WE have a lot of tolerances in the

[Alejandro Paravano]

WE have a lot of tolerances in the steel structures, for example during steel provision we have mill cross section tolerances (ASTM A6, AWS D1.1) during fabrication we have tolerances aobut length, holes, camber, etc (Aisc code) and finally we have erection tolerances. If anybody have commentaries about tolerances in steel structures and what happen with accumulate tolerances and how I can control them in the end of erection I`ll be very grateful.

 

[jdonville]

Alejandro,

Constructive criticism: Please use meaningful titles when you post. When you don't, it is difficult to understand what you want to discuss without first opening the post, and you could miss out on a lot of helpful advice from others skipping over your post entirely.

I am not a steel expert, but I work with experienced fabricators, so here's my perspective/understanding:

If deflection is critical, then it seems that there are 2 possible approaches: 1) specify maximum distortions (more stringent than A6 if necessary) when you order raw materials, and/or 2) measure the distortion in the raw pieces that you receive, and either reject those that don't meet your requirements or choose orientations of the pieces to fabricate (if you can) that will tend to cancel out the overall distortion in the assembled condition.

I know that for welding, one may need to carefully select the order of welds (and may need to redesign the welds) in order to minimize distortion from residual stresses.

Finally, perform test fittings or even create partial full-scale mockups (if the budget permits) to check for fitment and accessibility issues prior to mass fabrication and/or erection at the project site.

I would think that distortions from "true" within the limits of A6 should not normally be of concern unless the tolerance for distortion within the design itself is unusually tight.

 

[r13]

This question should be posted on "Standards & Codes" forum, where more persons have first hand knowledge on how standards and codes are developed and evolved. My personal believe is all modern codes are developed based around "engineering theories", "current state of strength of materials and construction practices", "past experiences", and the last, maybe one of the most important tool - "statistics and probabilities".
Yes, there are many tolerances for practical concerns, yet there are many more factors included in the design codes to guard safety. To be a responsive erector, I suggest to stay current on code issues, and observe/obey the code body recommended "standard cares and practices" should suffice.

 

[SlideRuleEra]

Alejandro - IMHO, the answer is to address tolerance stack-up at all levels, starting with design:

1) Create a design with features that allow reasonable variations in member length, fastener location, etc.

Some tolerances, such as mill cross section that you mentioned, rarely cause problems.

Put dimensions on drawings so that tolerance stack-up does not happen and never "double dimension" anything. To do this the engineer has to think.

Avoid the false precision which software can create. Just because software output states that, say, a beam needs to be 25' 6 3/16" long does not mean the engineer should expected a mill, fabricator, or erector to make that happen.

Include language in contract documents that make Items 2), 3) and 4), listed below, take place.

2) Retain a fabricator who knows what they are doing. Example: If a double-dimensioned drawing slips through, a qualified fabricator will ask for clarification.

3) Use an erector who, also, knows what they are doing. Example: Beams are allowed some camber when produced by the mill. When possible, a qualified erector will orient the beam so that natural camber does not cause a problem... in fact natural camber, used correctly, can actually be an advantage.

4) Have qualified, third party inspection of steel erection as it takes place... and listen to them, then act on their reports in a timely manner.

If all this seems "too expensive"... you will be asking this same question about the next project, too.

http://www.SlideRuleEra.net

 

[jimstructures]

SlideRuleEra,

Would you give an example of what you mean by "double dimension" in drawings.

Thank you in advance,

Jimstructures

 

[r13]

jimstructures - I think SRE meant "show dimension once only on the drawings" to avoid mistake after dimension changes (one element with two different dimensions).

 

[SlideRuleEra]

Jimstructures - retired13's interpretation of "double dimension" is spot-on and has caused many needless errors.

There is another type of "double dimension" error, also. That is, dimensioning a drawing in a way that can be interpreted in more than one way. I have sketched a simple example, a beam with holes in the web, shown below:

Case 1: Is the beam supposed to be 160" + 1/4" long?

Case 2: Or, are the holes supposed to be the specified dimensions apart (shown in red) with a tolerance of + 1/4" for each spacing?

If Case 2, say that each of the (red) dimensions is -1/4"... the tolerance deviation for each measurement is acceptable but the deviations are cumulative. The net result is the beams is not 160" long, it is 1 1/4" shorter (158 3/4" long).

On the jobsite, the short beam is blamed on fabricator error... but the engineer could have prevented the error by proper dimensioning.

Surprising thing is that both case 1 and case 2 could be the proper interpretation and the solution is simple:

Case 1: If the beam is supposed to be 160" + 1/4", make one of the spacing, say, the 12" dimension a reference dimension.

Case 2: If each spacing is supposed to be as specified + 1/4", make beam length (160") a reference dimension.

This is the engineer's decision and should be made and stated on the drawings.

Note: I know this is kind of nit picking, and on real projects rarely matters, but the OP's question concerns how to manage cumulative tolerance deviations. My point is that by thinking during the design process, the engineer can create an unambiguous design as the first step toward a solution.

http://www.SlideRuleEra.net

 

[r13]

SRE - Excellent/interesting example. I recall my supervisor advised us "not to put notes on the drawings unless absolute necessities". I think this would be one he meant to address.

 

[SlideRuleEra]

R13 - Thank you.

http://www.SlideRuleEra.net

 

[jimstructures]

SlideRuleEra & retired13,

Thank you both for your illuminating responses. I appreciate you both taking time to reply.

Jimstructures

 

[ACtrafficengr]

This isn't from personal experience, but occurred in a bridge project near me. It was a very wide, asymmetrical and complex bridge for a Single Point Urban Interchange.

All the steel components for phase 1 were within spec, but they were all on the small side, with no compensating errors. As a result, the contractor had a hard time assembling them in the field. For phase 2, the fabricator assembled and disassembled the components in their yard, so they could correct any problems before shipping.



My glass has a v/c ratio of 0.5

Maybe the tyranny of Murphy is the penalty for hubris. -

http://xkcd.com/319/