Since everyone leaned how to calculate rotor temperature rise per second in a single panic stop in previous posts, calculating the convection cooling vs speed per second should be almost as easy.
"Rotor rotational speeds of 342, 684, and 1025 [rpm] yielded fin convection heat transfer coefficients of 27.0, 52.7, 78.3 [W m−2 K−1], respectively, indicating a linear relationship. At the slowest speed, the internal convection represented 45.5% of the total heat transfer, increasing to 55.4% at 1025 [rpm]."
https://ir.library.dc-uoit.ca/bitstream ... thesis.pdf
Obviously the faster the velocity the better ducts and deflectors work, especially when you are only down to 100 mph in the apex of a curve