HSS moment connections with torsion

[chmich]

Hi, I'm designing a chin-up bar made of circular and square HSS sections. I am at a loss as to how to design the connections.

Joint A : Two round HSS are welded together at their ends with one member at a 30deg angle.

Joint B : A round HSS branch (CHS 3) is welded to a round HSS chord (CHS 2). The applied loading causes torsion about the branch member (CHS 2). Also, the applied loading causes bending of the chord (CHS2).

Joint C : Same as in Joint B, but this time the branch (CHS 3) connects to a square chord (SHS). Would I be able to convert the CHS dimensions into equivalent SHS dimensions, as described by Dr. Packer in this link?

http://www.hss-steeltubing.org/connections.html

These HSS connections do not appear to be discussed anywhere in the book "Hollow Structural Sections - Connections and Trusses - A Design Guide", 1997, by J.A. Packer and J.E. Henderson, so any advice would be greatly appreciated.

 

[BAretired]

JOINT A: Provide an elliptical compression plate at the junction of the two CHS members.

Joint B: Provide a fillet weld all around.

Joint C: Use Yield Line Theory to determine adequacy of square HSS wall. If necessary, reinforce with plate or increase HSS thickness. Provide fillet weld all around CHS at SHS wall.

BA

 

[chmich]

Thank you BA. How would you know what to do for these joints? There is torsion in the branch member, which causes torsion in the plane of the welds and I cannot find this case in literature.

 

[IsaacStructural]

I'm having trouble understanding how this is used. Are there other members that are omitted from those diagrams.

Is this designed to be a chin up bar with only 1 arm coming out from the wall? It seems like that would put a lot of torque on the arm and probably bend/spring quite a bit when in use.



M.S. Structural Engineering
Licensed Structural Engineer and Licensed Professional Engineer (Illinois)

 

[BAretired]

At Joint C, there is torsion and bending in the CHS. Torsion causes circumferential shear in the weld which is always tangent to the CHS. Bending causes variable shear normal to the HSS wall in the weld.

Factored shear resistance for various angles to the axis of the weld are given in the CISC Handbook of Steel Construction and I assume a similar table is provided in the AISC document.



BA

 

[ishvaaag]

Joint C quite likely can be checked like if torsion was non-existent, for normal stresses amplified by

1+(tau/tauyield)^2

where tau is the maximum extant torsional stress and tauyield the shear strength stress of the wall of the SHS, if and when the SHS wall is not overly thin.

 

[ishvaaag]

That said, a quote from some table would be far more reliable since this kind of approach uses to be for reduced deformations, and at some level of solicitation it will turn less and less adequate.

We have always open the possibility of making a FEM model with initial imperfections and material and geometrical nonlinearities that would reassure us of the adequacy of the decision taken.

 

[chmich]

That is in fact an incomplete drawing of the chin-up bar. There is another CHS 3 next to the one shown and the handlebar continues on the other side. I only showed half of it to simplify the drawing.

Where does the amplification factor for normal stresses come from?

 

[IsaacStructural]

Ahh ok great, got it. That makes a lot more sense.

A full FEM model seems like overkill, with a second CH3 member acting as a back span, the system shouldn't have all that much torsion.

not that different than this sort of system, right?

http://www.ultimatebodypress.com/wa...ctid=wall-mounted-pull-up-bar&channelid=FROOG

M.S. Structural Engineering
Licensed Structural Engineer and Licensed Professional Engineer (Illinois)

 

[dhengr]

Chmich:
It is always a good idea to “to simplify the drawing,” if that confuses the description and explanation of the problem. Witness the questions you got. You are asking for free advice here, and many people or more than willing to be helpful. But, they shouldn’t have to play 20 questions, and always remember they can’t see what you are imagining or looking at. Always read your work picture (description) and sketch, and ask yourself, will someone who knows nothing about what I’m trying to do, only draw one picture, one free body diagram (the correct one) from what I’ve told them? Maybe a center line, with the note “symmetrical about the centerline,” and some load and reaction directions and values would have cleared your picture up.

 

[TLHS]

Why don't you just duplicate CHS 3 some distance sketch-left? You can then take the torsion out as a couple between the two members and you get a significantly stiffer system?

 

[dozer]

Joint A and B cannot be an all around fillet weld. I would do a full pen weld for both of these. By the way, the fillet weld symbol is backwards on joint A. Also, the plan view of CHS 1 looks like a square tube at the end instead of a pipe. I agree with Dhengr, don't worry about simplifying the drawing. Show us what it really looks like.

 

[chmich]

IsaacStructural : Yes that looks quite like what I'm trying to accomplish except that instead of framing it on to a stud wall, I want to connect it to a square HSS beam 2 meters long and this beam will be supported by a concrete block wall.

dhengr : I was mainly interested in how to check HSS connections with the type of loads that are present in the sketch because they are not discussed in the book by Dr. Packer. As I understand it, HSS structures can fail at the connections even if the welds and members have enough capacity. But I do realize that I caused confusion with an incomplete drawing and I should have at least indicated that the drawing is symmetric. A new drawing is attached.

TLHS : The layout of the chin up bar is the way we would like to have it made, with functional and aesthetic purposes in mind. The CHS 3 bars also serve as pull-up bars so we do not wish to have them far apart from each other.

Thanks everyone for your suggestions!

 

[chmich]

Thanks Dozer, I uploaded the new drawing before I saw your post. How do you know that Joints A and B cannot be all around fillet welds?

 

[RareBugTX]

I think Blodgett's Design of Welded Structures has examples with tubular members.

 

[dozer]

For joint A the surfaces are flush so this would be a butt weld. It says CHS 1 and 2 so assume these are two pipes even though you show the end as a square. A butt weld would need to be made with some kind of groove weld not a fillet weld. For joint B since the pipes are the same diameter (that's how they look) you could not get a fillet weld on the top or the bottom. You could fillet weld the sides.

 

[BAretired]

Dozer,

The square section on Section 2-2 is member SHS. CHS1, CHS2 and CHS3 are circular hollow sections.

A butt weld at Joint A would need a backing plate if the CHS walls are thin. Perhaps the elliptical stiffener plate at the miter joint could be used for this purpose.

A fillet weld could be used at Joint B if the wall of CHS3 is cut normal to the axis of the member at the curve where the ID of CHS3 meets the OD of CHS2.

BA

 

[ishvaaag]

The formula is just an interaction formula. It is one of the kind shown several times in the Galambos' IV and V books, corresponding to a fit for a safe limit for the interaction between shear and normal stresses.

Every time we talk of interaction, we talk of a limit condition, sometimes as a curve, others as a surface, and a fit to the safe limit of the behavior. Tresca, von Mises are examples.

In the case I refer to, excessive deformation by whatever way we discover it will mean that a more complex field of stresses has developed (say, tension field stresses may be extant at that moment) and hence a fit to the limit that only considers St Venant shear, extant shear and elastic normal stresses may be inadequate; there may be some warping restrained by tension field, large deflection stresses in place etc. An experimental formula (or sets of formulas) covering the failure of the connection would cover for all the significant aspects within the contemplated ranges of dimensions.

However the stated formula when the deflections remains quite small (mainly out of moderate stresses and/or relatively thick walls in the members) should give a good start to get a safe design.

Attached, Mathcad 2000 Professional for the design of a circular fillet weld subject to moment, shear and torsion. Note that taking the shear stress in a plastified statement as done in the worksheet is conservative, since it checks the points where the interaction is worst, namely where the normal stresses are bigger. The interaction check is as per Pilkey's and the deformations in the ring fillet weld assumed small.

 

[ishvaaag]

The attachment is a pdf printout of the worksheet.

 

[ishvaaag]

Well, taking the plastic distribution of the shear stress as in the worksheet remains conservative as long the shear does not become predominant, in which case checking at midheight with an elastic evaluation of the shear stress would be more conservative (and other formula should be used for the shear stress).