Buried concrete box culvert deflection

[TehMightyEngineer]

Quick question as I don't have the PE around to find out what he wants to do until later tomorrow.

Anyone have any thoughts on the total (dead + live) long-term deflection limits for a buried, precast concrete box culvert with 2 feet of soil cover? It's a long span culvert that will be under a highway (non-DOT and there's little to no spec from the EOR) and when I size the structure for strength I'm getting about L/200 (1.3 inches) for total deflection (L/400 live load deflection). AAHSTO only really covers deflection limits for decks directly under vehicle loads. Given the 2 feet of fill the ACI 318 deflection limits don't seem quite applicable.

Thoughts? My guess is that this is an acceptable deflection and will be unnoticeable under the 2 feet of fill and pavement.

Maine EIT, Civil/Structural.

 

[JAE]

Probably OK - we usually use L/240 for total deflection but for a culvert there's not much reason otherwise.


Check out Eng-Tips Forum's Policies here:
faq731-376

 

[jayrod12]

The way I see it, they'll install the culvert, then install the backfill, compact it down. The only thing you'll have left once the highway is back open is live load deflection and long term creep.

 

[TehMightyEngineer]

That's my thinking as well jayrod, but I'm wondering if someone in the future may inspect the culvert and see a 1.3" deflection and think that the structure is unsafe.

Maine EIT, Civil/Structural.

 

[dicksewerrat]

I would call the Box culvert supply company. tell them what you have and ask for a quote. Let them design it.

Richard A. Cornelius, P.E.

http://WWW.amlinereast.com

 

[TehMightyEngineer]

....

We're the designer for the box culvert supply company...

Maine EIT, Civil/Structural.

 

[TehMightyEngineer]

Actually, let me expand on that. This is about the 40th precast concrete box culvert I've designed, it's quite routine by now. However, the vast majority of the deflection issues we've seen are live load deflections, as you would expect. However, we sometimes get deep buried box culverts which have no live load effects per AASHTO, in that case it's either deflection clearly wouldn't control or deflection was clearly unacceptable and we increased the slab thicknesses to bring it up to L/240 or so. This culvert has hit an annoying grey area where the live load deflection is acceptable for a buried structure (L/400) but due to the long span of the culvert (19.5 feet) the box has a significant dead load deflection as well. Thus, I've calculated the dead + live long term deflection as L/200. This is just so close to the line of what we've typically called acceptable and thus I'm curious what you guys would think is acceptable for a buried structure.

In the end my boss is clearly the determining factor, if he says thicker then thicker it is.

Maine EIT, Civil/Structural.

 

[JAE]

Can you camber up the top slab for 90% of the dead load?

Check out Eng-Tips Forum's Policies here:
faq731-376

 

[TehMightyEngineer]

Not without derailing the project which they need to pour next month, the precaster has standardized box culvert forms and such a camber would require some custom formwork.

Maine EIT, Civil/Structural.

 

[KootK]

I think you're fine. Some additional comments:

1) What %rho_b is your reinforcing at? I find the rule of thumb about keeping under 50% to be pretty reliable at preventing deflection issues.
2) If you wanted to do something to improve the deflection, you could add compression reinforcing to cut down on creep and shrinkage. That would add cost but not necessarily mess with the precast forms. Adding depth would probably be more economical, however, if that's an option.
3) On a 19' span, I don't see how having 2' of overburden will help you any.

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[TehMightyEngineer]

Thanks KootK.

1) I'll have to check when I get back in the office, I think it was greater than 50% as I remember playing with the bars seeing what it would take to get L/240 and my spreadsheet kept kicking me into compression failure modes in flexure for some of the trial bar configurations.

2) That's a good point, I actually have compression reinforcement as the exterior (negative) corner reinforcement in the box is continuous over the top, but I was conservatively ignoring it as I often reduce it down to minimal values above the points of high positive moments. For this box however the exterior reinforcement is actually somewhat large due to minimum reinforcement requirements and I should probably see what this does for my long-term creep. As for adding depth, yes I can do that. Though the precaster had me semi-commit to sizes very early on in the design so I only want to increase thickness if it's absolutely necessary as it doesn't make me look good (but I'm not afraid to and have had to give them a last minute thickness change before with no ill results).

3) Not sure what you're saying. Do you mean the 2 feet isn't going to help "hide" deflection issues or do you mean I should take the 2 feet of cover off and just place the roadway onto the top slab? If the latter; the two feet gives you a lot of "bonus" distribution perpendicular to the span of the culvert per AASHTO. It also kicks you into many of the box culvert simplified design criteria per chapter 12 of AASHTO and reduces impact loading. There's a great graph in the AREMA Railway design manual that shows total load on top slabs of box culverts subject to rail loads. The total load actually decreases from 0 to 10 feet before dead load takes over and starts increasing the load again. We've found that 2 feet of overburden is a good minimum design value for box culverts. Also, the EOR for the project specified 2 feet so changing that would require approval from them as it will mess with all their grading plans.

Maine EIT, Civil/Structural.

 

[KootK]

You obviously know a lot more than I do about box culvert design TME. As far as the 2' business goes, my thinking was this:

1) I don't see how it "hides" deflection in the roadway. If the box culvert moves down 1.3", doesn't the roadway move 1.3" too? Sure the 2' improves load distribution but I wouldn't expect that to change the deflection number much, particularly if much of it is coming from dead load.

2) I was thinking in terms of the soil arching around the culvert to some degree. The 2' certainly wouldn't help with that over a 19' span. Based on your last post, however, it sounds as though the analytical benefits of the 2' are of a different nature altogether.

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[TehMightyEngineer]

Yeah, we do a lot of these consulting for a local precaster and my boss hands them to me as I'm the most familiar. I've done almost 20 so far this year alone.

1) True, but I imagine the "curve" of the deflection will be spread out a bit and will also not be as obvious under the pavement while driving over it. I wonder if what the deflection limits for pavement to avoid cracking are?

2) Yes, soil arching will be minimal at 2 feet. However, AASHTO starts giving you credit for the distribution of the load through the soil at the 2 foot depth, any less and the specification gives you no credit for it and then it's just adding dead load.

Maine EIT, Civil/Structural.

 

[KootK]

From the top, I imagine that the most noticeable thing would be the change of slope as you transition from culvert to not culvert. And the 2' surely would help with that. Is it an asphalt driving surface? I'd think that would be pretty forgiving of deflection.

Is the work lucrative once you get it figured out?

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[TehMightyEngineer]

Yep, asphalt.

The work is reasonable. We charge a flat rate per culvert which caused us to break even at first while we developed our design spreadsheets and got more efficient at the design. Recently I added up all the hours we've put into culverts over the past year and it's worked out to be the equivalent of our time and materials rate we usually charge. It's also fairly steady and has allowed us to expand into DOT work.

Maine EIT, Civil/Structural.

 

[precast78]

TehMightyEngineer,

I am just curious, if you order a box culvert from a precaster, aren't they the one who design it? Also, doesn't ASTM C1433 already has it all designed already with truck loading?

 

[jayrod12]

As Teh mentioned earlier, he's the pre-caster's engineer. So he is the "others" in "designed by others"

 

[TehMightyEngineer]

To finish up the topic, the boss had me increase the rebar slightly to bring the stress ratio to 90% and give us around L/210 for the deflection. Roh_balanced = 0.0335 and Roh = 0.0113 so I'm at 33% of roh_balanced.

The precaster we work for does not have a PE in-house. They hire us to perform design work requiring a PE stamp and for general consulting. And, yes, ASTM C1433 does allow you to do precast culverts without a PE but many projects we do require stamped calculations submitted for review. Also, the precaster has found that they can save significant money by having us engineer their culverts regardless of ASTM C1433 as we can fine-tune the design to match their needs. In addition we update our designs to match the latest AASHTO as required by the state DOT, while ASTM C1433 lags behind a few editions.

For this culvert we designed it as a "clam-shell" design involving shear key joints at the mid-point of the walls. See below for an example of a precast clam-shell culvert we just completed.

The clam-shell design allows for easier casting and shipping of the parts and allows them to place the sections with a bucket loader rather than bringing in a crane. In addition they often have weird requirements such as wall penetrations, manhole openings, and wing walls which need to be designed by a PE. Obviously these non-standard requirements can't be designed per ASTM C1433.

Maine EIT, Civil/Structural.

 

[KootK]

Thanks for the update TME. She's beautiful. How significant is the shear that ends up getting transferred across the keys? Also, what's the purpose of the upstand thingy at the bottom?

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[TehMightyEngineer]

Yeah, these precasters do an amazing job. I encourage more engineers to look at precast concrete for many structures, I think most people would be amazed at the quality difference verses cast-in-place construction.

The shear across the keys hasn't controlled in any design we've done so far and even if the shear key were to fail I doubt the structure would even notice as the it's confined by soil on all four sides. For the culvert with the deflection issues we had a maximum shear across the key of 5,800 lb/ft with an axial load of 7,690 lb/ft on the key. This is about the largest we ever see on these clam-shell culverts. Roughly the factored strength of the key is about 7,000 lb/ft, or so. Plus, the friction alone should take care of most of the shear load. With a conservative friction factor of 0.4, 3,000 lb/ft of shear could theoretically be taken by friction. Thus you can see why it rarely controls.

The weird "weir" thingy on the bottom is there so they can fill the bottom of the culvert with natural riverbed material and keep it from washing away. Makes a nice habitat for fish and such in the streams. Keeps DEP happy when we can do that and adds very little cost to the culvert. Usually they're not sloped like that one, though. DOT specified the shape, not quite sure what they were trying to do.

Maine EIT, Civil/Structural.