So, after a afternoon of testing with the boat-tail:
SCx shows an visible improvement, of around 0.1 to 0.15, which is not quite bad (around 15% drag reduction). Visible on both back and forth runs as expected.
red plots are without boat tail, and black ones with boat-tail installed. Abscissa is number of samples at target speed (90km/h)
So there is improvement, but not as much as expected.
Tuft testing reveals that flow is properly attached throughout the top side of the boat tail (12° angle), but on the sides this is not as good, and on the bottom side flow is completely detached, with strong recirculation bubble (tufts angled toward front)
tuft test videos (remove space between https and :// after pasting) are here :
https ://drive.google.com/drive/u/0/folders/145GXU43Vr1g5UUUGFSZIjnWNE0dYWV1b
I redid the tuft testing after having installed the undertray and the rear wheel cover (assuming the undertray would provide more stable flow undercar, and same for rear wheel cover on the side).
However the tufts remained the same.
For the car side, I had checked previously that flow are conveniently attached up to the rear of the car.
So most probable cause is:
- for underside, the undertray is not efficient and/or the boat tail angle is too important (it is around 15°, while literature consensus is more around 7°). Problem of the undertray is that the transverse bar is so low that if I want the diffuser to go from the bottom of that bar up to the rear bumper it requires an angle of 15° for the diffuser, which is too large according to literature.
- for side, maybe the slant angle is too large (I think it is around 20° in average). there is not recirculation (tufts general direction is aftwards), but it is visible that the flow is perturbated by something.
I will attempt with some friends to add smoke in the boundary layer to see how it develops over the car.