adding a uniaxial material

[jwaugh]

Ok, so I coded up my new material model. if I assign it to a truss element I can cycle it and generate appropriate stress strain diagrams. The materia works just fine in this situation. You give it a particular strain it will give you the stress that goes with it based on the history.

Now if I take it and use it as part of a fiber based beam-column element abd run it, the analysis well not converge. I am at a complete loss at how to attck this problem. Thanks for all the help

[fmk]

a single truss may be too simple to check .. maybe a few truss elements might fail .. if that fails try a cantilever with the dispBeamColumn (no internal iterations at each step) .. the trick is to come up with as simple a model with as few materials that fail and then to look at the stressstraintangent in each material at each trial step .. instead of recorder command try algorithmRecorder

one thing to check is to make sure you determine new state based on last committed step and not the last trial.

[jwaugh]

Frank when I do that I get a very interesting output, but I have one question. Say I have a column element with 2 integrations points. They will be at the top and bottom of the column. Now first I apply an axial load., I print the convergence at each step so I get

Output for 1 IP
Output for 2 IP
Convergence information
(repeat until convergence is reached for the axial load)

Then I start tot laterally displace the column. So I get

Output for 1 IP
Output for 2 IP
Output for an IP
Output for an IP
Convergence information
Output for 1 IP
Output for 2 IP
Convergence information

Why do I get 4 outputs when I only have 2 integration points before the firsst convergence flag. But after that if only kicks out an output for the IP and then checks convergence. Thanks I hope that is clear. By output I mean it is printing to the screen whenever certain places are called.

[fmk]

are you referring to the 2 extra at the start of each step .. you are using disp control, so there is a initial predictive step at the start before the iterative corrective steps to achieve convergence.

[jwaugh]

Yeah, I am using displacement control, thanks Frank. That helps clarify things greatly.

[jwaugh]

If you have a beam column element with 2 integration points they are located at the top and bottom of the column correct?

If I take a cantilever column and apply an axial load to the column, and then laterally displace the top of the column. Shouldn't the fibers at the integration point at the top be uniformly stressed to resist the axial load, assuming a linear transformation so no P-Delta effects?

Thanks for all the clarification