Moment Connection

[IH1980]

Easy one for somebody I reckon:

I have a chs which cantilevers out from a flat plate. It is fully welded to the plate with a fillet weld around its perimeter. How do I calculate the moment capacity of this connection? I have done 'what I think' is right, but it has now become the critical location which limits the load capacity so I need to check my method. I have a fixed weld size and the known weld transverse capacity.

 

[jayrod12]

moment arm between the flange welds will give you the tension that length of weld needs to resist. The web weld can be checked for your shear.

 

[IH1980]

But it is a CHS (circular hollow section) - I would be ok with most sections but this one has me stuck.

 

[PSEPK]

Did you check AISC Steel Design Examples?

 

[jayrod12]

missed that sorry,

Moment arm = 2* centroid height of a half circle. determine your tensile force. see how much weld that requires, see if remaining weld works for shear.

This is just a rough calc that would let you know if it was going to govern. Some of the other designers here are much more adept at steel connection design than I.

I'll bet someone here might send you to "Design of welded structures" by Blodgett

 

[IH1980]

I have done something similar to the moment arm method. I have also worked out the modulus of a chs of 'unit' thickness and then multiplied by the weld allowable kN/mm force (UK based so metrics over here!) - the units work out I think to give a moment, but I might have made this method up completely! The two different methods give different answers as you might expect so not sure which way to jump with it.

Have a copy of Blodgett and can't see this scenario, also don't have access to AISC examples (being UK based).

 

[KootK]

It would be good to know more about the flat plate. Can it be considered rigid? What's the situation?

Another option.

You can work out the shear and bending stresses in the CHS at the support and then multiply those values by the CHS thickness to get required weld capacities in tension and shear. Unless you really need to whittle the welds down, I'd just combine the max tension demand and the max shear demand, even though they'll occur at different locations. This solution is reasonable when the supporting plate is rigid.

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[Brad805]

The CISC has also has a good book for HSS connections. I thought I would mention it.

http://books.google.ca/books/about/...nnections_an.html?id=abLZAAAAMAAJ&redir_esc=y

 

[PSEPK]

You may download AISC Steel Design Examples, for free, here:Link

 

[Gumpmaster]

Treat the weld as a line. The section modulus of the circular line is pi*d^2/4. This will give you a result in load per inch. You can then size the weld for the load.

That would be an elastic section modulus. You could also size the weld for the plastic moment capacity of the section using a plastic section modulus of the circular line.

 

[Gumpmaster]

d is the circle's diameter. Use the outside diameter of the CHS.

 

[JedClampett]

Are you sure you're using the right weld terminology? If you're welding a round section to a flat section, it's a partial penetration groove weld and the calculation is subtlety different.

 

[KootK]

My solution and Gumpster's are fundamentally identical. I prefer to tackle it from a section stress starting point for two reasons:

1) Often, I'll already have those numbers available from the design of the section and;
2) I find it to be a more intuitive way to incorporate the shear stresses into the weld design.

Six of one, half a dozen of the other.

You could also size the weld for the plastic moment capacity of the section using a plastic section modulus of the circular line.

This I'm curious about. I believe that Jayrod's solution implies plastic capacity as well. While steel sections are often ductile, welds themselves are much less so. Weld groups can't really deform plastically as far as I know. Do we trust that section yielding adjacent to the welds effectively shields the welds themselves from brittle failure under plastic strain conditions?

Are you sure you're using the right weld terminology? If you're welding a round section to a flat section, it's a partial penetration groove weld and the calculation is subtlety different

Can you elaborate on this Jed? We fillet weld baseplate for columns in tension all the time.

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[JedClampett]

I guess when I thought of cantilever, I thought of a horizontal tube and a horizontal plate. Now I see that is not what is intended.

 

[IH1980]

Gumpmaster - that is the method I had hit upon myself, so I am pleased somebody else (and KootK!) think the same. I am happy to go with that.

Thanks for the help guys, and for the links to the publications. I will get hold of those for future reference.

 

[jayrod12]

Although my method does imply plastic behaviour, The simplification results in conservative loading on the weld and so I feel there is still a bunch of excess capacity.

To be honest, no one had ever shown me anything different. Now that I've got a new way to check welds on round sections I will definitely give 'er a go. It will be the thing I learned today.

 

[Gumpmaster]

For the plastic capacity of the weld group, I think if you look at where the stress on the weld is coming from (the CHS), then maybe it makes more sense. The stress in the CHS can't exceed the yield stress, at the plastic moment capacity the entire section will be at yield. I think it would then be fair then to use the plastic section modulus of the weld.

I agree that you shouldn't use the plastic section modulus for the weld for just an applied load where the attached section isn't also developing its plastic moment.

 

[canwesteng]

I think the only case you can justify using the plastic section modulus of the weld group is when you detail the weld group to develop the entire plastic capacity of the attached section - potentially quite conservative as deflection, member stability, and other concerns may govern the member size.

 

[Gumpmaster]

agreed.

 

[KootK]

For cases where the member is expected to remain elastic under load, I believe that using a weld stress distribution that mirrors plasticity will result in an underestimate of peak weld tension demand at the extremities of the section. Then, since welds are not particularly ductile, there is danger of an unzipping failure.

This is just semantics, but the phrase "plastic weld capacity" and the like are a bit misleading for fillet welds in my opinion. Fillet welds themselves do not have reliable plastic capacities due to their lack of ductility. Rather, when members are expected to go plastic, welds must be designed to resist the applied plastic stress distribution while themselves remaining plastic. Semantics.

Having done some noodling on this now, I can think of two common examples of welds designed to deal with flexural plasticity in supported members:

1) AISC extended shear tab connections.
2) Moment connections in seismic frames.

Theses examples lead me to believe that welds should be designed to resist at least 1.25 Fy when the supported member is expected to go plastic. Essentially, the welds should be capacity designed using over strength of the supported member. Due to strain hardening etc, I would expect welds at the extremes of the section to apply tensile stresses to the welds in excess of Fy


The greatest trick that bond stress ever pulled was convincing the world it didn't exist.