[Anna Tian]

Hi,

I'm looking for CFX shortcuts summary to accelerate my operations on CFX. Could anyone provide it or let me know where could I find it?

[JuPa]

In steady state you can set individual physical timesteps for each equation to accelerate convergence, rather than having one universal timestep for all equations.

In transient you need to have converged residuals and global imbalances within every coefficient loop for every equation solved. If you care more about the psuedo steady state final solution then you can afford to have a few coefficient loop iterations to not converge.

Lastly, a better initial guess can reduce convergence times.

[Anna Tian]

Originally Posted by RicochetJ

In steady state you can set individual physical timesteps for each equation to accelerate convergence, rather than having one universal timestep for all equations.

In transient you need to have converged residuals and global imbalances within every coefficient loop for every equation solved. If you care more about the psuedo steady state final solution then you can afford to have a few coefficient loop iterations to not converge.

Lastly, a better initial guess can reduce convergence times.

When I say 'shortcuts', I meant 'hotkeys'. For example, 'ctrl+s' can save the project faster than clicking on 'save'. Any CFX hotkeys summary?

[Anna Tian]

Originally Posted by Anna Tian

When I say 'shortcuts', I meant 'hotkeys'. For example, 'ctrl+s' can save the project faster than clicking on 'save'. Any CFX hotkeys summary?

Could anyone help on this question?

[singer1812]

I am not sure of any hot keys that are available, not that I have looked for them. You must be making models at lightning speed in order to really want this feature

[monkey1]

RicochetJ is saying "In transient you need to have converged residuals and global imbalances within every coefficient loop for every equation solved"...meaning I would need only ONE coefficient loop per Time step iteration?
Or was it just a mistake and you wanted to say I need convergence within every Time Step iteration?
One Time step iteration consisting each of several coeffiecnt loop iterations.

Just wondering because I never managed to have transient runs with convergence within 1 coefficient loop iteration for each time step...but it could also be that my simulations are just not accurate enough

[JuPa]

Originally Posted by monkey1

RicochetJ is saying "In transient you need to have converged residuals and global imbalances within every coefficient loop for every equation solved"...meaning I would need only ONE coefficient loop per Time step iteration?
Or was it just a mistake and you wanted to say I need convergence within every Time Step iteration?
One Time step iteration consisting each of several coeffiecnt loop iterations.

Just wondering because I never managed to have transient runs with convergence within 1 coefficient loop iteration for each time step...but it could also be that my simulations are just not accurate enough

What I wrote isn't 100% true. It will be nigh on impossible to converge every coefficient loop.

What I should have said is: you can afford to have loose convergence (not every time step converging within 1% global imbalance) if your problem permits.

[monkey1]

Feeling better and reassured now