gross plastic deformation vs material definition

huangyhg

gross plastic deformation vs material definition
i have problem with practical using of finite element method in header analysis. i would like to use gross plastic deformation (gpd). in this case linear elastic - perfectly plastic material model should be used. then criterion of failure is 5% plastic strain. unfortunately i cannot use perfectly plastic material because of problem with convergence. insted of that i used material with tangent modulus 500mpa. after that i have no problem with convergence. i am thinking whether computed plastic strain (for material with tangent modulus) can be used to check gpd criterion (plastic strain should be less than 5%). what will be difference between computed plastic strain for perfectly plastic material and for material with tangent modulus?  
no experience in the thing but i think looking at the diagrams contains the answer. i would say that using a stress diagram where the plastic behaviour is not a flat region but still shows a small pent places the use of such diagram in the unsafe side where critical state is measured by the attainment of some deformation. over the yield strength the true plastic diagram wouldn't admit (nor need) any growth of the stress to produce deformation, where you admit that some growth of stress is required in the other diagram to still gain deformation; hence the deformations with the elastic-plastic diagram would be sooner bigger with the growth of stress and this places the use of your scheme in the unsafe side.
one thinks there must be some way to mathematically attack the elastic-plastic model for your case. it is everyday routine for the steel and reinforced concrete structural design, if masked by parctice.
often when you have problems with convergence when using a plastic or other nonlinear material, you can obtain your solution by decreasing the load increment used; say you applied 1000 lbs to something, no convergence..try applying 100 lbs over 10 steps, see if you can converge then.
is gpd some kind of code requirement? do you have to use elastic-perfectly plastic material model (which uses the usual linear elasticity slope to model the linear, then a horizontal slope for the plastic strains), to be conservative? or can you use whatever elastic-plastic material model fits the data (such as ramberg osgood)?
yes, it is en 13445 requirement to use perfectly plastic material, and for this material model plastic strain should be checked. prost,  you are right that applying more steps will be helpful.  
prost, returning to your question - using ideal plastic material is en 13445 requirement (if i use gpd method). see
section b.8.2 where gross plastic deformation method (gpd) is described.