Notched LVL beam configuration problem

[focuseng]

I need some fun opinions on this one. I have an architect situation which requires some out of the box thinking and blah blah blah. Anyway, it has been proposed to design a series of beams shaped like the attachment. The beams would get cut out of large LVL stock (about 20" deep members) and be shaped with a "tray". I know the limits of notching are vastly exceeded here but on one hand this doesn't seem so crazy. I am sure it can hold some load but how much? Obviously not simple beam theory going on and it is wood...

So I propose it to the consortium: Any experience with this, tips, tricks, analysis theory here would be appreciated. I don't need a bunch of "why are you doing that?" etc. It is what it is and I am trying to either refute it completely or work with it with some attempt at engineering principles. My personal opinion is to drop it out the window but I am trying to entertain the possibility.

Thank you in advance

______________
MAP

 

[msquared48]

Depending on the loading, I think it is doable. A bit weird and architectural, but doable structurally.

Mike McCann
MMC Engineering

 

[Triangled]

the bottom face notches seem to be the weakest spot, so what kind of stresses might occur there? what if you looked at it like a 3 member truss consisting of (2) 1.75 x 9.25 top chords and (1) 1.75 x 12 bottom chord and a single point load 2xP on the apex. what then would the connection of the bottom chord to the top chords be?

 

[msquared48]

More like a haunched frame or modified arch,,,

Mike McCann
MMC Engineering

 

[JAE]

I think you might have problems at the two lower re-entrant corners. If the loading isn't too bad, I might consider adding some steel side straps across those to prohibit any cracking propagation.

 

[jdgengineer]

Instead of shaping the LVLs can you use a kinked steel plate (ie flitch beam) at some regular interval (say 10-15') and then frame more conventional framing between the kinked beams?

I'd also recommend contacting iLevel. They could probably give you some suggestions for stress concentrations etc if you need to go this route.

 

[focuseng]

Yeah, the lower inside corners are my weak point is what I figure too. Worried about tension on bottom edge and splitting. The trick with modeling it as a truss or arch would be that the tensile fiber strength is not the same once you leave the flat part and go to the angled heels because I am cutting through the grain of all the fiber below the flat. Then the heel is in tension across the grain. I think LVL is not cross laminated like plywood.

Fabricated steel beam is not an option.

Some of my ideas:
Treat it like a standard beam & reduce the section some to some arbitrary amount at the ends? +Safety factor of say 1.5?
Model is as a 12" beam with reduced section on ends, sitting on corbel/haunches? Maybe the steel plate gussets have merit for this analogy so if it splits a little, so what.

I haven't even got loading on this yet as it is still conceptual.


Ultimately I guess it will come down to some guesswork and the only way to really know is testing but it seems like there is still an inherent amount of strength in this just waiting to be tapped. After stewing on this more I'd really want to make this work just to shove it in my Arch's backside. I just need a rational approach/model to thinking about this. If not then I go a different route after yelling at everyone who doesn't know, to show me how it is supposed to calc out.

______________
MAP

 

[jdgengineer]

I'd also be very concerned as to how the beam is notched. The rip in the joist would need to be done carefully not to over cut. The contractor would want to drill holes in the corners to prevent over cutting.

Personally, if iLevel won't sanction it (and they do allow bigger notches / holes with further analysis) I probably wouldn't go with it unless the loading is really light. You'd be on your own if anything were to fail.

 

[Triangled]

As jdg notes, expert workmanship at cabinet making level would be important for that notch.

What about using a trapezoidal shaped beam with 14'1" unnotched bottom, bearing on two haunches with a concealed connection. Still some tricky design, but... Again, if loads are relatively small...

 

[focuseng]

Triangled, sounds like one of my thoughts mentioned above (second one). Or maybe I am not understanding?

The key for me is to come up with a reasonable model that is defensible with some engineering methodology and then play it safe a little after that. I am just brainstorming here as I know it will carry some load but can't quite design (prove) it.

______________
MAP

 

[Triangled]

yes, just as you mentioned above (second one), sorry i missed that....
and yes, doing the brainstorming thing only here too....

so as you said (second one ) above, then you have a 12" beam with tapered ends, granted rather radically tapered, but some tried and true design methodologies exist for these, although perhaps not this radically tapered, i haven't walked through it in awhile.....AITC Timber Construction Manual 4th edition starting page 5-216, and perhaps the architect could adjust the geometry a little to provide a nicer heel if needed.

the corbel wants to rotate inward with a moment of approximately P ft-lbs (guessing moment arm as 12"). so either need to a) secure the corbel to the wall with a moment connection (difficult with moment arm somewhere around 1 3/4" ... maybe extend corbel beyond the outside face of wall to increase moment arm) or b) secure corbel to bottom face of tapered beam. in either case, as the corbel rotates, it'll raise the upper-outer corner of the corbel and the 12" beam with it... so you get a little help there too.....

just my 2 cents, hope i didn't make a dumb error, but at least those aren't usually as bad as my smart errors....