Help regarding Meaning of Usage of Scaling while defining the displacement variables in MOOSE SolidMechanics Physics type problem #28453
-
Hi Moose Team, I am a new Moose user and currently learning to use Moose. I have a doubt regarding defination of variables in the Moose input file. Ref : https://github.com/apatra6/rhocp/blob/main/examples/tantalum/strain_rate_effect/1e-4ps/bcc_pxtal.i
I am unable to understand what is the function of scaling parameter while defining the variables for FE problem. I have referred the MOOSE documentation and source files still didn't understood the purpose of scaling parameter in the defination of variables. And How do we decide ideal value of scaling parameter. Kindly help me understanding this concept. |
Beta Was this translation helpful? Give feedback.
Replies: 2 comments 2 replies
-
Hello Scaling is used as a simple preconditioning of the equations. When there are multiple equations, we weight the residuals of the equation for each variable by this scaling factor to make the convergence criterion more fair and/or help with matrix conditioning This page on porous flow has an explanation
Experimentally most of the time. You look at the residuals and see if they are orders of magnitude apart, maybe it's worth bringiing them closer together. |
Beta Was this translation helpful? Give feedback.
-
Thankyou Guillaume, for your help |
Beta Was this translation helpful? Give feedback.
Hello
Scaling is used as a simple preconditioning of the equations. When there are multiple equations, we weight the residuals of the equation for each variable by this scaling factor to make the convergence criterion more fair and/or help with matrix conditioning
This page on porous flow has an explanation
https://mooseframework.inl.gov/modules/porous_flow/convergence.html#46b5c8c6-401d-44a8-80b8-1ae589dcd695
Experimentally most of the time. You look at the residuals and see if they are orders of magnitude apart, maybe it's worth bringiing them closer together.