Plate with 1 free edge - DNV?

[DDP]

I am trying to do buckling strength analysis of an unstiffened plate with lateral pressure and inplane stress. DNV RP-201 looks to be appropriate but I am concerned it only applies to plate supported at all 4 edges.

Anything I am missing, or a correction factor?

I was thinking of maybe using symmetry and assuming a mirror image on the other side to account for this, IOW double the width of the plate so it buckles in the middle.

 

[ishvaaag]

I in the end found between my books one that had plates with in-plane loads and some free sides, the book is

Pandeo de Estructuras
Félix Escrig
Publicaciones de la Universidad de Sevilla

Given that the section that cares if these things refers to coefficients given by some chart that seems be of Pfluger autorship, I assume it is there tha will be better represented anyway

Stabilitatsprobleme der Elasotstatik
Springer,Berlín 1975

In any case, inspection of the charts woud give a Ncr for the plate in-plane forces, no account of concurrent flexural behaviour in the plate.

You might take ths sigma_cr implied by this Fcr as some limit strength for the out of plane check in some VonMises check in combination with the actual applied stress (in-plane).

But given that the pressure (we can assume) will only amplify the lateral defect of straightness of the plate, you may as well model it directly in some FEM program like RISA 3D or SAP 2000 with the initial imperfections to be enhanced by the pressure and apply P-Delta AND material nonlinearities (reduced E modulus according to some proper law of loss of stiffness, there's one implied in the LRFD code, and others extant like CRC etc) then check if the check is valid. Of course in RISA that is only elastic you will have only a code-like check (if stresses are under the defined value or not); in SAP 2000 you may do better if know how to well deal with nonlinear steel behaviour.

 

[msquared48]

I thought that Roarke had some solutions of this too.

Mike McCann
MMC Engineering

 

[msquared48]

I guess all Roark's "Formulas For Srtress And Strain" has is three edges fixed and one edge simply supported, not unsupported.

Mike McCann
MMC Engineering

 

[Denial]

I don't think your "symmetry" approach is valid, as it involves the implicit assumption that the shape of the buckling mode (of your "half" structure alone) is symmetric.

The buckled shape won't be symmetric, as it will involve rotation at the single unsupported edge. Therefore any prediction of the buckling load that came from your proposed "double" model will be an over-estimate, possible a gross over-estimate.

 

[ishvaaag]

I think the more simple way o obtaining a safe design with no proper references at hand is analysis with builtin initial imperfections, P-Delta and material nonlinearity corrections; this way your design your design directly falls -if it meets the limits- within code allowances. If you need more perfect evaluation and still have no references, you need more perfect FEM model and analysis.