Pipe Weld to Stringer

[Banks]

I'm looking for an accurate way to check the weld between a pipe and the top of a C-channel stringer. This is for a guardrail system and the pipe is assumed to have a fixed base (I don't know how else to get it to work otherwise). I know that when contractors are doing this type of connection they are just using a 3/16 or 1/4 inch fillet weld around the entire circumference of the pipe. The best that I could come up with is getting the centroid of each half of the pipe, figuring out the tension and compression forces based on the moment and using that to check the adequacy of the weld (also including the shear force...). The only thing I could really find in the AISC manual was for the design of the member at the connection (chord plastification, punching shear), but not for the actual weld. Does anyone have any other / additional suggestions?

Thanks...

 

[connectegr]

This is more difficult than it may first appear. For most stringer channel sizes the flange width is 1 1/2", which is equal to the 1 1/2" OD of most handrail posts. Therefore it is not possible to provide a fillet weld around the entire circumference of the pipe. Also if detailed as fillet welded, the least weld is provided in the regions of most stress. If a channel is selected with adequate flange width a fillet weld can be designed with adequate strength, using the allowable increase for fillet welds loaded axially. If insufficient width is provided, an insert can be used, providing backing for a square grove full penetration weld. We deal with these conditions frequently, and this is often a debate between the fabricator and engineer.

http://www.FerrellEngineering.com

 

[Banks]

Thanks for the input connectegr!

We're actually using a C12x20.7 so we have a 3" flange, which is good. I have checked it based on finding the centroid of each side and going from there and found that for this case a 3/16 inch fillet works. My concern is that I have to submit this to an EOR and the calculations were rejected in the first place for not checking the weld at the base of the pipe. I'm just trying to make sure that I have all my bases covered so that the calcs are not rejected again (and hopefully it doesn't become a big debate between us and the EOR).

Whether or not it's the right thing to do, I have never provided the calculations for the weld at the base of the guardrail post, so this is the first time I've had to check it.

 

[paddingtongreen]

I don't understand, "finding the centroid of each side and going from there".

If you work out the stress in the pipe, multiply it by the thickness, you get a load per unit length to use in designing the weld.

Michael.
Timing has a lot to do with the outcome of a rain dance.

 

[unclesyd]

You might want to checkout Anchron's program "Welds" if you need repetitive calculation.

http://www.archoneng.com/

These a book "Design of Weldments" by Blodgett available from Lincoln Electric for a very small cost that you might want to get.

https://ssl.lincolnelectric.com/foundation/store.asp?PID=16&cat=8

 

[Banks]

Thanks unclesyd - I downloaded the demo of the software and was getting the same results that I was getting by checking it by hand. I'll have to check out that book and see what they have to say.

@Paddintongreen: Basically what I was doing was using the ole T=C=M/d. For my "d" distance I was finding the centroid of each individual side and combining them for the "d" distance. I then figured the tension (and compression) force and used the arc length of 1/2 the pipe to find out the required weld.

My boss said I was putting too much thought into this, and I'm not really worried about the welding at all, but, like I mentioned before, someone is checking this over and I don't want there to be a single issue.

Thanks again everyone!

 

[Banks]

@Paddingtongreen

Sorry about misspelling your name in the above post.

I also meant to ask about your response, because that confused me a little bit. I know the stress in my pipe, but I'm confused on how multiplying it by the thickness gives me the load per unit length.

Unfortunately I have to run out of the office right now so I can't expand my confusion, but if you could clarity a little bit i would appreciate it.

 

[paddingtongreen]

The stress in the pipe, f=M/z, force/area, force/(thickness by length) If you multiply it by thickness, you get force/length.

Thus t*f_max gives the design load for the weld, in American units, (kips/sq.in.)in. = kips/inch.

Michael.
Timing has a lot to do with the outcome of a rain dance.

 

[Ron]

Banks...your procedure to check the weld stress based on the resolved moment is correct. One side of the post will be in tension, the other in compression...so you'll have a weld with only half of the area capable of withstanding the lateral load (induced moment), since compression doesn't help you in this case.

 

[dvd]

You should think about more than resolving stresses on the railing. Are you shipping steel with handrail already welded to it? If you are field welding the posts to the stringers it will be expensive (but the ironworkers will appreciate the work). Weld a plate with holes to the post and make a bolted connection. You need to ask someone how to best attach the railing before you concern yourself with the analysis.

 

[Banks]

Thanks again for the responses.

@dvd: The guardrail will be welded in the field - that seems to be typical procedures for any of the fabricators we work with. Due to the width of the flange I can't really get a bolted connection in there that would be sufficient to resist the moment, but I appreciate the suggestion.

@paddingtongreen again: The only worry that I think I would have with your suggestion is that I would get an unconservative value based on the post being a pipe. The weld at the center would be doing less than the weld at the extreme edge. This was why I was looking at taking the centroid of each side of the pipe to figure out where the tension force would be acting and figuring out the weld from there.

Anyway, I think I have at least confirmed that what I'm doing should be okay. Like I mentioned before, I wasn't really worried about the weld in the first place and the calculations that I have figured out confirm that I my assumption. I guess from here on out I'll make sure to include that calculation in my submittals to prevent anymore issues.

 

[paddingtongreen]

@Banks, I was describing the actual behavior. You do get lower stress in the center. But by the way I describe, you get enough weld to hold the highest stressed part of the pipe, but you get it all round.

You cannot just assume a different load distribution than the pipe wants to deliver.

Michael.
Timing has a lot to do with the outcome of a rain dance.