open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi all

1. Does anyone have a definitive list of the tracers in a TELEMAC3D steering file when coupling to AED2? The notes in the example say that 22 initial values are mandatory (in the AED2 manual 24 state variables are listed). I've been trying to match values in the T3D steering file with those listed as initial values in the aed.nml (aed2 steering) file and have successfully identified some, but not all. A list would be really helpful.

2. Is a value for every tracer also needed when the source values are listed, or are the source values matched to the "NUMBER OF TRACERS" keyword.

I would be grateful for any help with these values.
Thanks
John

Hello John,

If you read the balances in the listing file, you can have the name of tracers coming from AED2.
Depending on what you activate in the AED2 STEERING FILE (aed2.nml), the tracers from AED2 are automatically added in TELEMAC-3D (names etc.) and the number of tracers is updated.

I have not tried sources and T3D-AED2 but I think you need as many values of tracers for every source.

Hope this helps,

Chi-Tuan

Hi Chi-Tuan
Thanks for your reply.

If I understand you this means that if I only use the models for aed_sedflux, aed_oxygen, aed_carbon, aed_silica I would write Tracer values for Temp, Salinity (listed tracers in Telemac3d), oxy_initial, dic_initial, PH_initial, RSi_initial and then put zero values for the remaining aed initial values n as they are not used?

Maybe I am confused by the comment in the telemac3d steering file that 22 values are mandatory?

John

Hi John,

Could you please share your experience of compiling AED2 with T3D? Actually I did go through the compilation on Ubuntu 18.04 and had tested a few T3D examples. But when it comes to AED2 examples (waq3d_aed2 & waq3d_aed2_flume), errors emerge: 'Segmentation fault' for serial mode and 'PARTEL: Couldn't split your file' for parallel mode. Just wondering if you've ever met these errors when testing the AED2 examples, any information is appreciated!

Best regards,
Yunhao

Hi Yunhao
We did compile AED2 and ran the flume example but not in parallel. We used the original example setup and experimentsed with this for several trials but our further development into a live domain was stalled due to difficulties with tidal flats in V7p3r0 as described in post #28940.

We hope to restart this work in the next few weeks.
John

Hi John,

Thanks for the information, could you share your configuration file? (e.g. systel.cis-ubuntu.cfg)

Regards,
Yunhao

Hello Yunhao

This seems to be not a problem of AED but an issue with your parallel installation. Did you run other examples in parallel?
For AED examples, can you comment the coupling with WAQTEL and then run the case in serial and in parallel to see if it works?
For your information, AED coupling works only with gfortran compiler (no NAG or INTEL).

kind regards

Riadh

Dear Riadh,

Thanks for the reply, before the AED tests I've tested a few t3d/waqtel examples in serial/parallel and nothing went wrong. I'm using v7p3r1 on Ubuntu with gfortran compiler. As you said I commented keyword 'coupling with WAQTEL' and ran the examples in both serial and parallel, it worked just like tests done before, but if so the AED was not included I suppose?

I turned on the debugger and found the segmentation fault emerged at 'CALL OF CALCS3D_AED2' step, could you look into it? Please see the sortie file attached.

Best regards
Yunhao

Dear Yunhao,

Do not use an existing libaed2.a. You need to install on your system AED2 (attached).

Best regards,

Dear Jacques,

I didn't use an old (existing) libaed2.a file, instead it's created by the 'make' command. For the file you attached I'd like to know which aed version it is, for mine it is v1.0, please check the attachment.

Thank you & best regards,
Yunhao