Solving of big models in FloWorks

[ArtemTat]

Hi everyone,

I have one problem. I need to calculate transient process in a pipe of 1000m along. But I've ran into a few problems:

First of all the SolidWorks forbid the creation of bodies such long. But I've avoided this problem using shorter parts in assembly.

But the FloWorks fails the calculation and shuts down the applications.

May be someone can help me...

Thanks in advance

P.S. Sorry for bad English, I'm not from english speaking country.

 

[ZillionM]

Hi,

Do you use 64-bit version? Because that would add to your capacity.
Thats what I think is the problem, no?
Its so big, there isnt enough capacity to run it all through.
Link:

http://www.youtube.com/watch?v=PDOf3IEr4FU

Greets
ZillionM

 

[ArtemTat]

Thanks for answer.

I've tried to make this model in x64 version of SolidWorks, but it's still a ban to create the feature along more than 500m.

Unfortunaly I haven't a ability to use the FloWorks x64.

Also I've tried to split the pipe to some parts (along less than 500m). And solve them separately by the next way: firstly solve first part. After get it outlet parameters in every time moments and make them the inlet parametres of next part of pipe and so on. But I'm not sure that this way is correct.


Thanks,
Artem

 

[ZillionM]

Very innovative!
Thats how you should work with sw-fw by my idea.
great job.

You can test your theory bij scaling down the preject so you can run long lengths of pipe.
After that, split them up and apply your way:
"And solve them separately by the next way: firstly solve first part. After get it outlet parameters in every time moments and make them the inlet parametres of next part of pipe and so on."
When both results match, your theory is true.
When not, try to estimate the deviation and research if you can tweak your settings until you have desired result.

thats what I would do
good luck

ZillionM

 

[ArtemTat]

Thanks for help!

I decide to listen to your advice and if it's not critical deviation I will use this method. In any case I think it's no another way to solve this problem (at least in the exist versions of FloWorks).


Thanks,
Artem

 

[morgan349]

if it was me, I would be tempted you analyse seperate sections using a transferred boundary conditions method. its similar to splitting up the model, but will give more consistancy in what you are doing...

 

[ArtemTat]

Can I transfer all the results for each moments of time (I use time-dependent problem). Have I to select the project file instead .fld?