open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi,

I have greater than 99 (i.e. 115) liquid boundaries. In order to get around this issue I have modified q3.f and tr3.f as shown attached.

The liquid discharge (q3) appears to be working fine (see attached) but I am having issues with the tracer flux (tr3) which is bringing in salinities and temperatures as 0 when they have a positive value in the liquid boundaries file (images attached).

Any advice would be gratefully accepted to resolve this.

Thanks,

Tom

Hi Tom,

You can try to use
MAXIMUM NUMBER OF BOUNDARIES: 200
in your .cas file.

It also can be useful to set this number slightly bigger (islands are treated as boundaries)

Alexander

Thanks for your reply, Alexander.

I had already set

MAXIMUM NUMBER OF BOUNDARIES: 300

There were no issues when the number of boundaries was 99 or less.

Regards,

Tom

Hi
I would say you should have IRANK = IRANK+3 instead of IRANK+2 when boundaries number are higher than 100...

Regards

c.coulet wrote:
And probably FMT='(I3)' instead of FMT='(I2)'

I very much appreciate you suggestions but I am still unsuccessful in resolving this.

I have stripped back the problem to one with 99 boundaries (success) and one with 100 boundaries (fail).

Attached are the log files that may shed some light on the problem, together with the liquid boundaries files. The tr3.f code using I3 caused the case to crash (log file attached). The relevant section of the tr3.f code is also attached.

Thank you very much for considering this.

Tom

in tr3 you still have I2 for values greater than 100!

Yes, I realise that. As I mentioned in the previous message the I3 caused the run to crash. I also realise that I3 is a 3 digit integer so it should work.

maybe a too long line?
Try to remove previous values just too check if indices greater than 100 works

I set the 99th value in the liquid boundaries file to be 3 indices. According to the log file attached it doesn't appear that the 3-indice value is recognised.

It may possibly be an issue with line length in the liquid boundaries file as you suggest - is there a way around this?

Thanks.

Tom