Wood Portal Frame Connection 3

[RFreund]

I should start by saying that I understand wood moment connections are not rigid and may not even be considered moment connections. However I was trying to design steel L-plates that would bolt to the faces of a column and beam to make a semi-rigid connection for a wood portal frame. The beams may notch into the columns as well. However I'm not sure that I'm distributing the load to the bolts correctly in the case of a lateral load. I tried to basically do a elastic vector analysis but the L-shape plate and connection is throwing me off or something. See attached.
This detail may not be the best solution for a problem but its the analysis that is really is bothering me.

EIT

 

[JoshPlumSE]

It's very difficult to make a wood beam-column connection into a moment bearing connection. I've never done it, but I have heard lots of engineers saying "BEWARE".

Have you considered using kicker braces or knee braces instead? this can get you portal frame behavior, but without every having to transmit moment through a connection.

 

[RFreund]

Yes, we may use kickers.
Its a small balcony roof and it may be possible to consider the roof diaphragm cantilevered, however I would like to have some support at the columns.

Even if the kickers were used instead I would like to know how to rationally design the connection.

Say it was a steel tube column and steel tube beam would the analysis I have shown be appropriate?

Thanks.

EIT

 

[Splitrings]

The load in any bolt is +/-(P/a) +/- (Mc/I). I for the bolt group is the sum of distances squared. The distances are measured from the C.G. of the bolt group.

 

[RFreund]

Thanks,
Thats what I was trying to employ however I'm just not sure I was doing it correctly. The L-shape of the plate and the lateral load is throwing me off. therefor I assumed a portal frame with a pin base and a hinge at the center of the beam. I used the shear/reaction at the hinge in the beam as the P and the resulting moment as M and Isolated the one side of the L. Then for the other side I used the reaction at the base as P and again used M and Isolated the other side of the L. I'm sure thats as clear as mud but hopefully the sketch clears things up a bit. Even though we are going to use kickers and plates, it bothers me that this seemingly simple analysis is confusing me...

EIT

 

[MiketheEngineer]

The analysis is generally rather straight forward - however - you will probably not like the answers. In other words - moment connections in wood are hard to show good.

 

[dik]

Have you considered using 'glulam rivets'?

Dik

 

[RFreund]

dik -
I will look into the glulam rivets however I believe we will end up using the knee braces or a combination. Why do you suggest rivets?
mike-
I'm not sure if you opened my attachment or if it made sense but is what i have semi-correct? or how would the analysis be done for an L plate on each side of the column beam intersection?

Thanks again guys.

EIT

 

[Gillespie]

I've chainsawed in steel plates (like your L-plate)
and bolted (exposed and countered) to the wood to get a rigid
connection for lightly loaded wood timber frames.
I'd recommend getting a rigid connection at the base (more than just a simple simpson post base to help stiffen up the frame.

 

[RARWOOD]

There are two main problems that you encounter in wood moment resisting connections. The first one is that the fabrication is often not precise enough that you get proper alignment of the bolts. As a result some of the bolts may carry more load then the other.

Based on my experience the major problem is that you get some small amount of rotation in the joint before the bolts start taking up the load.

 

[RFreund]

I appreciate the comments.
Any thoughts on the analysis of an L-shaped side plate?
Thanks again.

EIT

 

[dhengr]

Rfreund:
As suggested by Gillespie above, chainsaw a center kerf in the beam and the column, make on L pl. the same thickness as the saw kerf, and for my discussion, say 24" long on each leg. Weld .25" flange pls. with width same as your wooden beams and columns to both the inside and outside of the L pl., you now have a WF shape. Drill the web for some smaller bolts to hold the whole thing together, but they’re not really taking the moment induced load any longer. Tight fit the beam and column members btwn. the flgs., and the wood takes the moment induced loads in bearing perpendicular to the edge grain, with lever arms of about 16" btwn. the opposing load block areas. These load blocks are akin to the compression block in a conc. beam, with some lever arm btwn. them. Use dry wooden members and I won’t guarantee zero angular change btwn. the two members at the jnt.

 

[RFreund]

Wow, thanks dhengr and gelepsie that is a good detail. I like it. They may complain about the welding but I think this is a good solution and will use it to bring up with the bossman.

Regarding analysis (this is the part I really enjoy) I'm not going to get any comments on my proposed/attached detail am I (it would be greatly appreciated)?
As for dhenger/gelepsie detail if you were to try and put a number to it what would you assume for a compression area maybe 1/4L and moment arm (0.8L), just curious what/why you would assume.

Thanks again I really appreciate the input.

EIT

 

[Splitrings]

RFreund,
I attached my thoughts on the analysis. I am not certain my signs are correct. I can't spend any more time on it. Maybe this will stimulate conversation on your original question! I know one reason these connections are not recommended is that they induce tension perpendicular to the grain. I would suspect placing the angle on the underside and top sides would be a better connection for this reson.

 

[RFreund]

Thanks Splitrings, I appreciate your time and everyone else.

EIT

 

[cg3375]

This problem comes up in work often and we have never had a consensus happy answer. I was directed by several people to a Dr Dan Dolan out at Washington State. He was polite and more than helpful. His response to this problem was as below


"However, you could engineer a moment frame out of wood by using a glulam rivet (timber rivets in the NDS). What I teach for seismic design (and it could be used for wind as well) is to design the timber portions of the frame as if they were glass (remain elastic) and use sacrificial 1/4 inch steel plates and
timber rivets for the connection. Design the shape of the plate to insure it will be the weak link and yield and then the performance of the frame is similar to yielding steel moment frames. "

 

[HouseBoy]

I don't see how approximately 306# can be right:
There is a moment in the steel plate (at the corner) of 5000 ft. lbs. that is resisted by a bolt group that is max 12" wide. Wouldn't there be at least 2500# in each bolt?

The "problem" I usually find with bolted wood connections resisting moment is that the bolts must be spread out very far in order to resist the moment sufficiently.

Additionally, for a 1/2" x 5" plate (guessing at the size assumed), wouldn't bending stress be about 30 ksi?

 

[Gillespie]

Houseguy is right. It eventually all comes to the damn bolt couple. So try to space out the bolts (the couple). 12" is a joke. On another point, this is no different than calcuating the strength of a guardrail post. Point load at 42" high is difficult to resist thru a bolted connection with a 8 or 9" bolt couple distance.

 

[csd72]

you also need to check the shear in the wood member between the opposing bolt forces.