The speed dependent efficiency calculation here includes the factor
Math.cos(rotorSpeed / (45.5 * Math.PI)).
That makes no sense. Did you mean to write
Math.cos(rotorSpeed / 450 * Math.PI)?

The calculation constants are picked so the wave cycles around 900rpm.

The period of the function is given by 2π/(1/(45.5π)), which simplifies to 2π∗45.5π=91π^2 = 896.79.

Your correction of using 450 would mean the peak rpm is around 8882rpm.

Um no. Using
Math.cos(rotorSpeed / (45.5 * Math.PI))
means as you say, that the maxima are at multiples of 91π², which is irrational.
Using Math.cos(rotorSpeed / 450 * Math.PI)
would mean that the maxima are at exactly multiples of 2π/(π/450)=2*450=900.

Thats why I consider it illogical. Why have the maxima at random irrational points when even the ingame documentation specifies nice round intergers.