Shear Reinforcing of a Box Culvert

[JSanchez88]

Hello,

I have been asked to design a new CIP box culvert that is to replace an existing culvert. The new culvert is 13'(W) x 15'(H) clear inside dimensions. I have already looked at several DOT (TX, KS, and MN) guides along with ASTM C857-12a and applied loads to it accordingly:

1. Vertical Earth Load on the roof
2. 3 and 4 axle HS20-44 Loads traveling across the culvert (applied in separate conditions)
3. (2) trucks side by side traveling parallel to the culvert
4. Lateral earth pressure when the water level is both above and below the culvert
5. Corresponding hydrostatic pressure when water level is above the culvert
6. Surcharge of the wheel loads
7. Interior hydrostatic pressure when the culvert is full


When figuring the slab and wall thicknesses without shear reinforcing, they become excessively large (21" - 28"). When looking at the plans for the existing culvert(which is dated around 1960), their design calls for 12" thicknesses. I've also seen several box culverts being designed to have wall thicknesses between 12-18" and slabs up to 24" - but not up to 28!

I know recent designs are made to where the thicknesses are just large enough to not require any type of shear reinforcing but are not needing to be 21-28".

My question is: Am I taking something into account that I shouldn't? If not, are there any recommendations to yield a more suitable design thickness?

 

[ztengguy]

How are you designing it? are you looking at hte AASHTO buried structures section? That shot span, we can get a 8" section to work, but we are arched top, not flat top. How much cover you have?

 

[JSanchez88]

Yes. I am using AASHTO. The roof on the culvert is flat and the cover varies between 6 and 9 feet.

 

[ztengguy]

What deisgn software you using? I would expect that a span like that, 10-12" max thickness.

 

[JSanchez88]

I am using Risa-3D.

 

[TehMightyEngineer]

Your thicknesses seem way too high.

Some things I would check:

The biggest savings in shear will be keeping track of what load actually can cause shear failure. Taking your shear loads at d away from the wall will save you a lot of load for the shear design. Keep in mind that your haunches will likely not control so there's 8 or so inches more you save as well. Do two checks, one for the haunch shear and one for the slab shear.

Also, refer to AASHTO LRFD article 5.14.5.3 if you have more than 2 ft soil cover. This is a simplified shear calculation for buried single cell box culvert slabs cast monolithic with walls. Should give you a small boost in shear strength.

Make sure you're figuring your equivalent strip widths correctly, refer to article 4.6.2.10 for slabs with less than 2 ft cover or distribute through the soil per 3.6.1.2.6.

Calculate the impact factor in article 3.6.2.2 for buried components, should reduce the load a lot.

Note that lane loading does not apply per article 3.6.1.3.3. Also, note that culverts shall be designed for single lane loading per article 12.11.2.1 (unless you have a local requirement otherwise).

For negative slab moments you can take the design moments at the intersection of the walls and the slab at the haunches per article 12.11.4.2.

Fatigue and seismic are not required for buried culverts, article 5.5.3.1.

For other loads check the walls for a construction load case assuming they have only backfilled the walls with no load on the top slab. Otherwise your interior wall reinforcement may be too small. Also, check for unbalanced loads on the culvert (surcharge and full lateral soil load on one side, reduced lateral soil load on the other side with no surcharge).

Maine EIT, Civil/Structural.

 

[TehMightyEngineer]

Oh, almost forgot the most important thing. Live load can be ignored at depths greater than 8 ft and greater than the span length per article 3.6.1.2.6.

Maine EIT, Civil/Structural.

 

[JSanchez88]

TehMightyPirate,

Thanks for the input. I double checked all those and were properly taken into account. Right now, the load case that is governing the design is when the culvert has 9' of fill and the water table is assumed to be at the top of the ground surface. I'm thinking my calculation of the saturated weight of the soil may be a little off.

 

[CELinOttawa]

Saturated soil is significant, and remember that culverts do fail... You may very well be designing it the way is *should* be designed, but usually isn't. Clients often want you to practice at the state of the profession, and as such to do what most similarly qualified and suitably trained & experienced engineers would do. Frustrating, but true...

Also, depending on the cover to the culvert, you can account for a spreading of the load from wheel type point-load (more or less) to a more distributed load on the culvert surface. Have you used some of this "UDLing" of the load? Does your code prohibit doing so?

 

[TehMightyEngineer]

CELinOttawa: In AASHTO article 3.6.1.2.6 (mentioned above) you can distribute the wheel loads out from their contact area. But if he checked all the items I mentioned he should be doing this.

Sanchez: Make sure you aren't just adding the water directly to the soil dead load. Saturated soil weights should only be about 10% higher than dry soil weights. Also, if this really is a culvert you wouldn't usually overtop it and, if you did, you would have equal and opposite water pressures inside and out as it's an open structure.

Maine EIT, Civil/Structural.

 

[ztengguy]

Are you modeling the RISA file with something to resemble the soil? We use CANDE software, and it will definitely help reduce your shear load.

 

[JSanchez88]

I am modeling it similar to TxDOT's "Level 2: RISA 2D With Springs" example. They place 10 compression springs equally spaced along the bottom slab. I've played around with the spring constants and they don't seem to make a significant difference.

 

[ztengguy]

I think you need to look into something along the sides. The frame will be supported by the soil, thus reducing the load on the frame, reducing the moment.

 

[TehMightyEngineer]

What ztengguy said, make sure you model the lateral soil loads in combination with the vertical soil loads as box culverts get a lot of help from the confining action of the soil on the frame.

Putting springs on the bottom isn't a bad idea, that will model the bottom slab soil reaction a lot better. I personally just use a pin/roller at the corners and apply opposite upward soil pressures evenly distributed on the bottom slab. Either works.

Maine EIT, Civil/Structural.

 

[ztengguy]

Also remember though your Phi factor will change as you become more compression controlled also. thats the latest in AASHTO, if LRFD

 

[TehMightyEngineer]

I've never seen concrete box culverts be compression controlled as the flexural stresses are usually too high to have compression control. Doesn't mean it couldn't happen, though. Deep culvert with a small height and span for example.

Maine EIT, Civil/Structural.

 

[JSanchez88]

So, after double checking all the loads and conditions and every possible input field, I cannot find anything that is wrong with my design. Having said that, I've decided to call for shear reinforcing to bring the thicknesses of the walls and slab down a little.

I always enjoy reading what others have to say about various situations.

Thank you all for your input!

 

[ztengguy]

What was your final thickness?

 

[JSanchez88]

I came up with a final thicknesses of 12".

 

[ztengguy]

Wow, makes you wonder why everyone that specs a box culvert isnt outbid by 3 sided arches.