[kahjoonyong]

Hello,

I am currently running a simulation using AMR. Previously I had a problem with the subgrid scale criteria for velocity and temperature to trigger the AMR which is related to this post: How to apply AMR settings

Yijin Mao had advised me to include 2 outputs as passives which are vel_sgs and temp_sgs to determine the subgrid scale variables to be applied in the AMR settings.

Recently I found out that the simulation results varies as I decrease the subgrid scale temperature from for instance 0.1K to 0.02K. When I checked the temp_sgs output, the mean value (TEMP_SGS_MEAN) is around 1E-4, whereas the value for TEMP_SGS is around 5E-7, and the TEMP_SGS_StdDev is around 5E-3.

This makes me wonder what those values actually mean and how can I make you of those values to determine the AMR settings. More importantly why do those 3 values vary from each other so differently.

Any clarification is very appreciated. Thank you very much.

Regards,

Kah Joon Yong

[ymao]

Hi Kah Joon Yong,

The physical meaning of sub-grid temperature is an error between actual temperature and resolved temperature at each cell. Converge calculates this value as its simulation. If this error is larger than user specified value, Converge refines the cell automatically. There is a time if the total cell count has already reached to the maximum cells. The cell refinement operation will NOT proceed. If the error is smaller than 20% of user-specified value, embedding on this cell will be released. By giving different user-specified value, the results can change as the error introduced to the simulations system varies.

Going to back to the question on the judgement of amr setting, comparing temp_sgs with the criteria (user-specified value) will give an idea whether the domain is well resolved or not.

These 3 values could vary, as TEMP_SGS_MEAN is mean value of temp_sgs at each cell, TEMP_SGS is instantaneous value at each cell, and TEMP_SGS_StdDev represents the deviation of temp_SGS at each cell. These 3 values can be very different.

Originally Posted by kahjoonyong

Hello,

I am currently running a simulation using AMR. Previously I had a problem with the subgrid scale criteria for velocity and temperature to trigger the AMR which is related to this post: How to apply AMR settings

Yijin Mao had advised me to include 2 outputs as passives which are vel_sgs and temp_sgs to determine the subgrid scale variables to be applied in the AMR settings.

Recently I found out that the simulation results varies as I decrease the subgrid scale temperature from for instance 0.1K to 0.02K. When I checked the temp_sgs output, the mean value (TEMP_SGS_MEAN) is around 1E-4, whereas the value for TEMP_SGS is around 5E-7, and the TEMP_SGS_StdDev is around 5E-3.

This makes me wonder what those values actually mean and how can I make you of those values to determine the AMR settings. More importantly why do those 3 values vary from each other so differently.

Any clarification is very appreciated. Thank you very much.

Regards,

Kah Joon Yong

[kahjoonyong]

Hi Yijin Mao,

Thanks for the reply.

Going to back to the question on the judgement of amr setting, comparing temp_sgs with the criteria (user-specified value) will give an idea whether the domain is well resolved or not.

Which value should I compare my user-specified value to? Or better put how can I utilize the output TEMP_SGS value to evaluate my settings? The smaller the value, the better the results? If it is possible can you please include a simple example?

Regards,
Kah Joon Yong

[ymao]

The value user specified is the error criterion (Sub-grid criterion). Users should decide the error they can accept. The temp_sgs is cell-wise value in post*.out file, you can compare each cell's temp_sgs with the criterion you specified and decide whether each cell meets the criterion.

Theoretically, the smaller the better, but it will need more computing resources to involve as the criterion decreases.

Regarding the example case, I would suggest you open any of example cases available in Studio, for example, "File > Load example case > Internal_Combustion_Engines > Gasoline_spark_ignition_PFI > SI8_engine_PFI_SAGE"

Originally Posted by kahjoonyong

Hi Yijin Mao,

Thanks for the reply.



Which value should I compare my user-specified value to? Or better put how can I utilize the output TEMP_SGS value to evaluate my settings? The smaller the value, the better the results? If it is possible can you please include a simple example?

Regards,
Kah Joon Yong

[kahjoonyong]

Thank you for your reply. Now I understands.