A novel dynamic model is introduced for the modeling of the wind turbine behavior. The objective of the wind turbine is the electric energy generation. The analytic model has the characteristic that considers a rotatory tower. Experiments show the validity of the proposed method.
1. Introduction
Researchers are often trying to improve the total power of a wind turbine. The dynamic model of a wind turbine plays an important role in some applications as the control, classification, or prediction.
Some authors have proposed the equations to model the dynamic behavior of the wind turbine as shown in [1–5].
This paper presents a novel dynamic model of a wind turbine; the first dynamic model is for the wind turbine, the second is for the tower, and both are related. Because the rotatory tower can turn, it may help the wind turbine to increase the air intake. Some companies propose wind turbines that consider rotatory blades; however, in this study, a rotatory tower is proposed instead of the rotatory blades because the first dynamic model is easier to obtain than the second.
The proposed model is linear in the states as the works analyzed by [6–10] show. The technique provides an acceptable approximation of the wind turbine behavior.
The paper is structured as follows. In Section 2, the dynamic model of a windward wind turbine of three blades with a rotatory tower is introduced. In Section 3, the simulations of the dynamic model are compared with the real data obtained by a real wind turbine prototype. Finally, in Section 4, the conclusion and future research are detailed.
2. Dynamic Model of the Wind Turbine with a Rotatory Tower
This section is divided into four parts: the first is the description of the mechanic model, the second is the description of the aerodynamic model, the third is the description of the electric model, and, finally, the fourth is the combination of the aforementioned models to obtain the final dynamic model.
The dynamic model of the wind turbine is, first, the equations that represent the change between the wind energy and mechanic energy and, second, the equations that represent the change between the mechanic energy and electric energy.
2.1. The Mechanic Model
A windward wind turbine of three blades with a rotatory tower is considered. First, the Euler Lagrangian method [11, 12] is used to obtain the model that represents the change from the wind energy to mechanic energy for the wind turbine and the change from the electric energy to mechanic energy for the tower. The masses are concentrated at the center of mass. Consider the lateral view of Figure 1 and upper view of Figure 2.
Lateral view of the wind turbine.
Upper view of the wind turbine.
From Figures 1 and 2, it can be seen that
(1)x2=-lc2S2(-C1)=lc2S2C1,y2=-lc2S2(-S1)=lc2S2S1,
where S1=sin(θ1), S2=sin(θ2), C1=cos(θ1), C2=cos(θ2), θ1 is the angular position of the tower motor in rad, θ2 is the angular position of a wind turbine blade in rad, and lc2 is the length to the center of the wind turbine blade in m. Consequently, the kinetic energy K1, K2 and potential energy P1, P2 are given as
(2)K1=0,K2=12m2lc22S22θ˙12+12m2lc22θ˙22,P1=m1glc1,P2=l1+m2glc2C2,
where m1 is the tower mass in kg, m2 is the blade mass in kg, g is the acceleration gravity in m/s^{2}, l1 is the constant length of the tower in m, and lc1 is the length to the center of the tower in m. The torques τT1a and τT2a are given as
(3)τT1a=τ1a-kb1θ1-bb1θ˙1,τT2a=τ2a-kb2θ2-bb2θ˙2,
where τ2a is the torque of the generator moved by the blade in kg m^{2} rad/s^{2}, τ1a is the torque of the motor used to move the tower in kg m^{2} rad/s^{2}, kb1 and kb2 are the spring effects presented when the blade is near to a stop in kg m^{2}/s^{2}, and bb1 and bb2 are the shock absorbers in kg m^{2} rad/s. Using the Euler Lagrange method in [11, 12] gives the following equations:
(4)m2lc22S22θ¨1+2m2lc22S2C2θ˙1θ˙2+bb1θ˙1+kb1θ1=τ1a,m2lc22θ¨2+m2glc2S2+bb2θ˙2+kb2θ2=τ2a.
The angular position of the blade 1 is related to the angular position of the blades 2 and 3 as follows:
(5)θ3=θ2+23πrad,θ4=θ2+43πrad,
where θ3 and θ4 are the angular positions for the blades 2 and 3, respectively. Then, using (4), it yields the equations for blades 2 and 3 as a function of θ2 as follows:
(6)m2lc22S2+(2/3)π2θ¨1+2m2lc22S2+(2/3)πC2+(2/3)πθ˙1θ˙2+bb1θ˙1+kb1θ1=τ1b,m2lc22θ¨2+m2glc2S2+(2/3)π+bb2θ˙2+kb2(θ2+23π)=τ2b,(7)m2lc22S2+(4/3)π2θ¨1+2m2lc22S2+(4/3)πC2+(4/3)πθ˙1θ˙2+bb1θ˙1+kb1θ1=τ1c,m2lc22θ¨2+m2glc2S2+(4/3)π+bb2θ˙2+kb2(θ2+43π)=τ2c,
where τ1b, τ2b, τ1c, and τ2c are the torques applied to move the blades 2 and 3, respectively. Adding S22+S2+(2/3)π2+S2+(4/3)π2, S2C2+S2+(2/3)πC2+(2/3)π+S2+(4/3)πC2+(4/3)π, and S2+S2+(2/3)π+S2+(4/3)π, respectively, it gives
(8)S22+S2+(2/3)π2+S2+(4/3)π2=1.5,(9)S2C2+S2+(2/3)πC2+(2/3)π+S2+(4/3)πC2+(4/3)π=0,(10)S2+S2+(2/3)π+S2+(4/3)π=0.
Now, adding the three equations of (4), (6), and (7) and using (8), (9), and (10), it gives
(11)4.5m2lc22θ¨1+3bb1θ˙1+3kb1θ1=τ1,3m2lc22θ¨2+3bb2θ˙2+3kb2θ2+2πkb2=τ2,
where τ1=τ1a+τ1b+τ1c and τ2=τ2a+τ2b+τ2c. From Figures 1 and 2, τ2 of (11) is defined as follows:
(12)τ2=C1F2,F2=F2a+F2b+F2c,
where C1=cos(θ1)F2a and F2b and F2c are the force of the air received by the three blades. Equation (12) describes the assumption that the air goes in one direction; if θ1=0, then the maximum air intake moves the blades of the wind turbine, but if the tower turns to the left or to the right and θ1 changes, then the wind turbine turns and the air intake decreases. From Figures 1 and 2 and [12], τ1 of (11) is defined as follows:
(13)τ1=kmi1,
where km is a motor magnetic flux constant of the tower in Wb and i1 is the motor armature current of the tower in A. Equations (11), (12), and (13) are the main equations of the mechanic model that represents the wind turbine and tower.
2.2. Aerodynamic Model
The aerodynamic model is the dynamic model of the torque applied to the blades. The mechanic power captured by the wind turbine Pa is given by [2–5]
(14)Pa=F2aθ˙2=12ρACp(λ,β)Vω3,
where ρ is the air density in Kg/m^{3}, A=πR2 is the area swept by the rotor blades in m^{2} with radius R in m, Vω is the wind speed in m/s, and Cp(λ,β) is the performance coefficient of the wind turbine, whose value is a function of the tip speed ratio λ, defined as [2–5]
(15)λ=θ˙2RVω.
For the purpose of simulation, the following model of Cp(λ,β) is presented [2–5]:
(16)Cp(λ,β)=c1(c2λi-c3β-c4)e-c5/λi+c6λ,
where
(17)1λi=1λ+0.08β-0.035β3+1
and the coefficients are c1=0.5176, c2=116, c3=0.4, c4=5, c5=21, c6=0.0068, and β is the blade pitch angle in rad. Using (12), (14), (15), (16), and (17) gives the dynamic model of the torque applied to the wind turbine blades as follows:
(18)τ2=C1(F2a+F2b+F2c),F2a=12θ˙2ρACp(λ,β)Vω3,F2b=12θ˙2ρACp(λ,β)Vω3,F2c=12θ˙2ρACp(λ,β)Vω3,
where C1 is defined in (12), λ is defined in (15), Cp(λ,β) is defined in (16), and λi is defined in (17).
2.3. The Electric Model
Now, analyze the change from the mechanic to electric energy for the wind turbine generator and the electric energy to mechanic energy for the tower motor. Figure 3 shows the wind turbine generator and Figure 4 shows the tower motor.
The wind turbine generator.
The tower motor.
In [12, 13], they presented the model of a motor, and similarly, from Figure 3, a model for the generator is obtained by using the Kirchhoff voltage law. Therefore, the dynamic models of the motor and generator are as follows:
(19)V1=R1i1+L1i˙1+k1θ˙1,k2θ˙2=R2i2+L2i˙2+V2,
where k1 is the motor back emf constant in V s/rad, k2 is the generator back emf constant in V s/rad, R1 is the motor armature resistance in Ω, R2 is the generator armature resistance in Ω, L1 is the motor armature inductance in H, L2 is the generator armature inductance in H, V1 is the motor armature voltage in V, V2 is the generator armature voltage in V, i1 is the motor armature current in A, and i2 is the generator armature current in A. For the generator of this paper, V2=Rei2. Thus, (19) becomes
(20)V1=R1i1+L1i˙1+k1θ˙1,k2θ˙2=(R2+Re)i2+L2i˙2,V2=Rei2.
2.4. The Final Dynamic Model
Thus, (11), (13), and (18) that represent the change between the wind energy and mechanic energy and (20) that represents the change between the mechanic energy and electric energy are considered as the wind turbine dynamic model with a rotatory tower.
Define the state variables as x1=i2, x2=θ2, x3=θ˙2, x4=i1, x5=θ1, x6=θ˙1, the inputs as u1=F2 and u2=V1, and the output as y=V2. Consequently, the dynamic model of (11), (13), (15), (18), and (20) becomes
(21)x˙1=-(R2+Re)L2x1+k2L2x3,x˙2=x3,x˙3=-kb2m2lc22x2-bb2m2lc22x3-2πkb23m2lc22+cos(x5)3m2lc22u1,x˙4=-R1L1x4-k1L1x6+1L1u2,x˙5=x6,x˙6=-3kb14.5m2lc22x5-3bb14.5m1lc22x6+km4.5m2lc22x4,y=Rex1,u1=F2a+F2b+F2c,F2a=12x3ρACp(λ,β)Vω3,F2b=12x3ρACp(λ,β)Vω3,F2c=12x3ρACp(λ,β)Vω3,λ=x3RVω,
where Cp(λ,β) is defined in (16) and λi is defined in (17).
Remark 1.
The model of this research considers a direct current motor and a direct current generator. This wind turbine is proposed to be used on the house roof for the daily usage to feed a light as a renewable source. If more than one wind turbine is used, more electricity is generated. The proposed model could be extended to other kinds of machines by changing (20).
Remark 2.
Note that the dynamic model states of the wind turbine with a rotatory tower (21) is a linear system [6–10].
3. Experiments to Validate the Dynamic Model
Figure 5 shows the prototype of a wind turbine with a rotatory tower which is considered for the simulations of the dynamic model. This prototype has three blades with a rotatory tower which does not use a gear box. Table 1 shows the parameters of the prototype. The parameters m2 and lc2 are obtained from the wind turbine blades. The parameters R1, L1, and k1 are obtained from the tower motor. The parameters k2, R2, Re, and L2 are obtained from the wind turbine generator. The parameters R, ρ, Vω, and β are obtained from [2–5].
Parameters of the prototype.
Parameter
Value
lc2
0.5 m
m2
0.5 kg
kb2
1×10-6kgm2/s2
bb2
1×10-1 kgm2rad/s
k2
0.45 Vs/rad
R2
6.96Ω
L2
6.031×10-1H
R
lc2m
ρ
1.225kg/m3
g
9.81m/s2
Re
30Ω
km
0.09Wb
kb1
1×10-6kgm2/s2
bb1
1×10-1kgm2rad/s
k1
0.0045Vs/rad
R1
18Ω
L1
6.031×10-1 H
Vω
5m/s
β
0.5 rad
Prototype of a wind turbine with rotatory tower.
The dynamic model of the wind turbine with a rotatory tower is given by (21) with the parameters of Table 1. 1×10-5 are considered as the initial conditions for the plant states x1=i2, x2=θ2, x3=θ˙2, x4=i1, x5=θ1, and x6=θ˙1. The root mean square error (RMSE) is used, and it is given as [13–15]
(22)RMSE=(1T∫0Te2dτ)1/2,
where e2=euj2=(ujr-uj)2, e2=exi2=(xir-xi)2, or e2=ey2=(yr-y)2 and ujr, xir, and yr are the real data of uj, xi, and y, respectively, i=1,2,…,6, j=1,2.
3.1. Example <bold />1: The First Behavior
Figures 6 and 7 show the wind turbine inputs, outputs, and states. Table 2 shows the RMSE for the errors.
The RMSE for Example 1.
RMSE
eu12
0.0204
eu22
1.1835
ey2
0.0171
ex12
5.7158×10-4
ex22
0.9843
ex32
0.0470
ex42
0.0658
ex52
0.0872
ex62
0.0163
Inputs and output of the wind turbine of Example 1.
States of the wind turbine of Example 1.
From Figures 6 and 7, it can be seen that the dynamic model has good behavior described as follows: (1) from 0 s to 2 s, both inputs are fed; consequently, the tower moves far of the maximum air intake, the generator current is decreased, and the wind turbine blades stop moving; (2) from 2 s to 4 s, both inputs are not fed; consequently, current is not generated and both the tower and wind turbine blades do not move; (3) from 4 s to 6 s, both inputs are fed, but the air intake is positive and tower voltage is negative; consequently, the tower returns to the maximum air intake, the generator current is increased, and the wind turbine blades move; (4) from 6 s to 8 s, both inputs are not fed; consequently, current is not generated and the tower and wind turbine blades do not move. The dynamic model is a good approximation of the wind turbine behavior because the signals of the first are near to the signals of the second. From Table 2, it is shown that the dynamic model is a good approximation of the real process because the RMSE is near to zero.
3.2. Example <bold />2: The Second Behavior
Figures 8 and 9 show the wind turbine inputs, outputs, and states. Table 3 shows the RMSE for the errors.
The RMSE for Example 2.
RMSE
eu12
0.0288
eu22
0.9423
ey2
0.0280
ex12
9.3212×10-4
ex22
1.7209
ex32
0.0766
ex42
0.0523
ex52
0.1324
ex62
0.0162
Inputs and output of the wind turbine of Example 2.
States of the wind turbine of Example 2.
From Figures 8 and 9, it can be seen that the dynamic model has good behavior described as follows: (1) from 0 s to 2 s, the input air is fed and the tower input is not fed; consequently, the tower remains in the maximum air intake, the generator current is maximum, and the wind turbine blades have motion; (2) from 2 s to 4 s, the air is not fed and the tower input is fed; consequently, current is not generated, the tower moves far of the maximum air intake, and the wind turbine blades do not have motion; (3) from 4 s to 6 s, the air is fed and the tower input is not fed; consequently, the tower does not move, the generator current is minimum, and the wind turbine blades almost do not move; (4) from 6 s to 8 s, the air is not fed and the tower input is fed with a negative voltage; consequently, current is not generated, the tower returns to the maximum air intake, and the wind turbine blades do not have motion. The dynamic model is a good approximation of the wind turbine behavior because the signals of the first are near to the signals of the second. From Table 2, it is shown that the dynamic model is a good approximation of the real process because the RMSE is near to zero.
4. Conclusion
In this paper, a dynamic model of a wind turbine with a rotatory tower was introduced; the simulations using the parameters of a prototype showed that the proposed dynamic model has an acceptable approximation of the wind turbine behavior. The proposed dynamic model could be used on control, prediction, or classification. As a future research, the modeling will be improved using other interesting methods as the least squares in [16–26], neural networks in [14, 27–32], or fuzzy systems in [15, 33–35].
Conflict of Interests
The authors declare that there is no conflict of interests regarding the publication of this paper.
Acknowledgments
The authors are grateful to the editor and the reviewers for their valuable comments and insightful suggestions, which helped to improve this paper significantly. The authors thank the Secretaría de Investigación y Posgrado, Comisión de Operación y Fomento de Actividades Académicas del IPN, and Consejo Nacional de Ciencia y Tecnología for their help in this paper.
FanL.-L.SongY.-D.Neuro-adaptive model-reference fault-tolerant control with application to wind turbinesŽertekA.VerbičG.PantošM.Optimised control approach for frequency-control contribution of variable speed wind turbinesMuhandoE. B.SenjyuT.YonaA.KinjoH.FunabashiT.Disturbance rejection by dual pitch control and self-tuning regulator for wind turbine generator parametric uncertainty compensationTangC. Y.GuoY.JiangJ. N.Nonlinear dual-mode control of variable-speed wind turbines with doubly fed induction generatorsVepaR.Nonlinear, optimal control of a wind turbine generatorDehghanM.HajarianM.Modied AOR iterative methods to solve linear systemsHajarianM.Matrix form of the Bi-CGSTAB method for solving the coupled Sylvester matrix equationsHajarianM.Effcient iterative solutions to general coupled matrix equationsHajarianM.Matrix iterative methods for solving the Sylvester-transpose and periodic Sylvester matrix equationsRubioJ. J.OrdazG.Jiménez-LizárragaM.CabreraR. I.General solution of the Navier-Stokes equation to describe the dynamics of a homogeneous viscous fluid in an open tubeLewisF. L.DawsonD. M.AbdallahC. T.SpongM. W.HutchinsonS.VidyasagarM.RubioJ. J.PachecoJ.Pérez-CruzJ. H.TorresF.Mathematical model with sensor and actuator for a transelevatorRubioJ. J.Perez-CruzJ. H.Evolving intelligent system for the modelling of nonlinear systemswith dead-zone inputRubioJ. J.Evolving intelligent algorithms for the modelling of brain and eye signalsDingS.DingR.YangE.A filtering based recursive least squares estimation algorithm for pseudo-linear auto-regressive systemsDingF.Combined state and least squares parameter estimation algorithms for dynamic systemsDingF.Coupled-least-squares identification for multivariable systemsDingF.Decomposition based fast least squares algorithm for output error systemsDingF.Two-stage least squares based iterative estimation algorithm for CARARMA system modelingHanL.WuF.ShengJ.DingF.Two recursive least squares parameter estimation algorithms for multirate multiple-input systems by using the auxiliary modelHuH.DingF.An iterative least squares estimation algorithm for controlled moving average systems based on matrix decompositionLiX. L.ZhouL. C.DingR. F.ShengJ.Recursive least-squares estimation for Hammerstein nonlinear systems with nonuniform samplingPanS.ChenJ.-S.A least-square semismooth Newton method for the second-order cone complementarity problemWangD. Q.DingF.ChuY. Y.Data filtering based recursive least squares algorithm for Hammerstein systems using the key-term separation principleWangD.DingF.XimeiL.Least squares algorithm for an input nonlinear system with a dynamic subspace state space modelChenJ.DingR. F.An auxiliary-model-based stochastic gradient algorithm for dual-rate sampled-data Box-Jenkins systemsChenJ.Several gradient parameter estimation algorithms for dual-rate sampled systemsDingF.LiuX.ChenH.YaoG.Hierarchical gradient based and hierarchical least squares based iterative parameter identification for CARARMA systemsPratamaM.AnavattiS. G.AngelovP. P.LughoferE.PANFIS: a novel incremental learning machinePérez-CruzJ. H.RubioJ. J.PachecoJ.SorianoE.State estimation in MIMO non- linear systems subject to unknown deadzones using recurrent neural networksSunJ.ChenJ.-S.KoC.-H.Neural networks for solving second-order cone constrained variational inequality problemBuchachiaA.Dynamic clusteringLughoferE.Sigle pass active learning with conflict and ignoranceMacielL.LemosA.GomideF.BalliniR.Evolving fuzzy systems for pricing fixed income options