Applied Bearing Pressure @ Base of RC Box Culvert

[damorim]

Hello all,

I am currently designing a cast-in-place RC twin-barrel box culvert to replace a deteriorating bridge. I have modelled the culvert based on the unit meter methodology from AASHTO and have supported the culvert on spring supports to which I assigned a vertical subgrade reaction modulus recommended by the geotech engineer. I have minimal fill overtop the culvert which I chose to ignore when determining the live load distribution though fill. Completed my entire analysis and all of the shear/bending/etc. checks out.

Now, I am trying to check the applied bearing pressure under the culvert. Again, I am using the unit strip method. For the LL component, I simply took the maximum reaction in my spring supports and divided that by the spring trib area. Once I consider all of the loads, I am getting an applied bearing pressure of 160 kPa which is greater than the geotech's recommendation of 100 kPa at ULS.

I am trying to determine how to proceed... One thing that I haven't wrapped my head around yet is that the LL is obviously being distributed along the longitudinal length of the box culvert which I'm assuming would be reducing the overall LL I am currently seeing in my spring supports from the unit meter model.

Any advice?

 

[robyengIT]

You can use (downloadable free) the FEA program "CANDE" (Culvert ANalysis and DEsign)

 

[damorim]

Thanks RobyengIT, I had not heard of the program that you referenced. However, it looks like I need quite a bit of soil information to model that (which can be acquired if really needed)... What I was hoping is that someone has some experience in the live load distribution through the concrete culvert in the longitudinal direction. I.e., if the truck is on the top slab of the concrete, can I assume that the load distributes at 1:1 to the underside of the slab?

 

[TehMightyEngineer]

What's your span and height? 160 kPa sounds about right for a large twin-barrel culvert and 100 kPa sounds reasonable for poor-moderate soil.

As box culverts are designed based on an equivalent unit strip I've never taken into consideration any spreading of the live load on the bottom slab (kept the same equivalent strip width all the way down). You could probably justify increasing the strip width on the bottom slab if you needed to, though.

That said, I'd opt to reduce the subgrade soil bearing pressure by putting the culvert on a "raft" of crushed stone wrapped in stabilization fabric. This will allow you to distribute the soil pressure from the culvert out wider due to the raft and lower your final pressure on the undisturbed fill. Contractors usually like this as they typically have to excavate out below the culvert and replace with crushed stone anyway to get a level, non-frost susceptible soil surface to place the box culvert on.

Professional Engineer (ME, NH, MA) Structural Engineer (IL)
American Concrete Industries

http://www.americanconcrete.com

 

[damorim]

Thanks TME.

It is a twin-barrel box with internal opening dimensions of 3.5m wide x 3.35m tall for each cell. Top slab = 400mm thick, interior wall 300mm thick, and all ext walls/floor slab = 350mm thick.

I have about 800mm of fill overtop. To be conservative, I ignored the fill for live load distribution purposes but included it as a vertical load. The base of the culvert is to be founded on 450mm granular fill, 100mm rigid polystyrene insulation, and a 75mm concrete working slab. I have asked the geotech for some advice, and they basically say it needs a more refined FEA analysis (aka, extra $$). Trying to avoid that right now.

 

[TehMightyEngineer]

Well, I'd take every credit you can get so if you can consider the live load distribution through the fill then I'd do so. This is of course assuming your calculated equivalent strip width isn't already considering the fill, don't want to double count the effects of the fill on live load distribution.

Just out of curiosity what is your strip width?

Also, as you have a 450 mm layer of fill + 100 mm of poly insulation + a leveling slab you'll want to consider the load distribution through that for sure. I bet you can reduce your soil pressure on the undisturbed soil by ~15% if you consider the load spreading out in the fill/insulation/slab below the culvert.

Professional Engineer (ME, NH, MA) Structural Engineer (IL)
American Concrete Industries

http://www.americanconcrete.com

 

[damorim]

I added the fill to reduce the live load effects. Fill of only 800mm didn't seem to have too big an impact on the overall force effects but like you say, might as well take every advantage I can get.

I'm working with a unit meter strip.

I mis-typed on the last post, I meant to say 950mm of granular. I agree 100%, I will have some distribution through the sub-base which hasn't been accounted for yet!

 

[damorim]

Should I be using "wheel" loads or "axle" loads in my analysis if fill is under 2ft? AASHTO LRFD seems to use these interchangeably

 

[TehMightyEngineer]

Yes, AASHTO isn't exactly the clearest as much of the provisions are written for bridges and then "modified" to work for buried culverts.

For fill with less than 2 feet I'm fairly confident that wheel loads are appropriate as below 2 ft you're pretty much just taking the wheel contact area (10" x 20" for HL-93) and spreading it out through the soil. So, for this case you'll have only 16 kips per wheel (32 kips per axle) for HL-93. I actually design all my culverts for wheel loads (helpful when I have a culvert that has a variable fill both below and above 2 ft).

Professional Engineer (ME, NH, MA) Structural Engineer (IL)
American Concrete Industries

http://www.americanconcrete.com

 

[robyengIT]

may be it can help looking at "DESIGN DATAxx" or at "DESIGN MANUAL" in

http://www.acpa.org

(American Concrete Pipe Association)

 

[damorim]

Thanks TME and robengIT. AASHTO seriously needs to clarify some of their clauses...

LRFD C3.6.1.3.3 "The design load is always an axle load; single wheel loads should not be considered."
LRFD 4.6.2.10.2 "The axle load shall be distributed to the top slab for determining moment..."

Not to mention the number of State DOT examples that say "wheel" but use axle loads and vice versa. Judging from the American Concrete Pipe Association Design Manual, a wheel is clearly used in their manual.

What I will do is check that the bearing area on the top slab doesn't intersect with that of the other axle wheel. If it doesn't, I'll just use the wheel load. If it does, I'll use the summation of the overlapped pressures.

Thanks for the help on this guys. Appreciated.

 

[TehMightyEngineer]

I've always taken LRFD C3.6.1.3.3 with a grain of salt. If I could write in my interpretations it would be:

"The design load for bridges is always an axle load; except for buried structures, single wheel loads should not be considered."
"The vehicle load shall be distributed to the top slab for determining moment..."

This is supported by a few passages later in AASHTO. For example:
LRFD 12.11.2.1 "...Distribution of wheel loads and concentrated loads for culverts with less than 2.0 ft of fill shall be taken as specified in Article 4.6.2.10...
...Distribution of wheel loads to culverts with 2.0 ft or more of cover shall be as specified in Article 3.6.1.2.6."

Also, I remembered this provision:
LRFD 12.11.2.3 Distribution of Concentrated Loads to Bottom Slab of Box Culvert The width of the top slab strip used for distribution of concentrated wheel loads, specified in Article 12.11.2, shall also be used for the determination of moments, shears, and thrusts in the side walls and the bottom slab.

They explain that this is to account for multiple loaded lanes even though culvert analysis is always done with single lanes loaded.

Professional Engineer (ME, NH, MA) Structural Engineer (IL)
American Concrete Industries

http://www.americanconcrete.com

 

[damorim]

In the end, it is kind of silly really...

If I consider wheel loads, (the Province of Manitoba uses a modified AASHTO HS-20 called the HSS-30) which has a axle load of 216 kN or two wheel loads of 108 kN, I actually end up with the same results as considering axles.

The reason being that my span is 11.5'. When I calculate the distribution length parallel to the span (Eqn. 4.6.2.10.2-1), I get an E = 113" or 9.4'. Seeing as the wheels are spaced 6' apart in the transverse direction on the design vehicle, there is a width > 1' where the distributions overlap... As a result, although I'm using the wheel load, my pressure becomes based on the axle load.

Anyways, interesting discussion being had!

 

[TehMightyEngineer]

Wow, that's quite a load! Does that include impact? In northern New England we use 125% of the HS-20 load (40,000 lb per axle, or 180 kN).

Yeah, I run into that a lot where you can have wheel loads overlap and then you're just back to an axle load. Sounds like you've considered pretty much everything I can think of off the top of my head. About the only other thing I'd do is tell them to give you more soil cover. Something this is cheaper than any other solution as you also likely reduce the load on your box culvert allowing for less rebar/concrete.

Professional Engineer (ME, NH, MA) Structural Engineer (IL)
American Concrete Industries

http://www.americanconcrete.com

 

[damorim]

Nope, doesn't include impact. It is quite the load eh! I think Canada in general allows heavier vehicles on our roads. Manitoba is the only Province that uses AASHTO (everyone else uses CHBDC S6) hence why Manitoba came up with the Modified AASHTO - HSS-30 for our major highways.

Since I can't add any more cover (I'm tying into an existing roadway and don't have any room to modify the vertical profile), I will try widening the floor slab to increase my overall bearing area. Hopefully that'll help. Thanks TME.

--
David Amorim, M.Sc., P.Eng.