A direct conversion would use 132 turns of #17. usually you can manage with a bit less than double the turns as the diode drop is insignificant at 48v and the line loss will be 1/4
As you seem to be in a high wind area again you can manage with less turns, too low a cut in speed in a good wind area will murder your top end performance. Rather than increase air gap you will be better of with thicker wire and less turns and keep the gap smaller.
I suspect your 110 turns will be a good compromise.
From the pictures of stators I have seen there seems to be a fair bit of space not used on the otherpower stators. I would make the coils with nearer triangular holes as the coils only get close at one point. I think I would stick to the 110 turns and cram in the thickest wire that will fit with the coils touching.
You may have to trim the air gap and possibly the line resistance to get things out of stall but with the reduced turns that should be less of a problem.
You may need to check the furling I believe a lot of the control comes from stall rather than true furling, if you make it more lively the furling characteristics will be different. Check against a fixed reference that the prop is turning from the wind and the tail is not coming towards the prop with the prop staying in the wind.
If it furls properly it will drop power when the tail goes beyond 50 deg, if power is still rising it will be wind seeking and relying on stall.
Flux