I am wondering same, but I don't necessarily think there's a maximum bearing length.

However, I do want to believe that beyond a certain length it is unnecessary. Especially after you've developed your welds and connections for the required strength. Also, stress reduces as you go further to the ends depending on the load applied and the strength of your beam and bearing plate will be adequate for that.

I suspect that the further back you go, the stress diminished because of the curvature of the beam, not the increased bearing area.

-----*****-----
So strange to see the singularity approaching while the entire planet is rapidly turning into a hellscape. -John Coates

-Dik

@dik One possibility I am also thinking of is the reduction of stress due to "longer" bearing length. Longer bearing length reduces local stresses (web crippling and yielding) and not due to bearing area.

I agree with dik. You'll have to use some judgment to establish a bearing stress distribution. It's probably a triangle with max pressure near the right end of the plate.

Beyond that, you would need to do something crazy like build an FE model and compute the bearing pressure distribution and see how it falls off toward the left side of the plate.

Thank you for your responses. I agree that there must be some kind of limit on the bearing length that we can consider in the calculations. However, it seems there is no code guideline that clearly establishes the maximum allowable bearing length. It seems the way to deal with this issue is engineering judgment based on some sort of stress distribution or in the extreme case a finite element model.

One approach would be to compute the required bearing length similar to what's done for unstiffened seat connections in the AISC Manual Part 10. Then pretend like the bearing pressure distribution is uniform, again as is done for those seats.

Quote (StructuralAddict)

It seems the way to deal with this issue is engineering judgment based on some sort of stress distribution or in the extreme case a finite element model.

I agree. I think that this matters and know of no formally expressed way to deal with it.

What I do myself is this:

1) Set some limit that I'm willing to let the substrate crush below the bearing. I kind of vacillate between 1/16" and 1/8".

2) Use the design deflection limit (L/360) or whatever to work out the slope and use that to estimate the crushing.

3) Set my bearing length equal to the length that would satisfy my limit from #1.

See below for an example based on L/360, 1/16", and a centrally located point load.

One advantage of this is that the math winds up being stupid simple. And I find that the bearing lengths feel good to me. I feel that a stiffer beam ought to be able to avail itself of a greater bearing length.