Discussion Forum

Hi,

To cut the long story short, I am modelling an airfoil based on NACA 4418 profile.

My simulation set-up as follows:
Wind speed: 5m/s, 10m/s, 15m/s and 20m/s
Angle of Attack: 0, 5, 10, and 15 deg
Chord length: 5meters

By the way, I am using k-eps turbulence model due to Reynolds number of the order 10^6.

First, I simulated the model in 2D (only chord length is considered in geometry-i.e. cross section) for verification purposes with the book entitled “Theory of Wing Sections - including a summary of airfoil data” by IRA H. Abbott and Albert E. Von Doenhoff, 1959. For this 2D simulation, I used operator (in the expression under Line Integration) "p*ny" for Lift Force and "p*nx" for Drag Force. From here, I then manually calculated the Lift and Drag Coefficients in Excel and the results are all agreed with the book as mentioned above.

Once verified, I then extrude my airfoil (of 5meter chord length) to 1m, 5m, 10m, 20m and 50m in spanwise direction. No taper and no twist involved. It is just a plain extrusion from 2D to 3D geometrically. So now, it will looks like rectangular in planform view. This extruded (with 5meter chord length) geometry is inserted into air domain (looks like a box) whereby each side of it attached to the side walls as to ameliorate the effects of vortices. So that the air can only flow around the airfoil and not longitudinally across it. In other words, there are side wall attached to this extruded model. This is to replicate a wind tunnel test i.e. 2D flow field. Constructed model (solution cleared) in 4.3b version is attached for your perusal. I attached only for 10m extrusion due to similar set-up with other extrusion lengths.

According to COMSOL Support team, for this 10meter extrusion, they have suggested to either use the operator "spf. T_stressy" (i.e. Total Stress in y component) or "p*ny" for Lift Force and either "spf.T_stressx" (Total Stress in x component) or "p*nx" for Drag Force. From these, I then manually calculated its Lift and Drag Coefficients in Excel. Both operators yielded similar results for Lift and Drag Coefficients.

Furthermore, when I made a comparison between 10m extrusion and 2D results, results from Lift Coefficient of 10 meter extrusion agreed with the 2D results. However, results from Drag Coefficient (of 10 meter extrusion) seems to be 10 times larger (in terms of its decimal points) than the 2D results of Drag Coefficient.

For your information, this happened to all extrusion lengths (i.e. 1m, 5m, 10m, 20m, and 50m). In other words, all Drag Coefficients experienced 10 times larger than the 2D results of Drag Coefficients. Theoretically, since these extruded models attached to sidewalls on both sides, the results of Lift and Drag Coefficients should be the same with 2D results.

Based on your experience, could you please advise/suggest me where/which part that I did wrong?.

By the way, should you require more information as to help you understand the model, please contact me.

Thank you,

Regards,
Ahmad