Restart problems: cs_xml_reader.XMLError: Multiple instance

[Luca]

I am very new to Code_Saturne and I have some problems in restarting analysis. I activate the restart option from the GUI and point to the checkpoint folder under RESU. Then I lunch the analysis but I get an error:

cs_xml_reader.XMLError: Multiple instance of restartnodes

Could someone tell me what I'm missing? I know it might be very stupid. The version I'm using is 2.2.0

Thank for your help!

[Yvan Fournier]

Hello,

Did you generate the XML with a previous version ? If this is the case, re-building the XML is usually safer.

In any case, you may try to edit the XML file with a text editor, find all <start_restart> tags, remove them, then re-edit the file with the GUI to re-select the restart.

Regards,

Yvan

[Luca]

Thanks Yvan, it solved the problem!

[Luca]

Sorry to disturb you again!

Now I can restart simulations but I would like to reset the time to 0. Do I have to do it using a subroutine or there is a more convenient way? In case, which subroutine do you advise me to use?
I also tried to change the postprocessing frequency with a subroutine but in version 2.2.0 I cannot find usnpst.f90 in the SRC/REFERENCE/base directory. Is it normal? in the manual 2.2.0 it is referenced from what I can see.

Thanks again for the help!

Luca

[Yvan Fournier]

Hello,

To reset time to zero, you need to use a user subroutine (and do it after reading the restart file header), but do you need to reset time (except for logging/postprocessing legends) ?
If you really need to do it, I recommend cs_user_initialization.f90 (for recent versions/usiniv.f90 for older versions), as it is one of the first user subroutines called after reading the restart files.

The 2.2.x documentation is not quite up to date: usnpst.f90 is now replaced by cs_user_postprocess_activate, in cs_user_postprocess.c.

Best regards,

Yvan

[Luca]

Thank you again!
I need to restart time because I want to initialize the fields to start an ALE motion at time 0. The initial field is going to be the same for many different analysis. I think your reply anyway will solve the problem!