Box Culvert Analysis

[BadgerPE]

Hey all,

I am working on my first CIP box culvert design. It is a double barrel with 6'x3' tall flow path. I know this is a pretty small culvert, but I would like to spend some time understanding the analysis, design, and construction so that when a larger one comes up, I will be adequately prepared for it.

I have my state's bridge design manual, WisDOT, an they have an example of the loading and design of one in there.

My question for you all is, what is the prefered method of analysis for this type of structure? The methodology of analysis was not covered in the manual. I would like to anaylize this by hand if possible to learn it, and then be able to produce spreadsheets for future use. My thought was to anaylize the structure as if the upper slab and walls formed rigid frames. I would use moment distruibution to determine forces in the frame. Then verify that the base slab can withstand the fixed end moments at the bottom of the walls.

Thoughts, comments, concerns?

 

[TopKnot]

Don't over think it. Reference Colorado DOT's plans.

http://www.coloradodot.info/business/designsupport/standard-plans/2006-m-standards

If you try to save 15% on materials costs, you will end up later having to defend you design against being load rated by the DOT. Ask me how I know this.

I analyzed a double cell 10x13 multiple ways through RISA (beams, various plate elements, plastic hinges, etc.) These designs seemed about 15-20% too strong, but I spent a week defending my design two years after the culvert was built. Not worth it. They used an old fortran program called FloridaCulvert to analyze it, and they didn't even uderstad the programs assumptions. It was fine.

The other painful lesson I learned was to not trust the contractor to follow your CJ detail. They ignored it, and built a continuous 120' culvert, that surprisingly had shrinkage cracks every 30'. Either reinforce it not to have shrinkage cracks, or threaten them with their lives about the CJ detail. Water coming in through cracks in your culvert are ugly. Especially when it a pedestrian crossing.

 

[Qshake]

Agree with TopKnot...

Most DOTs have standard plans for that size so long as the fill isn't over 30' or so. Those plans were likely developed using ASD and not LRFD and just weren't changed for the new code. The main reason being is that ASD is a great code for crack control and the method is still used in the water/wastewater industry where retaining water is the goal (i.e. not letting it run through the cracks).

There are good resources out there for culvert design. IDOT, Caltrans, MoDOT, and Florida all come to mind. I recall Florida has a MathCAD sheet set up for culvert design based on matrix methods. There are also other dedicated culvert programs and WyDOT may have one as I recall. There is also a free precast culvert program good for culverts up to 18' x 18' and you may have to contact a culvert manufacturer to get it. You can design a CIP culvert using this program. I forgot what it's called though.



Regards,
Qshake

Eng-Tips Forums:Real Solutions for Real Problems Really Quick.

 

[BadgerPE]

Thanks for the info guys. However, I am more concerned with being able to perform the analysis just so the standard design details make sense to me. It is more of an exercise than an actual design. Any thoughts on how to analyze a structure like this?

 

[PEinc]

Some DOT's also have box culvert design software available on their web sites.

http://www.PeirceEngineering.com

 

[JennyNakamura]

For design, you should consider the bottom slab rigid compared to the subgrade, with vertical loads applied to the culvert assumed to be uniformly distributed reactions applied to the bottom slab. As you suggest, you will analyze the top slab and side walls as a rigid frame. You can use any standard frame program for this, I like RISA.

There is also a design software package called BRASS-Culvert. Some states are requiring use of this software for load rating purposes of existing culverts, as well as for design purposes. It is something to look into if your firm will be doing a lot of culvert designs or inspections/load ratings.

Practically speaking from a consulting production perspective, I would start with a typical state detail and do a back analysis to verify the standard design is adequate for your specific situation. As everyone else mentioned, most states have their own standard details available on the web.

 

[TopKnot]

RISA is very limited with plate element functionality. Unless you covered FEM plate element theory in graduate school and understand how limited RISA is in this application, I would not rely much on it.

However, I did use RISA. I modeled it four different ways, (plates, beams, columns, varying meshes, assumptions of fixed and spring supports, variability of applied loading, and adding plastic hinges.) If you can bound your solution on all sides, then RISA is okay. Otherwise, get a more capable FEM program. The better programs are very expensive and usually too complicated for things less than long span, complicated structures.

If you use plates, be sure to use a tight mesh and keep them relatively square. This is why they don't work well for culverts. a plate should be thin, relative to its overall dimensions. But if you modeled this with 18" square plates, that are 10-12" think, your model is crap. Most mistakes I see in modeling is to not use a tight enough mesh-which means the plate is too stiff and does not represent the real stress or deflection. I think I ended up trusting the 2-D beam and column model much more. My 12"x12"x12" think plates just did not behave in the model as they should have - too stiff.

It's a fun exercise, but it's really not worth it. Usually box culvert design contracts will not pay for this level of analysis. Don't over think it. You can spend a week on this analysis, and your experienced construction-savy boss will likely bump up your steel because it's not worth the risk.

The biggest issue you will have is tying the flexible cantilevered retaining wall wing walls to the rigid box. The wall has to rotate to engage the soil, but the box won't let it. So it cracks because the stress is much different than your cantilevered wall analysis. GOOD LUCK!

_________________________
Tony Krempin, PE
TopKnot Engineering

 

[TopKnot]

And for whatever it is worth, I always recommend playing with varying spring stiffness's for supports, as opposed to rigid supports. Let your support move 1/4" and see how different the stress in the frame above is. PEMB frame analysis is the most ridiculous. I have never ever seen a truly rigid support condition ever. Take your fixed frame and let the supports move 1/2" and it totally changes stress in the frame. So, I guess I disagree with a rigid base support condition. I'd model it with many spring supports, and play with the stiffness.

_________________________
Tony Krempin, PE
TopKnot Engineering

 

[DWHA]

What I would do is design it per the DOT manual and then check it with whatever method you feel is appropriate.

Follow the loads through the structure. If the boundry conditions are detailed such that you can transfer moments through them, then analyze the structure that way. If the details cannot transfer the moments, then analysis the structure that way.

If in the end it comes out with 15% overdesign and that is the way the DOT (client) wants it. Then I would be satisified.

 

[JennyNakamura]

Tony K., I think you misunderstood my post. From a production driven consulting perspective, if my boss walked by and saw me doing an 3D FEM analysis with plate elements for a 6x3 box culvert, he would probably show me to the door. That is just way over the top for such a simple structure that can be designed by hand in less than an hour using standard DOT guidelines. A 2D frame analysis is sufficient and RISA is just fine for that if you're not in the mood for doing by hand.

 

[dicksewerrat]

Is it really cheaper to use Cast-in Place than calling Hansen Pipe and have it delivered to site?

Richard A. Cornelius, P.E.

http://WWW.amlinereast.com

 

[Qshake]

dicksewerrat - Yes, pending on various factors. In the Midwest where I'm located, CIP culverts are very economical and contractors thus prefer them over precast. When a culvert is small, say in the 5' to 4' rise range, then precast is preferred as the forming and wrecking forms is a problem.

I'm sure there is a point at which really large culverts are best CIP but since precast only goes up to 18' x 18' I can't think of one right now.

Regards,
Qshake

Eng-Tips Forums:Real Solutions for Real Problems Really Quick.

 

[DWHA]

Additionally, CIP last longer than precast (in my experience).