Learning Forum

[pashazanousi]

Hello, everyone,
In a static analysis, I simulated a floating roof of a liquid gas storage tank, and the results agree with the analytical ones. Firstly, I thought I could transfer the static results to transient as its initial condition. But then I found that it is not applicable. After surfing the web, I found this forume and realized that I should define three steps in the transient analysis; The two first steps generate the initial condition for the 3rd step. Thus, I turn off time integration option in step controls for the first two steps to get a static state at the end of the second step. But now, I face some strange errors and wonder why the solution can not converge.

 

I attached some pictures that may be useful.

 

 

[Ashish Kumar]

Hi,

I cannot see the exact error. Can you try increasing the substeps or number of iterations? By default Mechanical uses 25/26 iterations. You can increase this value by issuing a command under setup branch. Use the command: neqit,50 (no. of equilibrium iterations =50).

Regards,

Ashish Khemka

[pashazanousi]

Hi,

I'll apply what you said and increase the equilibrium iterations to 50. As I said before, I did a static analysis successfully, which made this error confusing.

Thanks for your attention

[pashazanousi]

Hello experts; the problem remains unsolved. I think the frictionless contact cause this problem. Still, the only contact simulating the actual physical condition of the Floating body is the frictionless contact between the solid and fluid face (neglecting friction to validate the analytical solution). When I used bonded connection, the static results didn't match the analytical solution, but the transient analysis ran smoothly.

I tried to use Dynamic Explicit to unravel the issue. But there is no option to control time integration to generate initial conditions for transient analysis and no option to control timesteps. Another option for the initial state is using the static solution as an initial condition for Dynamic Explicit analysis, but it encounters errors because of frictionless and no separation contact.

Is there any idea?