Curvature on straight steel beam

[pba]

This may be of more relevance to UK engineers but anyone with something to say please feel free to do so!

Straightness tolerance on universal beam sections is given by BS EN 10034 as 0.3% x length for beams up to 180mm deep. So, for a 9000mm length the curvature can be up to 27mm.

National structural steelwork specification gives a straightness tolerance after fabrication of 0.001 x length, which is a third of that allowable on the member supplied by the rolling mill.

I take it that the steel fabricator has a duty to straighten beams. Is this correct? The question has been prompted because a tie beam was observed on one of our sites to show significant curvature and we discovered the above clauses in the documentation during our investigations. Fortunately (for the steel contractor) the actual curved beam checked out to be fine.

 

[btrueblood]

If I were a steel supplier, I'd never accept responsibility for the straightness of a beam that's been installed (welded, bolted, riveted) into a structural assembly. How would I know that your joints didn't load the beam into its deflected state?

BT

 

[HgTX]

I run into this all the time. We do allow the mill tolerances to apply to rolled sections, but for flat plate the mill tolerances are significantly looser than the tolerances we have on completed plate girders. The fabricators want us to loosen our fabrication tolerances to match the mill standard.

My philosophy is that there are two sets of tolerances for a reason. The as-rolled standard is intended to provide a certain quality standard for product emerging from the mill, regardless of intended application. For particular fabricated applications, every industry sets their own standards for what's required.

So I tell them there's no reason we should accept what comes directly out of the mill; that standard is irrelevant. What comes from the mill is not painted, it's not cut to length, it's not always cambered & curved as much as is required--this is raw material. Why on earth should the mill's out-of-straightness tolerances apply to our final fabricated product intead of our industry-specific tolerances, when nothing else about that plate other than its thickness is in final condition?

They have yet to convince me that I'm wrong.

Hg

 

[andyr]

I would argue in the UK that what comes out of the mill has to confirm to standards and their own quality assurance proceedure and should conform to BS EN docs.

Its certainly not down to the fabricator to bend it back to "straight" in the sense of the code, as this could have further implications if he wasn't skilled in doing it.

Its an intersting comment from Btrueblood, the joints, or connections, would be designed and detailed, at least in the UK by the fabricator as this is considered to dictate the erection procedure and he his better placed to do that.

 

[HgTX]

I'm in the bridge industry, which is much more tightly controlled than building. Heat correction is a routine part of bridge fabrication. They need to know how to do it; they're just looking to avoid this one application of it.

But the fabricator in any industry should know how to correct bends, if nothing else because there might be damage at their site.

If an industry chooses to set tolerances tighter than those at the mills, it's certainly not up to the mill to meet those standards. That leaves it up to the fabricator--who else can do it? The industry setting the standards needs to understand the consequences of setting these tighter tolerances--that someone other than the mill or foundry will be making the adjustments. If they're not comfortable with that, they need to create a new industry-specific standard to order to (and raise the price of steel for their application) or leave the tolerances alone.

In the case at hand, there's no legal question about it--the fabricator signed a contract to provide steel to the steelwork specification. The fact that the only raw materials available are made to the mill specification, with a lower tolerance, is an inconvenience but a fact of life.

If the facility owner is nervous about the fabricator's ability to make these corrections, they could choose to waive the requirement, but it is up to them to take responsibility for that decision. Unless the design code has a factor for initial out-of-straightness, there's no easy way to account in the design for using the lower standard.

But I've never liked exusing one deficiency in fabrication on the grounds that they have another deficiency in their repair skills. ("You shouldn't make us grind those welds--you know how we mess up grinding!" "Oh, so because you're incompetent, I should hold you to a lower standard than your competition?") On the other hand, there is a certain amount of "cure is worse than the disease" consideration that needs to be made for every noncomformance.

Hg

 

[Ron]

I agree with btrueblood....fabricator shouldn't be held to mill tolerances for installed pieces. There are many reasons for sweep and deflection after installation. Even after fabrication that might include angle clips or other features welded or cut into sections, there might be "out of tolerance" sweep, camber, or deflection.

 

[pba]

Thanks for the responses, especially HgTX. You confirm that straightening of beams is at least routine in some industries.
The case to which I referred is a minor tie beam which simply looked odd as installed and caused us to check the specification. There is no question; the member will perform adequately as installed. I was thinking of a potentially greater problem if the issue arose on more significant beams. We often call for member pre-set or camber to reduce the dead load deflection of framed buildings. The rolling tolerance of 0.3% could be as large as the specified camber. This means that if the rolling tolerances were to govern, we could specify a camber and have a straight beam supplied, still within tolerance!

 

[HgTX]

Looking back at trueblood's post...yes, tolerances should apply at time of fabrication, not installation. All kinds of things happen at installation. (And trying to correct the deflection while the member is under load is a risky proposition.)

So if you're going to hold the fabricator to a tolerance, the steel needs to be inspected before erection. (To me, "fabrication" is what happens between the mill and erection--cutting to length, installing stiffeners, drilling holes, applying designed sweep/camber, etc.)

I definitely wouldn't try to get the mills to do such correction. They don't habitually do much in the way of correction--they just turn around and sell noncompliant product to someone else who purchases under a more lenient spec. (And that's when they actually know they're out of spec--my understanding is they never flip plate over before it goes out the door, so they're routinely selling product with specified surface quality criteria that is 50% uninspected.) I take a very dim view of steel mill quality control, and it's gotten significantly dimmer since the out-of-control seller's market we've had for the last many months.

[Disclaimer to any mill reps reading this--of course I'm not talking about *your* mill!]

Hg