Spatial order of 1 instead of 2
 

Dear everyone,

I am simulating a 2D low reynolds (0.01) flow over a wavy sinusoidal floor, with moving top plate and left/right periodic boundary conditions using the icofoam solver. I tested my solutions on various meshes to test the spatial order of the solution method. Applying Richardson extrapolation resulted in a spatial order of 1, whereas I expected 2 based on the numerical schemes used for the spatial derivatives (see attachment of my fvSchemes): I used linear interpolations and corrected Gauss schemes. Furthermore, I used polyline for the sinusoidal boundaries.

I cannot explain the difference between the emperical and the expected order of accuracy. Do you have any experience with this problem?

Thank you guys!