[Numericer]

I've recently started learning OpenFOAM. As a source of practice, I thought I'll just simulate the example problems in textbooks (ex. Fersiger and Peric).



The first example is a demonstration of finite difference method applied to a canonical convection-diffusion problem, which is as follows. In a 1D geometry, phi (scalar) has a value of zero at the inlet and 1 at the outlet. What is its distribution across the interior, if the flow speed is a given constant of 1 m/s, diffusivity is 0.02 m^2/s and length of geometry is 1 m?


I've solved the problem as mentioned in the textbook. Two schemes, upwind and central difference, are compared against each other and against the exact solution. I've gotten the exact results as the authors.



I wanted to try this simple problem in OpenFOAM. So I cloned the pitzdaly case and made the necessary adjustments to set the case up. For the central difference scheme, I chose 'Gauss linear' as the divergence scheme, and the resulting 'plot over line' matches exactly. But strange things are happening with 'Gauss upwind' (or vanLeer too).


As is known, the upwind is a TVD scheme and as such is bounded albeit a lower order scheme. But the result of the simulation show oscillation as can be seen in these links: https://ibb.co/qg0M1kV and https://ibb.co/jDb6syJ.

Can anyone explain if I'm doing something wrong or what is OpenFOAM doing?

[Numericer]

Can anyone reply to this thread?