Learning Forum

[mmia1]

Hi

I tried to apply hydrodynamic water load at pipe as nodal load like below:

*LOAD_NODE_POINT
$# nid dof     lcid    sf
        1   1     1       1.00
        2   1     1       2.00
        3   1     1       3.00

.............................

  9570   1      1      1.00

I have a load curve (lcid) as below:

*DEFINE_CURVE_TITLE
$# lcid sidr
        1 0
$#   a1           o1
       0.000      0.000000
      1.000000 1.000000

 

The sf (scale factor) is basically my load that are applied as pipe nodes (please see image). I am doing implicit analysis and found that if I exclude these load steps (*LOAD_NODE_POINT) and include only hydrostatic load, my model ran completely. However, if I include *LOAD_NODE_POINT steps, then my model is not converging and aborted after just 0.1 sec (total time = 1 sec). Could anyone suggest me what to do. I am stuck at it for a while.

 

[Jim Day]

Rather than nodal loads, which may distribute loads in a less-than-ideal fashion, try using *LOAD_SEGMENT_SET_NONUNIFORM to apply the load as a traction (force per unit area) onto a set of segments.

[mmia1]

Thanks for the reply. I was trying to apply pressure on the elements of pipe before but the problem is that when water flows in X direction say some of the pipe elements are in positive pressure and some are in negative pressure (suction), which I think difficult since I have pipes and elbows as well. So, which elements are in suction is hard to apply. I have attached an image. So, I calculated the nodal force and applied forces at corresponding nodes. 

What I found that for larger diameter pipe(1' and 9-inch), the code converged, but for smaller diameter the code did not converge. What else, if I used the same code that converged in LSDyna but if water flows in different direction then also the code does not converge. I have added two k files as well. Very weird situation for LSDyna!

https://drive.google.com/drive/folders/1XK_wxAW6stMO7wFqNJE8DXqVVZ_AT2Kc?usp=sharing

 

[Jim Day]

We are not permitted to click on links or download files provided in this forum. I see you submitted a related thread to the forum ("Convergence issues with implicit analysis in *load_node_point") which provides further details into your implicit non-convergence issue. To prevent redundancy, that new thread should take precedence and this older thread should probably be closed.