Bug when using FRICTION DATA = YES with Finite Volumes in TELEMAC-2D

When using FRICTION DATA = YES (friction laws defined by areas, more than 1) and the Finite Volumes kernel of TELEMAC-2D, there is a bug with the use of KFROT variable (integer) as it is not the number of friction law but it is changed to the number of friction areas. KFROT is used in SOURCE_MOMENT subroutine. Bug raised by Rebekka Kopmann.

added module:telemac-2d type:bug labels

In SOURCE_MOMENT, KFROT should be replaced by NKFROT%I(IS) (friction value for node IS). Fix validated by @florent.taccone

Fix done in branch FV_frictiondata In case of KFROT = 0 or FRICTB and NKFROT%I(IS)=0 CF=0. The check that a correct friction law is used (NKFROT.GE.1.AND NKFROT.LE.5) is deleted. In case of friction data=yes this is checked in friction_read, and without it is checked in friction_calc.f.

Hello Rebekka,

I think the following lines:

          AKAP = 1.D0 + CF%R(IS)*DT*SQRT(QU%R(IS)**2+QV%R(IS)**2)/
     &           (2.D0*HN%R(IS)*UA(1,IS))
          IF(KFROT.EQ.0.OR.(FRICTB.AND.NKFROT%I(IS).EQ.0)) THEN
            AKAP = 1.D0
            CF%R(IS) = 0.D0
          ENDIF

can be replaced by

          IF(NKFROT%I(IS).NE.0) THEN
            AKAP = 1.D0 + CF%R(IS)*DT*SQRT(QU%R(IS)**2+QV%R(IS)**2)/
     &             (2.D0*HN%R(IS)*UA(1,IS))
          ELSE
            AKAP = 1.D0
            CF%R(IS) = 0.D0
          ENDIF

as in case of not FRICTB, NKFROT%I(IS) <-- KFROT That avoids additional tests.

Chi-Tuân

Hello Chi-Tuân, yes, your proposal is better! I have changed it in my branch and already push the modifications. Also the validation run through except the vnv_gouttedo_med.py (I have not compiled med) and delwaq there is a problem with merging T2DL11 file but the parallel run was correct. 

Best regards, Rebekka

Hello Rebekka,

Please find enclosed a better file with less declarations for USE DECLARATIONS_TELEMAC2D, ONLY (KFROT changed to NKFROT as not used anymore and no need for FRICTB for this version of fix). It works for a validation with tag telemac2d (no differences between result files). I can push this fix on the main if you are OK.

Chi-Tuân

source_moment.f

mentioned in commit 9b40a87d

Fix pushed in SHA 9b40a87d

closed

mentioned in commit 08fd121c

mentioned in commit 6928b439

mentioned in commit 8a65ad56

mentioned in commit 496d20e1