Lamellar tearing on flush end plate moment connections?

[abusementpark]

Has anyone seen any research on this? Design guide 16 doesn't talk about it all, unless I'm missing something. AISC warns about certain weld details that may be susceptible to lamellar tearing. These same detail would naturally be used to make flush end plate welds.

 

[LonnieP]

Not sure I'd lose any sleep over laminar tearing with an end plate. The end plate isn't restrained against rotation from weld thermal shrinkage like a beam/column moment joint is.

LonnieP

 

[abusementpark]

Is that because of the bolted part of that connection?

 

[LonnieP]

No, I've just never heard of a laminar tearing issue with a flat plate before. Probably something to do with how plates are rolled in the mill versus how column shapes are rolled. We had problems with Laminar Tearing in moment frames during the Northridge California Earthquake where the beam bottom flanges tore a chunk out of the column flanges. I assume that's what you are referring to as laminar tearing.

 

[JoshPlumSE]

AISC's Structural Journal had a good article (3rd quarter 1973) about lamellar tearing which I read in a graduate steel course some years ago.

I still have a copy of this article and glanced through it when I saw this posting. Based on my (extremely cursory) review of the article, I would think that AbusmentPark's concerns are justified. A number of the details described as "susceptible" to lamellar tearing would be similar to what we could see in flush end plate moment connections.

That being said, in these cases there is always an "improved" detail shown which shows a revised method of gouging the weld so that plate will not be susceptible to this type of lamellar tearing.

 

[dhengr]

Abusementpark:
You can get lamellar tearing on the end plate connection you are talking about. It is a cross grain, through thickness phenomenon, and can be caused when large cross grain stresses are built up, for whatever reason. The cross grain (perpendicular to the direction of rolling, or through thickness) mechanical properties are usually a bit inferior to the properties in the direction of rolling, due to work hardening and grain orienting during rolling. This difference is usually not significant enough for most Structural Engineers to worry about, on most structures. It goes unnoticed because of the way we design the structure and use or orient most of the material. We usually assume steel is homogeneous and isotropic and for most practical purposes this is true. The rolling process also orients the grain and any laminations if they exist.

You and LonnieP are talking about two slightly different details, although the causes and outcome (end result) of the problem manifests themselves in about the same way. Lonnie’s beam flange to column flange moment connection is very highly restrained, particularly in the region of the column web/flange/fillet radius, and the beam web/flange/fillet radius region, with each having their own residual stresses from rolling. Then the heavy welding and high restraint adds to the complexity of this residual stress picture, all in the same region. This same joint is much less restrained out at mid width of the flanges; but weld imperfections can cause problems again, out near the flange tips, a slightly different imperfection problem. Finally in the heavy column sections, if there are going to be slag inclusions, piping, or laminations they are most likely to occur near the column web/flange/fillet radius region. And, all of these combined with the high EQ moment stresses lead to the problem Lonnie is thinking of. Northridge was a real eye opener.

Your end plate detail is less restrained in the way Lonnie’s is, but your joint detail still can build up significant through thickness residual stresses due to the welding, and it does have high bending moment stresses in the same region. These will be acting perpendicular to any laminations in the plate, so they can cause tearing. In a rolled plate the laminations or slag inclusions will be flat, generally in a plane parallel to the surface of the plate, and of an irregular shape. The salvation in your case is that the shop should usually detect any significant lamination in the plate during the fab. process. These usually show up in shearing or flame cutting from a larger piece of plate, or in the hole drilling process, etc. The reason small slag inclusions and laminations go unnoticed is their orientation and location in a piece of plate. If you are using a UM plate for a plate girder flange, a small lamination will most likely be oriented in the same direction as the stress field in that flange, so it is generally of little consequence. Lamellar tearing is almost exclusively a cross grain, through thickness, perpendicular to the defect (lamination) phenomenon.

 

[Ron]

Essentially any rolled shape, plate or otherwise, has the potential for lamellar tearing. It is a rolling defect.

 

[LonnieP]

I agree with dhengr and Ron, there's always a tearing potential. Has anyone out there actually seen an end plate tear, and what was determined to be the cause?