[Paulo SPS]

Hi there, I keep getting the following error message when running a stiff chemistry combustion case:

********** ISAT_ABORT **************
routine = isat_leaf_init
location = 1
message = svd failed, INFO=
61

****** END ISAT_ABORT **************


I already set the aggressiveness factor to 0 and ISAT error tolerance to 0.1 (which I know is pretty high) but after a few iterations (less than 20) the chemistry integration fails with the same message. There are no other error messages so I'm a bit at a loss at what to try, any ideas?


Thanks in advance

[jllamas]

Hi Paulo. I am afraid I cannot help you, the same issue is happening to me. I am simulating a combustion process and from time to time I get the same error message. Every time in a different iteration.
In my case, agressiveneness factor is 0.5 and ISAT error tolerance 0.001. It does not matter the figures in the tolerance, I get the error message. It can be running for hours without any issue or it ca be running five minutes and find the error message again.
I am a bit loss as well. I have no idea why.

I would appreciate any help on this topic.

Thank you.

[Maruko]

Hi Paulo, I'm Maruko. I have a similar problem. I am wondering if you solved the problem. I am simulating a combustion process used the EDC model with GRI3.0. Once the case runs, I get the message
********** ISAT_ABORT **************

routine = phi2y
location = 1
message = non-positive species sum

****** END ISAT_ABORT **************
I have no idea why either like you once did. I will be really appriciate if you reply me.
My e-mail address is : 904670919@qq.com if you find it more convenient.

[Maruko]

Hi, Paulo, the same issue is happening to me. Could you please tell me how you solve your problem? I would appreciate any help on this topic.

Thanks in advance

[Paulo SPS]

Hi there, unfortunately I dont have an answer

[lei]

Hello everyone. I agree that the ISAT table together with EDC model is one of the key challenges in modeling detailed reaction mechanisms within reacting flow. As you already mentioned, it may show error and cause divergence from time to time, especially in simulations which is sensitive to convergence. Some best practices might be useful to my experiences:

1. A good initial condition is critical to the convergence of EDC/ISAT model. This good initial condition can be achieved by running the non-reacting flow with a robust turbulence model until it is fully converged, then patch your flame zone with high temperature to mimic a spark. Or, use other combustion model (non-premixed or partial premixed flamelet model) to get a rough temperature initial solution;

2. Use smaller URFs for momentum, turbulence for better convergence control, and start with small aggressive factor for EDC model, at least in the beginning of the simulation;

3. Try to use chemical calculation per 1 flow iteration in the beginning to stabilize the flame, then progressively change to more iterations for better convergence;

4. Some reaction mechanisms might be challenging for convergence if it has global reaction parameters (some may contains negative reaction orders which cause singular point), some numerical treatment might be needed for these reactions. GRI in general is robust but need attention for some complicated flows.

Hope these tips help.