Model Predictive Control

in Wind Turbine
Sisca Dina 2414201006
Nur Fitriyani 2414201010
Wind Turbine
Model Dinamik Nonlinear Wind
Turbine
Mekanika Turbine
Pemodelan Angin
 Kecepatan angin dipengaruhi oleh kecepatan
angin rata-rata dan kecepatan angin turbulen

 Dimana

 Model angin turbulen dapat dirumuskan

dalam bentuk state space
Modelling Wind Turbine
 Servo-Elastics: The first tower fore-aft
bending mode, the rotational motion, and the
collective pitch actuator are modeled linearly
as:
Modelling Wind Turbine
 Aerodynamics: The aerodynamic torque and
thrust acting on the rotor with radius R are
nonlinear static functions:
Model Design Controller
Pemodelan
 Persamaan differensial individu dapat dirubah
menjadi :

 Input control u dan disturbance d :

 Output sistem y model desain merupakan

percepatan tower atas dan kecepatan rotor yang
dapat diakses dari sensor yang biasa digunakan :
Mereduksi Model Linear
 In order to design a linear MPC, a linear
representation is needed. Therefore, the
obtained nonlinear reduced model (2)-(4) is
linearized at an equilibrium point defined by v0.
Then the linear state space representation is
ESTIMATOR DESIGN
 It can be separated into the prediction update part and the
measurement update part. For the prediction update part the
discrete nonlinear system is used

 The error equation is

 The prediction update part is then

 Together with the Kalman Gain

 the measurement update part is

Pitch controller estimator
 The pitch controller has a separate estimator
that is linear in it self and hence not
linearized around anything.
Full Sensor
 The full sensor estimator has available
measurements of all the variables being
estimated.


output and disturbance matrices are

The Optimal Control Problem
 The objective is to find the optimal control
inputs ui minimizing the cost functional J, which
is defined as the sum of all deviations from the
system outputs yi from the reference values r
over the prediction horizon P weighted with Q
and the sum of all changes in the control inputs
ui over the control horizon M weighted with R:
Pitch controller
 The pitch controller has the the following control variables

 The tuning parameters of the controller are

The optimization matrices are

Controller KI dan KIII
 Control variabel

 The tuning parameters of the controller

of the mounted tower are

 The tuning parameters of the controller of the floating tower

are

 Optimization Matrice