How to model frictional resistance using FEM Software

[PlasmaStruc]

Can anyone please advice me on how to model sliding friction of a retaining wall. Currency I presume that It can be done by applying lateral springs. But is it correct? Any standard FEA approach on how I can check the sliding resistance of a retaining wall? Thank you.

 

[SteelPE]

Sliding resistance of a retaining wall shouldn't be that complicated. Why can't you do this by hand? Last I checked the friction force is calculated a the vertical weight if the wall + soil above the footing times the coefficient of friction on concrete to soil.

 

[PlasmaStruc]

Yes I agree with you for a simple retaining wall. But say, how about a retaining wall with shear key. If I need to calculate the stress on shear key, excluding the friction.

Of course, I could manually calculate the soil load retained by the wall and subtract the friction force based on the vertical load. Then the the resultant load would be the unfactored shear force on the Shear key.

But how do I model this using a finite element approach. Is there a method in Finite element modelling tools?

Thanks.

 

[KootK]

Here's some advice on doing it in SAP. Link.

Like SteelPE has suggested, this problem wouldn't seem to warrant the heavy FEM. Few things involving soil do given the uncertainties involved. I'm an over-modeler myself though so I get the impulse. Sometimes an unnecessary FEM model that works just the way that you want it to is just good for the soul.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[IDS]

There are plenty of reasons for doing a retaining wall analysis with FEA, especially since there is soil involved.

The best way to model the friction depends on the FEA package you are using. The one I am familiar with (Strand7) has friction contact elements, but for a surface where the relative movement should be very small it should be adequate to model the soil with plane strain plate elements with Mohr-Coulomb properties.

Doug Jenkins
Interactive Design Services

http://newtonexcelbach.wordpress.com/

 

[XR250]

I guess you could model friction, but what about the elasticity at the soil/concrete interface. I imagine the soil will deform laterally under the sliding load which will tend to load up the key.
Sounds like a waste of time to me as you are trying to get too exact with this problem. I think I would stick with your original statement, " I could manually calculate the soil load retained by the wall and subtract the friction force based on the vertical load. Then the the resultant load would be the unfactored shear force on the Shear key. "

 

[XR250]

I guess you could model friction, but what about the elasticity at the soil/concrete interface. I imagine the soil will deform laterally under the sliding friction load which will tend to load up the key.
Sounds like a waste of time to me as you are trying to get too exact with this problem. I think I would stick with your original statement, " I could manually calculate the soil load retained by the wall and subtract the friction force based on the vertical load. Then the the resultant load would be the unfactored shear force on the Shear key"

 

[SteelPE]

Maybe I'm missing something but a retaining wall with a shear key shouldn't warrent a FEM analysis.... But to each his/her own.

 

[Splitrings]

In RISA 3D I use lateral springs on circular and rectangular tanks. I'll run my model with both really stiff and very soft lateral springs to see the affect.

A straight retaining wall with uniform backfill may not warrant a FEA analysis but one with different restraints, corners, free ends, differing backfill conditions certainly.

 

[PlasmaStruc]

I see that almost all of you all are in the view that FEM modelling is not a preferred approach for a retaining wall.

However since I am seeking for a friction modelling method using a FEM software approach, I found out this YouTube video (

https://www.youtube.com/watch?v=g-l2fuLNe_U&feature=player_detailpage#t=690

). The modeler is applying area springs using SAP2000 software. But is this method correct? How do a figure out the spring constants for lateral direction?

Thanks.

 

[Israrkhan123]

Hello Dear Friends,
I started modelling of pull out test technique in FEM Abaqus version 6.12-3.I use Connector damage evolution behavior for the rebar in the embedded region of concrete.I applied velocity at the top node of the rebar. I use (axial+Revolute) type of connection, however ,my problem is , Is the type of connection is Ok and in output I can see that there is damage (from the figure) but I do not know how to plot appropriate curve (Force vs Displacement).Could anybody help me regarding this issue.I am really faced with this problem.
Thank you in advance,

 

[IDS]

The modeler is applying area springs using SAP2000 software. But is this method correct? How do a figure out the spring constants for lateral direction?

Are you using Sap2000? If so, have a look at the link KootK posted, which seems to give good detailed advice.

If not, let us know what you are using, since how you go about it will depend on the facilities in your software.

Doug Jenkins
Interactive Design Services

http://newtonexcelbach.wordpress.com/

 

[Splitrings]

PlasmaStruc,
I use a trial and error approach. I choose a spring rate, solve the model and then calculate the coefficient of friction based on the displacement of my model . The last model I ran, I used a lateral spring of 5000#/in. When I solved the model I had a lateral displacement of 0.044in. Therefore a 220# lateral force exists per spring. The tributary wall section for that group of springs (there were 6 springs), weighed about 3000#. So six springs at 220# is 1320# of lateral force divided by the 3000# weight equals a coefficient of friction of about 0.44. I compare this value to a table and decide if it is acceptable based on the soil and concrete interface. Like I said earlier, I will also run my model with really stiff and really soft springs to see how sensitive the model is to the spring rate.

 

[Splitrings]

I had an engineer once tell me once that he expected lateral displacements to be between 0.01 and 0.1 inches. He didn't have much to support this other than experience with FEA. The few models I have run, this does seem to be somewhat valid.

 

[dcarr82775]

I do something similar to Splitrings. I model springs and then adjust the stiffness until I get results that 'feel' right to me.

 

[IDS]

Why use springs, where you have to guess/iterate to find the appropriate stiffness, rather than plate elements and friction link elements where you can directly use parameters related to the soil friction angle and the soil/concrete interface friction factor?

Doug Jenkins
Interactive Design Services

http://newtonexcelbach.wordpress.com/

 

[Splitrings]

I just noticed you were trying to model a shear key. That will be a bit more difficult to model correctly. You won't develop full passive resistance until some displacement occurs. The friction between the rest of the footing and soil will develop just as displacement occurs. I am sure someone with more experience than I could do it.

@ IDS,
I am not sure RISA has the ability to due friction. Maybe with solids but not plates, that I am aware of.