Moment Frame Base Condition....to Fix or not to Fix

[bigmig]

I was running some design on a moment frame (end fixity at beam and column connection) and realized that I was always taught to
leave one of the column connections pinned, and the other "released", or a roller if you will.

I started playing around with modeling both column bases as "pinned" and noted that the frame behavior improved significantly in terms
of story drift.

In the real world, we will bolt that column base to the beam or wall on which it stands....it is not rolling around.

what is the accurate way to model a typical, 3 member moment frame? Pinned on both column bases? Or Pinned on one, roller on the other?

 

[structSU10]

Pinned bases make sense to me. never heard of defaulting one to a roller. If the moment frame lands on a beam rather than a footing, I would be inclined to make sure that beam is modeled, or a spring support is used, as that will greatly affect the drift you could see, or the load the frame tries to attract.

 

[DaveAtkins]

Agreed--I have never heard of assuming one support is a roller. The usual decision is whether to make both bases pinned, or both bases fixed. This affects the footing design, of course.

DaveAtkins

 

[WiSEiwish]

I default to pinned. You will get more drift than a fixed base, but it greatly simplifies your base plate and foundation design. I haven't heard of roller connections at a moment frame being used.

 

[msquared48]

I have heard of it for upper story beam analysis where the beam is connected to a flexible column, but not at the foundation.

In thinking about it though, that could be the case if there was no grade beam or wall between the columns, but only isolated spread footings.

Mike McCann, PE, SE (WA)

 

[DaveAtkins]

msquared48,

I don't agree. For a support to become a roller, the footing would have to actually slide. I don't see that happening. Pinned--yes. A roller--no.

DaveAtkins

 

[JoshPlumSE]

You make this assumption when doing hand calcs for gravity loads... where you don't expect to have any real shear at the base of the columns. When I close my eyes, I can still see all the sketches from my early structures courses.

But, I've never done it in practice when lateral loads are applied.

 

[rb1957]

the roller connection is a theoretical abstract, no? if you rigidly constrain both frame bases then mathematically there's a redundant load ('cause the two reactions share the same line of action). in the trenches we'd normally say 50% of the load on each. I don't think anyone would really design a foundation on roller skates ?

another day in paradise, or is paradise one day closer ?

 

[msquared48]

Dave:

I agree in most cases, but if one footing is fully laterally restrained by soil and the other is not, as with a footing on or very near a downslope, I can see the footing want to move laterally without a tension tie. Hence I would use the roller without the tension tie to be conservative.

Mike McCann, PE, SE (WA)