Channel splice welded with another channel 1

[PSSC]

I had a bolted beam splice (channel), however due to things moving the bolted connection has to be removed.
My first option would be to simply weld flanges together in the field with a welded splice plate on the web.
For safety and other reasons I do not believe it would be possible to weld the bottom (tension) flange in the field.
My next option is to weld a smaller in height but equally strong channel to the backside of the original channel, basically a MC10 welded to the back of a C12.
The connection has to resist moment and shear but is braced at the center.

My real points of concern are two things
1, with a 1" gap above and below between the two channels flanges does the force transfer well from the C12 to the MC10? I had though of putting end cap plates on the ends of the MC10 that would extend to the edges of the C12. This is maybe overkill.

2, the weld, I have looked at this a few ways, I would rather just design the connection to handle the entire allowable moment for the C12. I have ran the weld calcs a few different ways, as a rectangle with the moment in the center and as two "C" shaped welds, these turn out to be the same thing after ran the numbers and actually looked at it.

Does anyone have any suggestion on the how to look at this weld or any other insight?

I have tried to attach a sketch.

thanks

 

[r13]

For safety and other reasons I do not believe it would be possible to weld the bottom (tension) flange in the field.

Please elaborate these issues, as I don't see good way out. The MC10 to C12 connection should be able to handle the strength requirement, but with other concerns, such as deflection, and the stresses in each member are everybody's guess, unless validated through a finite element analysis.

 

[PSSC]

This a piece that will ship in 2 pieces. The splice would be installed while on the ground with no loads applied and then lifted into place.
However if I can avoid having to require the field crew to block the pieces up and someone crawl under it to weld the bottom flange, I would like to.

 

[PSSC]

The C12 was satisfactory for strength and deflection, the MC10 has a slightly larger Z value and would be satisfactory. I am now trying to make sure that I am looking at weld correctly and that the 1" gap at the top and bottom between the flanges will not inhibit transfer of forces between the two channels
I chose the MC10 in order to allow space for the weld.

 

[r13]

Provide a backing plate/bar under the bottom flange, then butt weld from top side. Rather a simple process, any welder can do it.

 

[Retrograde]

with a 1" gap above and below between the two channels flanges does the force transfer well from the C12 to the MC10?

The flange force needs to transfer from the C12 flange, through the C12 web, through the weld, and into the MC10 flange. Therefore the overlap needs to be long enough such that the shear stresses developed in the web of the C12 due to this force transfer are not excessive. Then size the weld to suit.

Something like this. Force in flange = area of flange x yield stress. Force in web = 0.6 x web thickness x yield stress x overlap. Solve for overlap and add a bit for some conservatism.

Edited for typo

 

[PSSC]

Retrograde, when you say "overlap" are you referring to how long the MC10 overlaps on each side of the joint? I sized the overlap for the weld but hadn't considered this. Thank you

 

[Retrograde]

Yes that's what I mean by the overlap.

 

[PSSC]

Retired13, that weld procedure will work I believe. Thank you

 

[KootK]

I sized the overlap for the weld but hadn't considered this.

In my opinion your original approach, reiterated below as I understand it, was satisfactory or very close to it.

 

[KootK]

It is actually the out of plane issues that concern me most which is why I liked the sound of this and would improve upon it if possible.

The connection has to resist moment and shear but is braced at the center.

 

[KootK]

And just some detailing concepts because this stuff's super fun to talk / think about.

 

[PSSC]

After looking at the drawings again this morning, I remembered why I cannot use the weld with the backing bars. Part of the original bolted connection involved 1/2" plate at the end of each channel, these plates are not being removed and if they were there would be an inch gap to fill.
So I am back to my original design.

KootK,
First thanks for looking at this.
I designed the weld like you have shown above, I actually used Blodgett's book, it's old but it has served me well for a long time.

My question is more about what moment should I use to design to.
I have the moment and shear diagrams for this beam but I basically just want this to be as strong as an unspliced beam.
I used the 3 sided weld applying half of the allowable moment for the shape. Using the idea that half the moment will go into each side.
I believe this overkill but they need a fix, I need to work on other things.


Using Retrograde's check I needed roughly 8 1/4" length, checking the weld, I needed about 18" for 1/4" weld.

thanks

 

[PSSC]

KootK,

I cannot brace it from the MC10 side.
The toe side of the C12 is braced at the center (where the splice is) and bolted to a channel above it at 6" increments.

 

[KootK]

I have the moment and shear diagrams for this beam but I basically just want this to be as strong as an unspliced beam.

Retrograde's proposal to effectively develop the yield strength of the flange will be prudent then. Your design connection moment would be roughly the material yield stress times the plastic section modulus of the channel being spliced.

I used the 3 sided weld applying half of the allowable moment for the shape. Using the idea that half the moment will go into each side.

It's the same moment, and the whole moment on both sides of the joint that needs to be developed in the connection. Halving the moment puts you in violation of static equilibrium.

Using Retrograde's check I needed roughly 8 1/4" length, checking the weld, I needed about 18" for 1/4" weld.

18" sounds altogether reasonable to me. My gut check minimum would have been 1.5 x the member depth which, handily, works out to exactly that. Of course, if you need to double your moment now, this dimension is likely to grow.

 

[KootK]

Another concept that I like to make the joint more efficient if you need it.

 

[KootK]

Green dots? Given what we're doing here, I like the idea of being able to form a moment couple with the braces in order to counter some of the natural eccentricity in the joint.

 

[KootK]

You mentioned end plates on the splice piece at one point. Set up correctly, I can see some logic in that now. It would kind of create nominal, weak axis moment connections between the splice and the main members and make a single brace point more robust.

 

[PSSC]

It's the same moment, and the whole moment on both sides of the joint that needs to be developed in the connection. Halving the moment puts you in violation of static equilibrium

Ok this has always confused me, if there was a simply supported beam with a moment at the center, the reaction at each side would be equal and opposite but you wouldn't ignore what the other reaction was doing. You don't just assume the moment is going one direction.
I originally used a rectangle shape as the weld model with the entire moment at the center.

 

[JLNJ]

If the channels are not fully stressed (i.e. if the flexure can be carried by the web only) then there is less to worry about. Channels have pretty thick webs and those webs carry much more of the flexural load than, say, a wide flange beam.

This is the way I look at bent stair stringers, the welds at the knee, and whether or not stiffeners are required.