Hi

I simulate the time varying inlet velocity based on 2D model, the attack angle is 0, the model is square cylinder, when I monitor the Drag force, the force has negative value, in fact the drag force should be positive, so can anyone help me, thanks

Hi,

Quick, patent it. I think you have found a perpetual motion machine.

Or maybe you should check your simulation is properly converged, the physics correct and the mesh adequate.

Glenn Horrocks

hehe

well, the force cfx spits out is the force acting FROM the FLUID ON the WALL. you have to watch out in what direction the forces are acting.

for example: you have an airfoil with an upper and lower surface. the y-force on the upper surface is -35N and the y-force on the lower surface is +40N. The resulting y-forcing lifting the airfoil hence 5N.

hope I could help

in fact, i simulate the 2D square cylinder, and the drag force is whole force around the square cylinder

hi Glenn

how can I check the convergence, I use the coefficient loop 10, when i run solver, i saw the output windows, which shows the coeffcient loop is between 3-7, you mean mesh adequate, that means i need more fine mesh around the square cylinder, is that all right

Hi,

You need to do basic validation checks to ensure your simulation is accurate. If you are getting negative drag forces you can be certain it is not accurate.

Check whether your convergence is tight enough, your timesteps small enough (if transient) and the flow is mesh independant. Also check that the flow physics (turbulence model, any multiphase/chemistry/heat transfer stuff) is correctly set up.

Glenn Horrocks

Hi Glenn

thanks your answer, I check the flow model, cause i simualte the flow around the 2D square cyliner, when I did not consider the inlet time varying velocity boudary condition, which means I use the mean value, the result is right, but when I use the time varying inlet, it is not right, my domain is 24m(H)*34m(L), and the min grid size is 0.1m*0.1m, the cylinder is 2m*2m. you mention checking whether the convergence is tight, how can I make sure whether my simulation convergence is tight. casue I use the unsteady simulation, the RMS courant number is 0.12, and the max courant number 0.36 during the whole simualtion,my simulation time step is 0.01s, total time 180s, can you help me again, thanks a lot

Hi,

The way you check sensitivity to any parameter (mesh size, convergence tolerance, timestep size etc) is to run one model, refine the parameter a significant amount (say halve the mesh element length) and rerun. If the two results are the same within a tolerance you are happy with then you have established your simulation is independant of that parameter.

This needs to be done for all important parameters.

Glenn Horrocks