[diamondx]

I need help regarding refinement. It may seems long to read but it's easy to understand. Thanks for reading this:

I'm performing an analysis on a supersonic air intake, and i need to do an adaptive mesh refinement. As i have strong shock waves, the tutorial in fluent suggests me to select a gradient of static pressure, then in the normalization, select scale and 0.3 for coarsen threshold and 0.7 for refine threshold. I couldn't understand what's the meaning of these values ?
After reading another tutorial about the elbow, i noticed that they use the adaptive mesh refinement and they selected the standard normalization. then they assigned 10% of the maximum static temperature. What's the meaning of it again, does it mean that let's say this value is 100 degrees, fluent is gonna refine node where value is greater than 100 ?
I went with the first option and selected dynamic , so my cluster can refine after each 200 iterations. i got this:

Adapting mesh (Adapt Gradient of pressure)...
%mark-with-gradients:
According to Min/Max # of Cells to many/few cells marked for refinement.
Coarsen/Refine Threshold automatically re-adjusted on 2.89615/186.09221


%mark-with-gradients:
According to Min/Max # of Cells to many/few cells marked for refinement.
Coarsen/Refine Threshold automatically re-adjusted on 2.89615/186.09221

Is it working or is it a kind of error.

As you can see I'm interested to refine where i painted in red in the above image. I'm willing to select a gradient of mach number but i don't know what to mention in the threshold value.

thanks a lot for your help.

[Ananthakrishnan]

Hi,

As you have already understood, Fluent just splits the cell into four cells (non conformal meshes) wherever it witnesses a value greater than the threshold value.

You can know the min and max value of your physical variable of choice by clicking compute in adapt-gradient-compute.
You can know first hand about the number of cells above the threshold value through adapt-gradient-mark. Do not use coarsen as it is the opposite of refine!! its just to make your mesh lighter..uncheck the coarsen check box.
It is generally advised to keep the number of cells getting adapted around 30(not more than that), you can do this by trial and error by changing the refinement threshold and marking the cells each time. If the cell number >>>30, there might be drastic changes in cell values between successive calculations which might or might not be problem.
You can either "adapt" immediately or just "apply" and start you iterations. So when ever fluent encounters cells with higher values than threshold it will refine them.

You can visibly check the refinement by plotting a contour from time to time. It will be getting denser and denser in the areas of refinement.

I generally use the option adapt-gradient-"method"gradient.

I dont think you can decide the area where you can refine, fluent automatically decides that through the physical variable and its threshold value you had chosen.

Just out of curiosity what is that you are looking for from adaptive mesh refinement.

[diamondx]

i'm trying to get a better resolution around shock waves and also help fluent converge.
the gradient used for adaption was static pressure, then i used scale and values of 0.3 and 0.7 in the threshold.

[Ananthakrishnan]

of course you are correct in selecting the static pressure, but you need to make a smart choice of the threshold depending on your current range of values, if not its a problem