Design of steel column

[Bubik]

Hi everyone

I am starting a job in structural engineering and something is really nagging at me. The question is probably stupid but..well there it is.


In a Multi-storey steel braced building a column splices are located every third floor (as I read that sometimes it might be the case) just above beam support (let’s say within 500 mm). Are the splices considered to be pinned supports so that I design a pinned-supported column spanning over the two floors between connections?
If yes ..do the beams laterally brace the column?


Or a column is designed from floor to floor, and end-condition of the column depends on the beam-to-column connection type?

 

[human909]

A pretty basic question. But yeah it is better to ask than not.

Up to the designer to design the column. But under normal conditions a column is designed and considered as a continuous member. A splice is often done so full member capacity is achieved. You could design splice for a column as a pinned connection but then you significantly change your buckling capacity of your column.

 

[Bubik]

human909 thank you for answering.


So is the way to design a column that we take a fragment of it located between floor and ceiling?

let's say that we have a continuous column over multiple floor levels. Does the effective length of a column section in a single floor depend on beam-to-column conditions then?

 

[milkshakelake]

@Bubik: Yes. I'd design the splice as continuous. Then the effective length wouldn't depend on the splice. It would only depend on beam to column conditions.

 

[Bubik]

milkshakelake Thank you!

 

[chucklesNOLA]

I don't know that I agree that column splices are often done to develop full member capacity. Full axial capacity, generally yes. Full flexural capacity depends a lot more on what the role of the column in the system is. In fact, the AISC Steel Construction Manual actually discusses that column splices need not develop the full capacity of the column unless actually required. All of my copies of the SCM are in my office, so I can't lay my hands to it just now to sort out where it says that, however.

There are a variety of freely-available AISC references on splices (in general), several of which reference not assuming you have to fully-splice every column.

Boulanger et al (2016)

Muir (2015)

Matthew (2015)

 

[human909]

I don't know that I agree that column splices are often done to develop full member capacity. Full axial capacity, generally yes. Full flexural capacity depends a lot more on what the role of the column in the system is. In fact, the AISC Steel Construction Manual actually discusses that column splices need not develop the full capacity of the column unless actually required. All of my copies of the SCM are in my office, so I can't lay my hands to it just now to sort out where it says that, however.

There are a variety of freely-available AISC references on splices (in general), several of which reference not assuming you have to fully-splice every column.

Quite true. And I suppose it depends on where and what you are splicing and the scale.

But if you start deviating too far from full flexural capacity then you might have to more thoroughly check other aspects of you design. Often the easier approach is to aim for full or close to full capacity. If your building has 100 of these splices then the costs savings by doing a more efficient design might be well worth it. If there is only a couple maybe not.

I could't say immediately if my recent splice designs quite achieved full flexural capacity I'd have to got back and check. My aim is to achieve full or close to full capacity, while also placing the splice in a location away from any significant moments or mid points in compression members.

EG; A recent beam with a cantilevered section had the splice designed to be at the point of moment reversal under normal loading conditions. The flextural strength needed was minimal. Though varied loading conditions would change that. Also the detailed comes back wanting to move the splice be 3 feet. Sure no problem as the splice has already been designed to accommodate significant moment. (Not to mention a the section will be lifted as a single lift so you need decent rigidity during the lift too.)

 

[Bubik]

By the way..when it comes to creating a model for the analysis of a braced frame I noticed that in models columns span from floor to ceiling (pinned at top and bottom), no consideration for the accurate location of the splices. This tells me that it is assumed that the end condition depends on beam to column connection type and also it is assumed that splices are designed as a moment? Any thoughts if such practice is right?

 

[human909]

By the way..when it comes to creating a model for the analysis of a braced frame I noticed that in models columns span from floor to ceiling (pinned at top and bottom), no consideration for the accurate location of the splices. This tells me that it is assumed that the end condition depends on beam to column connection type and also it is assumed that splices are designed as a moment? Any thoughts if such practice is right?

I don't know what models you are referring to. Using pinned column ends is a modelling simplification which with computer modelling today is better to be not pretend they are pinned connections when they are not. This is unlikely to reflect the reality of a continuous column and thus you could end up with unintended consequence.

I would avoid assuming and instead design to how your intend the structure to be built. I would also seek better guidance from within your company. (That said for a braced frame there should be minimal moments in the columns)