GOFT/GENER-Error in the Restart/Continuation Run

[Matti]

I have been simulating mass transfer with diffusion only, without sources/sinks in the case of study the glacial erosion of the bentonite buffer.

Just like the effects of different waters to the erosion of the bentonite buffer. In this case, the mass transfer is assumed to be diffusive.

I have planned to take 2-3 steps with different water composition with RESTART/continuation Run –application.

In the calculations I have had a problem with GOFT & GENER –items.

The calculations have been failed however and the program gives following error message; “Element (*number*) in data block “ Goft†does not have a valid “Gener†item –ignore this and proceed “

In the calculation, however, with runs with short end time, ca. 1 year, the calculations are successful. But when you increase the end time, the calculations fail with above error message.

Matti Liukkonen

[AMCHE]

Dear Matt (and others),

I am using the PetraSim software for modelling CO2 sequestration. In order to look at both advective as well as diffuse scenarios affecting the caprock, I am, as a first step trying to equilibrate the rock fluid to a choosen mineralogy of both a reservoir and a caprock region in the model before adding a gas phase. I find no problems in doing that. However, when I add gas, and prepare to set up for a restart run, using the equilibrated system I get the same error of “Element (*number*) in data block “ Goft†does not have a valid “Gener†item –ignore this and proceed “, and the system chrashes. This is also the case when running the "gas-scenario" without equilibration of the fluid.

I have first started using the program; and suspect it is due to my unexperience. I would very much welcome any clue as to why set-up fails; thanks

Anne-Mette

[Alison]

Dear Matti and Anne-Mette:

The GOFT/GENER message you are encountering is actually a warning message, rather than an error message. This message will show up in all models, even for models that run. The problem will be addressed in a future version of PetraSim, but can be ignored for now. There must be another reason that your model is crashing.

In the case of Matti's error, we found that the problem was caused by the fixed file format used by TOUGHREACT to write and read the number of time steps (KCYC) and total number of iterations (ITERC) in the INCON file used to restart the analysis. This is the SAVE file written by the previous analysis. You can also view these values in the flow.out file.

Following is the SAVE file:

INCON -- INITIAL CONDITIONS FOR 6 ELEMENTS AT TIME 0.316000E+10

1 0.10000000E+00 0.00000000E+00 0.00000000E+00 0.00000000E+00

0.1013000000000E+06 0.2500000000000E+02

2 0.43000000E+00 0.00000000E+00 0.00000000E+00 0.00000000E+00

0.1013000000000E+06 0.2500000000000E+02

3 0.43000000E+00 0.00000000E+00 0.00000000E+00 0.00000000E+00

0.1013000000000E+06 0.2500000000000E+02

4 0.43000000E+00 0.00000000E+00 0.00000000E+00 0.00000000E+00

0.1013000000000E+06 0.2500000000000E+02

5 0.43000000E+00 0.00000000E+00 0.00000000E+00 0.00000000E+00

0.1013000000000E+06 0.2500000000000E+02

6 0.43000000E+00 0.00000000E+00 0.00000000E+00 0.00000000E+00

0.1013000000000E+06 0.2500000000000E+02

+++

********** 2 0.00000000E+00 0.31600000E+10

Notice the ********** after the +++. This is an overflow of the fixed format. This causes an error during restart when TOUGHREACT reads this file.

There are two solutions:

1. The easiest is to increase the maximum time step size so that the total number of time steps and interations remain below 99999.

2. You could manually edit the INCON file and set KCYC and ITERC to 0. These values are only used for output, so do not affect the solution. If you choose to do this, you would need to be careful to preserver the fixed format within the file. You would then need to manually run TOUGHREACT from the command prompt.

Regarding Anne-Mette's model, it is hard to know if this is caused by the same problem, or something else. If you would like to send your files to alison@rockware.com, I would be happy to take a look at this with the engineers at Thunderhead.

Best Regards,

Alison