Bolted Doubler Plate Design 1

[jseng9]

I am reviewing a steel beam that does not have requisite capacity for my design end reactions because of the coped section. The fabricator has provided a doubler plate that is bolted onto the beam web. Does anyone know of an AISC design example that shows how this works? The closest I can find is II.A-6 but this example provides a horizontal stiffener plate. Also, how can this connection work without a longitudinal stiffener? See a picture of the connection attached.

 

[JoshPlumSE]

Tricky -

If it were welded, I'd probably take it as the strength of the coped beam (14th edition manual pages 9-6 through 9-8) using the some modified geometry and the total thickness of the weld. Or, I'd use the strength of the double coped web (page 9-9) again using the total thickness.

For a bolted plate, I'm not sure. Certainly you can count on the coped beam strength (using only the thickness of the web) from pages 9-6 through 9-8. I'm tempted to say that you could calculate the plate strength (per page 9-9) using only the plate thickness. Then add those two strengths together for the full strength of the connection.

I'm tempted, but something makes me a little skeptical that that the plate will really provide that much strength.

 

[jseng9]

Thanks for your reply! I will try using the AISC examples and pages you have referenced to see if the plate strength and the coped beam strength are enough to meet the demand.

 

[CANPRO]

Need more info/better sketch. Is that a double angle connection, bolted to the beam? So the bolts nearest the support would pass through - angle/doubler plate/beam web/doubler plate/angle?

Either way, I'm going to say that I don't think you have a reasonably accurate way to determine load sharing between the beam web and the doubler plate. The load is transferred to the doubler plate via the bolts between the beam web and doubler plate, and then into the connection angle through another bolt group. The load transfer through the double plate causes rotation on the bolt group at each end of the plate - essentially the double plate has to rotate a little bit to pick up load.

The main issue here is that the beam web is going to pick up the slack until the double plate engages, and I'm not sure how you quantify that. Sounds like something that would need to be tested.

If you absolutely had to do this this for some reason, I would extend the doubler plate back into the beam further and provide more bolts...even just one more vertical line of bolts I think would make the double plate contribution a bit more believable.

 

[jseng9]

Hi CANPRO,

Thanks for your response. This is a W12x19 beam with an end reaction of 42 Kips LRFD. The W12x19 frames to a W14x149 girder. The beam to girder connection is a double angle connection with 3 rows of A325 bolts on the beam web. The doubler plate is a 1/4 inch A36 steel plate 8-1/2" x 11". The fabricator has specified the doubler plate on each side of the beam web.

I didn't quite understand your first sentence. Are you essentially concerned with the load path and transfer of forces from the beam web to the column? You are basically saying that there isn't a well defined mechanism for the forces in the beam web to be transferred to the doubler plates if I have that correct?

 

[CANPRO]

I think the mechanism for load transfer is fairly well defined, I'm just saying you don't know how much of the shear is going to be carried by the double plates vs the beam web.

 

[KootK]

Also, how can this connection work without a longitudinal stiffener?

1) It can work. The coped, top flange-less section just has to be able to handle shear and flexure without over-stress or web buckling at the top.

2) I think that the place to start here is by giving name to the failure mode that we're trying to address to begin with. Then we can meaningfully discuss whether or not the proposed detailing addresses that failure mode. Do we know what failure mode(s) would have doomed the un-reinforced cope?

3) I don't see this beam cope having a shear problem to begin with so I doubt that the intent is to move shear out of the web and into the doublers. And that's good, because I don't see the doublers accomplishing that feat with any efficacty anyhow.

4) My guess is that the plates are specified as a means of preventing local compression buckling at the top of the coped web. And they probably do a pretty good job in that role. You might check this by treating the top of the web as a longitudinal compression member and seeing if the doublers possess enough stiffness to restrain lateral buckling of said, faux column. Given that you're starting with a 1/4" thick web and effectively tripling buckling resistance with the addition of the two 1/4" stiffeners, I suspect that it works. But, of course, all this depends on my having guessed the failure mode correctly.

5) If it were me, and the coped section was really failing, I'd just ask for a longitudinal stiffeners in order to save yourself the time, effort, and uncertainty of evaluating this. Or tell he fabricator that they have to provide a calc for it and see if that steers them towards a more conventional solution. If there are 10,000 of these things then, sure, put the time in to make each unit marginally cheaper. Otherwise, some mousetraps just are not the worth the effort of bettering relative to the value of your own time.

 

[jseng9]

Hi KootK, thanks for your input.

To give a name to the failure mode I would be concerned with it would be block shear rupture of the beam web. The block shear capacity would be: maximum of (beam web shear yielding, beam web shear rupture)+ the tension capacity of the net section, correct?

As long as the coped beam web is not failing in block shear rupture then we wouldn't need a horizontal stiffener if I understand correctly.

I do see your point in asking for a calculation and I think that is probably a good idea. I am still trying to learn connection design and so part of this is a learning exercise for me.

 

[bhiggins]

The beam appears to be failing in shear buckling at the cope. If that is the case then I don't see the doublers as effective because they do not appear to completely restrain the beam at the buckled web:

 

[jseng9]

Hi Bhiggins. You are correct, the beam fails a web buckling check (web buckling limit is 32 kips while the load demand is 42 kips). How would you prove that the web doubler provided does not restrain the beam at the buckled web. Is there a calculation I can run?

 

[KootK]

How would you prove that the web doubler provided does not restrain the beam at the buckled web. Is there a calculation I can run?

W. T. F?

You might check this by treating the top of the web as a longitudinal compression member and seeing if the doublers possess enough stiffness to restrain lateral buckling of said, faux column. Given that you're starting with a 1/4" thick web and effectively tripling buckling resistance with the addition of the two 1/4" stiffeners, I suspect that it works. But, of course, all this depends on my having guessed the failure mode correctly.

 

[jseng9]

W.T.F?

Sorry for the oversight on my behalf. I think I get what you are saying now. Treat the coped beam web as a column and check buckling of the column. The buckling resistance of the beam web is effectively increased due to the presence of the two web stiffener plates which create a wider column section.

 

[Agent666]

I'd second what bhiggins is saying, as proposed I don't think the plate actually does anything at all, it's like tacking a simply supported plate on the side of the web with no other connection and expecting it to take some load and/or enhance the strength of the beam even though its pinned at both ends and can do neither.

Basically all the load is going into the original web and staying there in my opinion. Put three columns of bolts and you have the plate is able to act as a propped cantilever (actually start carrying load), but how much load it takes is dependant on the slip in bolts and relative stiffnesses (certainly you can provide slip critical bolting). By the time you factor in the bolting on site with slip critical, then welding is easier/cheaper potentially, hard to say without knowing the driving force for this non conventional detailing.

 

[JoshPlumSE]

I don't think the plate actually does anything at all, it's like tacking a simply supported plate on the side of the web with no other connection and expecting it to take some load and/or enhance the strength of the beam even though its pinned at both ends and can do neither.

Thank you. That's exactly what was giving me pause about the connection when I first looked at it. It's why I said a fully welded plate would work, but that the bolted one made me skeptical about the plate strength.