[longderson]

Dear all,

Is there any way to freeze some of the additional variables while solving the rest? All the additional variables are defined as transport scalers. Some of them are initialised with external files and actually act as data stacks, thus, they are preferred to be frozen.

I know the expert parameter might help when the model is overridden as "solve scalar = t/f". But the question is how could I disable the scaler equation selectively?

Maybe someone has an idea?

[Opaque]

Under the additional variable definition, i.e.

ADDITIONAL VARIABLE: MyAV
Option = Definition
...
Update Loop = None
...
END

You can do that by a right-click on the icon in the outline for that variable, and use Edit in Command Editor

[longderson]

Originally Posted by Opaque

Under the additional variable definition, i.e.

ADDITIONAL VARIABLE: MyAV
Option = Definition
...
Update Loop = None
...
END

You can do that by a right-click on the icon in the outline for that variable, and use Edit in Command Editor

Thanks for your remarks. May I ask what version of CFX do you use? I tried exactly as your suggestion with CFX 17.1, but it shows little difference.

[Opaque]

May I ask why you are using transported variables for stack variables, when you have algebraic variables already for that exact purpose?

[longderson]

Originally Posted by Opaque

May I ask why you are using transported variables for stack variables, when you have algebraic variables already for that exact purpose?

Yes, you are right. The algebraic variables work perfectly for data stacks. However, these variables are intended to be updated during the calculation with a user-defined function. If algebraic variables are used, the updated value will be automatically overridden by the specific expressions.

[ghorrocks]

Then don't you make the expression setting the algebraic variable point to your user function?

[longderson]

Originally Posted by ghorrocks

Then don't you make the expression setting the algebraic variable point to your user function?

Yes, I did. I use the GETVAR command to update the variables. The initial value of the algebraic variables setting by the expressions is supposed to be overridden.

[longderson]

My problem has been solved by setting the variable type with the algebraic equations and disabling the loop updating by the command line "Update Loop = None" as suggested by Opaque.

Under the additional variable definition, i.e.

ADDITIONAL VARIABLE: MyAV
Option = Definition
...
Update Loop = None
...
END

You can do that by a right-click on the icon in the outline for that variable, and use Edit in Command Editor

BTW, for transport scalers, this command seems invalid. Thus, the question is still open for how to disable the transport scaler equations selectively.

[ghorrocks]

No, what I meant was that you set the algebraic variable to UserFortran() where UserFortran is a CEL link into your user fortran routine. Then you set it to whatever you like in your user fortran expression. Should be no need to use GETVAR. Have a look in the CFX documentation on user fortran expressions for CEL.

There should be no need to use undocumented features like Update Loop for this. But at the end of the day you have to use whatever works.....

[longderson]

No, what I meant was that you set the algebraic variable to UserFortran() where UserFortran is a CEL link into your user fortran routine. Then you set it to whatever you like in your user fortran expression. Should be no need to use GETVAR. Have a look in the CFX documentation on user fortran expressions for CEL.

There should be no need to use undocumented features like Update Loop for this. But at the end of the day you have to use whatever works.....

Sorry for the misleading information, I use the junction box to read data from external files other than using CEL functions to define the variables.

[ghorrocks]

That is what I am saying. If you use a user CEL function and not a junction box then you won't get this problem. User CEL functions are much easier to get working, as well.