[SJRC]

I am using ANSYS 2020R1 to simulate a laminar flow around a spherical obstacle. I have chosen an axisymmetric domain, so my sphere is represented as a semi-circle next to the symmetry axis x. Simulations are run using Fluent, and all the files (geometry, mesh, case and post) are managed from the Workbench. My simulations contain a scalar representing solute dissolved in the fluid.

After I run a simulation, I wish to compute the force exerted by the solute on the sphere. This force comes from solute-sphere interactions at a distance, and I calculate it by integrating a volumetric force f over my domain.

#######################################
Using Reports Definition in Fluent to calculate the integral

Sum(f * 2 * PI,['corps_surfacique'],Weight = 'Volume')

I get 6.7558829e-14 N
#######################################

#######################################
Using Function Calculator in CFD-Post

Function: volumeInt; Location: corps_surfacique; Variable: f

I get 6.91156e-14 N
#######################################

Now I know these values are quite close. But this small difference affects my end results significantly. The field function f depends on scalar and position, so I have tried doing some things to figure out the exact reason for my issue.

1) Integrating the scalar
Results are the same in Fluent and CFD-Post

2) Integrating abs(x) (x is azimuthal coordinate)
Results differ by only 0.0084%

3) Integrating abs(y) (y is radial coordinate)
Results differ by 0.45%

My guess from these outcomes is that mesh elements in Fluent and CFD-Post have slightly different coordinates. Is that correct? In any case, which option should give me the most accurate result: Fluent or CFD-Post?