The 'shorting' currently
is your brake.
You gotta get that thing tied down before something besides "nicely on the ground" is found in the description of the aftermath.
As far as charging goes, yes, the mill will need to reach a Voc (open circuit voltage) that is higher than the battery's OTV (colloquially the same as Voc) PLUS the drop in the blocking diode(s).
So, if you have yer standard everyday silicon diode, there is a typical voltage drop of 0.6-0.9V (depending on current, usually "counted" as the nominal value, or about 0.7V) across the diode.
Let's assume your 12V (nominal) battery is resting at 12.6V and you need to charge it.
For this to happen, your genny needs to hit (approx.) 13.3V just to
begin charging. But at that point, the current flow will be nil, because the genny+diode combination is only
matching the battery's voltage.
To actually charge, the mill needs to increase speed from this point (called "cut-in").
This is where the limitations of the motor/blade combination will become apparent. A rotor that can spin fast enough to get above cut-in for your particular motor needs to be small, so that the centrifugal forces at the blade tips don't rip it apart at operating speeds. The problem with a smaller rotor is, it can't grab as much wind. Less wind to grab = less power to harvest.
Steve
EDIT - Question - What did you have it loaded with when you were measuring the 18V @ 3A?