Learning Forum

[İlker]

Preconditioned conjugate gradient solver error level 9999.  Please send  the data leading to this operation to your technical support provider,   as this will allow ANSYS, Inc to improve the program.

 

 

 

 

How can I solve these two errors?

 

[Nanda]

Hello Hasan,

To answer your first error, it simply states that your system doesn't have enough memory for solving the model. Theoretically ANSYS needs about 15 GB RAM for solving 1 million DOF's.

1. Reduce the model size, you can either change the element formulation to linear or make a coarsh mesh if possible.
2. If that error persists, kindly free some space from your available memory.
3. If you need to solve with your current mesh settings, and you can't free any space, you need to upgrade your hardware and increase available memory

Coming to your second error, the error message itself means that the PCG preconditioner matrix size overflowed. You can try to reduce the level-of-difficulty by using command PCGOPT or change it to the sparse solver. It does not suffer from preconditioner matrix size overflow.Theoretically Distributed ANSYS also helps because the preconditioner matrix will be split.

Regards,

Nanda.

Guidelines for Posting on Ansys Learning Forum

How to access ANSYS help links

 

[İlker]

Hello,
Firstly, thank you. I am using 16GB of ram. I have space on my hard drive. There is more than 300 GB of space.
How do I use the command you said? I would appreciate it if you could explain a little more in detail. Thank you.

[Nanda]

Hello Hasan,

Can you check the mesh statistics in your model and see if number are nodes are falling inside the criteria I mentioned earlier, DOF can be calculated by number of nodes*3 for solid elements and nodes*6 for shells. Have you tried changing to linear element formualtion or making a coarse mesh. Your message tells you have around 14GB available memory, whereas this model requires about 25GB memory. You can switch solver type to iterative, this needs less memory to solve.

Regards,

Nanda.