Question about reinforcing existing beams

[jay156]

I have a question. My structural software is giving me answers that don't seem to make sense. I'm trying to add a W section to the bottom of some existing W16x40's to reinforce them for a mill that's in the room above.

Now I first tried adding a WT, and a little WT6x13 worked. But the contractor wants to use a W instead, so I'm trying some. One with an equal flange width and thickness is a W6x20, so I tried that but it failed, by a lot. I've been trying bigger and bigger sections, but they're not working. What is so fundamentally different about having the W section's extra flange's worth of steel in there that would make it fail like that when the WT worked?

 

[JAE]

What software are you using?

Is this an analysis program?

 

[JLNJ]

It's quite simple to run one by hand to check the software results.

 

[jay156]

The software is Ram Advanse.

You would think it would be easy, but there must be some compactness criteria or some other thing that I'm missing which is causing the beam to be analyzed by a different equation or something, because it's so counterintuitive that a beam with a higher moment of inertia would be failing in bending when a beam with a lower I is passing.

 

[Lion06]

Are you welding this to the bottom of the existing WF? Are you jacking the existing beam such that the new composite section is taking all of the load and not just the load applied after the new WF is welded?

All of these questions will help you get a better answer.

For problems like you're doing now, I typically use RAM Elements (used to be RAM Advanse) as an analysis tool only and design the reinforcing and welds by hand based on the analysis results provided by Elements.

 

[a2mfk]

Start with a quick hand analysis, then go to your software to fine tune it. Provide a sketch showing us exactly what you are wanting to accomplish, what your assumptions are in terms of loading, bracing of the top flange, etc. So far this is a pretty foggy question...

 

[jay156]

Ok, here's a sketch. I don't understand why that extra little bit of flange from the W6x20 is making it fail so bad when the WT6x13 (which is otherwise very similar) is passing so easily. It must be using a different formula to calculate the allowable stress, because the moment of inertia of the W is slightly higher.

There are no shear studs in the concrete, so it's just the beam acting on its own.

Another question I guess is how do I even go about designing this by hand? The code is very clear in telling you how to design a single beam, but when you're stacking two different beams, I don't see anything regarding that in section F. Is it in a different chapter? If not, which member's tw do you use in checking it for compactness and all that?

 

[csd72]

I did a spreadsheet for this when I was in the US although it didnt really go into lateral torsional buckling as the top flange was full restrained.

If I can find it I will post it on the Web, but it is really not a difficult calc.

 

[Lion06]

I would use the smaller tw when checking for compactness, though this is rarely a problem. If that got to be too limiting, use the existing beam tw, as that is the web that will see compression.

AISC 360-05 doesn't really address this for LTB, but if you have an small unbraced length I wouldn't worry about it. If your unbraced length is something longer, you can use the green book to determine an allowable bending stress.

How did you intend to check RAM Elements output if you don't know how to design this by hand?

 

[a2mfk]

So you are designing this as one new big built up section?

 

[jay156]

PAStructural, That's what I'm trying to learn here.

The way I would normally design it by hand is I would find an allowable stress based on section F. Then I would check the actual stress and compare it to the unity check in section G. Right? Just like a regular beam.

I must be overlooking something though, because I still don't understand why adding the W section causes RAM to think the beam fails when the WT does not. Adding the W gives it a higher section modulus than the WT.

 

[paddingtongreen]

You have moved the combination Neutral Axis further down a section with the same depth, That can increase the compression flange stress. The additional flange is near the NA and contributes little resistance to moment.

Michael.
Timing has a lot to do with the outcome of a rain dance.

 

[msquared48]

A couple of points to consider:

With the WT section added, the top and bottom flanges are comparable. With the W6X20 section added, the top and bottom flanges are not. As a result, with the WT section, the NA is likely to be pulled lower than the W6X20 section, not decreasing the stress in the top flange as much.

How did you model the connection of the W6X20 to the existing beam? Not being truly an integrated member with only welding at the periphery of the flanges could be a problem, even though there is not much contribution to the composite I from these members. I wonder if the program is reading a buckling of the top flange of the W6X20 member?

Mike McCann
MMC Engineering
Motto: KISS
Motivation: Don't ask

 

[Lion06]

Jay-
Are you jacking any of the stress out of the existing beam before you reinforce it? That is one of the biggest factors in the analysis of the section.

 

[jay156]

Any of the stress still in the beam is just from the dead load of the floor. The live load is a lot higher.

Paddington and Msquared, thanks, that seems plausible.

 

[Lion06]

If you're not jacking the beam then here's the procedure you need to use.

1) Determine the stress in the beam with the DL, call this fb1. This will be the moment that is currently in the existing beam divided by the section modulus of the unreinforced steel
2) Determine the section properties of the new, composite section
3) Determine the stress at the top of the composite section (top flange of existing beam), call this fb2. This will be the superimposed moment divided by the section modulus (to the topo of the section) of the reinforced, composite beam.
4) Determine the stress at the bottom of the composite section (bottom of the reinforcement), call this fbt. This will be the superimposed moment divided by the section modulus (to the bottom of the section) of the reinforced composite section. Verify this is less than the allowable. This rarely, if ever, controls the design.
5) Add fb1 and fb2, this is the net compressive bending stress in the existing steel. Verify this is less than allowable (you'll probably have to use the green book to get an allowable). This is what usually controls the design.
6)Determine the required cutoff point for the reinforcement. This will be the location on the bending moment diagram where the total moment (existing and superimposed) does not exceed the capacity of the bare, unreinforced section.
7) Design the skip welds for VQ/I (stagger the welds on each side of the web of the WT reinforcement
8) Extend the WT past the cutoff point by a length that will develop the WT and design the weld to develop the WT in that extension zone.

 

[Lion06]

Or I should say the procedure that I use.

 

[JAE]

One other thought here - does the RAM program allow you to actually have it design a built-up section like this? How are you modeling the beam? As one single member or two?

If you are simply popping in two WF beams from two nodes, the program may be only loading one of the beams and not the other. Software usually doesn't have the ability to deal with non-common shapes like that. Can you explain how you are modeling this in RAM?

 

[Lion06]

JAE-
RAM Elements allows you to input a lot of custom sections - one of them is two WF stacked as a single, composite member.

Jay-
I just popped one in as a simply supported beam and did a design check. The design report references F4.2, which is appropriate. Now that I look at F4.2, it actually uses the same radius of gyration of the compression portion of the member as does the 9th edition (green book). The overall equation isn't identical, but it gives similar results.
My guess is that the WT was just barely working and the WF was just barely not working. I'm sure it has something to do with the section properties of the built-up section just working out a little differently.
The WF will drag the ENA lower on the composite section, which will put more of the web in compression. This will result in a lower rt in eq. F4-5, which results in a lower Fcr.
This is only true for LTB. If you have closely spaced bracing this shouldn't be governing your design. What are your bracing conditions?

 

[ToadJones]

does RAM have an output file for the beam with the code check in it?
STAAD has a similar facility for stacked beams.
The output in STAAD shows you exactly the equations and values used in the code check.

Like M^2 is suggesting, if the NA is actually such that the top flange of the W6x20 is in compression (doubtful by looking at it) and your parameters for the W6x20 are such that the unbraced length is the full 20' beam length that could be your problem.
Does RAM allow you to assign a parameter to "tell" it the W6x20 top flange is braced?
Does Ram see this as truly a single member?