Field splicing of long steel beams 2

[ajk1]

Can 56-foot span steel beams be field spliced at the mid-length without welding, by using bolted splice plates to the flanges, and I suppose the same would have to be done to the web to take the compressive force in the web above the neutral axis as well as the web shear? Right?

Are the flange splice plates to be designed for compressive force only, or do they have to take any moment as well, and if so, how to calculate that moment?

The welding would have to be done in stages, some at the top flange, then some at the bottom flange, then back to the top flange, etc. to avoid warping the beam as the weld cools. Is there any document that specifies this procedure and how many stages it must be done in?

I expect field welding would be very very expensive and slow. Right?

 

[StrucPEng]

There are a lot of unknowns here that may change the answer like, what is this beam being used for, what is the loading conditions just to name a few. I'm also not sure I fully understand what you want to do here. Are you asking can a beam be fabricated with a splice in the middle and then fit up in the field via welding or bolting? All else equal, I would say welding a full moment connection in the field would be much more expensive to do than to field bolt the beam together.

For forces at the splice location (note I'm not the best with my statics, so hoping someone can confirm my thoughts) I would think you would look at the beam with your specific end conditions and with an internal hinge at your splice location. The splice location would have moments and shears associated with it that you could calculate to design your connection.

Take a look at the Steel Bridge Design Handbook Splice Design document (

https://www.fhwa.dot.gov/bridge/steel/pubs/hif16002/volume14.pdf)

I see bridge girders with splices in them all the time, so this might be a good resource to reference.

 

[CANPRO]

ajk1, are you designing the beam? I would suggest making a note that the field splice is by the fabricator and provide the design loads at that location - they can decide the most economical way to splice the beam to suit their method of installation. You might want to consider installing the longest piece practical instead of splitting the beam in 2 equal pieces so that splice is away from the maximum moment (assuming this occurs mid-span).

If the beam is designed by others and you're designing the connection, you need to ask the EOR for the design loads.

Bolting is definitely cheaper than welding, but you need to make sure you have room for your splice plates. If the top of the beam is tight to something above, you can't have flange plates and web plates, you need to go with a flush end plate splice...which depending on your loads could require really thick plates and large bolts.

 

[ajk1]

Thank you all.

The splice is required because the beams are most likely too long to be manouvered into the garage, as the garage is open one side only to a very busy street in the city.

Steel beams are being used to support a floor where existing concrete beams have tendons that are corroding and breaking.

Since the repair will be by a garage repair contractor, I would not trust him to retain a large steel fabricator who has an engineering department who would be experienced in this. Safer to design it ourselves and apply appropriate caution. That is what the client expects of us. No arguments later with the contractor. Leaving it to a small contractor to do structural design is a route to trouble, argument and waste of time, as I have found out from long experience. Different of course if it is a large project with a reputable steel contractor with his/her own engineering department, in which case your advice would be good.

Thanks for the Link. This is particularly useful.

 

[jayrod12]

Do you not mandate that all steel is prepared by an approved fabricator and sealed shop drawings are submitted? This is the norm where we are, no matter who the general contractor is or how big/small the project. It should be an automatic.

 

[BridgeSmith]

See Section 6.13.6 of the AASHTO Bridge Design spec. The 7th and 8th Editions have a much simplified method to design bolted field splices. The old method applied a portion of the moment on the beam at the point of the splice, the moment due to eccentricity of the bolted web connection, and all of the shear at the splice, to the web splice, while the flange splices carried the rest of the moment. In the new method, the web splice carries only shear, so long as the flange capacity is adequate for the full moment at the splice. The requirement for the splice components' capacity to be a minimum of 75% of the capacity of the corresponding girder components, is retained. For most of our bridge girder splices, that 75% minimum governs the design, since we typically place them at low moment locations as 'encouraged' by the spec.

 

[BridgeSmith]

If we had to splice a girder in a simple span, we would move the splice as far away from the midpoint as possible, probably splicing it in 2 places, if necessary, to avoid splicing at the midspan, where the highest moment occurs.

Edit: I should add that, in my world, 56' is short, and we wouldn't consider splicing it at all. I assume the reason for splicing it has to do with physically getting the beam in place? If that's the case, consider splicing at the 1/4 and 3/4 points of the span, perhaps using a smaller beam for the end pieces.

 

[ldeem]

I always try to design splices for the maximum moment capacity of the beam - regardless of actual moment. It's more expensive but in my world of mining and heavy industrial robustness wins over cost.

As for welding - I deal with a lot of contactors who prefer welding to bolting. Mostly because it is easier for them when you consider field fit-up issues. In your case if you are dealing with a small contractor be very careful of their welding skills. I have seen a lot of very poor field welds.

 

[spieng89]

Typically for a simple span beam with continuous distributed loading I'll splice the beam at the 1/3 length. It'll be away from the max moment and it's easier to design the connection for shear than moment