Optimizing a Tail for Low Drag: Part 4

Curvature

Now that I know from my first round of testing which panel orientations have attached flow, how the pressure behaves with changes in angle, and a rough prediction of pressure drag reductions from all that information, I’ll move on to a larger test buck that will start to approximate a full tail. This buck is almost as long as my maximum length requirement and, rather than a flat panel like my first board approximations, has some curvature in it. Specifically, the extension here bends from an angle of about 20° from horizontal at its front to 23° at its trailing edge, in between the shallowest and middle angles I tuft- and pressure-tested:

“Conventional wisdom” says to bend the tail in a convex curve like this for lowest drag. But is that “wisdom” correct? Only testing will answer that. I’ll test this buck in this configuration as well as with an added spoiler that, when affixed to the tail, is horizontal:

More fun with the miter saw.

All in, I’m still sitting at $0 invested in this project since I’ve been using exclusively materials on hand. Some of these are leftovers from other projects, such as my stability testing fins, which I cut apart to use as the side plates here:

I also repurposed the dowel support brace from that project, to stiffen the trailing edge of this buck:

With everything screwed together, the tail is now stiff enough to shake the whole car if I push on it.
 
Tuft Test
 
I’m hoping that the moderate angle and gentle curvature of this test buck will ensure attached flow, and that curving the end of the tail back up will increase pressure over the sloped surface without causing any separation. To check this, I first ran a tuft test using the bare tail, flat spoiler, and then a bit of leftover lip spoiler from previous testing.
 
These tests were conducted on a N-S road (near the UIUC Energy Farm), and winds were out of the SW at around 10 mph.

As you can see, the flow stays attached in both directions, with only a few tufts in the middle of the board starting to lift as reverse curvature is added. Hopefully this means the pressure is increasing there as well.
 
I also taped some tufts to the side plates, just to see if there was any flow attachment there. Not really—some images showed reattachment at the end of the plate, but for the most part they were in separated flow. I was curious if I could leave the taillights open, but it looks like that isn’t the case; they’ll need clear fairings if I want them to be visible with a full tail.
 
Pressure Test
 
On another day, I measured pressures along the centerline of the tail board and car body. I used the same road as my previous pressure test; this day, winds were out of the SSW at 15 mph, so a bit windier than before.
 
First up: the curved tail board without versus with spoiler, southbound.

Position	No Spoiler	Spoiler Added	Difference
Rear	0 Pa	+10 Pa	+10 Pa
Center	-20 Pa	0 Pa	+20 Pa
Front	-60 Pa	-40 Pa	+ 20 Pa
Backlight	-30 Pa	-30 Pa	0 Pa

Next, the tail board without versus with spoiler, northbound.

Position	No Spoiler	Spoiler Added	Difference
Rear	-10 Pa	0 Pa	+10 Pa
Center	-20 Pa	-10 Pa	+10 Pa
Front	-30 Pa	-30 Pa	0 Pa
Backlight	-20 Pa	-20 Pa	0 Pa

And finally, pressures on the base, with and without the tail board, southbound and then northbound.
 

Position	No Tail	Tail	Difference
Upper	-20 Pa	0 Pa	+20 Pa
Center	-20 Pa	-10 Pa	+10 Pa
Lower	-30 Pa	-20 Pa	+10 Pa

 

Position	No Tail	Tail	Difference
Upper	-10 Pa	-10 Pa	0 Pa
Center	-20 Pa	-10 Pa	+10 Pa
Lower	-20 Pa	-10 Pa	+10 Pa

Adding the spoiler, it turns out, does increase the pressure acting on the top surface of the tail—so this is something I will want to incorporate into my final design. The reason this happens is because of the change in direction of the airflow as it follows the surface; the change in direction of the spoiler changes the momentum of the airflow (“bending” it up toward horizontal from its downward direction before; remember, momentum is a vector—that is, it has both magnitude and direction), resulting in a force in the opposite direction (down) acting on the body. This force acts on the surface in part through pressure*, which is measurable as an increase in pressure with the spoiler compared to without it. Because of the orientation of the panel, inclined from horizontal, that pressure increase reduces drag.

(*Aerodynamic force is exerted by air on a body through only two mechanisms: tangential stress, which acts along the body surface, and normal stress, which acts perpendicular to the body surface and which we call pressure).
 
This partial tail reduces drag in another way, by raising the pressure acting on the base of the car. I alluded to this in a comment on another post, but here it is verified with a real example. Even lower down, in the center of the rear bumper cover, measured pressure was higher with the tail board on the car than without.

This is something I plan to explore further. If I can get a significant drag reduction from just a Bonneville-style spoiler, that might be a better option than a full tail. It would be cheaper, easier to construct, require no license plate relocation or additional lighting, and allow continued use of the hatch. Always leave yourself open to exploring other options like this, especially if you discover something surprising during testing. There's no rule that you have to stay locked in to whatever device you expected to build going into a project, and you may find that some other solution is almost as good and mitigates some drawbacks. That's one form of optimization.

Previous: Optimizing a Tail for Low Drag: Part 3