Buckling of pipe on nonlinear elastic foundation

[CivilEngineerNOR]

Hi,

I am trying to model axial buckling of a pipe buried in sand, in Abaqus. The pipe is modelled with 2D beamelements B21 and the sand is modelled with springelement SPRING1 attached to the beam in one end and the "ground" in the other. I have used about 100 springs along the 3 m long pipe and have given the springs a nonlinear behaviour by changing the input file. The analysis is static general with nonlinear geometry on and with a displacement at the end of the pipe.

The problem is however that the analysis will not go through with the nonlinear springs when the stiffenss in the spring is set to zero after 0.2 mm displacement upwards. Any ideas what to do?

I have also tried to modell with connector elements, but the same happends.

 

[inline6]

instead of stipulating zero stiffness can you just use a very small value which might make it numerically more stable?

 

[CivilEngineerNOR]

I have tried with a small stiffness after the breaking point at 0.2mm, but have only been able to use a stiffness down to 50% of the original stiffness (way to high). I have also tried to modell the springs force/displacement relation as a piecewise curve with decreasing stiffness, but it did not work.

 

[GregLocock]

Taking a step back, once you've gone from a finite stiffness to zero you've competely rewritten the Jacobian.

It seems to me your options are

1) switch to a code that is specifically designed to handle weird changes in the system as it goes. Perhaps LS Dyna, certainly it would be possible in Open FEM, which runs in Matlab. Basically you'd write a new model out each time a grain of sand moves. Another option is ADAMS, which has a robust solver and has non linear beam elements.


2) hand analysis


Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?