This is the same radar-curtailment model as the 3D version, drawn as a top-down radar screen instead of a rendered scene. Each turbine sweeps a rotor disc of diameter D at angular speed Ω (rad/s). A bird crossing that disc is struck with probability from a simplified Band-model strike risk:
P_strike = min( 1, σ · Ω · D / (2 · v_bird) )
σ = blade solidity ≈ 0.045
Ω = 2π·RPM / 60, D = rotor diameter
Slowing Ω toward idle drives P_strike toward zero almost linearly — turbines don't have to stop, they only have to spin slow enough that a crossing bird has time to pass between blade sweeps.
The radar tower tracks every bird inside its detection radius and extrapolates its straight-line heading. For a bird at position P with velocity v and a turbine at T, the predicted closest approach is:
t* = (T − P)·v / (v·v)
closest distance = | P + v·t* − T |
If that predicted distance is inside the rotor radius and t* falls within the next few seconds, curtailment ramps that one turbine's target speed down to idle — the rest of the farm keeps running at full output. Delivered power scales with the cube of the speed ratio, P ≈ P_rated·(Ω/Ω_rated)³, which is why even a partial slow-down costs real energy — the loss percentage you see trades directly against the strike counter.
- Because this is a flat radar-screen view, every bird is treated as flying at rotor hub height — the 3D version also gates a strike on a height band around the hub, which this simplifies away since there's no vertical axis to draw here.
- Drag the screen to pan, scroll/pinch to zoom.
- Turn curtailment off to watch the same migration pass through at full rotor speed — the strike counter starts climbing instead of the loss percentage.