Non-linear analysis of an inflatable air-beam model not converging.

I am analyzing an inflatable air-beam with a baffle (similar to the ones used for inflatable wings). 

Geometry and Mesh.

Both bodies are modeled as surface bodes and meshed by shell-181 elements.

The boundary conditions are - Fixed support at 1 end, force over the cylindrical surface and internal pressure, to simulate the beam.

Shortly after 27-28 converged sub-steps, the solution fails and I get an error stating - "Solution not converged at time 5.47E-02 (load step 1 substep 30). Run terminated."

Also, the Newton-Raphson residual plots show barely any residuals - 

Things I have tried - 

1. Refining the mesh.

2. Increasing number of steps and sub-steps.

3. Adjusting % tolerance and MINREF load conditions.

4. Changing the material.

When I changed the material to structural steel the analysis did converge but I cannot converge it at all with the required nylon fabric material. I even tried reducing the overall load to about 1-2 N but that didn't help. 

Are there some irregularities in my material properties or am I missing something else?

Thanks in advance

Comments

  • peteroznewmanpeteroznewman Member
    edited August 2019

    Looking at the Force-Convergence plot, it looks like the model was close to converging three times but reached the 26 iteration limit when a bisection was used.

    Try inserting a Command Object into the model and use the command

    NEQIT,100

    which will let the solver do 100 iterations before bisection instead of the default 26 iterations.

  • Zaber5Zaber5 Member
    edited August 2019

    Thanks for replying!

    I tried the command you suggested and it seemed to work when the shell thickness was set to a high value of about 5-10+ mm. My models thickness is 0.4mm.

    This is the force convergence plot I am getting after using the command - 

    The above is for nylon fabric and the bottom one for steel.

    The error has changed to excessive element distortion and I am getting very high residuals as well.

    Assuming material properties are correct, is it safe to conclude that for the current thickness and cross sectional moment of inertia the load is too much to handle? 

     

Sign In or Register to comment.