***ERROR: TIME INCREMENT REQUIRED 1

[mambo5]

I am modelling on composite beam with concrete slab on the top subjected to heat, when I use element S4R, the analysis run OK, when I change the element to S8R, ABAQUS could not complete the analysis and gave me the following error;

***ERROR: TIME INCREMENT REQUIRED IS LESS THAN THE MINIMUM SPECIFIED

I had my *.INP step as follow;

*STEP, AMPLITUDE=RAMP, NAME=MIN 1, INC=500, NLGEOM=YES
*STATIC
0.01, 1, 0.00001, 1

what should i do? please advice.

 

[stone05]

Change the "Maximum time increment allowed" (the 4th parameter) to 0.1.

 

[corus]

I don't think changing the maximum allowable time increment will do anything as the analysis has cut back the time increment to be less than 1e-6 and has never reached the maximum.

You could have problems with the meshing and the positions of the mid-side nodes (with distortion) to give you bad results, even if it were at only one node. Without looking at the mesh it's difficult to say. If you are looking to improve your accuracy and can't reason why 8 noded elements won't work then simply use a higher mesh density for your 4 noded linear elements.

If you have been able to get results at some time within the step, then look at these results. You might find that you have gross distortion at some point which causes the analysis to fail at the next time interval.

corus

 

[mambo5]

Ya Corus, yor are right, by changing the maximum allowable time increment does not help. well, I think I understand your suggestion, I am looking into it. Thanks.

 

[mambo5]

I had already checked the meshing and the positions of the mid-side nodes, they seems alright. I suspect is the concrete properties of the RC slab causing error during the thermal anlysis. any advice?

 

[corus]

It's difficult to say what the problem could be. It can be useful to remove some of the non-linerity in the problem, such as the NLGEOM you have and then re-run the analysis to see if the error still occurs.

corus

 

[mambo5]

i am now try to define the concrete properties myself, ie. not using *concrete but using *elastic, *plastic... as how i've done for steel. and i not included the concrete failure part (after yield), i made it remain constant after yield (as the steel properties, fully ductile), the analysis ran longer than before.

 

[corus]

I can only say I've had similar problems with quadrilateral elements where there is non-linearity in the solution, such as in contact, and now tend to keep to linear type elements but with a higher mesh density. Models I've had using refractory type materials (similar to concrete) I've also used the standard elastic/plastic properties or used cast iron type properties to define the differences between compression and tensile yield.

corus