Flitch Plate Beam Question 1

[kxa]

I have a situation where the beams (2-11 7/8" LVLs) are already in place and there is about 5" of clear height on the bottom right below the connecting floor joists on each side. I am thinking of adding two 5" wide by 1/2" thick plates on the bottom. One on each side to create a composite section. I am using a transformed section and calculating the top and bottom Sx based on the total Ix.

Should I just calculate the top and bottom stresses and compare them to the LVL's allowable stresses. For some reason I get very small stresses.

 

[Lion06]

You are going to add a 5" high plate (on each side) to the bottom of a 11 7/8" high LVL, is that correct? The first question I would ask is why are you doing this? Was the LVL not designed right to begin with, can it not take the load?

 

[enginerding]

You will need to check both the stresses in the steel and those in the LVL. If the stresses in the LVL are low, it is because the steel is taking more of the load.

I think the more difficult part of this task is getting this beam to act as composite. What type of connections would you use to ensure the members are deflecting together? Typically when a flitch plate is designed, the wood on either side of the plate is ignored and the steel is assumed to take all the load. The wood acts only to brace the steel.

 

[miecz]

Agree with enginerding. Calculating a composite section is one thing. Providing the connection is entirely something else. I would treat the steel as being at the neutral axis of the LVL and simply add the transformed moments of inertia.

 

[MiketheEngineer]

Not sure I completely understand your problem - the 5'' thing does not make sense. Anyway - most flitch beam calcs will show that the wood carries very little load -- wood deflection too high versus the steel. A lot of times I just assume that the steel will need to carry all the load and the wood is used to stabilize the steel (Lc and Lu). Makes calcs much easier and is conservative but not overly.

Beamchek.com has software that will analyze that - if you want.

 

[apsix]

Why not use 2 channels (are 5" channels available in the US)? You get greater stiffness per weight of steel.

I also agree with the others. Forget about composite action and assume the steel is taking all the load.

 

[kxa]

Thanks for all the comments. Unless I am missing something here, I have two 12" deep LVLs that can carry over 500 plf. Why should I disregard them? If I convert the steel to the equivalent wood section, the neutral axis comes down but the over all section strength increases. I can even jack the beam for the amount of dead load and then attach the steel plates or the channels.

Anyway, if I disregard the wood the steel won’t work for the 18’ span and over 800 plf. Based my calcs, it will work as a composite section. I just wanted to make sure I was doing it right. Any input is greatly appreciated.

 

[PMR06]

kxa- Of course you can analyze the section as composite. The material is there, why not use it? Plus, it sounds like it's not as simple as a steel plate on the same neutral axis as the lvls.

The other suggestions of ensuring composite action occurs are valid though. And your idea of jacking the beam is exactly right. Otherwise the existing lvl would be stressed with the existing dead load, then the new live load would add stresses proportional to the new transformed section.

Regarding the stress check, you are correct in transforming the section and calculating I and Stop, Sbot. You can transform the section twice, once to equiv LVL and once to equiv steel to check the stress in each material. I find this easier than keeping track of a modular ratio. Deflections are easy. You should get the same deflection whether you choose equiv steel or lvl.

 

[miecz]

In order to provide composite action between the wood and the steel, you need to ensure no slip between the two materials. I don't see how you can do that. I there a reference in the literature that shows how to make the connection?

 

[PMR06]

jmiec- Correct me if I am wrong, but I think you are picturing this beam similar to composite action as one would have with a steel beam and conc slab. In that case, you are correct, there cannot be slip between the materials. However, flitch beams are not composite in that sense. They are essentially three beams bolted together side by side. kxa's example is slightly different in that the NAs do not line up, but that just means a more complicated calc of I and both an Stop & Sbot. Search this site for "flitch" you'll find numerous forums on this topic.

 

[ataman]

To ensure composite action you would need to calculate the shear flow and ensure that there are enough connections (probably bolts) between the LVL and the plates. You should use a timber design manual to calculate the capacity since the timber component of connection will probably govern as opposed to the steel component.

 

[ataman]

Alternatively, you can design the beam without considering composite action and design based on deflection compatibility (relative stiffness). Since it is just a plate and only 5 " tall the stiffness will not be that great but might enough to take the balance of load that the lvls can't handle.

 

[enginerding]

ataman is right. Forget the composite action and use deflection compatibility. Place bolts (or lag screws, etc.) as needed to brace the plates. Hopefully this will be stiff enough to take the balance of the load. Remember to check the resulting stresses in both the steel and the wood when you have distributed the loads among the beams.

 

[miecz]

PMR06-

You are right, if kxa is simply adding the transformed Is (see my original post), one does not need to calculate the shear flow as in VQ/It.

However, kxa pointed out that the plates were "on the bottom". This implies that the difference in neutral axis is a consideration. Further, you stated "the NAs do not line up, just means a more complicated calc of I", and both an Stop and Sbot.

If one is simply adding the transformed Is, then it doesn't matter that the NAs do not line up, and the calculation of I is not effected, and Stop=Sbot.

Anytime one locates a composite neutral axis and then starts adding A*d squared terms to the transformed Is, there can be no slip between the two materials.

 

[kxa]

Maybe I wasn’t that clear. I am not adding the steel plates to the bottom of the LVLs. The plates will go on the sides of the LVLs right below the floor joists. There is about 5” clear area.

This should not make the clac’s too complicated. Once the new neutral axis is calculated which will be below the LVL’s N.A., Sx top and bottom are figured out and stresses are determined.

 

[miecz]

Doesn't matter whether the plates are on the bottom or on the sides. Once you start adding A*d^2 terms to determine your total moment of inertia, you have to prevent slip between the two materials so they act as a unit. I expect the connection would be some variation on VQ/I. Is there an example in a textbook or anywhere else that shows how to do this?

 

[xxpegasus11]

Flitch beams are commonly assessed based on deflection compatibility. Load taken by timber/plate is based on their stiffness. In calculating load taken by steel plate its best to use the minimum timber E for maximum load on plate. Conversely, timber is checked based on its mean E value. You need enough bolts throughout the span to transfer the load from timber to plate, and bolts at ends for load to be transferred back to timber if only timber is bearing at supports. ( as would be by making plate slightly shorter than timber beam to allow for shrinkage)

 

[Lion06]

Please tell me if I am interpreting this problem correctly or not.
Don't you have to ensure composite action for this to work? Even if you put the top of plate directly under the joists framing into the lvl, you can't count on bearing to transfer the load into the plates because the bearing distance will be too small. Can you use a 5" channel to achieve a greater bearing distance?

 

[xxpegasus11]

Myerges,

In terms of looking at it as sharing the load through bearing transfer , no I dont think you are interpreting it correctly. The thickness of the plate is therefore not the factor as to how much of the load the plate will take. Transfer of load to the plate is via bolting of the timber/plates and the amount of load taken calculated on deflection compatibility.

 

[haynewp]

If you are transforming the steel into an equivalent amount of wood (or vice versa) and calculating a new neutral axis location, then this is called composite and adequate connections have to be made to account for shear flow. This is the most efficient use of the beam and plate material.

Otherwise, the steel and wood will share load relative to their E*I's, each material having its own neutral axis location. It is possible to make it work either way.