One of the noticeable change that has been made in 2019 solver is its capability to handle large models. SIMULIA has noticed that in recent past customers have shown increased interest in dealing with very large models with 2M degrees of freedom or more. There are multiple reasons for this requirement. First is scalability. More and more customers are interested in large system level simulations compared to part or assembly level. Second is fidelity and accuracy. Mesh size is getting finer to capture behavior at micro level instead of macro level.
When it comes to solving very large models, iterative solver has many advantages compared to direct solver such as less memory consumption, scalability and speed. A new hybrid iterative solver scheme has been introduced that offers more flexibility for choosing number of MPI ranks and number of threads per rank for a given node. These parameters can be defined in the abaqus_v6 environment file. This is equivalent to DMP+SMP allowing efficient memory management.
The iterative solver is available in 3DExperience 2018x FD05 and 2019x FD01 release. It supports many more Abaqus features compared to conventional iterative solver such as gaskets, friction, plasticity, creep, periodic boundary conditions etc.
This blog is a part of series “what’s new in SIMULIA 2019”. Please follow our blog site regularly for next blog article on this topic.
Latest posts by Ankur Kumar (see all)
- From point solution to collaborative solution: Are we alone! - August 13, 2020
- Meshing in 3DX vs. Meshing in Abaqus CAE: Does comparison makes sense! - July 21, 2020
- 3D Experience simulation compute power - July 16, 2020