Partial beam splice 1

[Xairi]

Hi everyone, new user of the forum, long time reader! So here is my problem, I have to do a beam splice, using only the top flange and about half the web of the beams, the other parts are not accesible.

Something like the following image:

It's an strange setup, but I guess it should work. It's a cantilever part of the beam, light load, the top flange will always be on tension.

So, usually for splice beam joints, you would leave a gap between the two beams, I am thinking in this case, it would be best to leave no gap, so the lower flanges will be compressed, and that way the load on the bolts will be lower, as the moment arm will be higher, something like this terrible sketch:

Would you guys agree about that?

To calculate the bolt forces, I would make a moment/force equilibrium problem according to those load distributions and try to verify with a FEA model.

 

[JAE]

I would be tempted to add a horizontal stiffener plate on each side of the webs - to create an intermediate flange at mid-height - then treat the top and middle flanges as the spliced member.

Check out Eng-Tips Forum's Policies here:
faq731-376

 

[Xairi]

Thanks, that's a good idea, there is not option no weld, only drilling is allowed.

 

[SlideRuleEra]

Xairi - Perhaps using angles (as long as necessary, even full beam length) for the web splices, instead of plates, will created the mid-height flange JAE mentioned, without welding.

http://www.SlideRuleEra.net

 

[Xairi]

Hi, that's an interesting idea, but for this case, I don't think we would gain stiffness on the joint or get to lower the bolt load.

Another problem that I have here is the lack of friction on the beam surfaces, so a pretension on the bolts won't be effective.

I am going to calculate the joint with the most conservative hypothesis, considering the rotation centre on the web, and check if that is enough against the external load.

 

[Agent666]

The first 'full depth case' would only work effectively if you were able to mill the beam ends to achieve the end bearing condition similar to the requirements for say an end bearing column splice. If you can achieve those tolerances on site for the bearing surface then this would be my preferred approach for the reasons you note (however you didn't give us the load, design actions or location relative to the end of the cantilever to make an informed decision on whether the 2nd part height option would calculate out to work, even if it's ultimately a less desirable arrangement). It's not clear if the reason for the lack of access is due to one of the beams being already in place or something, which would effectively mean you couldn't mill the beam end cut for achieving the required end bearing.

The angle cleat proposed by SlideRuleEra only helps with the cleat strength, loads transfer into the beam is otherwise essentially the same as a flat plate. I think Jae was more heading in the direction of welding a longitudinal stiffener to the web to aid in that part of the loadpath and effectively turn it into a shallower equivalent I-section at the connection location.

I don't think you need an FEA model to tell you the forces, simply working on the basis of the elastic distribution of loads as you sketched to the various bolt groups would be sufficient. It's no different in the 2nd part height case as splicing a beam tee.

One other alternative would be to use a longer channel sandwiched each side of the web, one set of bolts at channel ends, one adjacent to the gap (either side), resolve the bolt forces into vertical forces transferred through the channels.

 

[IFRs]

If your critical load path is bolt bearing make the distance between the bolts greater.
Is there even a remote a possibility of load reversal that would put the lower end of the joint in tension?