One question that everyone discovering open-source CFD using OpenFOAM can ask is:
“Is OpenFOAM reliable compared to commercial softwares?”
The quick answer to this question is:
“Of course yes!”
Maybe the first proof of that is that OpenFOAM is known to be used for years in many universities and commercial companies.
As a second proof, one can find numerous successful validation studies on the Internet, of small list of which can be found below:
As an additional proof, we propose here to make a quick OpenFOAM test on a 1m long NACA0015 profile and to compare results with experimental data found here:
http://smiller.sbyers.com/temp/AE510_03%20NACA%200015%20Airfoil.pdf
As usual, we used cfMesh to generate the mesh. We produced here a 2D cartesian mesh with a single refinement area enclosing the profile, coupled to small refinements on the leading and trailing edges. The final mesh had 65,000 cells.
Prismatic layer parameters were chosen so as to ensure acceptable y^{+} values for a flow of air at 80m/s.
To compute the flow in OpenFOAM, we chose the simpleFoam solver (for steady incompressible cases). We selected the most robust available discretization schemes (“upwind” first-order divergent schemes). We set a “freestream” boundary condition on the farfield patch. We also activated the k-omega-SST turbulent model with wall functions.
Below are some charts and plots from the case: AoA = 10deg.
Residuals
C_{L} Variation
Y^{+}
Velocity
Pressure
Comparison between C_{P} values along the chord obtained from OpenFOAM (circles) and experimental measurement (lines) can be visible in the following charts:
As we can see, the pressure distribution along the foil perfectly fits experimental results. Small irregularities inside the CFD data are due to the cartesian nature of the mesh. The vertical jumps near the trailing edge are due to the thickening of the trailing edge surface in the mesh, obtained by trimming the foil at x/c < 1.
Integrated values of C_{L} and C_{D} can be seen in charts below. C_{L} values from OpenFOAM (circles) are very close to experimental ones (line).
C_{D} value seems less accurate when looking at the line, but the black cross at AoA = 0deg, which indicates a more precise value of the drag measured in the wake (cf. pages 6 and 13 of the report), shows a very good agreement. Estimated drag values after correcting them by the ratio C_{D wake}/C_{D} are represented by gray crosses.
We could now see that OpenFOAM could give excellent results by using a simple mesh and robust schemes.
Source files for the present OpenFOAM case can be downloaded by clicking on the following link:
OpenFOAM is of course only one of the numerous open-source tools that I can use for you. So, what about these other tools: are they also reliable?
Once again, the response is “yes”, for two reasons:
In conclusion to this page, one can be sure that I will use for you only validated and maturated open-source tools so as to provide you high quality results in the best time.
The question is then:
“Why do people use commercial softwares instead of open-source CFD tools when they are reliable?”
The response is simple:
Nevertheless, having access to an open-source CFD expert allows you to get rid of complexity and to have customized support even with open-source tools.
Besides, open-source tools have valuable advantages:
In consequence, working with me is the guarantee that:
Tougeron CFD © 2017