Wind Turbine Pitch Control: Power Curve Optimization

The Two Regimes of a Wind Turbine

A modern horizontal-axis wind turbine operates in two fundamentally different regimes, and each requires a different control strategy:

Region II (below rated wind speed): Maximize energy capture. Let the rotor speed track the wind to maintain optimal tip-speed ratio.

Region III (above rated wind speed): Limit power output to rated capacity. Use pitch control to spill excess wind.

This regime switching is why wind turbine control is significantly more complex than a simple PID loop. The Twinsys simulation captures both regimes.

Wind Turbine Aerodynamics

The mechanical power extracted by a wind turbine:

P = ½ · ρ · A · Cp(λ, β) · v^3

ρ   = air density (1.225 kg/m³)
A   = rotor swept area = πR²
v   = wind speed (m/s)
Cp  = power coefficient (max ~0.59, Betz limit)
λ   = tip-speed ratio = ΩR/v
β   = blade pitch angle (°)

The power coefficient Cp is a nonlinear function of both tip-speed ratio λ and pitch angle β. At optimal λ (typically 7-9 for modern turbines), Cp peaks around 0.45-0.50 when pitch β = 0°.

The Cp-Lambda-Beta Surface

Cp
0.50|                           ****
0.45|                       ****    ****
0.40|                    ***            ***
0.35|                   *                  **
0.30|                 **                     **
0.25|               **                         **
0.20|              *                             *
0.15|            **                               **
0.10|          **                                   *
0.05|      ****                                      *
0.00|******
    └───────────────────────────────────────────────── λ
         2       4       6       8       10      12
                                 ↑ optimal λ ≈ 8 at β = 0°

As β increases (pitching blade to feather):
The entire Cp curve shifts down and flattens.
This is how Region III power limiting works.

Twinsys's Lookup Table block is one-dimensional, so the Cp surface is implemented as a Function block with λ and β as inputs. It evaluates a widely used empirical fit; the curve above is this formula at β = 0°:

Cp(λ, β) = 0.5176 · (116/λi − 0.4·β − 5) · e^(−21/λi) + 0.0068·λ

1/λi = 1/(λ + 0.08·β) − 0.035/(β³ + 1)

Region II: Maximum Power Point Tracking

In Region II, the target is to maintain λ = λopt by adjusting rotor speed. Since pitch is fixed at 0°, the only actuator is the generator torque command.

The optimal torque reference for maximum power tracking:

T_opt = K_opt · Ω²

K_opt = ½ρπR⁵ · Cp_max / λopt³

This is a speed-squared torque controller — it needs the rotor speed, but no wind speed measurement. In Twinsys: a Math Op block computes Ω² (Ω × Ω), a Gain block multiplies by K_opt, and this becomes the generator torque setpoint.

Region III: Collective Pitch Control

When wind speed exceeds rated (≈9 m/s for typical 2 MW turbines), power is limited by increasing blade pitch angle. The control architecture:

Power setpoint         Generator speed
P_rated                Ω_rated
    │                     │
    ▼                     ▼
┌─────────┐           ┌─────────┐
│ Power  │           │ Speed  │
│  PI    │           │  PI    │
└───┬───┘           └───┬───┘
        │ (select max)   │
        └─────▼─────┘
              │
       ┌──────────┐
       │  Pitch    │  β_cmd (0-90°)
       │  Actuator │───────► Rotor
       └──────────┘
       (rate limited
        ±10°/s)

The "select max" logic (Twinsys Switch or Min/Max block) activates pitch control only when the power or speed PI output demands pitch-up. In Region II, pitch stays at 0°.

Region Transition: The Critical Challenge

The transition between Region II and Region III is the most challenging point. At the transition wind speed (~9 m/s):

Region II torque control is still active
Pitch starts moving from 0° toward feathering
Generator speed is near rated

Poor transition management causes power spikes and mechanical stress. In Twinsys, model this by running a wind step from 8 m/s (Region II) to 12 m/s (Region III) and observe the transition:

 Power (MW)       Pitch Angle (°)

 2.5|              20|
 2.0|──────────    15|
 1.5|    **         10|         **
 1.0|  *            5|        *
 0.5|*              0|────────
    0  100 200s       0  100 200s

Wind speed steps 8→9→10→11→12 m/s

Blade Pitch Actuator Model

Pitch actuators are not instantaneous. They are electromechanical or hydraulic with limited rate (±10°/s) and limited range (0-90°). In Twinsys:

Rate Limiter block: limits dβ/dt to ±10°/s
Saturation block: clamps β to [0°, 90°]
First-order lag (Transfer Function): actuator bandwidth ~1 Hz

These actuator limits are critical for realistic simulation. Without rate limiting, the controller can demand instantaneous pitch changes that are physically impossible, leading to optimistic simulation results.

Tower Foreaft Vibration Coupling

An advanced consideration: blade pitch affects tower side-to-side and fore-aft vibration. Aggressive pitch control can excite tower modes. The tower foreaft frequency for a 2 MW turbine is typically ~0.3-0.5 Hz.

In Twinsys, model the tower as a mass–spring–damper from the Mass, Spring and Damper blocks, with the aerodynamic thrust applied through a Force Src block:

Mtower·ẍ + Ctower·ẋ + Ktower·x = Faero(t)

The aerodynamic thrust Faero couples the pitch controller to tower motion, creating a feedback loop that can cause oscillations if the pitch controller bandwidth overlaps with the tower natural frequency. This is why modern pitch controllers include a tower damping filter (notch filter around the tower frequency).

Conclusion

Wind turbine pitch control is a beautiful case study in regime-switching control, nonlinear aerodynamics, and actuator constraints. Twinsys captures all of these elements: the Cp surface, the Region II/III switching logic, the actuator rate and saturation limits, and the tower dynamics coupling.

Running this simulation before deploying a new control algorithm can prevent costly surprises. A pitch controller that looks perfect in a linear analysis may excite tower resonance or cause power transients at the region transition — effects only visible in a nonlinear time-domain simulation like Twinsys.

Try this: add turbulence to the wind input using a Sum block fed by a Random (white noise) block. Observe how the pitch controller handles stochastic wind gusts versus the smooth step inputs used in initial tuning.