High ERR% with no contact or high forces
Altair EmployeeHi ,
I´m Javi from Spain.
I´ve some problems with a Radioss Analysis and after searching during a whole day in the Altair Forum I didn´t find a solution.
1) Radioss has hidden all the BCS and CLOAD but even clicking on Show or touching the graph options I´m not able to view in the graphics area the constraints and loads.
2) When I ran the model, it takes a lot of time to solve it and the energy error is too high. I read that it might be because of the penetration or high deformation but I don´t think my model has a high displacement field. My model is a dynamic force test with AL7075-T6, and it tries to solve a piece attached to a chassis with the force of the shocks applied from 0 to its maximum value (I´m doing all in 2 seconds, ie: 0N in T=0, 500N in T=0.5s … Fmax at T=x and F=0N at Tstop, all of this in a Force curve)
Could you check the model and tell me If I´m doing something wrong? I´m going nut because of these 2 problems and I´m trying to solve it through the forum and the manual but I´m not able to do it.
Thank you very much !
Kind Regards,
Javier
Answers
-
Hi,
1. Utility>BC's manager select the BC you want to display and hit Update
2. If the initial energy in the system is low, then it is possible to have large Energy Errors early in a simulation that reduces as energy is added to the system and this is fine.
This is because small numerical differences in the energy cause large percentage Energy Errors. Hourglass and contact can also cause energy errors because they are not part of the energy balance equation. To avoid hourglass energy use fully integrated element or underintegrated with physical stabilization (Isolid=12 or 24). That being said, I was not able to correct high energy error in your model, which goes to 99.99% right from the start of the simulation and never reduces.
3. The model you have shared is not suitable for explicit simulation, because of small element size that penalizes time-step. You should remesh the model, paying more attention to minimum element length (check with F10>time). Simulation can be sped up using DT/NODA/CST or Advanced Mass Scaling. Time scaling can also be applied to obtain a quasi-static solution as long as kinetic energy is less than 1% of total energy. If loads, geometry and BC's are symmetric the model can be appropriately reduced. Given the load is applied over 2 second period, there are negligible inertial effects so explicit is not necessary. There are also no geometric, contact or material non-linearities that can only be solved by explicit.
Check the remeshed model attached below:
0 -
Altair Forum User said:
Hi,
1. Utility>BC's manager select the BC you want to display and hit Update
2. If the initial energy in the system is low, then it is possible to have large Energy Errors early in a simulation that reduces as energy is added to the system and this is fine.
This is because small numerical differences in the energy cause large percentage Energy Errors. Hourglass and contact can also cause energy errors because they are not part of the energy balance equation. To avoid hourglass energy use fully integrated element or underintegrated with physical stabilization (Isolid=12 or 24). That being said, I was not able to correct high energy error in your model, which goes to 99.99% right from the start of the simulation and never reduces.
3. The model you have shared is not suitable for explicit simulation, because of small element size that penalizes time-step. You should remesh the model, paying more attention to minimum element length (check with F10>time). Simulation can be sped up using DT/NODA/CST or Advanced Mass Scaling. Time scaling can also be applied to obtain a quasi-static solution as long as kinetic energy is less than 1% of total energy. If loads, geometry and BC's are symmetric the model can be appropriately reduced. Given the load is applied over 2 second period, there are negligible inertial effects so explicit is not necessary. There are also no geometric, contact or material non-linearities that can only be solved by explicit.
Check the remeshed model attached below:
Thank you very much @Ivan !
I was processing your message during the weekend, and I was able to understand almost everything, but I think I must read a lot of FEA theory…
Also, thank you for the remeshed model, I have checked all the parameters you have changed and I will take care of my mesh next time
/emoticons/default_smile.png' srcset='/emoticons/smile@2x.png 2x' title=':)' width='20' />
To conclude, (I know this is not the post to ask for that, but know you answered my question, its a good chance to ask that) do you know any great book to learn all you said to me like how to know the best time-step or element size, anything to know what I´m doing when I mesh a model or when I run a solver more deeply. I would like to learn a lot of theory at the same time I´m learning FEA with Altair. (I already have got Altair's Radioss book, but I need something to beginners like me)
Kind regards !
Javier
0 -
Hi,
for beginners I would recommend Practical Aspects of Finite Element Simulation (and other free ebooks from Altair).
Practical Finite Element Analysis is another good place to start (here is a preview).
There is also a theory manual section in Radioss help.
When/if you are ready to dive into depths of FE: The Finite Element Method for Solid and Structural Mechanics.
For practical demonstrations I would also recommend the following Youtube channels:
Altair University
Altair India Student Contest
ELEATION By Apoorv Bapat
0 -
Altair Forum User said:
Hi,
for beginners I would recommend Practical Aspects of Finite Element Simulation (and other free ebooks from Altair).
Practical Finite Element Analysis is another good place to start (here is a preview).
There is also a theory manual section in Radioss help.
When/if you are ready to dive into depths of FE: The Finite Element Method for Solid and Structural Mechanics.
For practical demonstrations I would also recommend the following Youtube channels:
Altair University
Altair India Student Contest
ELEATION By Apoorv Bapat
Hi,
again, thank you very much. I´m sure it will help me to learn a lot of FEA.
Have a nice day!
Javier
0