Intermittent Support

[msquared48]

Anyone know of any structural program that will allow a beam, say a simple beam for example, to deflect a certain amount before another support at the center is engaged, becoming a three support system?

This would be helpful to allow a maximum of 50% negative moment capacity over the intermittent support for existing glulams for remodels with periodic extra live loads - full use of a 24FV4, not having to use a 24FV8.

Right now I am modifying an old link span system from the 50's where it liks like they actually reduced the snow load with the LLR equations - a big nono now - beam system just does not work now under current loadings. Snow was considered a live load back then.

Mike McCann
MMC Engineering

http://mmcengineering.tripod.com

 

[dhengr]

Mike:
Why not calculate the load and deflection which will give you the max. bending and/or shear stress on the longer beam. Set your mid support just at/or above that deflection elevation, at the max. delta location. Now add more load to this new two span beam system, until you get a max. negative moment at this mid support or a max. positive moment in one of the two new side spans; using superposition to combine the two loadings and stresses. Just do the problem in two steps (two load and span conditions), almost any beam program, and coordinate the node points on the two beam systems so you can do the superposition additions at each node.

I don’t know a specific program that’ll do this in one step.

 

[JAE]

Mike -
For your simple span example, could you do this:
1. Apply 100% of the simple span beam loading to the beam.
2. Also apply a point load on the beam where the limiting support will be located...in your example at the center.
3. This point load would be taking the place of the support reaction at the center and would be an upward force.
4. Manually increase the point load sequentially until the deflection at the point load is the limiting deflection you are considering.
5. The moments and shears in the beam should mimic what you would have if you had an initial gap that eventually closed and became a reaction against all subsequent loads.

 

[gwynn]

I think most programs which have non-linear spring capability will do this. S-Frame does, likely SAP. With S-frame you could assign a compression only spring with a "gap" to the node at the middle of your beam. The other way to do it would be to assign a vertical support to the mid-point and give that node a support settlement as one of your loads.

 

[msquared48]

Thanks guys. I'll check it out further.

JAE:

Yea, I know that old trick. Just trying not to do it if I don't have to. Being a link span system, it's not quite as straight forward as a simple span, but it is doable.

Mike McCann
MMC Engineering

http://mmcengineering.tripod.com

 

[DST148]

This is what I would attempt even without a software, similar to what dhengr has posted.
1) Find out the loading on a simple span which produces a given amount of deflection at the midspan.
2) For any additional loading on the beam, I would introduce the middle support and analyze as two-span beam for that additional loading.
3) Superimpose 1 & 2 to get the final values.

 

[NS4U]

I believe you can do this in SAP2000 using their "gap" links. It's obviously a non-linear problem so you need some sort of software that can do non-linear analysis.

 

[dhengr]

Mike:
What is a linked beam system? What I call cantilevered beam system with hanging beams btwn. the cantilevered beams; or what I think some (the Canadians) call a Gerber Beam system?

He doesn’t want to overstress an existing beam. So, he must start with the simple span length and a loading, then to bending moment and max. allowable bending stress; then determine the beam deflection at that load/stress, and stop the deflection so the bending tensile stress doesn’t increase any more. Install the support at the max. deflection point and actually reduce that deflection by a small amount (your FoS, comfort level), then treat any added loads on this new two span beam system with superposition.

Also, since he is talking about snow loads, he may have drift loading, so I’m not sure he knows that the max. delta is at the center. Snow loading has certainly never been considered a dead load, but it is not the same as other transient LL’s which we often think of either. We did reduce snow load for roof slope, but I don’t recall reducing it on flat roofs. And, we have been considering drift loading for years too, that came out of Canada, in about the 70's I think. Although, we didn’t complexify the process into a 10 page calc. for every roof plane the way the new codes do. I’m not sure about LL reduction, Mike..., maybe, but take a look at the fact that they may not have used unbalanced loading from span to span, as this might relate to high mid span stresses. This last assumes you are talking about cantilevered construction, or some form of beam continuity.

 

[msquared48]

Cantilevered, linked, all the same.

Problem I'm running into is that the whole system is overstressed in bending. Do something to the simple spans to help that, such as add a column, and the load to the tip of the cantilever is reduced, causing the positive moment to be worse in the adjacent span, necessitating yet another column. Kinda a column of errors if you know what I mean.

Right now, I'm leaning very strongly at post-tensioning the simple span condition where the minor additional loads are going to be, which will not reduce the point load to the adjacdent cantilever, and not touching anything else.

Damn thing's like a house of cards... but it is work...

Mike McCann
MMC Engineering

http://mmcengineering.tripod.com

 

[dhengr]

Mike:
The post-tensioning of the shorter hanger beams sounds like pretty good solution, if it will work. Since it doesn’t change the canti. beams, if they are not the main problem. There was usually a gap btwn. the beams at the hangers, probably enough to put a 3x3 or 4x4 angle over the top corners of the hanger beam. Vert. leg down, and end connection hardware for a .5" dia. rod +/-, threaded and tensioning nuts each end, pre-bent at mid length. Or, a piece of cable, it’ll conform to the saddle. Then a saddle at mid span made from a piece of pipe, split lengthwise (cuts two saddles). A bearing plate to the GlueLam and welded to the lengthwise pipe cuts, and some semi-circular end (retainer) plates. The tensioning rods are on both sides, of course.