Hello Abaqus users,
I'm writing UMAT for a fatigue analysis. I use direct cyclic step for calulations. My mat.properties depend on N cycles or total time. Do you know how to define the total time step or number of cycles in umat?
Now, I calculate total time by KINC parameter (increment number). It works fine until my results converge. If not, the totaltime reset to zero, but the increment number not. I tried to call TIME(1) and TIME(2), but they just show step time not the total time.
Abaqus experts, how to address this problem?
Even though I didn't fully understand your intention - writing to Time(:) does not make sense as this variable is only passed into the subroutine. If it changes within the routine it will not affect the solution as it is not an output parameter from the subrountine. Also you will have a hard nut to crack if you're trying to define the cycle jump of the direct cyclic approach.
We move to static general (or explicit) analysis for fatigue subroutines with the total time representing the cycle number. I know this won't answer your question, but it might give you some hints.
Raffael Bogenfeld, I hoped that time(2) is smth like total time. Now, I understand that it can't be used for this.
Am I correctly understood that for fatigue you use static general? so, is it like set time period =1000, max. iter= 1000, and increment size =1 (as equivalent to 1000 cycles)?
Yes, that's the way we do it with user subroutines. However, most of my personal work is with explicit simulation (explicit dynamic step) and VUMAT. Here we do the same thing.
Of course it has some disadvantages, as you can't easily realize a cycle jump as it's done by the direct cyclic step.
Raffael Bogenfeld Thank you. I tried a fatigue analysis with static general, and it works fine. Is it possible to use Explicit solver for fatigue? It messages me that too many steps required even for 1000 cycles. I would like to use it, because buckling can be considered.
The time increment in an explicit simulation is usually very small and you're limited with the maximum number of increments before the simulation becomes unstable (Abaqus claims it's around 2Mio for a double precision analysis)
But you are free to choose the correlation between the time and the cycles as you wish.
Hi Aleksandr Elkin,
I didn’t understand your question. However, when you want to use a subroutine to simulate fatigue in the composite structures, you define steps, and for each step, you define a specified number of cycles. Then, you can observe the effects of fatigue on material properties in each step.
I recommend checking the link below. You will learn many things about implementing subroutines to simulate composite fatigue. https://caeassistant.com/product/composite-fatigue-simulation-with-subroutine/
Best regards.
- Badges
- Science topic