Oscillating force plots

Hello there
I was simulating an incompressible flow over a flat plate and although my mesh looked good, the residuals didn’t converge and my force plot were oscillatory. I have attached a link to my workbench below. Could you please look at it and give your opinion. Some extra details-

-the plate has a chord of 15cm and span of 30cm

  • the flow is incompressible and the velocity is 15m/s

  • I have incorporated an angle of attack of 9 degrees in the geometry

  • I was calculating the coefficient of lift from pressure force in z and coefficient of drag from pressure force in y

  • to neglect the effect of the tip vortices, I have restrained the bounding box in x direction

-I want to obtain the Cl and Cd value for this setup, the ideal values for which are 0.82 and 0.14 respectively(for reference)

The force plot, residuals and mesh pictures have been attached below;

link to my workbench-

Congratulations, this is a very well set up project!

I think that your residuals are converged as well as you can expect for this mostly separated flow simulation run.

In the future you might reach convergence quicker with Simulation Control>Potential Foam Initialization = ON

In actuality, your oscillating Forces and Moments Results are oscillating around quite stable values.

In the case of your Pressure Force Z, the value it is oscillating around is -4.815 N and which is 1% stable over 1500 iterations using a 1500 iteration Moving Average filter. This was determine by my ORSI program which you can read about and download here.

Here is an image to show the above program in operation for your Simulation Run results data:

I am sure that you realize this, but the accuracy of your results will be significantly improved if you add boundary layering to your mesh.

By the way, your full solution set download for this sim run was 2.5GB. Since this is a steady state sim run you could lower this 2.5BG download to about 1GB without losing any significant data. If you change your Simulation Control ‘Write Interval’ so that it equals your ‘End Time’, only results for Time 0 and End Time will be saved (~1GB download). Then when you download and unzip the 1GB data for use in ParaView, you can further save 0.5GB of disk space by deleting the Time 0 folder.

Since you have reference values for CL and CD, you should properly setup a ‘Forces and Moments Coefficients’ results item and then we can watch as you try to verify the experimental values with your CFD project :slight_smile:


Hey Dale!
Thanks for the valuable input firstly. I tried adding inflation in the mesh and with the default settings for inflation I got an improvement in the force plots, although, the value which its oscillating about changed from what you got. My concern now is that I am unable to get inflation in the mesh if I change the default settings for inflation. The moment I change the settings, the boundary layer doesn’t form which I know from the meshing log. I’m once again attaching a link to my project below. Mesh 1 was the original mesh. Copy of mesh 1 has inflation with default settings. Copy of copy of mesh 1 has changed settings in inflation where the boundary layer isn’t forming. Please share your opinion.
Much thanks!