[Sahus]

Hello Community,
I am getting many warnings in my IC engine simulation (GDI RANS, SAGE). Can anyone help me to get rid of these warnings? which parameter I should look in CONVERGE?

1. (Rank 1) WARNING: Stuck film parcel: ncyc=8111, RANK=1, location (3.902065e-02,-6.000000e-03,-3.442074e-02)

This warning starts 20 CAD after spray and lasts 20 CAD after ignition. I checked the locations and they are at the intersection of piston and liner. I tried compressing small area in piston but it didn't help.

2. recovering .... because transport equations did not converge or energy extrapolations

I tried changing relaxation parameter but again didn't get rid of these warning.

3. WARNING: max iterations exceeded on pstar, continuing for compressible simulation

increased max no of iterations for pressure to 700 from 500 default value, but it still doesn't converge.


Although I am still able to get results but still I would like to know the solutions for these.

Thanks

[jetcheve]

Hello!

Let me address these one by one.

1 - This literally means a film parcel is stuck. It likely has something to do with the geometry near the location it provides. Check for sliver triangles, overlapping triangles, etc. - things that CONVERGE can run with but can sometimes cause problems. However, if you're not getting a large number of these warnings it should be safe to run with this warning.

2 - Recoveries indicate that the code wants to run at a lower timestep for stability. Plot num_recovers in time.out vs cycle number. How many times do you see a recovery? Are they clustered around an event occuring in the simulation? Valves/ports opening, spark firing, combustion, etc.

3 - This can be caused by taking large timesteps or poor initial conditions. Again, you can run through it, though it would be ideal to fix the root cause of it. How often are you seeing this warning and how long does it show up for?

- John

Originally Posted by Sahus

Hello Community,
I am getting many warnings in my IC engine simulation (GDI RANS, SAGE). Can anyone help me to get rid of these warnings? which parameter I should look in CONVERGE?

1. (Rank 1) WARNING: Stuck film parcel: ncyc=8111, RANK=1, location (3.902065e-02,-6.000000e-03,-3.442074e-02)

This warning starts 20 CAD after spray and lasts 20 CAD after ignition. I checked the locations and they are at the intersection of piston and liner. I tried compressing small area in piston but it didn't help.

2. recovering .... because transport equations did not converge or energy extrapolations

I tried changing relaxation parameter but again didn't get rid of these warning.

3. WARNING: max iterations exceeded on pstar, continuing for compressible simulation

increased max no of iterations for pressure to 700 from 500 default value, but it still doesn't converge.


Although I am still able to get results but still I would like to know the solutions for these.

Thanks

[Sahus]

Originally Posted by jetcheve

Hello!

Let me address these one by one.

1 - This literally means a film parcel is stuck. It likely has something to do with the geometry near the location it provides. Check for sliver triangles, overlapping triangles, etc. - things that CONVERGE can run with but can sometimes cause problems. However, if you're not getting a large number of these warnings it should be safe to run with this warning.

2 - Recoveries indicate that the code wants to run at a lower timestep for stability. Plot num_recovers in time.out vs cycle number. How many times do you see a recovery? Are they clustered around an event occuring in the simulation? Valves/ports opening, spark firing, combustion, etc.

3 - This can be caused by taking large timesteps or poor initial conditions. Again, you can run through it, though it would be ideal to fix the root cause of it. How often are you seeing this warning and how long does it show up for?

- John

Hi John,
thanks for the reply.

1. I have checked through diagnosis tab, there are no overlapping or any other geometrical issues. I don't know what does silver triangle means and how to check it. There are however small area triangles in the piston but the location of stuck parcel is at the intersection between piston and liner. The warnings are actually quite large in numbers, I guess once the parcel is stuck, it moves with piston till it gets evaporated/burned and it shows warning till then for every location it travels.

2. I use minimum timestep of 1e-8 s. It occurs during combustion event for around 30 CA duration. There are 50+ or so recoveries which I think is not good.

3. Again minimum timestep is 1e-8 s and I use default simulation time as well as solver parameters. This warning is also quite large in number there is no specific event it occurs. As far as initial condition is concerned, would it matter if go for multi-cycle simulation?

Thanks

[Qm260]

Hi,
Have these issue been solved as i encountered the same warnings in my simulation. If it has, please post some solutions. Many thanks
Kind regards,
Qiutong

[rvigneshwar]

Hi Saket and Tom,
1) The issue is due to triangulations (geometry), try understanding how the triangulation looks in the provided location at a given CA and modify accordingly. Can't comment on what needs to be changed unless we see the geometry, it's case-specific.
2 & 3) This might be due to the initial condition or only during specific stages in the simulation. This can be confirmed as John mentioned checking the recoveries vs CA from time.out to see how frequently the recoveries occur. If it's because of the initial condition in a multi-cycle simulation allow the case to run for 1-2 cycles completely so that the initial conditions are washed out, else modify the initial conditions such that the solver can handle better.
If you still facing the issues and not able to resolve them, please share the case setup with details on the case to support@convergecfd.com. We need to take a look at the case to identify the root cause.

[ad1314]

Originally Posted by rvigneshwar

2 & 3) This might be due to the initial condition or only during specific stages in the simulation. This can be confirmed as John mentioned checking the recoveries vs CA from time.out to see how frequently the recoveries occur. If it's because of the initial condition in a multi-cycle simulation allow the case to run for 1-2 cycles completely so that the initial conditions are washed out, else modify the initial conditions such that the solver can handle better.

same error showing throughout the simulation run. My simulation in version 3.0 i am getting "recovering because transport equation did not converge or energy extrapolations" after every three/four cycles. initially the error would show up when dt increased beyond 2e-4 and after running for some time it shows up then dt start to increate beyond 4e-4.

Since i am using the variable time stepping algorithm, does this mean when dt increase beyond certain value the solver has convergence issues and decrease dt? If so this should not affect accuracy of the result since max CFL limit is still maintained. So can we just ignore this warning if we are using variable time stepping algorithm? If not what parameters should I focus on. There are no problems with the mesh.