Wind Turbine Control Systems - Wind Turbine Technology: Fundamental Concepts of Wind Turbine Engineering - page 746

wind disturbances, accounting for turbine flexibility, and mitigating loads and deflections.
When additional control objectives must be met, additional SISO control loops must be add-
ed to the classical controllers. These additional loops can destabilize the machine if they are
not very carefully designed.
Modern control designs using state-space methods more adequately address these issues,
because the controller uses a model to determine system states. Controllers can be designed
not only to maximize power and regulate speed but also to add damping to important flexible
modes. Integrating all the available turbine actuators in a single control loop to maximize
load-alleviating potential is advantageous. One of the most prevalent advanced control meth-
ods that have been applied to wind turbines is called full state feedback [Kwakernaak and
Sivan 1972]. This method allows multiple control actions to be performed in a single loop,
including speed regulation, adding active damping to low-damped turbine dynamic modes,
and mitigating loads caused by stochastic wind disturbances.

Some Advanced Control Design Methods

Most advanced controls are based on linear control design methods and linear time-in-
variant models [Kwakernaak and Sivan 1972]. These linear models can be represented in the
following form:

D x = A D x + B D u + B d D u d (14-3a)

D y = C D x + D D u + D d D u d (14-3b)

where D x is the state vector, D u is the control input vector, D u d

is the disturbance input vector,

is the measured output, A represents the state matrix, B the control input gain matrix, B d
the disturbance input gain matrix, C relates the measured output D y

D y

to the turbine states, D
relates the measured output to the control input, and D d relates the measured output to the dis-
turbance states; D x represents the time derivative of D x ; D x , D x, D y

(perturbed
values) represent small perturbations from the calculated operating point values x op

, D u , and D u d

, ˙x op , y op ,

u op

.
Values in the state vector D x might represent generalized coordinates describing the
flexible modes of the turbine, such as blade flapwise and edgewise motions, tower-top fore-
aft and side-to-side motions, and rotor or generator speeds. They may also include states to
describe the control actuator dynamics [Wright and Fingersh 2008]. The values in D y

, and u d op

are the
measurements from the turbine, such as generator or rotor speed, tower-top acceleration, and
blade-root flap bending moments obtained from strain gages.
Values in the vector D u represent the control inputs. An example is blade pitch for con-
trol of turbine speed in Region 3. For Region 3 speed regulation, each blade is pitched identi-
cally, so that only one pitch value is commanded. This is called rotor collective pitch control .
In other control actions, the pitch of each blade may be commanded separately, which is
called individual or independent pitch control . Individual pitch control becomes important
when the controller is trying to mitigate the effects of asymmetric variations of wind speed
across the rotor. In such applications, the pitch of one blade may be different than the others
because of differences between the local wind speeds at each blade.
In full state feedback, the control law formulates D u as a linear combination of the tur-
bine states, as follows:

D u ( t ) = G D x ( t )

(14-4)