[Pyotr]

Hi everyone,


I have a similar question to the one made by Felipe on post #19.


I’d like to know if it is possible to get the value of a variable at iteration n, pass this value to an expression that would be constant for the next n iterations, and, at iteration 2*n, get the value of that same variable and pass again to the expression, and so on, until the value of the variable doesn’t change above a certain acceptable error, i.e., the value of the variable at iteration (a+1)*n be different from the value at iteration (a)*n in less than, say, 1%.


Is it possible to get this done using CEL?


Or is it only possible using user fortran? If so, does anybody have a fortran code that does a similar job, so I can use it as a guide? I’ve never used user fortran (just read the tutorial) and I have little experience with fortran programming, so any help is precious.


Thanks

[NewToAnsys]

In such a case, where one of the material properties of the domain is dependent on another variable (which has been added to the dependency list of the property in the RULES file), an analysis that normally could be run at steady state would have to be transient right? Since the material property of the domain is changing, at least initially?

[ghorrocks]

I have not looked at the discussion in this thread in detail, but I can say that convergence will be similar for steady state or transient. The fact the material property changes affects convergence regardless of whether it is SS or transient, but once made numerically stable you will be able to get SS and transient simulations to converge with it.