non linear shell case

sofien_73

Hi everyone, I'm looking for tips to push a classic case of a shell which has it's four sides simply supported and submited to pressure to the collapsing. Non linear analysis is not enough to do this (equivalent plastic deformation over 12% is the higest limit for this kind of analysis with steel plate) maybe there's somebody who had made this kind of analysis with realistic failure criteria with openradioss.
Here the model for example.

BR

Fatmac

I was not quite sure ur model was simple supported on 4 sides, so i fiddled around with the BCs but then with QuasiStatic steps I got to some high plastic strains

disla

2 Simply supported isotropic plate under uniform load benchmark.
Page 4/53.

Authors. Biswajit Banerjee[1]
Jeremy Chen, Raj Das, Anjukan Kathirgamanathan

https://www.researchgate.net/publication/258848858_Comparison_of_ANSYS_elements_SHELL181_and_SOLSH190

Use symmetry BC and reduced integration S8R.



That is to start assuring the right BC and minimum number of elements.
Then you can switch to Plasticity (

sofien_73

Thanks to both of you, that's ok for BC and element type.
Disla, what is the next step the main objective is to evaluate failure pressure in accordance with real material curve which I don't have. Is it possible to use failure criteria with non linear analysis we can only simulate elastic and plastic domain I just watched for equivalent plastic strain but it goes higher than 100% which have no physical sens but it's ok because we don't have any failure criteria computed.

disla

From my perspective, two fundamental issues should be clarified before progressing further. Otherwise, there is a risk of developing the analysis around assumptions that may later prove inconsistent with the actual physical behaviour.

1. Boundary conditions

A model intended to reproduce the real behaviour of a component can rarely be considered genuinely simply supported. Although this boundary condition is widely used for theoretical verification and benchmark problems, it is difficult to reproduce accurately in a physical system.

Constraining displacement in the z-direction does more than prevent the plate from moving downward or “falling” through the support. It also prevents any local separation or uplift within the constrained region. In this particular case, allowing the corners to lift off the supports appears to produce a significant difference in the structural response.

A contact formulation that permits separation while preventing penetration may therefore represent the actual support conditions more realistically.

2. Definition of the applied load

The physical origin of the applied pressure should also be clearly defined. Pressure is a distributed load acting normal to the current surface, and its direction may change as the plate deforms. This is relevant for a finite, initially flat plate with an unconstrained perimeter, since a follower-pressure load will tend to promote a curved or dome-shaped deformation.

It would therefore be useful to clarify what physical mechanism generates the load and how it is applied at the plate boundaries. Depending on the actual loading arrangement, the model may be subjected to a prescribed surface traction rather than a true pressure load. The distinction is important, particularly in a geometrically nonlinear analysis, because pressure and fixed-direction traction do not generally produce the same response.

sofien_73

thanks disla, this model was build only for general knowledge of thin plate behavior within plastic deformation. This is not a real case it's only for learn and improve. this case show how plastic strains evolve with elastic and plastic phase and the membrane effect which increase considerably the bearing capacity of this thin plate. Now I wonder how can we go deeper to the failure and estimating ultimate pressure that this plate can handle.

disla

@sofien_73
Look for NAFEMS Benchmark Challenge Number 2. Assessment of a Simply Supported Plate with Uniformly Distributed Load. I referenced it above. Read about the different Engineers approach to this subject , different element types limitations and complexity of reproducing simply supported BC condition for 3D elements.

With a quick search you will realize how much debate there has been arround your problem.
This is a free access webpage containing reference values:


https://www.esrd.com/resource-library/product/nafems-benchmark-problem-02-solution/