0

# termination criterion following two time steps for which convergence on ITER = 1

Hello,

It looks like I am running into this exact issues described on page 44 of the T3 user manual (copied below). What suggestions do you have when this occurs? Reduce the time step duration by implementing a smaller value for DELTMX?

Thanks

Yannick

=============================

From T3 manual

Another termination criterion is monitored when automatic time step control is in effect. In this case, execution will be stopped when convergence failure is encountered following two time steps for which convergence on ITER = 1 was achieved. One benign situation where this case occurs is when a simulation approaches steady state. As conditions approach steady state, the rates of change in thermodynamic variables become small, and time steps will grow to very large values when automatic time step control is in effect. The off-diagonal elements in the Jacobian matrix are proportional to time step size, and may become so large that convergence of the linear equation system may no longer be achieved because of numerical roundoff. Thus, we may have a situation where for a somewhat reduced time step convergence is achieved without requiring a linear equation solution (ITER = 1), while no convergence at all can be obtained when a linear equation solution is required. The aforementioned stopping criterion was designed to prevent such “no progress” calculations to go on for a potentially large number of user-specified time steps.

=============================

From my output file

CYCIT ..........  KC =2786  KCYC =2786  DELT = 0.776506E+08  SUMTIM = 0.293568E+10

QQQQQQQQQQ SUBROUTINE QU QQQQQQQQQQQQQQQQQQQQ --- [KCYC,ITER] = [2786,  1]
...ITERATING... AT [ 2786, 1] --- DELTEX = 0.776506E+08  MAX. RES. = 0.841066E-01 AT ELEMENT 01M51  EQUATION   3

QQQQQQQQQQ SUBROUTINE QU QQQQQQQQQQQQQQQQQQQQ --- [KCYC,ITER] = [2786,  2]
...ITERATING... AT [ 2786, 2] --- DELTEX = 0.776506E+08  MAX. RES. = 0.194497E+02 AT ELEMENT 03y83  EQUATION   2

QQQQQQQQQQ SUBROUTINE QU QQQQQQQQQQQQQQQQQQQQ --- [KCYC,ITER] = [2786,  3]
...ITERATING... AT [ 2786, 3] --- DELTEX = 0.776506E+08  MAX. RES. = 0.544579E+02 AT ELEMENT 03y86  EQUATION   1

QQQQQQQQQQ SUBROUTINE QU QQQQQQQQQQQQQQQQQQQQ --- [KCYC,ITER] = [2786,  4]
...ITERATING... AT [ 2786, 4] --- DELTEX = 0.776506E+08  MAX. RES. = 0.662639E+02 AT ELEMENT 03y85  EQUATION   1

QQQQQQQQQQ SUBROUTINE QU QQQQQQQQQQQQQQQQQQQQ --- [KCYC,ITER] = [2786,  5]
...ITERATING... AT [ 2786, 5] --- DELTEX = 0.776506E+08  MAX. RES. = 0.419441E+05 AT ELEMENT 03y78  EQUATION   3
TOUGH STATUS: Failed in EOS. IGOOD=           2
+++++++++ REDUCE TIME STEP AT ( 2786, 5) ++++++++ ++++++++   NEW DELT =0.194126E+08

QQQQQQQQQQ SUBROUTINE QU QQQQQQQQQQQQQQQQQQQQ --- [KCYC,ITER] = [2786,  1]
01M51( 2786, 1) ST = 0.295510E+10 DT = 0.194126E+08

CYCIT ..........  KC =2787  KCYC =2787  DELT = 0.388253E+08  SUMTIM = 0.295510E+10

QQQQQQQQQQ SUBROUTINE QU QQQQQQQQQQQQQQQQQQQQ --- [KCYC,ITER] = [2787,  1]
01M51( 2787, 1) ST = 0.299392E+10 DT = 0.388253E+08

CYCIT ..........  KC =2788  KCYC =2788  DELT = 0.776506E+08  SUMTIM = 0.299392E+10

QQQQQQQQQQ SUBROUTINE QU QQQQQQQQQQQQQQQQQQQQ --- [KCYC,ITER] = [2788,  1]
...ITERATING... AT [ 2788, 1] --- DELTEX = 0.776506E+08  MAX. RES. = 0.841066E-01 AT ELEMENT 01M51  EQUATION   3

QQQQQQQQQQ SUBROUTINE QU QQQQQQQQQQQQQQQQQQQQ --- [KCYC,ITER] = [2788,  2]
...ITERATING... AT [ 2788, 2] --- DELTEX = 0.776506E+08  MAX. RES. = 0.194537E+02 AT ELEMENT 03y83  EQUATION   2

QQQQQQQQQQ SUBROUTINE QU QQQQQQQQQQQQQQQQQQQQ --- [KCYC,ITER] = [2788,  3]
...ITERATING... AT [ 2788, 3] --- DELTEX = 0.776506E+08  MAX. RES. = 0.544351E+02 AT ELEMENT 03y86  EQUATION   1

QQQQQQQQQQ SUBROUTINE QU QQQQQQQQQQQQQQQQQQQQ --- [KCYC,ITER] = [2788,  4]
...ITERATING... AT [ 2788, 4] --- DELTEX = 0.776506E+08  MAX. RES. = 0.661837E+02 AT ELEMENT 03y85  EQUATION   1

QQQQQQQQQQ SUBROUTINE QU QQQQQQQQQQQQQQQQQQQQ --- [KCYC,ITER] = [2788,  5]
...ITERATING... AT [ 2788, 5] --- DELTEX = 0.776506E+08  MAX. RES. = 0.412518E+05 AT ELEMENT 03y78  EQUATION   3

QQQQQQQQQQ SUBROUTINE QU QQQQQQQQQQQQQQQQQQQQ --- [KCYC,ITER] = [2788,  6]
...ITERATING... AT [ 2788, 6] --- DELTEX = 0.776506E+08  MAX. RES. = 0.104062E+06 AT ELEMENT 03w73  EQUATION   3

!!!!!!!!!!!!!! EXCESSIVE RESIDUAL =  1.04062E+05

+++++++++ CONVERGENCE FAILURE ON TIME STEP #2788 WITH DT = 0.776506E+08 SECONDS, FOLLOWING TWO STEPS THAT CONVERGED ON ITER = 1
STOP EXECUTION AFTER NEXT TIME STEP

3replies Oldest first
• Oldest first
• Popular
• Hi Yannick,

You could maybe try setting MOP2(1) to 2 in the MOMOP block? This will force TOUGH to perform at least 2 iterations and update the primary variables. This sometimes helps me when I run into this type of problem, but I get the problem you are describing a lot - especially during hysteresis runs and often I just cannot find a way to progress the simulation.

I would be interested in any thoughts from the TOUGH experts in this forum, as I have very limited knowledge of the code.

Good Luck,

Gareth

Like
• Gareth Williams Thanks very much for that tip, my run was able to finish properly. I wonder what others' experience might have been on this issue?

Like
• Yannick Champollion No problems Yannick, I am glad it worked for you.

All the best, Gareth

Like
Like Follow
• 7 mths agoLast active
• 3Replies
• 37Views
• 2 Following