[sherif35]

I’m trying to calculate the drag and lift coefficient along the vertical axis of the cylinder and I’m using the following functions in the controlDict file (by changing CofR (0 0 -2.0) to the desired vertical location). However, I’m getting the same exact Cd and Cl (and forces for that matter) along the length for a 3-D calculations with turbulent flow.

How do I get the variation of forces (or Cd) along the cylinder in the vertical direction? Thank you.

forceCoeffs02
{
type forceCoeffs;
functionObjectLibs ( "libforces.so" );
patches (cylinder);
rhoName rhoInf;
rhoInf 1000;
UName U;
magUInf 1.0;
log on;
outputControl timeStep;
outputInterrval 1;
CofR (0 0 -2.0);
lRef 1.0;
Aref 10.0;
liftDir (0 1 0);
dragDir (1 0 0);
pitchAxis (0 0 1);
}

forceCoeffs03
{
type forceCoeffs;
functionObjectLibs ( "libforces.so" );
patches (cylinder);
rhoName rhoInf;
rhoInf 1000;
UName U;
magUInf 1.0;
log on;
outputControl timeStep;
outputInterrval 1;
CofR (0 0 -3.0);
lRef 1.0;
Aref 10.0;
liftDir (0 1 0);
dragDir (1 0 0);
pitchAxis (0 0 1);
}

[sherif35]

Hello all,

I figured out how to get the sectional CD and CL by adding the binData section to the forceCoeffs function in controlDict as follow:

forceCoeffs
{
type forceCoeffs;
functionObjectLibs ( "libforces.so" );
patches (cylinder);
rhoName rhoInf;
rhoInf 1000;
UName U;
magUInf 1.0;
log on;
outputControl timeStep;
outputInterrval 1;
CofR (0 0 -5.0);
lRef 1.0;
Aref 10.0;
liftDir (0 1 0);
dragDir (1 0 0);
pitchAxis (0 0 1);

binData
{
nBin 10;
direction (0 0 1);
cumulative no;
}
}


I'm getting two distinct groups of Cd curves based on the vertical location of the section.

The simulation is for classic flow over a cylinder with Re = 200, (the flow is 2-D) with the vertical distance of the domain is 10D divided into 24 cells of equal height along the axes of the cylinder.

Looking at the Cd curves, sections 0, 3, 6 and 9 (from the bottom to top) are all in one group with Cd around 0.172 (multiply by 10) vs. the remaining sections with Cd around 0.114.

For higher Re (3-D flow) same issue but with random signal.

I understand that sectional Cd might be different for different sections but having 2 distinct groups of curves doesn't look right to me.

Can anyone help explaining this issue? Any contribution to the discussion of this issue is welcomed, Thank you.

[sherif35]

The issue I explained above is resolved now, it’s simple geometry issue. Basically, I have divided the grid into 48 or 24 cells in the vertical direction. And for the purpose of calculating the sectional drag coefficients, I've divided the domain into 10 sections in the vertical direction.

When openFoam calculates the forces in each vertical section, it adds the forces from all the cells that their centers lie within this section vertically. Although, all the sections have the same delta z, some sections may include more cells than other sections. hence, higher forces in some vertical sections.

This explains why the 2 grids with 24 cells in the vertical direction have similar patterns regardless of the Reynolds number and the grids with 48 cells have different pattern. So I think this issue is resolved now.

Just wanted to share with others.