Lamellar tearing on moment beam/column web connection 2

[Italo01]

Hello,

I specified the connection depicted in the first image attached, in which the tension of the plate is transmitted through the web of the column. The fabricator just started to fabricate the connections and i realized that the connection between the plate and web is not good for lamellar tearing. I could use the connection depicted in the second image, but since the fabricator already started to fabricate the connections, i'd like to change the connection only if necessary.

The moment on the connection is not high and the stresses on the plate for the ULS are below the yield stress. The plate is A36(8mm) ,the column is A572(5mm) and the weld will be a 5mm fillet weld. I know that thinner plates are less suscetible to lamellar but this is vague and i wish to know what parameters can be used to determine this susceptibility.

Thank you.

 

[Italo01]

This is the other option that i could use.

 

[HTURKAK]

In general , lamellar tearing could be a risk in steels over 20 mm thk. You may ask ( UT ) ultrasonic testing to detect .

 

[Italo01]

Thanks HTURKAK.

Do you know any article that treats this question of thickness influence in a detailed way?
Since the structure is not built yet, it would be better change the connection before than build and test it. I would like to know if i can rule out lamellar tearing completely for this case in order to maintain the connection how it is.

 

[DrZoidberWoop]

Regarding the flange plates: You can control the load path of the flange force to avoid putting severe tension loads into the column web. In some cases, you can simply not connect the flange plate to the column web, and only weld to the inside flanges, up to the toe of the column's inner fillet (k1).

Regarding the shear connection: An extended shear plate would likely be preferable. Your current connection is likely the least preferred configuration for the erector. End plates and double angles to column webs, with flange plates, must be stabbed in and swung into position at the other end of the beam. Make sure your assembly is erectable.

 

[dhengr]

Italo01:
Lamellar tearing usually involves highly restrained welded joints and thicker, heavier plates and rolled sections, and much heavily welded details. Then, it involves stresses/strains which generally cause tension through the plate thickness. And, it is this through pl. tensile stress/strain which can cause the lamellar tearing. We usually ignore this, but heavier stl. pls./materials are not perfectly isotropic, they are slightly inferior/weaker in the through pl. direction. In good part this is caused by the rolling process, and then some slight rolling imperfections and residual stresses caused by the cooling process can compound the problem. In a good share of our structures the tension stress field runs parallel to the material planes which might be tearing prone, so this causes no problem. The stress field flows right along with the slightly weaker material planes, around them, not through them.

As mentioned above, your connection detail should be reviewed, reconsidered and probably changed for several reasons. The difficulties of erecting the beam with your details is a big one. That is a very congested connection area, is another. Tell us the actual sizes of the various members and pls. so we can get a better feel for the actual proportions. Show us some real details of those connections, with dimensions, etc. What are the moments, shears and stresses, do you really need the moment connection or would a slightly larger beam allow a better solution all around, at no more cost? The proportions and thicknesses in your drwgs. don’t scream lamellar tearing to me. But, your det. #1 shows the flg. end pls. (8mm thk., A36) going into the web inside the web/flg. radii (the k area) and you say you need a 5mm weld. This will cause pl. bending in the col. web pl., and concentrate the tension reaction out near the tips of your weld, near the stiff k areas, and tend to unzip that weld. Your det. #2, with clipped corners, at least allows you to weld to the insides of the col. flgs. to take most of that tension reaction, as Dr.Z suggests, lessening any web weld issues.

Google ‘lamellar tearing’. AWS and Lincoln Electric (both Omer Blodgett & Duane Miller), and other welding suppliers, and various Unis. all have had some good writings on lamellar tearing and its potential and problems. This all really came to light after the Northridge EQ in Calif., when so many highly restrained moment connections failed. It was not unknown before then, just not paid much attention to. If you take care in detailing it is usually not an insurmountable problem.

 

[Italo01]

Thank you for your replies.

With respect to the erection, i'm in touch with the erector to deal with this connection.

dhengr, the reason why i'm using this connection is not because of the beam capacity but because i cannot brace the structure on this direction and need some moment connections. Since the wind load is small and i'm in a region with very low probability of earthquake, the moments are low and the tensile stress on the plate is about 12ksi, while the plate is A36 and the column is A572.