Braced Frame- Single angle-to-gusset 2

[NS4U]

For braced frames that use single angles for the bracing, do you need to consider the eccentricity between the angle and gusset in the design of the gusset plate? If so, are there references that cover how to do this?

I know that for the single angle design, I need to use E5 to check the angle, which considers the eccentricity.

Thanks.

 

[Agent666]

Strictly speaking I would say yes as there would be some bending in the gusset, to some degree though the angle restrains the gusset if it has a few rows of bolts and this is why the angle is designed for the full moment. To be honest I never have taken account of it though when working from 1st principles in the past. I have seen a failure occur in a previous company I worked for due to it being ignored.

Remember any other type of section also has an eccentricity even if its only 1/2 the web thickness on a beam for example, but I have never seen anything in any codes/design procedures which require it to be considered. Usually if its a shear type connection between two bolted plates the full capacity of the plates is considered with no reduction due to any moments that are generated. Angles are a little different as the ecccentricities can be much larger.

The only advice I am aware of is the following report which does allow for this reduction in capacity, it is the result of some work that came about due to a number of failures in NZ and Australia of eccentric cleats connections including the one I referenced (primarily due to the member buckling sideways due the moments in the gussets causing an unconstrained sidesway motion when in compression in which the cleat goes into double bending). Its written in a New Zealand context though so may not be too useful to you. Link

There is a stiffened connection detail covered in the procedure that could be interpreted as being similar to a connected angle, the gusset in this connection procedure is designed for the greater of the eccentric moment or 2.5% of the axial load in the member acting perpendicular to the gusset applied at the centre of the connection (this is the restraint force required to be considered to provide restraint to the end of he member). The first order moment is factored up by a moment magnification factor to account for second order effects, then a combined actions check is undertaken on the gusset. Check attached except from this document.

The steel code I work with in New Zealand (NZS3404) also requires the angle to be designed for the compression and moment generated when the slenderness (length over radius of gyration) is less than 150. The tension capacity is also factored down when its connected by one leg by a factor of 0.75-0.85 depending on how its connected to any cleat. So it makes sense that the gusset should be designed for some moment.

 

[NS4U]

Thanks, Agent666. Great info.

 

[JoshPlumSE]

There is a good MSC article by Bo Dowswell on horizontal bracing that addresses this exact subject. Horizontal Bracing: An Overview of Lateral Load Resisting Systems and How to Implement Them. See figures 2 and 3 of the article for what I believe is essentially the same as your situation.

In it, they basically say that you design the BRACE itself for the eccentric load / moment. I was a bit skeptical when I first read this, so I built a couple of simple FEM models just to see what would really happen. Sure enough the vast majority of the moment (something like +90%) went into the brace and the gusset didn't see much flexure at all.

 

[paddingtongreen]

The simple view is that the gusset can only bend if the brace bends, so you check the brace for the eccentricity.

Michael.
"Science adjusts its views based on what's observed. Faith is the denial of observation so that belief can be preserved." ~ Tim Minchin

 

[Agent666]

Article Josh posted doesn't specifically address the gusset question as far as I can tell, only the angle which it's generally accepted in codes needs to be specifically checked.

The failure I mentioned was directly attributed to a published design model accepted at the time having not specifically considering the eccentricity in the connection and a sidesway failure mechanism in which the gusset hinges at the member end and at the supporting plate forming a mechanism that is not capable of restraining the end of the member from moving perpendicular to the gusset. The longer the unrestrained length of gusset the worse the issue is.

I found this article from SCI in the UK which reinforces the view that bending in the gussets should be considered/not ignored depending on the configuration.
Link