attached link model of shell element using surface springs feature. It has known not working properly. This is not a problem when using solid element.
boef_spring.pmx (32.1 KB)
Hi synt, Do you have a link to the huang et al 2005 paper you are using as reference?
I can see many tests but I can’t figure which is the reference value for the Slab uplift.
Bu other hand, you are referencing Elastic foundation and Compression Only support indistinctly. ¿Are you comparing both results?
I have been working on your Slab and found the analitical solution with the expected values.
Edited:
| ELASTIC FOUNDATION | |||
|---|---|---|---|
| Bending of beams on elastic foundations | |||
| Assumption : beam is always connected to the foundation | |||
| Finite Concrete Slab (E = 20GPa) , square cross section (b =1000 h = 300mm) rests on Winkler foundation of modulus ko = 0.018 N/mm2/mm. Downward uniform q= -1.8 KN and puntual forces applied to the center -400KN. No weight efects are considered. | |||
| Analytical result | PREPOMAX Solid | Deviation | |
| wmax | 6.503 mm | 6.645 mm | 2.18% |
| x(wmax) | 0.000 mm | 0.0 mm | 0.00% |
| wmin | 1.333 mm | 1.329 mm | 0.30% |
| x(wmin) | -5000 mm | -5000 mm | 0.00% |
| syymax | 11.94 MPa | 11.63 MPa | 2.60% |
| x(s max) | 0 mm | 0.0 mm | 0.00% |
| szzmax Soil Pressure | 117.05 KPa | 118.47 KPa | 1.21% |
Say that there are some comments to do:
Inherently, the ‘‘Beam on Elastic Foundation’’ theory implies continuous, elastic support for the foundation or mat/slab strip being analyzed.
Thus, there must be full bearing (compression), as soil cannot take tension (negative bearing). This is not the case here so we can expect some small discrepancies.
On beam areas where there is uplift the Elastic support, (springs) are still acting ¿isn’t it?.
Si I think we can’t compare Only Compression directly with Elastic Foundation.
Only compression is nonlinear and the springs on the extremes of the slab gets disconnected when there is uplift.
One would expect large values of uplift and sinking using this last kind of Support.
I have check it in that’s the case. Maximum uplift of the tips of beam are 5.51mm against the values with Elastic Suport 1.33mm and maximum sink is 6.99mm for only compression against 6.645mm for elastic.
References:
- “Foundation Analysis and Design” (5th Ed.) - by Joseph E. Bowles
- “Principles of Foundation Engineering” (5th Ed.) - by Braja M. Das
please, open new threads regarding target values. Here, i’m exploring and observing PrePoMax implementation at both Compression Only and Surface Springs feature. So, it’s about general behavior, not in significant digit result values.
try updating to version 1.5.5dev, it seems surface spring in shell element has not been fixed.
below of two models in single files for clarity. First part (left) using Stiffness per area (N/mm3) and second part (right) using Stiffness only (N/mm).
this problem did not occur when the model using solid element.
boef_surface_spring.pmx (40.0 KB)
How do you know?¿. Compression Only support can only be imposed as N/mm. That is mesh dependent ¿isn’t it?. Each gap element has the same stiffnes no matter how populated of springs is the area.
Until we don’t have Compression Only as N/m3 we can’t confirm if result is as expected and reasonable.
Shells don’t work and they have an additional problem. I don’t think we can remove the averaging on shells.
as i guessed and notify before, it may lead to a bit of confusion since the same units of individual spring and total spring stiffness. Probably, many users may not to check further INP files generates by PrePoMax
it’s not related to this, but the value of individual spring stiffness when Surface spring is in pressure per length units (N/mm3). When the mesh of shell surface and solid face is exactly the same, individual spring stiffness should have the same value in numbers.
The stiffness in [N/mm^3] is not pressure per length it is stiffness per area. That means that if you want to replace a surface spring in N/mm with a surface spring in N/mm^3 you have to divide the stiffness of N/mm with the area of the surface the spring is attached to.
The N/mm^3 value is a normalized stiffness (per area) that is used to describe the properties of soils. So, if you use N/mm^3 and have two supports of different sizes with the same stiffness value and then apply the same load to them, the displacement in the support will be the same no matter the surface spring area.
different in explanation, however it’s actually the same in equality and meaning as shown both lead to unit of N/mm3. In rigid pavement or foundation design, there’s known field plate loading test to derived and know the value. Recorded and display in graph of pressure load and deflection response.
However, Surface spring and Compression only is not always related to soil and plate foundation. It’s general problem may exist in elastic interaction and contact by simplification, different of material or not.
regarding problem that still exists, below another examples, both surface spring is in N/mm3 units. Using solid element in model is working properly, but it’s not for shell element model.
I was unaware of this, so thank you for sharing the info.
And I am sorry it took me so long to understand the problem. I was able to reproduce it and I will try to find the error.
1 Like
I fixed the problem. First, I added an additional support in the y-direction on the other side of the slabs (top corner), just to be sure.
And using the fixed stiffness evaluation the results are the same.
Since you work with such structures, would it make sense to add the behavior “stiffness per area” also to the Compression-only constraint?
1 Like
From my point of view yes.
it could be great and become useful when option of “Stiffness per area” also available in “Compression only” feature.