Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

Retaining Wall Design with ETABS & SAFE?

Status
Not open for further replies.

hotmailbox

Structural
Dec 2, 2006
90
0
0
US
Just wanted to pick some smart brains here and see how other tackle this design problem. Please let me know if someone addressed this before. Most likely someone asked this before but I shoot it anyway.

I am dealing with retaining wall design every day (flat lots, hillsides) and I always think that using software like Enercalc or the like to design a cantilevered wall or restrained wall individually without considering the effect of adjacent return walls is too conservative. For example, a basement box 25ft x 25ft with 10ft retaining wall and if you design reinforcement and its footing as an indefinitely long wall, it's way too much.

So what I have been trying to do is:
1. Model all the walls in ETABS and apply soils loads (static and seismic).
2. Assign vertical loads from the structural above if needed.
3. Assign fixity at the base of the wall.
4. Analyze the model to get the deflection.
5. Use section cut to get design moment where needed to design reinforcement.
6. Transfer all support reactions to a SAFE file for foundation design.
7. Model footings in SAFE and assign estimated soils modulus for those footings.
8. Delete all forces in X & Y (assuming friction/passive take care of that after few checks).
9. Create non-linear uplift load cases and combo to check uplift and soils pressure.
10. Create design combos for footing design.

So far this procedure give me much more economical and aggressive design than using Enercalc or the like.

Do you see any flaws in this workflow? What would you do if you have to design a 30'wide x 50'deep house built into an existing 50% up slope hill?
 
Replies continue below

Recommended for you

It might be worth posing a similar question to the Earth Retention sub-forum. There may be more suitable 3D FEA software.

CSI are slowly phasing all SAFE features into ETABS to create the one-stop shop. At present, I think you can achieve everything you mention above within ETABS.

In regards to your proposed methodology a few thoughts:
- How are you accounting for horizontal diaphragms? Do you tie up your basement slab into the retaining walls? Can the ground floor slab transmit loads over to the opposite side of the site?
- The amount of fixity between your slabs and walls will be crucial for load distribution.
- How are you accounting for cracking and long term deflection? Adding a fudge factor of say 5 to the uncracked elastic deflections?
- How wide do you take the section cuts? Per metre (3.2ft)? You'll probably FEA hotspots along the way, which will need local consideration.
- Instead of transferring data from ETABS to SAFE, which can lead to data handling errors and wasted time, why not design the foundations within ETABS?
- You can apply non-linear compression springs to the foundations and can bearing plots out of ETABS.
- Lastly I would still use basic/conservative Enercalc modules to sanity check all of the answers you're getting from this high-powered aggressive approach. Perhaps you'll find the savings aren't all that great in the big scheme of things and not worth additional time/effort.


 
Trenno,

Thank you for your input. See my comments below:

I know CSI is trying to incorporate SAFE features into ETABS but it'll take them a while to make ETABS slab design work like SAFE (in terms of matching long-term deflection estimate).

- If I have solid floor (concrete slab) then I could model it also to tie the walls together at the slab level. But since it's hard to simulate the passive pressure on the wall at the opposite side, I typically assign pin connection at slab level then use the reaction to check the shear and compression in the floor diaphragm. It's more or less the same when you use Enercalc or RetainPro to design restrained walls.

- Yes, the fixity at the base is important. At the moment I assumed 100% fixed at the base.

- To account for cracking, I use stiffness modifier of 0.35 for out of plane bending.

- I use the same concept as structural slab to divide the wall into design strips like column strips and middle strips to determine width of section cuts.

- Transferring data from ETABS to SAFE is auto and it takes just a few clicks. I have to look in to foundation design within ETABS but as far as I know, we have to manually mesh the foundation then assign spring properties to all the nodes. ETABS is not ready to utilize area spring yet.

- Same as above, I would love to do this if ETABS supports area spring.

- Checking FEM results are ok but the problem with Enercalc and the like when designing retaining walls that for tall retaining walls, you would never be able to make the foundation work realistically without relying on the effect of return walls at each end.

 
You can automesh shells and apply non-linear compression springs in ETABS. View the bearing pressure plots under Display > Force/Stress Diagrams > Soil Pressure...

I'm fairly sure slab design (strip based approach) is included in the standard version (not Ultimate) of ETABS these days.



 
hotmailbox said:
8. Delete all forces in X & Y (assuming friction/passive take care of that after few checks).
Why would you do that and how?
I don't see any problem keeping them as they are..
Will it create any issue if I don't delete the X and Y reactions?
 
MSUK90 said:
Why would you do that and how?
I don't see any problem keeping them as they are..
Will it create any issue if I don't delete the X and Y reactions?

You either do this or create restraints in X & Y for those forces.

MSUK90 said:
Why ignore slabs?
You mean you are just making a box like structure without top and bottom cover right?

If you have a solid concrete slab then you want to model it. In case I have a wood framed floor then I would model a line of springs or rollers at the floor level then use the reactions to check framing members for bending and compression.
 
hotmailbox said:
You either do this or create restraints in X & Y for those forces.
You didn't still answer my question: What if I don't remove reactions or provide restraints?
hotmailbox said:
If you have a solid concrete slab then you want to model it. In case I have a wood framed floor then I would model a line of springs or rollers at the floor level then use the reactions to check framing members for bending and compression.
Correct. I didn't know that you have a wood framed roof.
 
Status
Not open for further replies.
Back
Top