Composite column in fire with concrete damaged plasticity 1

[szymeksqc]

Hi everyone,
I am trying to model a composite column which is subjected to fire according to standard-fire curve and validate the obtained results against the test that had been carried out previously. I have, unfortunately run into some problems.

I am using a concrete damaged plasticity model for concrete and plastic for steel, both of them being temperature dependent. My model consists of 3 steel parts merged into one, i.e. steel tube, bearing and support plates. Tube is filled with concrete.

I am doing a sequentially coupled thermal-stress analysis. After obtaining buckling mode and heat transfer results, I import these results into the final model, which consists of two steps – one being load introduction, while the second is reading temperatures in the nodes from the previous analysis.

I was unable to reach the desired failure time, with CDP analysis aborts after 450 seconds at best. The same model with plastic material formulation for concrete is far closer to the test result, offering more than 2000 seconds until failure (2400 is the desired time). I am aware of the deficiency in using pure plastic behaviour for concrete and I am willing to proceed with CDP, if onlq the results might be improved.

Here are some observations that I made in the process:
Neglecting thermal expansion of steel results in failure after 2327 seconds, while dividing it by 10 - 2170 seconds.
For some reason, computed time is sometimes dependent on imperfection size – after introducing imperfection of 1 mm Abaqus aborts analysis after 230 seconds, for 3 mm it is almost doubled, being 450 seconds, as mentioned above.

As contact property I use normal behaviour (“Hard” contact, Default), for tangential - friction coefficient 0,3 with a specified shear stress limit equal to 0.55. Other contact formulation (frictionless, rough of tied) did not end up with better results.

For the most part I am getting error Time increment required is less than the minimum specified after several warnings plasticity/creep/connector friction algorithm did not converge at n points.

I enclose msg file for the last analysis and 3 input files that I am using in the process.
Hope somebody encountered such problems before and might help me.

Best regards,
Szymek

 

[manarfadul]

hello,
I hope that you soled your problem. I am working on similar thing,but still stuck in the heat transfer analysis as I don't get enough convergence due to water migration problem in concrete. have you validated your heat transfer results? I noticed that the specific heat you defined for concrete is not the same as for the Euro code. what moisture content you assumed in your model?

 

[szymeksqc]

Hi,
sorry for the delayed response, hope you have already overcome your problems.
Specific heat I used in the enclosed model was for 2% moisture content in concrete (with a peak at 115C). Only because of constant step between temperatures I did not catch peak value of 1653 J/kgK for 115C in the input file. Basically, for the validation most important thing was to introduce proper gap conductance between steel tube and concrete core, as in this manner I could influence the slope of temperature rise.