Comparing VSPaero to NACA Fowler Flap Data, TN-808

[Dennis Haimerl]

My first objective with using OpenVSP is to obtain a better approximation of how close the variable incidence horizontal stabilizer used on two specific experimental STOL airplanes are to experiencing a tail stall when in an adverse trim condition with flaps fully extended. I’m trying to find a method that will yield better results than NACA Report 648 without resorting to a complex CFD model or somewhat dangerous flight test scenarios. Since the downwash angle is directly effected by the wing lift coefficient, I am interested in determining how reasonable VSPaero can predict the lift coefficient with flaps deflected.

Square one has been to determine the validity of using VSPaero to estimate the lift a Fowler flap produces by comparing results to NACA TN-808 that utilizes a NACA 23012 section in conjunction with a Fowler flap. I first started by comparing section results obtained using the USAF DATCOM used when developing the airplanes and comparing them to data in TN-808. The cl_max and alpha_cl_max derived using DATCOM is a little low compared to TN-808, which in my experience is not too unusual for these type of airplanes.

Secondly, I applied the section data from both the DATCOM method and that reported in TN-808 to a wing with full span flaps and an aspect ratio of 20 to reduce the amount of 3D effect for the comparison (using equivalent Reynolds Numbers). As expected, the data using TN-808 out performed the DATCOM method.

I then executed VSPaero’s Panel Method using the geometry specified in TN-808, while assigning 2D cl_max values for each flap setting obtained from the DATCOM calculations. The results from flaps fully retracted to a flap extension of 20 deg. are in reasonable agreement. The 30 deg. deflection results were a little high and there was no obvious sign of the 2d cl_max having been accounted for at flap deflections at 30 deg. and above. The lift curve of the 40 degree deflection was significantly above what would be expected, so I also included a 35 degree deflection curve. The difference between the VSPaero lift curves (delta cl) of 35 and 40 are similar to that between the 30 and 40 degree deflections using the DATCOM method, though at significantly larger cl values.

Finallly I executed the Vortex Lattice Method. The flaps fully retracted condition again was reasonable, but the 10, 20 and 30 deg. deflections exhibited somewhat lower than expected cl values.

Wind tunnel data, as well as my past experience with flight tests show that any deflection of a Fowler flap much past 40 degrees will cause major flow separation on the upper surface of the flap. This could explain the disparity at the 40 deg. deflection setting, but it doesn’t explain the large delta_cl at 35 degrees.

I am thinking I might evaluate the downwash at the tail using the Panel Method for flap setting up to 20 deg. Then us the VLM for 30 deg. and the VLM at 35 deg. to represent the 40 deg. deflection. 2D cl_max values similar to those in TN-808 will be applied to all cases. Once I evaluated the actual wings using the proper airfoil sections (one similar to the NACA 23000 series and the other being a highly modified variant of a NASA MS section) along with the actual flap span ratios, I may change my mind.

Lift curve plots are attached below. Comments would be appreciated.

[Brandon Litherland]

The first thing I would look at is the Load Dist. plots for these cases and see if the set 2D Cl_max is being captured appropriately.  In the Results Manager, you should see the local Cl smash up against that limit for most of the wing, particularly when flaps are applied in VLM mode.  What I suspect is happening, and I've seen before in VSPAERO, is that the stall model using local Cl limiting cannot account for the losses in flap efficacy at higher deflection and alpha due to separation, etc.  This is also why you'll see VSPAERO polars plateau rather than drop at stall.  The VLM control surface model enacts a superposition of velocities to account for the flap/aileron deflection.  I've not gone digging into the source code to determine where the maximum effective deflection is determined (if at all) but taken to the extreme you will see diminishing returns from the flap.  Try setting a VLM flap to 80° and you'll see what I mean.

If you slice a VSPAERO solution, you can see the Cp and velocity vectors by selecting your slice in the Cut Planes tab of Viewer.  This will give you an idea of what the velocities and pressures are doing for a given case.  If you look at the trailing wakes, you should find that the VLM or Panel flaps impart inboard and outboard vortices from the wing, as expected, so you should be able to capture the effects on the tail.