[Springyuu]

Hi, when I was processing data in CFD-POST, I found that the minimum temperatures obtained using two methods on the same location were different. A minimum temperature is obtained in the contour plot, T_min1=2.58 K, shown in fig.1. Another minimum temperature is obtained in the function calculator, T_min2=1.96 K, shown in fig.2. what's the difference between the two minimum values? and which one should I choose as the minimum value? please give me some advice, thanks a lot.
fig.1.JPG
fig.2.JPG

[LuckyTran]

The 1st uses temperatures interpolated onto your surfaces selected


The 2nd minVal/maxVal uses the cell centered values adjacent to the boundary selected. So these are min and max values don't correspond to the actual boundary.

[Springyuu]

Originally Posted by LuckyTran

The 1st uses temperatures interpolated onto your surfaces selected


The 2nd minVal/maxVal uses the cell centered values adjacent to the boundary selected. So these are min and max values don't correspond to the actual boundary.

Thank you very much! I understand it now.
Specifically, the 1st uses the vertex values to interpolate onto surfaces selected, the 2nd minVal uses the cell/face centered values. When choosing face as the variable location in the contour, the minimum value is the same as fig.2, T_min3=1.96 K, shown in fig.3.
Some relevant information found in the help document also provides more detailed explanations, shown in fig.4.
But, why the 2nd minVal don't correspond to the actual boundary? In other words, is it more reasonable to use 'Vertex' instead of 'Face' as the variable location to determine the min or max values in practice?
fig.1.JPG
fig.2.JPG
fig.3.JPG
fig.4.JPG

[LuckyTran]

minVal and maxVal are functions designed to return values actually solved by the solver (and stored on the mesh), it is the minimum and maximum of the actual solution without any additional interpolation and extrapolation and "interpretation."

CFD solvers never need boundary values, they only need boundary fluxes because is solves transport equations. Hence, the boundary values you are looking for don't actually exist and they can only be obtained by additional interpretation.

It's up to you how you want to interpret an interpreted result via outside knowledge from your process. We call it a best practice because it is actually based on practice (i.e. you go and practice it and find out what is the best for your practice). You might want to use max vertex values, max face values, or the maximum of the maxima (i.e. the infimum). Best practices will have a methodology and then throw a safety factor on top of it and these are determine from practice in a industrial setting, you don't ask some rando on the internet for these safety factors, you ask your boss or the company president.

[Springyuu]

Thanks, you're so nice!