tolerances for welded steel fabrications 1

[ssylvest]

I am looking for a good source to define recommended tolerances for welded steel fabrications (large steel flood gates), something I can cite in the project specification I am writing to impose practical limits on the final product. Most of my prior experience has been with structural steel which uses the AISC Code of Standard Practice, but it is more suited for columns and beams, not fabrications.

 

[KENAT]

Hang on, are you looking for 'process capability' information, or recommended tolerances.

Tolerances should be set by the design function with due consideration given to the appropriate process capability.

To some extent it will vary on what process (more detailed than just 'welded') you're looking to use.

Self fixtured (tab & slot), permanent fixtures...

Posting guidelines faq731-376

http://eng-tips.com/market.cfm?

(probably not aimed specifically at you)
What is Engineering anyway: faq1088-1484​

 

[MikeHalloran]

Angles can be held pretty close, especially right angles, e.g. using 3/4/5 triangles and such, so they'll be limited by linear measures over the length of a typical member.

Linear measurements are generally uncertain to the width of a soapstone or a Sharpie mark.
You can do better, but it costs more, e.g. to use Dykem and scratch marks.



Mike Halloran
Pembroke Pines, FL, USA

 

[drawoh]

ssylvest,

I have asked welders what they can do, and they say ±1/16". Of course, I am talking about small space frames. You are talking about structural steel.

This would be a good question to post in forum507.

You can always call your fabricators and ask them what they are willing to guarantee.

JHG

 

[ssylvest]

I am looking for dimensional tolerances for the final fabricated gate to confirm to for acceptance. It will be up to the fabricator to choose what means and methods he uses or processes and jigs to achieve the specificaitons.

But I want to make sure the requirements are within practical limits, not impacting the cost unrealistically relative to the function. My initial thoughts are tolerances of 3/16" in key lenght dimensions, on a gate of 32' x 7' x 2' thick, made up of plate with various beams and stiffeners.

 

[MintJulep]

Tolerances are a design function. The purpose of a tolerance is to define the amount of acceptable deviation from the "perfect" dimension that will still allow a part, component or assembly to function as intended.

As such, a specifier such as yourself cannot and should not attempt to define tolerances in a specification. This is the job of the design engineer who is designing whatever it is that you are specifying.

Your specification should state the required function. Specifications should not impose arbitrary constraints on design.

 

[ornerynorsk]

Check out "Design of Weldments" by Blodgett. It's an old book with a lot of timeless principles. Worth a look, and very inexpensive from Lincoln Welding.

It is better to have enough ideas for some of them to be wrong, than to be always right by having no ideas at all.

 

[mfgenggear]

ssylvest

as suggested by all.

Tolerances will depend on the function of your gate.
& the method of manufacture or fabrication.

The thickness can be Flame cut, sawed or water jet ,

Water Jet can hold +/- .032 or .062, ,
there must be weld prep, +/- .062" ,
& there will be weld shrink that needs to be addresed.
tolerances can certainly be more liberal.
as suggested it's best to discuss with who is going to manufacture your gates, just to minimize the drawing changes later.

if linear tolerances are tight, It would require test to verify the amount of contraction after welding. or any post weld treatments if any.

if Tolerances can be more generous flame cut can be used but requires lots of hand prep. if closer tolerances are required then it will require machining, but theodolite inspection will be required.

these allowances are subject to Fit, form & function, experience & opinions, so this has been my observations.

http://en.wikipedia.org/wiki/Theodolite

Mfgenggear

 

[drawoh]

MintJulep and KENAT,

I am interpreting that the OP wants to know what tolerances he can specify. If it is ±.005" or ±1/2" or ±10mm, he is going to have to make his design work with it.

Asking what tolerance he should specify is dumb, unless he provides us a complete set of functional drawings, and pays us as consultants to analyse his design.

JHG

 

[KENAT]

Yeah, but we have had the odd dumb question on here where people really don't get the difference between what the typical processes can achieve V what they can/should specify for function and all related factors. I'm not familiar enough with the OP to be sure.

Posting guidelines faq731-376

http://eng-tips.com/market.cfm?

(probably not aimed specifically at you)
What is Engineering anyway: faq1088-1484​

 

[ssylvest]

For the most part, I feel like I have arrived at a practical set of tolerances to specify at this point. As stated by one above, looking at the function lead to some logical compromises to keep from over specifying.

For example, around the permiter I have rubber seals that can be unbolted and slid to adjust up to 3/8" or in some instances 5/8". On the ends, I need a tight straightness where the hinges go.

The gate has standard W-beams making the framing on the back of the skin plate. ASTM A6 has permissible tolerances on beams for camber, sweep, and depth, etc. Therefore, I did not want a dimenion that made it impractical (or impossible) to use standard beams (with no straightening or modifications beyond stock A6 limits).

A few examples of what I set:
bow outward at mid-span perpendicular to skin plate: +/- 3/8".
Out of square on diagonal or length variation along any side: +/- 3/16".
Warping out of plane at one corner: +/- 3/16"
Out of straight along one edge: +/- 3/16"

The fabricator will be required to use jigs, assembly methods, etc. to meet the tolerances.

 

[dhengr]

Ssylvest:
I would think the tolerances you have shown should be easy to meet on a fabricated steel flood gate; 32'x7' x 2' thick, made up of a face plate and structural shape stiffeners backing up the plate. Which is the length and which is the height? The bigger question might be that you define how your tolerances are to be interpreted, beyond what you have stated; how are they measured and what do they mean? And, that you pay some attention in your design and these tolerances, as they relate to the tolerances and potential movements of the structure which this gate will fit into and must mate with upon closure. That structure is probably some sort of a reinforced concrete wall section, jambs, header beam and sill/found. beam, on what kind of a foundation, in an overall earthen flood wall. Excessively tight tolerances on these kinds of structures are usually expensive for what they achieve. The trick is to design your gate and its details to tolerate some reasonable variations and to provide some means of adjustment for settlements and movements over time. The same tolerancing and adjustability considerations should apply to the structure surrounding the gate, so the two mate well and allow max. long term adjustment.

 

[JStephen]

I think you could run into some major issues in a couple of different ways.

One issue is weld distortion. If you take a perfectly flat plate, weld doohickeys or stiffeners or whatever on it, it isn't flat when you get done with it. That has nothing to do with how accurately it's measured. You can change up welding procedures and the sequence of welding and that has some effect on the distortion, but doesn't eliminate it. Another remedy is to start straightening things afterwards, but that is not a perfect science either.

A second issue is the contractual/bidding end of it. If you write a specification with unreasonable requirements, and send it out for bids, you're either not going to get responsive bids, or you'll get bids based on a cost-is-no-object way of thinking, which is hopefully not what you're after. Or you'll get bids from people that have no intention of complying with the specification in the first place, which is also not a good situation.

Tolerances on a welded object like that would normally be based on past experience with the same kind of thing. Different people build them, some look good and work good, some don't, and you measure them and see what the issues are and set your tolerances accordingly. If it's something no one has ever done before, you don't have a good basis to work from. If you're designing it and don't know what the tolerance is for functionality, the fabricator may not have as to what tolerances he can hold on the finished product, either.

From a fabricator's perspective, if that tolerance is viewed as unrealistic, or if the proposed remedies for that tolerance are viewed as unrealistic, there aren't good options. They can guess how much re-work or straightening might be required and just start throwing money into the estimate and hope they have it covered. Or they can no-bid the job and let some other sucker deal with it.

It might make more sense to set an operational tolerance, to say that the thing should seal all the way around, or should seat with a certain pressure or something like that, rather than specifying dimensional tolerances.

 

[MintJulep]

Well, it's possible to fabricate something this size and hold those tolerances.

Do you want this thing to work when it needs to work too?

Take a look at some videos of the Japanese tsunami to gain some appreciation for the forces associated with moving water.

Then think again about designing something to work in that environment. How much deflection due to load? How to close and seal when there is silt, or rocks, or trees where you don't want them.

Something that won't work if it's 3/8 of an inch off won't work in that environment.

 

[ssylvest]

dheng and JStephen posing some useful thoughts to consider.

For Concrete jambs and foundation we set the tolerances per ACI 117, but tightened them in some respects (since ACI tolerances are fairly loose, particularly for foundations) to make it more consistent with the function, yet achievable for concrete construction with embeds of steel.

Regarding the comments on distorsion, straightening, and "reasonable versus unreasonable tolerances", that is the whole reason I posted this question in the first place. The tolerances I proposed above are deemed to be consistent with the seal adsjustment and built-in variation of the members, and will not hinder the operational function. But the reason for the post is to see if someone else in this forum has experience welding up a similar fabrication to confirm they could control the straightness.

[Some of the other replies seem like smart-aleck naysaying as opposed to useful experience.]

 

[ssylvest]

Attached is my current draft for tolerances, showing each to the amount we can tolerate in that direction.

 

[mfgenggear]

if the welding is kept to a minimum, & all Assembly is held in a fixture during the welding & if the parts are heated before welding this will reduce if not eliminate most of the distortion. of course this is oversimplified. there has to be weld engineer involved, addition to a manufacturing engineer.

The condition of the material before welding will also either reduce the distortion or add on to the issues.

all the parts must be tacked in place in a fixture, in the restrained condition, then once tacked the sequence of welding can also reduce or add to the distortion issues.

The tolerances given should be good start far as manufacturing.
but I am used to aviation & aerospace tolerances.

ssylvest

find a manufacturer of good reputation, take your drawings not released yet but at the preliminary stage & do concurrent engineering with them.
this will alleviate most of your issues.

I do this for my customers many times.

Mfgenggear

 

[dhengr]

Ssylvest:
There are hundreds of these types of structures fabricated every day, and a good fabricator shouldn’t have a problem with a good design. Alternatively, a poor design can make for difficult fabricating. Straightening is often part of the process and a good fabricator can weld or heat camber and straightness into or out of the gate. You would do well to talk with a few fabricators about your requirements. You want to talk with plate and misc. structure fabricators, not someone who does volume tonnage sales of 8" columns and W16 beams. I’ll bet there are some fairly std. designs and criteria for these gates, from State agencies or the Corp. of Engineers or some such. Have you looked for any of these? I still ask which dimension is water depth, and which is the gate width? You probably have your stiffeners oriented the wrong way for this kind of structure and the way it might be designed, assembled and built. You want to be able to adjust their spacing and the plate thickness as you move down in the water column. I would want to know more about the hinging and locking systems. Also, does the water pressure press the gate closed or tend to push it open. Looking at the ACI tolerances was one of the points I intended to make. And, I would try to design more than the 3/8" gasket adjustment into the gate. This might be within the gasket system itself, at its attachment to its support and/or at that support’s attachment to the gate structure. A big issue might be protecting the gasket mat’l. from vandalism and weathering btwn. periods of need. Within reason, you don’t mind that the gate and its surround rotate or tip in unison, a foundation consideration; but you must adjust for or control any differential movement of the gate and the conc. opening structure.