steel column reinforcing

[bjb]

I have an existing steel column that I need to reinforce when the column is loaded mainly with dead load. I want to use an LRFD approach, but I am new to this method. My column slenderness parameter is less than 1.5. Is is appropriate to just calculate the design strength of the reinforced section and make sure that this strength is greater than the new factored load? This means that I am ignoring the existing dead load stress.

 

[Taro]

Be careful! If the slenderness parameter is less than 1.5, inelastic buckling will control the design strength. And if the existing column is currently heavily loaded with just dead load acting, then you cannot simply design reinforcing to carry the remainder of the load. The existing column will reach yield and become ineffective first, then the reinforcing will become overloaded and collapse will occur. Either unload the existing column before the reinforcing is added or conservatively design the reinforcing to carry the full load.

 

[JAE]

Taro,

Respectfully, I guess I'd question your response - and I might first say I am not claiming to be any kind of expert on this but in LRFD, I don't believe that existing load necessarily needs to be a concern.

The concept, as I understand it, is that LRFD is an ultimate design method whereby the original section takes X amount of load and is found under stress. Now you add plates or whatever, change the section properties, including the propensity to buckle, and you add Y load to it. X+Y would create strain in the total section such that the stress in the original section exceeds Fy.

The original section goes plastic but the new steel doesn't. BUT, the total cross section MUST strain together now (presuming you've properly welded them together) and they proceed to further strain until the total X+Y load is added.

The propensity to buckle generally based on the combined composite section (even though the original shape is beyond Fy).

The new "AISC Rehabilitation and Retrofit Guide" has in its Chapter 4 the following:

"Columns may also have to be reinforced to accommodate greater loads. Generally this can be accomplished by welding on plates or other sections.....It is often necessary to make such column reinforcements while they are loaded, although the loading can usually be reduced. Several authors have addressed design considerations for this condition, but a consensus is not evident. Some have contended that the geometry of the reinforcement and the initial load can affect column capacity (see Brown and Ricker, 1988) However, a well-known authority has stated that the strength of columns reinforced under axial load and reinforced under no load is identical. (Tall, 1989)...."

So there is some uncertainty and your recommendations to be conservative are not disagreeable to me.

 

[bjb]

Thank you for your replies. Because I am not very familiar yet with LRFD, I reverted to analyzing with ASD methods. I use the approach given in Ricker's AISC engineering journal article, which I know is conservative.

Taro, I respectively disagree with you when you say that the existing section will become ineffective when it reaches yield. I think what happens is that the existing section goes plastic and has constant stress at increasing strains. The increase in strain will be resisted by the newly added elements that have not yielded yet. However, I seems that the ratio of existing stress to new stress should be a factor in designing reinforcement. I know that the AISC equations for critical buckling stress take into account the effect of residual stress. In terms of the reinforced combined section, the existing stress in the original might act like a residual stress, the magnitude of which might differ significantly from the assumptions that went into the equations from AISC. I don't know for sure, but I wonder if this could effect the applicability of the AISC equations for critical bukling stress for some ratios of existing stress to new stress. This is the theory that I have, and I'm not saying that it's definitely the "way it is".

I hope that more research will be done, and that AISC can give some authoritative guidelines. Thanks for your replies.

 

[Taro]

The preload in the existing column and differential yielding will absolutely make a difference. This can be proven by looking at the equation for buckling resistance. For an Euler column, the buckling load is:

Pcr = pi^2 * EI / L^2

Now, when the existing column reaches yield, it's stiffness, E, goes to zero (assuming elastic-perfectly plastic behavior). Therefore, it has zero buckling resistance. The reinforcing must resist the entire load at this point.

This behavior has also been shown experimentally by Dr. Joseph Yura. I would highly recommend obtaining a copy of AISC's stability seminar notes by Yura and Helwig and attending the seminar itself if you have a chance.

 

[JAE]

Taro,
I've got the notes, and attended the seminar and do not recall anything in it about built-up columns under staged load.

Is there a particular page number or reference in Yura's stuff you could point me to? I've got the booklet he handed out called, "Bracing for Stability"

 

[Taro]

JAE,
Did you attend the old bracing stability seminar or the new column & frame stability seminar? Yura covered reinforcing columns in the new seminar. I'm not in the office right now, so I'll try to give you a page number reference later.

 

[JAE]

It was the bracing/stability course he taught with Todd Helwig, Univ. of Houston.

 

[Taro]

That's why it's not in your notes. Yura and Helwig (along with Ted Galambos of Minnesota and Perry Green of Florida) have a newer stability seminar that deals with column and frame stability issues. They cover column buckling and frame issues such as P-Delta effects with leaner columns. Yura explained that he has performed experiments on reinforced columns. If the existing column was heavily loaded when the reinforcing is added, the buckling strength of the reinforced column is equal to that of the reinforcing acting alone. This was quite an eye opener to many engineers who have done this design wrong for many years.

 

[JAE]

Well, it makes sense. - Previous years with ASD, I would calculate the stress in the original section, then add the plates, and re-calculate the new buckling characterstics, add the live load (or subsequent loading after reinforcing) and limit the stress in

1. The original section to the max. allowable stress based on the combined built-up section.
2. The new steel to the same allowable stress...

Usually, the original section stress plus the new stress controlled in the orignal section areas.

But with LRFD - there was no stress calculations and you deal with Pu and phiPn - so staged loading has to be reverted back to stress.

But what do you make of AISC's comment in my post above? Seems curious that they say different than Yura, no?

 

[Taro]

The paper by Lambert Tall (The Reinforcement of Steel Columns, AISC Engineering Journal, Q1 1989) is not very convincing. Tall makes the claim that "The strength of reinforced columns is identical for the conditions of reinforced under load and reinforced under no load". However, there is no supporting evidence presented anywhere in the paper.

The paper presents a handful of test results showing that reinforcing increases the capacity of columns, but there is no mention of preloading the column prior to reinforcing. The paper also contains a section titled Reinforcement Under Load, but this section only discusses the temporary capacity reduction caused by the heat of welding.

I think I'll have to go with Yura on this one. He probably knows as much about column buckling as anyone on the planet.

 

[JAE]

Agree with the high regard for Yura.

So if we use LRFD - do we break down the factored loads into stresses and approach it similar to that of ASD, with a max. stress based on the new x-section, considering previous stress of the unreinforced condition?

 

[Taro]

If the column cannot be unloaded, that would be the approach if you really want to "sharpen your pencil". It gets a little complicated because you have to make some assumptions about the residual stresses in the column, but it can be done.

But unless space is really tight, I would tend to be conservative on the design of the reinforcing. Even with the high price of steel these days, the labor to install the reinforcing will probably be the biggest cost. So you might as well add a bit more steel to get that "warm fuzzy" feeling.