Post and Beam Contruction - Mimimum overlap?

Factual Questions
1

Driving to work today, I was sitting at a stoplight under an overpass they are building near my house. There are some beam that span the roadway 3( lanes in either direction) that are help up by posts on either side.

I noticed that the beams are held up by posts on either side, and from my perspective, it appears that perhaps as much as a foot of of the beam is resting on the posts.

I wondered - is a foot enough? Would three feet be better, or just wasteful? How about just 6 inches? How is the amount that rests on the posts determined?

2

Maybe I am misunderstanding your question, but in post and beam construction the amount of beam resting on the post depends on how wide the post is. Determining how wide the post is is a matter of calculating the size necessary to carry the load.

3

Wouldn’t the type of material make a significant difference?

It seems to me that multiple factors such as creep with vibration, elasticity of the material (steel v concrete) and composition of the column would all factor in. If the column is concrete, the compressive strength of the entire column is not the same as just the outside four inches, for instance. The interior rebar design for a concrete column would affect how far the overlap would have to be to prevent shearing off an edge. If the beam is concrete, there is interior steel to prevent the underside of beam from deforming under tension (since concrete is strong under the compression along the top of the beam and weak under the tension along the underside); the mechanism for anchoring this reinforcement at each end of the beam would affect how much of the beam end needs to bear on the column.

If the whole thing is steel then the anchoring mechanism for the beam and post would figure in more prominently b/c steel is so much more elastic than concrete.

4

On a related note, I’ve always wondered why motorway bridges, which have massive thick columns supporting them, always seem to rest on a relatively thin “point” at the top of each column, rather than resting on the whole width of the column.

You can kind of see what I mean in this photo.

5

Your eyes are pretty good, I’d say. It really isn’t that complicated. If the beams are steel, the bearing plate must be long enough (parallel to the beam) so that the web of the beam does not undergo crippling and also that the web doesn’t buckle. A 12 inch plate isn’t that atypical, I’ve seen say 9" to 12" being on the long side. In addition, very often there will be bearing stiffeners on either side of the web to prevent web buckling.

The other piece of the equation is the allowable bearing stress on the concrete surface. So the plate that rests on the concrete surface (we call them masonry plates) will generally be a few inches longer and wider than the bearing plate on the steel. Or sometimes you’ll have a rocker bearing or elastomeric pads. The allowable bearing on concrete is 0.3f’c, so for typical 3000psi concrete you need to keep the bearing stress at 900 psi. As long as the bearings are big enough to meet these parameters, they need not be bigger. A lot depends on the superstructure design and whether you have continuous or simple spans or steel or concrete beams as far as what type of bearing is needed.

6

Those are bearings. They allow the roadway to move a bit with load or temperature variations without pushing the top of the column sideways. The concrete columns handle the verticle loading (compression) nicely, but not the sideways loads.

7

I have tried and tried to fine a better picture, but this is the best i can do to illutstrate my question:

http://oldcooperriverbridge.org/small/aug_12_overpass_before_3072.jpg

excluding the fact that these spans are centered on thier supports, my question is just how are you able to determine how much surface area is nessasary to be in contact with he supports

I guess to me, this seems a lot easier to visualize the load-bearing process if the beam were shaped like this

/\

in this case i can clearly see the load being transerred down each arm of the

8

sorry, got cut off there…
/\

So if a load was placed directly in the center, it would transfer event down both sides of the beam.

But when it is flat, and I image enough of a load being placed in the center, so that it bows. I’d think that the inner edge of the support would take most of the load.

If this is the case, then to me, this just appears to be an extrapolation of normal load. So, to me, it seems that overlapping quickly reaches a point of deminishing returns. im wondering how that is calculated.

9

IANAE, but I believe part of what is happening their, bcullman is that it is a cantilever design. Its like picking up a stack of books standing on end, by pressing on the otter books.
–push–>[book][book][book]<–push–
The middle book does not fall because it is supported by the other books beside it.
The center support columns simply hold up the set of beams while the bulk of the load is transfered along the entire span to the edges (which is likely held up by the earth).
Engineers, correct me if I am wrong, but that is how I understand a cantilever system to work.

10

I think that what bc is trying to ask is: “how big of a footprint does the resting part of the beam need to have?” which BobLib answered in his own way. Perhaps it would be wise to ask BobLib “how are the dimensions of bearing plates or masonry plates determined”? I remember vaguely enough from Structures to guess that it depends upon the size of the beam and the expected live/dead loads from the superstructure. BobLib knows waaaaaaaaay more than this hydrologist about that. :wink:

11

Yes, this is fairly close to what I am trying to ask.

Additionally, I wonder if the load is spread evenly throughout that footprint, or is it at a maximum at the edge.

12

That’s a pretty standard prestressed beam/reinforced bent bridge. (BTW, it’s not a cantilever and I don’t think that’s waht you’re talking about). BobLibDem basically answered your question: it’s a function of allowable bearing strength in the concrete and your loads. COnfession, I’ve never calculated the required area, in Texas things like bent size and elastomeric bearing pad are pretty standardized.

If you’re really interested, here is the TxDOT standard for elastomeri bearing pads. (Warning PDF) What size pad we use depends on the type of beam and the skew angle (the angle of the beams to the bent). Also, if you look at the table at the bottom of page 2 you can find your minimum required cap width (the part of the bent the beams rest on).

So in the bridge above, those look like Type C beams (very common until Type IV beams - almost twice as much possible length - became popular). A Type C beam at zero skew needs a cap width of 2’9" and a 7" x 16" elastomeric bearing pad. The pad does two things, it distributes the load evenly over the pad’s area and it help with small thermal shifts.

OK, I started to calculate out loads to prove to you that it’s more than adequate and such but, I’m not at work and I don’t have my manuals, calculator, graph paper (essential to any engineering calc :slight_smile: ) , etc. so I stopped. If Bob doesn’t step in, I’ll do quick calc on Monday and post it.

I should note that I’ve done very little steel design but the principal is the same. Generally, we use bearing pads for steel beams as well (which are slightly different, prestressed conc. beams are always “simple spans” meaning they only span one span. Most steel beams are continuous spans meaning they span two or three spans.)

Hmm, at first I thought those were steel beams with shoes or rocker bearings like Bob mentioned, a different type of bearing design (pads are used more frequently nowdays) but now I think it’s just a beam resting on a bearing pad directly on the column. (BTW, when looking for a picture of the shoes I found this nifty bearing and girder guide from the Indiana DOT, also PDF). Usually, you’d see a cap like in the pic the OP linked. It’s an odd design, to my eye, those columns look much stouter than you’d need for a bridge that doesn’t have water flow forces. B

13

OK, if anyone’s still interested, I ran some calcs.

I assumed a very conservative load:

90’ spans (the usual span of a Type C is 75’, 90’ is really pushing it)
8’ beam spacing (also the max and if the span really was 90’, would be more like 6’ spacings)

General assumptions:
Type C beam = .516 kips/ft
8" slab plus 10% for diaframs (which we dont’ use much anymore but leave in this bump) at .15 kips/cubic ft
2" future overlay (asphalt resurfacing) at .14 k/CF
Single Slope Rail (the heaviest) at .376 k/ft

Total dead load = 76.9 kips per beam

Live load - I assumed a single beam takes a whole lane of traffic. Very, very, very conservative, usually this load would be distrbuted to multiple beams. I used HS-25 loading which is 25% more than standard loading and usually only used where you have very heavy truck traffic or possible military movement.

Total live load (including Impact factors) = 58.1 k per beam

Throwing in a DL factor of 1.3 and a LL factor of 2.17, my total load per beam is 226 kips (or 113 tons).

Using the above bearing pad size of 7" x 16", I have 2.02 ksi.

Allowable bearing stress = .3f’c but if the bearing pad is resting on a much larger surface, I can multiply the allowable by 2. Standard bent concrete is f’c = 3600psi so I get an allowable bearing stress of 2.16 ksi.

So even with my unrealistically overloaded beam, I have more than adequate bearing capacity.

As for the beam, prestressed beams use a much higher strength concrete (like f’c = 7500psi and up) and have more than adequate bearing capacity.

Anyway, way more than you probably wanted to know.