Reinforce existing steel beam for deflection

[twinnell]

I have a 3 span continuous steel beam, uniform load on top flange. Contractor wants to remvoe the two interior beams. Beam has adequate strength without the two interior columns, but fails the deflection criteria. Does anyone know of any references or articles that discusses how to calculate the amount of reinforcing needed to reduce the beam deflection.

Does the reinforcing have to be the full length of the member?

I think I can do a 3 member model with the middle member being the reinforced part. And by changing the middle member properties and moving the two interior nodes out, I can determine the reinforcing required and how long it needs to be. Does anyone know of a better mouse trap?

 

[DaveAtkins]

Sounds like you are on the right track. I find on this type of problem, trial and error works best. I don't know of any formula that will lead you to the correct answer quickly.

My experience is the amount of reinforcing needed will be more than you expect. For example, don't start with a cover plate on the bottom flange--start with a WT welded to the bottom flange!

DaveAtkins

 

[twinnell]

This is residential. Beam is supporting a loft. Loft floor framing is sitting on top of bottom flange, which puts the bottom of bottom flange roughly at ceiling level. Can't add much to bottom of bottom flange due to head clearance. Maybe a thin HSS. Or might just have to add angles welded to web and top of bottom flange.

 

[trainguy]

twinnell

You need to drastically increase the beam inertia, and whatever you do, this will require you to make the beam deeper. This thing is going to get real heavy, real quick, if you can't increase the depth.

Don't forget to check (and most likely reinforce) the beam's end connections, and maybe the remaining supporting columns.

With all due respect, this is much more than a "moving the nodes" exercise. Your reinforcement needs to extend past the nodes in your analysis, in order to develop the strength of the reinforcement. You should discuss this with a more experienced engineer.

tg

 

[Lion06]

I've often wondered about doing this. I wanted to do this to my own house in the basement but don't know how to go about beginning to check the concrete basement walls or footings without knowing what is in the walls or how big the footings are or what the allowable bearing pressure of the soil is.
Would you shore the entire basement framing at the beam with makeshift walls while you remove the post(s) and reinforce the beam?
It just seems like much more of a task than simply reinforcing the beam.

 

[csd72]

Can you leave a cantilevered length at the end past the interior column? This will provide you with some continuation and reduce the deflection.

Alternatively moment connect to the columns and beef these up as necessary.

 

[twinnell]

I have to apologize. My original post should have said...the contractor wants to remove the two interior columns...not beams. there is only one beam.

There is hardly any load on this beam...beam end reactions after the interior columns are removed is only 5 kips. The problem is that the beam is a W8x35 and is 29 feet long. It is supported at the ends by steel columns..not by walls.

I know how to reinforce steel for strength, never had to do it for just deflection.

 

[msquared48]

I see three options...

1. Weld two 8" deep plates to the flanges, one each side of the beam to create a box section, thickness dependent on inertia requirements, and,or

2. Add more beams parallel to the existing beam to reduce the tributary width, and, or

3. Add another one or two beams transverse to the existing beam with moment connections at the splice points to serve as supports where the columns used to be. This would be two way beam action, with equal deflections at the points of moment connection for load analysis.

Mike McCann
McCann Engineering

 

[zaes73]

This rolled steel beams can be strengthened with partial-length, adhesive-bonded carbon fiber-reinforced polymer (CFRP) plates. This is the easiest solution to the problem, with no distrubance at all.

 

[DaveAtkins]

zaes73,

Will that help with deflection? My hunch is that the steel modulus of elasticity is a lot higher than the CFRP modulus of elasticity, and so the CFRP would not help much with stiffness.

DaveAtkins

 

[csd72]

You could still look at making it into a portal frame and beefing up the end columns to take the additional moments.

If this doesnt quite work then combine it with stiffening the beam.

 

[UcfSE]

I would stiffen the beam and be done with it. There may be other mitigating factors we don't know about but the gist of it is deflection doesn't work so you have to do something.

There are many ways to stiffen the beam. You need to consider all the factors that should be weighed when choosing what you want to do. These can include efficiency, cost, accessibility, availibility of the steel shape and so on. It takes a little grunt work to investigate several options but that's our job. you may try altering the connections also, as suggested above, but again there may be other factors that rule this option out.

I would reinforce the whole length unless you check the deflection as a non-prismatic beam with varying moment of inertia.

 

[lkjh345]

Agree with previous posts: Unless you add depth to the beam it will be hard to 'move the needle' with regards to significantly increasing the moment of inertia of the beam. Flange cover plates are inefficient ways to increase the I of the beam.

You have indicated that you do not want to add depth, or much depth, to the bottom of the beam due to clearance reasons. Is there a way to add anything to the top of the beam to increase its depth? Or this is occupied space as well?

I have come to this, or similiar problems, many times over the years. I agree with DaveAtkins, the amount of reinforcing required will be a more than you would expect, particularly if you cannot add to the beam depth significantly.

I also ususally resort to a trial and error process. Fortunatley, many software packages have simple routines to calculate built up section properties. They take a lot of the grunt math work out of the process.

 

[minorchord2000]

You could externally post tension the beams with externally applied cables.

 

[twinnell]

I'm not going to worry about trying to on reinforce where needed (my original question); I have looked at reinforcing the entire length and I would need a PL 1"x8.5" full length of member welded to the bottom flange.

Contractor came back and said can move the interior columns instead of removing them completely. Now the options are remove the columns and add 1" plate to the bottom of a 30' long beam. Or relocate the interior columns and not have to reinforce the beam. But two new footings will have to be installed through an existing sleeper floor joist and slab on ground.

Which is cheaper? I guess that is for the contractor to decide.

 

[rb1957]

can you reinforce the ends to make the beam doubly cantilevered ? ... particularly since there isn't much load, this might be do-able. of course, then you'd have to account for the end moment in the supporting columns, but again that shouldn't be much

 

[ctcray]

From a performance stance, I would think moving the columns and adding new footings would be a better solution. What's your target deflection for stiffening the beam? That loft is probably very stiff right now with 2 posts under a W8x35. You'd have to do some MAJOR upgrading to get something with a similar "Feel" for the loft floor. Just meeting L/360 or L/480 over 29' will probably still feel pretty bouncy to the homeowner, compared to what it was.

Especially if this is where they keep their treadmill!

 

[msquared48]

I thought it was the waterbed!

Mike McCann
McCann Engineering

 

[geoffdale]

Is it also possible to add knee braces at the columns to reduce this apn- even by a bit? You would still probably have to reinforce-but maybe not as much? There's a free spreadsheet at Alex's Corner-I think theres a link to it through the SEOSC website, calculates reinforcing for a beam by a bunch of different methods!