Hi All:

I want to run fluent for optimisation using batch approaches. Before exit a fluent run, some parameter data need to output from a fluent case. I want to know how to take data from batch run. If anybody has these experiences it is appreciated to share his/her experiences.

Depends what data you mean, but if it's something you can do within the existing Fluent text menu system (like exporting text values as a file), you can do that from a journal file. You can also use Scheme to create a string that encodes the name of the file, for example you can step through 10 velocities and have filenames that include the characters V1, V2 ....

Thanks for your reply. Do you have some references to do it?

Cheer!

Wang

Wang, an early chapter in the User Guide describes the text interface (this is also the chapter that introduced the GUI interface). Look there. I posted some information to this board a week or so ago that is on this topic, so I cut-and-pasted that posting below.

It takes a bit of learning but now that I'm using Fluent that way I am very happy with the result. They could have made this cleaner, I think; but it does work. I'd say the same for Gambit, in re setting things up under program control (though of course you're not looking for data output from Gambit).

For years I've used another product, SAS, which is a statistics and general purpose analysis package. Now I have it creating journal files for Gambit and Fluent, running sessions of Gambit and Fluent based on these files, and analyzing the text files that Fluent writes. I can combine the methods of "experimental design" and CFD. Very powerful.

Here's the earlier posting, which I hope helps you get started writing programs in the form of journal files:

>>>>>>What the manual didn't make clear enough (in my opinion) is that there are practically two different languages in play.

One is Scheme, which is a dialect of Lisp. It's distinguished by lots of parentheses and by prefix notation. You can find several web references or buy books. The funny thing about Scheme and Fluent is that Fluent obviously wrote many functions and defined many variables in Scheme, but the only one you're supposed to know about is ti-menu-load-string, which passes a string to the Text Menu / Text Prompt system. But that should be enough.

The other is Fluent's Text Menu System and Text Prompt System. These are sort of text versions of the menu commands, that prompt you for whatever they need. There's a command reference in Fluent's docs that lists all the Text Menu System items and what they do, but it doesn't explain all the prompts that each item can elicit. AFAIK you just have to experiment to find these out so you can anticipate them and string all the answers together in your journal file. Warning - the prompts are often situation dependant. You can reduce the situational dependancy with /file/set-batch-options y y y y (or something like that).

You can mix the languages together on a line by line basis. Actually, Fluent wrote the Text Menu System and Text Prompt System in Scheme, so you're really only dealing with Scheme running their program. But the feel is like switching between two.

Generally getting Fluent to do anything involves a TM / TP system command, whereas complicated programming mechanics like running loops or using conditionals or assembling strings out of variables always involves Scheme. So you can write a Scheme loop that includes several ti-menu-load-string commands to feed Fluent the commands it needs, and for example you could test your heart valve at 10 different velocities.

Thanks!