A new fuzzy logic proportional controller approach applied to individual pitch angle for wind turbine load mitigation

Accepted Manuscript A new fuzzy logic proportional controller approach applied to individual pitch angle for wind turbine load mitigation Zafer civelek, Murat Lüy, Ertuğrul Çam, Hayati Mamur PII:

S0960-1481(17)30380-4

DOI:

10.1016/j.renene.2017.04.064

Reference:

RENE 8759

To appear in:

Renewable Energy

Received Date: 31 January 2017 Revised Date:

17 April 2017

Accepted Date: 28 April 2017

Please cite this article as: civelek Z, Lüy M, Çam Ertuğ, Mamur H, A new fuzzy logic proportional controller approach applied to individual pitch angle for wind turbine load mitigation, Renewable Energy (2017), doi: 10.1016/j.renene.2017.04.064. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT 1 2 3

A NEW FUZZY LOGIC PROPORTIONAL CONTROLLER APPROACH APPLIED TO INDIVIDUAL PITCH ANGLE FOR WIND TURBINE LOAD MITIGATION ABSTRACT In the world, efforts to increase the resource diversity in electric generation have accelerated lately. So, the great improvements have been achieved in wind turbines (WTs). The dimensions of WTs have grown for more electric generation and their energy productions have increased. Depending on these developments, it has become more important to reduce the WT load mitigation. Thus, a tendency to pass an individual pitch angle system control rather than a collective pitch angle system control employed to stable the output power of WTs over nominal wind speeds has whetted in recent studies. However, in literature, a controller proposal relating to how to incorporate the blade moments used for providing the individual pitch angle system into the output power control system has not yet been offered. Therefore, in this study, a new fuzzy logic proportional control (FL-P-C) approach has been recommended to mitigate the moment load on blades and tower to a minimum possible value while keeping the output power of WTs at a constant value. The offered FL-P-C has also been verified by MATLAB/Simulink. Through the first application of the FL-P-C on a WT, a significant improvement of 33-83% has been managed for the blade and tower moment loads. Furthermore, the grid fluctuations have been reduced because of the stabilisation of the output power of the WT. Ultimately, by the offered FL-P-C, not only the WT load mitigations and maintenance costs of WTs could be reduced, but also electric costs could be decreased owing to increasing lifetimes of WTs.

20 21

Keywords: energy, wind turbine, individual pitch control, wind turbine load mitigation, blade load mitigation.

22

INTRODUCTION

23 24 25 26 27

Nowadays, wind turbines (WTs) are become more popular with a growing interest in renewable energy resources. It is thought that this interest would be alive as environmental concerns are present and also a need for electricity increases. In parallel with technological developments and the demand for electricity, both the capacity of WTs is grown and their control methods employed are enhanced and changed.

28 29 30 31 32 33

Wind, which is energy source of WTs, suddenly varies in nature. Therefore, the sudden fluctuations are observed electric energy generation obtained from WTs [1]. In order to minimise the effects of the sudden fluctuations, also to increase the generated power and to stable the output power of WTs, maximum power transfer methods are applied on WTs when wind speed is lower than nominal values. On the other hand, the output power of WTs is kept at a nominal output power or the nearest to the nominal output power by means of a blade pitch angle control when wind speed is nominal values.

34 35 36 37 38 39 40 41

In literature, a number of control methods have recommended for the pitch angle control of WTs. These are control methods such as the proportional-integral (PI) control, the fuzzy logic control (FLC), the fuzzy logic proportional-integral-derivative (FL-PID) control and the artificial neural networks (ANNs) [2-5]. Great deals of studies have been conducted on the advantages and disadvantages of these methods [6, 7]. Recently, due to increases in the physical size of WTs, it is necessary to rethink the effect of the WT load mitigation. Moreover, owing to the fact that these turbines are manufactured for large powers and also allow the individual pitch angle control, the individual pitch angle control method has come to the agenda.

AC C

EP

TE D

M AN U

SC

RI PT

4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

1

ACCEPTED MANUSCRIPT The greater and more flexible of WTs enhance the significance of WT load mitigation. Because of the modern control methods such as the FLC, the ANN, the intuitive optimisation (IO), the WT load mitigation has become more attractive. Moreover, by means of the developed software’s and sensors, the understanding of the mechanical loads has made easier for researchers and scientists. Thus, they are allowed researchers to develop the different and suitable control algorithms. They have also been used in studies on the reduction of edgewise and flapwise moments in the wind turbine blades [8-11].

48 49 50 51 52 53

While the efforts are being made to reduce the energy costs and to increase the power quality, a control system should not be designed to exactly follow an ideal power curve being possible. Other control parameters have to be taken into account [12, 13]. One of these parameters is the mechanical loads in WTs. The mechanical loads cause a fatigue on various turbine components. The fatigue shortens the useful life of a WT, then consequently the average cost of the WT increases and the energy cost of it goes up [14].

54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74

When literature is evaluated, some close studies could be seen. In a carried out study by Bossanyi in 2003, the periodic mechanical loads, which have a phase difference of 120ᵒ in WTs with three blades, were mentioned. As a solution to these loads, the individual pitch control was recommended [8]. Again, in another fulfilled study by Bossanyi, a feedforward filter was added to the individual pitch angle control and thus the compensation of 3P frequency loads was provided. In here, P was the number of harmonics [9]. Moreover, in an executed study by Larsen et al., it was showed that the mechanic load mitigation was possible by means of the based on local blade flow measurements [15]. In addition, in a recommended study by Selvam et al. in 2009, they recommended a linear quadratic Gaussian (LQG) controller and a feedforward disturbance rejection controller (FDRC), which they were individual each other and had multi variables. By means of these control methods used for the individual pitch angle control, they realized the mechanical load mitigation of a WT [10]. Moreover, Zhang et al. extended the classical PID control method a bit further and formed a proportional integral plus (PIP) control method and then they used the method on the individual pitch angle control of WTs for the WT load mitigation [16]. Furthermore, in another study by carried out Liu et al., the moment control below nominal wind speeds and the pitch control over nominal wind speeds were conducted. The pitch control was realized by using the individual pitch control and the proportional integral (PI) control [17]. In order to control the output power of a WT, a FLC was used and also in order to obtain the individual pitch angle, two different FLC in d and q axes were utilized [18]. Lastly, in a study performed by Pan and Ma, a fuzzy PI controller (F-PI-C) was managed to acquire the collective pitch angle and the individual pitch angle values [19]. Jureczko and colleagues examined the aerodynamic loads on the turbine blades [20].

75 76 77 78

Another method of reducing the mechanical load on the wind turbine is the smart rotor. The smart rotor consists of control devices placed along the blades of the wind turbine and allows real-time adjustment of the aerodynamic characteristics of the blades [21, 22]. A lot of work has been done in this regard [23-27].

79 80 81 82 83 84 85 86

In order to keep the output power of WTs at a steady-state, the pitch angle control is utilized for wind speeds, which is over nominal wind speeds. Many control methods have been tried for the output power control. In this study, a fuzzy logic PID controller (FL-PID–C) has been operated at wind speeds being over nominal wind speeds. Since the FL of the FL-PID-C is suitable for nonlinear system control and also the PID of it has a response speed. For this reason, the FL-PID-C gives better results than the other control methods in WTs [6]. Several studies have been carried out to adjust the coefficients of the PID controller with fuzzy logic. These studies describe the basis of the fuzzy PID method [28-31].

AC C

EP

TE D

M AN U

SC

RI PT

42 43 44 45 46 47

2

ACCEPTED MANUSCRIPT On the other hand, in this study, the output power optimisation has been executed for the first time by the FL-P-C taking into account the addition rate of blade moments to the produced collective pitch angle. Additionally, the FL-P-C recommended for the output power quality and moment value controller in this study can be applied on all conventional and modern control systems that produce the collective pitch angle value. According to the offered FL-P-C, the individual pitch angle value is obtained by means of adding the blade moments to the collective pitch angle at a certain rate.

93 94 95 96 97 98 99 100 101 102 103

In previous carried out studies, the addition rate of the moments on blades has been entirely left to a designer preference. However, in this study, this rate has been determined and automated by the offered FL-P-C. The FL-P-C has two inputs. One of them is the blade moment average and the other is the output power quality. These two values are continuously evaluated at particular intervals by the FL-P-C. While the evaluating is being carried out, the FL-P-C slowly increases the k-coefficient value, which is the addition rate to the collective pitch angle, from starting zero. As this value increases, the moment load of the WT decreases. As a result, some changes in the power quality occur. These changes are sometimes increasing and sometimes decreasing according to the preferred control and control strategy. However, the FL-P-C defines the k-coefficient value to keep the moment and power quality values within the limits that the designer desires. Thus, the WT operates by providing the power quality and the moment load improvement like as the designer’s desire.

104 105 106 107 108 109 110

The presentation of this work is as follows. In the “Material” section, which is the second part of the presentation; wind energy and turbine mechanics, WT working zones, and power quality-moment variation are given. After, in the third part "Method" section, individual pitch angle control of WTs and power quality-moment balance, output power controller design and power-moment balance controller design are explained. The obtained results and the evaluation of the results are made in the fourth section "Simulation Results and Discussion". In "Conclusions" section, the outcomes are also written.

111

MATERIALS

112

Wind Energy and Turbine Mechanics

113

The wind power is given in Equation (1) as follows:

SC

M AN U

TE D

EP

115 116


= 0.5

(1)

where, P, ρ, A and  are the wind power (W), the air density (kg/m3), the swept area by blades (m2), and the wind speed (m/s), respectively[32].

AC C

114

RI PT

87 88 89 90 91 92

122

WTs can convert a certain part of the wind power into electrical energy. The certain part of the wind power cannot exceed a Betz limit. The Betz limit is 0.59. Therefore, a WT can convert up to 59% of the wind power into electrical energy [33]. The power amount that can be taken from the WT is determined by a power coefficient ∁ . The power coefficient ∁ is a function of the blade pitch angle

and the blade tip speed ratio λ (TSR) [34].

123


 =
∁  , .

(2)


 = 0.5 ∁  , 

(3)

117 118 119 120 121

124 125

The mechanical power that can be obtained from WTs is explained in Equation (2):

When Equation (1) is substituted in Equation (2), Equation (3) is obtained as follows:

3

ACCEPTED MANUSCRIPT

129 130 131

The turbine power coefficient ∁ is given Equation (4) as below: ∁ = 0.5176 

 − 

 '

=

 ().)*+

=

).) , +(

-. / 0

where, 1 is the turbine rotor angular velocity (rad/s) and 2 is the blade radius of the WT (m).

where, 3 is the aerodynamic moment (Nm).

3= 4

141

When Equation (2) is substituted in Equation (7), Equation (8) is obtained as below:

142

3 = 5 62  5

146 147 148 149

(5)

(6)

TE D

∁7 +, 

.

(7)

(8)

The ratio between ∁  , λ and λ in Equation (8) is a new unitless parameter and is known as a moment coefficient [35].

EP

145



This value is also described by the following equation:

AC C

143 144

(4)

An aerodynamic moment is the power amount per unit the angular velocity [34]. The aerodynamic moment that drives the rotor is given in Equation (7):

139 140

−

M AN U

137 138

+ 0.0068.

The TSR is the ratio of the blade angular velocity to the wind speed and is given in Equation (6) as follows:

135 136

!"# $

Where λi is a transitional variable used to simplify the formula. It does not represent any physical parameters in the turbines. The λ' value given by Equation (5) is calculated by using Equation (4):

132 133 134

0.4 − 5

RI PT

128

where, ∁  , λ, and λ are the turbine power coefficient, the blade pitch angle and the TSR, respectively.

SC

126 127

∁8  ,  =

∁7 +, 

.

(9)

According to Equation (6), any change in the rotor speed or the wind speed varies the TSR. Then, this changes the power coefficient. As a result of the power coefficient change, the power amount that can be obtained from the wind speed will change.

150 151 152

According to Equation (4) and Equation (5), the power coefficient change ∁ also depends on the blade pitch angle change. Ultimately, the output power control of WTs is carried out according to this principle.

153

Wind Turbine Working Zones

154 155 156 157

In the WTs with variable speed and variable pitch angle, the output power is change [36]. In Figure 1, the working zones of WTs are showed. According to Figure 1, these zones are separated by four. I. zone is a region where a WT has not begun to operate for electrical energy generation. The wind speed values in this region are not economical to operate the WT. Therefore, the output power in I. zone is 4

ACCEPTED MANUSCRIPT zero. The wind speed value while passing from I. zone to II. Zone is called a cut-in point. After this wind speed value, the WT starts to operate for electrical energy generation. II. zone is the region between the cut-in point and the nominal wind speed. In here, the generated power amount is below the nominal power. The maximum power point tracking (MPPT) is performed in this region. For the MPPT, the generator moment is adjusted to keep the power coefficient ∁ of the WT at a maximum value by the various power electronic circuits [35]. For this reason, the zone is an important region. III. zone is the region between the wind speeds over nominal wind speeds and the cut-out point that WTs are shutdown for security reasons. In here, the output power of WTs is attempted to keep at a nominal value by controlling the pitch angle [34]. As the last zone, the WT in IV. zone is closed for security [37].

RI PT

158 159 160 161 162 163 164 165 166 167

Figure 1. Working zones of WTs.

169

M AN U

SC

168

In WTs, there are mainly two types of mechanical loads; static and dynamic. These static loads occur as a result of turbine interaction with the average wind speed. On the other hand, these dynamic loads are more important for a control point. These dynamic loads are caused by the spatial and temporal distribution of the wind speed field on the surface swept by the rotor. Also these dynamic loads include the variations in the aerodynamic moment that spread downward from the drive-train and in the aerodynamic moment that is the effect on the mechanical structure. These are also called as the drive-train and the structural loads.

177 178 179 180 181 182 183 184 185

There is also a common classification of these dynamic loads. It is also the transient loads caused by storms and fluctuations and has low frequency mainly. In particular, the transitional loads used to calculate the WT component values are very important at high wind speeds. These transition between the MPPT in II. zone and the power regulation in III. zone directly affects them limited power at wind speed values over the nominal wind speed. It is inevitable that inadequate control strategies lead to strong transition loads. In addition, a controller design and setting affect the transition loads. Indeed, in a closed loop system that closely follows the steady state control strategy curve, there will be heavier transition loads during storms. Therefore, when planning a control strategy, these loads have to be taken into account[38].

186

Power Quality and Moment Variation

187 188 189 190 191

The rotational sampling causes the high frequency loads that intensify at times of the rotor speed and also spectrally peaked. In a WT with n blades, the spectral loads are dominant in nP, in which P is called as a harmonic. When they spread to the drive-train and the structure, they activate some unwanted damped vibrations in the system. Thus, the flexibility of WT components and the importance of control systems raised in large WT systems. By means of a controller design and its

AC C

EP

TE D

170 171 172 173 174 175 176

5

ACCEPTED MANUSCRIPT adjustment being good, the defined control strategy influences the mentioned periodic loads at high rates. While the pitch angle control is affecting the structural loads, the generator control influences the expanding of the drive-train loads. Therefore, a poor control design can cause the vibration mode in a WT. Then, this may lead to some unwanted mechanical part damages such as a gear box or blades. A controller to be designed has to prevent the vibrations in order to mitigate the high frequency loads and the fraction risk depending on fatigues as much as possible. In other words, a control strategy has to avoid from any operation point that may cause a vibration that cannot be absorbed by the controller [14].

200 201 202 203

In WTs, individual pitch angle values are obtained from blade moment values. These moment values are evaluated and then added to a collective pitch angle value. The adding ratio of these evaluated moments to the collective pitch angle is an important process. According to the defined value of the ratio, the moment load mitigation and the output power quality are identified.

204 205 206 207 208 209 210 211 212 213 214 215 216

In this study, the moment and the power quality measurements have been fulfilled for different values of the multiplication ratio. The average of the moment values at the blade angles has been measured in real time, and then multiplied by the multiplication factor and lastly added to the collective pitch angle value. The multiplication factor has been gradually increased starting from zero during simulations. According to the defined value of the multiplication factor, the average of the blade moment values and the fluctuation rate at the output power have been measured. After, the multiplication factor has been increased until a value allowed by the FAST WT simulator and the obtained results are illustrated in Figure 2. The multiplication factor or k coefficient linearly increased is presented in Figure 2 (a). Also, the average moment load on the blades as a function of this increase of the multiplication factor is depicted in Figure 2 (b). The value, in which the multiplication factor is zero, is an important point that the collective pitch angle control has been carried out. Other values correspond to the individual pitch angle control. When the multiplication factor is low, the angle difference between the blades is low. The angle difference between the blades is high when the multiplication factor is high.

TE D

M AN U

SC

RI PT

192 193 194 195 196 197 198 199

218 219 220 221

AC C

EP

217

(a)

(b)

(c)

Figure 2. The obtained results by the FAST WT simulator; a) the multiplication factor or k coefficient, b) the average moment load on the blades and c) the power poverty.

222 223 224 225 226 227 228

The output power quality seriously deteriorates when the angle difference between blades pass a certain value in the individual pitch angle control. The power poverty is seen in Figure 2 (c). In Figure 2 (b), when the multiplication factor is at zero, which represents the collective pitch angle control, the moment load has been taken above 3,000 kNm. On the other hand, when the multiplication factor is at 0.006 value, the moment load has been measured under 2,000 kNm. For these values, as shown in Figure 2 (c), the amount of the output power oscillation has dropped below 6

ACCEPTED MANUSCRIPT 0.1%. By means of this, a very good output power quality has been achieved. This is already the expected results from the individual pitch angle control. The output power quality, which is the opposite of power poverty, is high for some values of the multiplication factor and low for some values. Here, the purpose of a good control system is to look for the multiplication factor values that have the low moment values and the high power quality, in which the power poverty is low.

234

METHODS

235

Individual Pitch Angle Control of Wind Turbines and Power Quality-Moment Balance

236 237 238 239 240 241 242 243 244 245 246 247 248

A WT model with a power quality-moment balance consists of four basic sections. The WT model with the individual pitch angle control for moment and power stabilisation is given in Figure 3 in this study. The first is a FAST WT simulator. This simulator can be run in MATLAB software. In addition to classical WT simulators, it allows the calculation of the individual pitch angle control and the mechanical loads on various parts of the WT. Here, the FAST has been adjusted to work with the appropriate programs in MATLAB/Simulink. The second part is a collective pitch control unit that determines the collective pitch angle to keep the output power at a steady state level at wind speeds over the nominal wind speed. The third section is an individual pitch control section, which produces three different pitch angle difference values by using the blade moment values and the phase angles. The fourth part is a FL-P-C, which takes the blade moment and the output power values from the FAST WT simulator in real-time and also evaluates a fuzzy logic. This FL-P-C evaluates the moment and the power quality and then notify the individual pitch controller how to convert the blade moments to the individual pitch angle difference values.

M AN U

SC

RI PT

229 230 231 232 233

AC C

EP

TE D

249

Figure 3. The WT model with the individual pitch angle control for moment and power stabilisation.

250 251 252 253 254 255

As shown in Figure 4, the individual pitch angle control system unit is the section that produces the individual pitch angle difference values from the moment values obtained from the blades. These moment values of three different channel Myc1, Myc2 and Myc3 that taken from the FAST WT simulator in real time are converted to Md and Mq values of the d-q axis by a Coleman cycle. The Coleman cycle is given in Equation (10) as follows: 7

ACCEPTED MANUSCRIPT

256

5C DC :IJ :; 5 cos B cos B +  cos B +  : 9: = = > 5C DC G H IJ5 K < sin B sin B +  sin B +  :IJ



(10)

257 258 259 260 261

The high frequency components are separated by applying a low pass filter Md and Mq converted to the d-q axis. Through multiplying these moment values to k coefficient coming from the FL-P-C, the pitch difference angle values ∆βd and ∆βq in the d-q axis are reached. After, these values are passed through the inverse Coleman cycle given in Equation (11) to obtain the blade pitch angle difference values ∆β1, ∆β2 and ∆β3.

262

(11)

RI PT

SC

After that, these values are summed with the collective pitch angle values to obtain the individual pitch angle values. Lastly, these calculated values are applied to the FAST WT simulator by passing a saturation limiter and a rate limiter. While the saturation are keeping the pitch angle between 0ᵒ and 90ᵒ, the rate limiter restricts the pitch angle to 9 ᵒ/s.

M AN U

263 264 265 266

cos B sin B ∆  5C 5C ∆ ; L∆ 5 N = Ocos B +  sin B + P 9∆ = < DC DC ∆ cos B +  sin B + 

TE D

267

EP

Figure 4. The individual pitch angle control system.

268

The suggested FL-P-C gives positive, negative or zero numbers according to the moment and power values coming to its inputs. But, the general process is to produce a positive number until the desired moment and power. When the desired power quality and moment value is reached, the FL-P-C stops the k- coefficient changing by means of making its output zero. Providing that the system enters an undesired power quality or moment value in the operation, the FL-P-C will give an output to correct it.

274

Output Power Controller Design

275 276 277 278 279 280

In this study, Sugeno-Takagi type FL-PID-C is preferred because this is a simple mathematical operation and it is also faster than Mamdani. When the other control processes are scanned, it is seen that this control is widely preferred. As shown in Figure 5, the collective pitch angle control system has been carried out a first order Sugeno-Takagi type FL-PID-C. Here, the output power has been taken as a controlled variable. After, the output of each rule has been found by linear combination of the input values. After that, the sharp output values have been found by taking the weight average.

AC C

269 270 271 272 273

281

8

SC

RI PT

ACCEPTED MANUSCRIPT

Figure 5. The collective pitch angle control system.

M AN U

282

In order to control the output power of the wind energy system, the designed Sugeno-Takagi fuzzy logic controller has five rules;

285

1. Rule: z1 = a.x + b.y + c (a = 1, b = 0.5, c = 0.5),

286

2. Rule: z2 = d.x + e.y + f (d = 0, e = 1, f = 0.5),

287

3. Rule: z3 = g.x + k.y + l (g = 0, k = 0, l = 0),

288

4. Rule: z4 = m.x + n.y + o (m = 0, n = 1, o = 0),

289

5. Rule: z5 = p.x + q.y + r (p = 0.5, q = 0.5, r = 0).

290

Where; z value is taken as follows:

TE D

283 284

EP

Q=

291

R# .S# (R" .S" (RT .ST (RU .SU (RV .SV S# (S" (ST (SU (SV

(12)

The x and y letters in the fuzzy rules indicate the error and the derivative of error, respectively, which are the inputs of the FLC. The a, b, ..., r letters show the coefficients. The w in Equation (12) represents the weight of each rule at the output. In the applied system, all weights are equal to each other and the value is 1. In the MATLAB / Simulink block, the function that defines these rules has been defined as below:

297

function z = suge1(e, de, w, a)

298

a = [1 0, 5 0, 5 0 1 0, 5 0 0 0 0 1 0, 5 0 0, 5 0, 5 0];

299

w = [1 1 1 1 1];

300

z = 0; j = 1;

301

for i = 1 : 3 : length(a)

302

z = z + (e*a(i) + de*a(i + 1) + a(i + 2))*w(j);

AC C

292 293 294 295 296

9

ACCEPTED MANUSCRIPT j = j + 1;

304

end

305

z = z/sum(w);

306 307 308

The difference between the desired output power value and the output power is the error. The change of the error in unit time is the derivative of the error. The inputs of the FLC are the error and the derivative of it.

309

Power-Moment Balance Controller Design

310 311 312 313 314 315 316

A Mamdani type FLC has been utilized for the power-moment balancer task in this study. In the simulated system, the power quality and the blade moments have been averaged at determined time intervals. The time interval should be defined by the person designing the control system. Furthermore, the time interval should be long enough in order to ensure that the control output will have an effect on the system and at the same time short in order to manage the control quality of the controlled system. The FL-P-C system for the moment and the power stabilisation in the executed work is showed in Figure 6.

SC

RI PT

303

M AN U

317

TE D

Figure 6. The FL-P-C system for the moment and the power stabilisation in the executed work.

318

AC C

322

The input fuzzy clusters utilized for the output power quality, the input fuzzy clusters managed for the moment, the output parameter and its fuzzy clusters, and the variation between moment-power quality and output parameter are given Figure 7 (a, b, c, and d), respectively.

EP

319 320 321

323 324

(a)

(b)

10

325 326

(c)

(d)

RI PT

ACCEPTED MANUSCRIPT

Figure 7. a) The input fuzzy clusters utilized for the output power quality, b) the input fuzzy clusters managed for the moment, c) the output parameter and its fuzzy clusters and d) the variation between moment-power quality and output parameter.

330

Nine fuzzy rules have been assigned to the designed system and these are listed below.

331

1. If moment is good and power is good then dkp is z

332

2. If moment is good and power is medium then dkp is z

333

3. If moment is good and power is bad then dkp is ns

334

4. If moment is medium and power is good then dkp is z

335

5. If moment is medium and power is medium then dkp is z

336

6. If moment is medium and power is bad then dkp is ns

337

7. If moment is bad and power is good then dkp is pb

338

8. If moment is bad and power is medium then dkp is ps

339

9. If moment is bad and power is bad then dkp is ps

M AN U

TE D

EP

340

SC

327 328 329

SIMULATION RESULTS AND DISCUSSION

342 343 344 345 346 347 348 349 350 351 352 353 354 355

The generated wind speed variation from the FAST program and used in the simulation is shown in Figure 8 (a). Since the wind speed data repeats themselves every 100 s, in Figure 8 (a), only 100 s is depicted between 500 s and 600 s. The controlled output power depending on the data is given between 0 and 2,000 s in Figure 8 (b). The output power has been stabilised at 1.5 MW because the WT has an output power of 1.5 MW. For passing of the initial unstable state of the WT, the FL-PID-C simulation has been activated at 33th s in the simulation. Because of the fact that the FL-P-C has been interfered in the system control at 100 s, a time has been given to stabilize the output power to the FLPID-C. Then, both controllers have been operated to determine the blade pitch angle. One side, the FL-PID-C has produced the blade pitch angle value that absorbs the wind speed variations, the other side, the FL-P-C has generated the individual pitch angle value that are both increase the output power quality and mitigate the moment load. In the offered system control, the FL-P-C has the ability to finetune the output power control as well as the multiplication factor to mitigate the moment load. Thus, the offered individual FL-P-C output power in the system control is better than the collective output power as shown in Figure 8 (b). Moreover, the parameters have been modified with respect to the

AC C

341

11

ACCEPTED MANUSCRIPT 356 357

fuzzy rules to keep the output at the most stable state. Thus, some mistakes that may be caused by nonlinearity of the system have been corrected.

RI PT

358

(b)

SC

(a)

Figure 8. a) The generated wind speed variation and b) the controlled output power.

(a)

(b)

TE D

360 361

M AN U

359

(c)

Figure 9. a) The average moment load variation on the blades, b) the k coefficient variations of the powermoment stabilizer FL-P-C and c) the power poverty variations in the power fluctuations.

364 365 366 367 368 369 370 371 372 373 374 375 376 377 378

The average moment load change values on the blades during the simulation are displayed in Figure 9 (a). Also, the k coefficient variations of the power-moment stabilizer FL-P-C is shown in Figure 9 (b). The power poverty variations in the power fluctuations are also given in Figure 9 (c). In the system control, the FLC has increased the k coefficient until finding the most suitable value for the output power and the average blade moments. When it finds a desired value at a wanted level, it has stopped the coefficient increase. If there is a worsening of the power or moment, which may be due to changes in operating conditions or turbine parameters, the FLC increases or decreases the k coefficient with respect to the fuzzy rules to reach a better point. It depends entirely on the fuzzy rules. Therefore, the k coefficient does not have a particular value. The WT has been initially operated with a collective pitch angle because the k coefficient, which is the output of the FL-P-C controller, is zero at first. Then, the k coefficient has been increased as shown in Figure 9 (b). This has provided a smooth transition from the collective pitch angle to the individual pitch angle for the pitch angle control of the WT. During the smooth transition, the power poverty has changed as shown in Figure 9 (c). When the k coefficient becomes a stable position, the moment and the power quality have come at the desired level limits.

379 380 381

The variations of the Kp, Ki and Kd parameters of the Sugeno-Takagi type F-PID-C that designed for the output power control are denoted in Figure 10. After a certain period of time from the start of the simulation, these parameters have gained the values that will give the best output power of the WT.

AC C

EP

362 363

12

ACCEPTED MANUSCRIPT

(a)

(b)

RI PT

382

(c)

Figure 10. The output power of the FL-PID-C; a) Kp, b) Ki, c) Kd parameter variations.

383

M AN U

SC

For the mechanical loads examined in this study, the moment value of blade I and the other moment values of blade II and blade III have been the same as because of having the balanced and defined intervals loads. Hence, the showing of the moment loads of the blades II and III has not been a need. In Figure 11, the comparing of the collective and the individual pitch angle controls in the moment on blade I is presented. Also in Table I, the moment values on blade I are given. The percentage reduction amounts of these values in Table I are indicated. These obtained values could be interpreted as follows. By means of the power-moment balance controller, a reduction of 61% in the edgewise moment (RootMxb1), a reduction of 39% in the flapwise moment (RootMyb1), a reduction of 83% in the pitching moments (RootMzb1 and RootMzc1), a reduction of 72% in the plane moment (RootMxc1) and a reduction of 34% in the out of plane moment (RootMyc1) have been recorded. These percentage reduction amounts in the moments on the blade I from 34% to 83% have testified that the power-moment balance controller operates successfully.

AC C

EP

TE D

384 385 386 387 388 389 390 391 392 393 394 395

Figure 11. The comparing of the collective and the individual pitch angle controls in the moment on blade I.

396 397

Table I. The moment values on blade I. Moments (kNm)

Individual

Collective

RootMxb1 RootMyb1 RootMzb1 RootMxc1 RootMyc1 RootMzc1

839.3918 2826.885 32.56761 734.1797 2865.467 32.56761

2150.767 4614.758 191.6559 2599.363 4345.918 191.6559

% Reduction 60.97244 38.74252 83.00725 71.554 34.06533 83.00725

13

ACCEPTED MANUSCRIPT 398 399 400 401 402 403

In Figure 12, the collective and the individual pitch angle controls and the moment variations on blade I as a function of time are illustrated. The moments on the blade I obtained from the simulation; the edgewise moment in Figure 12 (a), the flapwise moment in Figure 12 (b), the pitching moment in Figure 12 (c), the in plane moment in Figure 12 (d), the out of plane moment variations are demonstrated in detail.

SC

RI PT

404

(b)

(d)

AC C

EP

(c)

TE D

M AN U

(a)

(e) Figure 12. The collective and the individual pitch angle controls and the moment variations on blade I as a function of time; a) the edgewise moment, b) the flapwise moment, c) the pitching moment, d) the moment in plane, e) the out of plane moment.

405 406 407 408 409 410 411 412

The comparisons of the tower moments in Figure 13 (a) and the yaw moments in Figure 13 (b) are depicted. In Figure 14, the tower moment variations as a function of time also are shown in detail. In addition, these moment values obtained by means of the individual pitch angle and the collective pitch angle controls and the percentage reduction amounts of these values are given in Table II. These given values could be evaluated as follows. Through the power-moment balance controller, a reduction of 71% in the tower base roll moment (TwrBsMxt), a reduction of 84% in the tower base yaw moment (TwrBsMzt), a reduction of 62% in the local tower roll moment (TwHt1MLxt), a 14

ACCEPTED MANUSCRIPT reduction of 40% in the local tower pitching moment (TwHt1MLyt) and a reduction of 83% in the local tower yaw moment (TwHt1MLzt) have been carried out. Moreover, according to Table II, the rotating with nacelle tower to/yaw bearing roll moment (YawBrMxn), the rotating with nacelle tower top/yaw bearing pitch moment (YawBrMyn) and the tower top/yaw bearing yaw moment (YawBrMzn) have been dropped 71%, 34%, %83 in ratios, respectively. Furthermore, the nonrotating tower top/yaw bearing roll moment (YawBrMxp) of 60% and the nonrotating tower top/yaw bearing pitch moment (YawBrMyp) of 39% changes in direction of decrease have been also calculated. Finally, these results have confirmed that the power-moment balance controller has successfully mitigated the tower loads. The WT and generator system parameters used throughout the study are lastly given in Table III.

RI PT

413 414 415 416 417 418 419 420 421 422

M AN U

SC

423

TE D

(a) (b) Figure 13. The comparisons of the tower and yaw moments; a) the tower moments and b) the yaw moments.

AC C

EP

424

(a)

(b)

15

RI PT

ACCEPTED MANUSCRIPT

(c)

(d)

425 Table II. The tower moments individual

collective

735.4624

2573.347

TwrBsMzt

32.34672

197.7878

83.64575

TwHt1MLxt

838.8293

2231.766

62.4141

TwHt1MLyt

2822.565

4711.805

40.09588

TwHt1MLzt

32.34672

197.7878

83.64575

YawBrMxn

729.4143

2538.148

71.26195

YawBrMyn

2865.228

4324.47

33.74382

YawBrMzn

32.37449

187.0314

82.69034

YawBrMxp

841.6205

YawBrMyp

2826.588

M AN U

Moments (kNm)

2112.477

60.15954

4620.097

38.81972

Table III. The WT and generator parameters

EP

427 428 429

TwrBsMxt

% reduction 71.42

TE D

426

SC

Figure 14. The tower moment variations as a function of time; a) the roll moment, b) the yaw moment, c) the bearing roll moment and d) the bearing yaw moment.

Value 1.5 70 3 84 90/0 9 45 3/25 1.5 690 50

AC C

Properties Rated power of WT Rotor diameter Number of blade Hub height Max/min pitch angle Max pitch rate Rated rotor speed Cut-in/cut-out wind speed Generator rated capacity Rated stator voltage Rated frequency

Unit MW m m degree degree/s rpm m/s MW V Hz

430 431 432

CONCLUSION

433 434 435

In the study, a new FL-P-C to be implemented in order to optimize the balance between the output power and the mechanical load via the individual pitch angle control of WTs was suggested. Also, the suggested FL-P-C was verified by MATLAB/Simulink software. The FL-P-C become the first 16

ACCEPTED MANUSCRIPT application of how to include the blade moments used to provide the individual pitch angle in the output power control system of WTs. The moment values on blades of a WT were compared for the individual and the collective pitch angle controls. Moment improvements of about 34-83% were succeeded by means of the FL-P-C. On the other hand, in the WT system, a FL-PID-C was utilized for output power control over nominal wind speed values since the FL-PID-C is suitable for nonlinear system as well as giving fast response. The suggested FL-P-C for the output power quality and the moment value stabilisation would be applied to all conventional and modern control systems that produce the collective pitch angle value. Ultimately, in the study, the individual pitch angle value was defined through adding the blade moments to the collective pitch angle at a certain ratio as a new and effective control approach. In earlier executed studies, the addition ratio of the moments on blades was entirely left to designer preferences to decide. However, in this study, the ratio was determined and automated by the suggested FL-P-C.

448

REFERENCES

449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486

[1]

[6] [7]

[8] [9] [10]

[11] [12]

[13]

[14] [15]

SC

M AN U

[5]

TE D

[4]

EP

[3]

T. Senjyu, R. Sakamoto, N. Urasaki, H. Higa, K. Uezato, and T. Funabashi, "Output power control of wind turbine generator by pitch angle control using minimum variance control," Electrical Engineering in Japan, vol. 154, no. 2, pp. 10-18, 2006. P. Idowu, "Fuzzy Speed Regulation of A Wind-Turbine DC Generator," Intelligent Automation & Soft Computing, vol. 2, no. 1, pp. 27-36, 1996. A. Musyafa, A. Harika, I. Negara, and I. Robandi, "Pitch angle control of variable low rated speed wind turbine using fuzzy logic controller," Int. J. Eng. Technol, vol. 10, pp. 21-24, 2010. X. Jian-jun, Y. Li-mei, Q. Xiao-ning, J. Chun-lei, and W. Jian-ren, "Study of variable-pitch wind turbine based on fuzzy control," in Future Computer and Communication (ICFCC), 2010 2nd International Conference on, 2010, vol. 1, pp. V1-235-V1-239: IEEE. J. Qi and Y. Liu, "PID control in adjustable-pitch wind turbine system based on fuzzy control," in Industrial Mechatronics and Automation (ICIMA), 2010 2nd International Conference on, 2010, vol. 2, pp. 341-344: IEEE. Z. Civelek, M. Lüy, E. Çam, and N. Barışçı, "Control of Pitch Angle of Wind Turbine by Fuzzy Pid Controller," Intelligent Automation & Soft Computing, pp. 1-9, 2015. B. Xiao and S. Yin, "Velocity-Free Fault-Tolerant and Uncertainty Attenuation Control for a Class of Nonlinear Systems," IEEE Transactions on Industrial Electronics, vol. 63, no. 7, pp. 4400-4411, 2016. E. Bossanyi, "Individual blade pitch control for load reduction," Wind energy, vol. 6, no. 2, pp. 119-128, 2003. E. Bossanyi, "Further load reductions with individual pitch control," Wind energy, vol. 8, no. 4, pp. 481-485, 2005. K. Selvam, S. Kanev, J. W. van Wingerden, T. van Engelen, and M. Verhaegen, "Feedback– feedforward individual pitch control for wind turbine load reduction," International Journal of Robust and Nonlinear Control, vol. 19, no. 1, pp. 72-91, 2009. E. Bossanyi, "Wind turbine control for load reduction," Wind energy, vol. 6, no. 3, pp. 229244, 2003. H. De Battista, R. Mantz, and C. Christiansen, "Energy-based approach to the output feedback control of wind energy systems," International Journal of Control, vol. 76, no. 3, pp. 299-308, 2003. W. Leithead, S. De la Salle, and D. Reardon, "Role and objectives of control for wind turbines," in IEE Proceedings C (Generation, Transmission and Distribution), 1991, vol. 138, no. 2, pp. 135-148: IET. F. D. Bianchi, H. De Battista, and R. J. Mantz, Wind turbine control systems: principles, modelling and gain scheduling design. Springer Science & Business Media, 2006. T. J. Larsen, H. A. Madsen, and K. Thomsen, "Active load reduction using individual pitch, based on local blade flow measurements," Wind Energy, vol. 8, no. 1, pp. 67-80, 2005.

AC C

[2]

RI PT

436 437 438 439 440 441 442 443 444 445 446 447

17

ACCEPTED MANUSCRIPT

[21]

[22]

[23]

[24]

[25]

[26] [27]

[28] [29] [30] [31] [32]

[33] [34] [35] [36]

[37]

RI PT

[20]

SC

[19]

M AN U

[18]

TE D

[17]

D. Zhang, P. Cross, X. Ma, and W. Li, "Improved control of individual blade pitch for wind turbines," Sensors and Actuators A: Physical, vol. 198, pp. 8-14, 2013. H. Liu, Q. Tang, Y. Chi, Z. Zhang, and X. Yuan, "Vibration reduction strategy for wind turbine based on individual pitch control and torque damping control," International Transactions on Electrical Energy Systems, 2016. B. Han, L. Zhou, F. Yang, and Z. Xiang, "Individual pitch controller based on fuzzy logic control for wind turbine load mitigation," IET Renewable Power Generation, vol. 10, no. 5, pp. 687-693, 2016. T. Pan and Z. Ma, "Wind turbine individual pitch control for load reduction based on fuzzy controller design," Proceedings of the Institution of Mechanical Engineers, Part I: Journal of Systems and Control Engineering, vol. 227, no. 3, pp. 320-328, 2013. M. Jureczko, M. Pawlak, and A. Mężyk, "Optimisation of wind turbine blades," Journal of materials processing technology, vol. 167, no. 2, pp. 463-471, 2005. T. Barlas, G. Van Der Veen, and G. Van Kuik, "Model predictive control for wind turbines with distributed active flaps: incorporating inflow signals and actuator constraints," Wind Energy, vol. 15, no. 5, pp. 757-771, 2012. M. A. Lackner and G. van Kuik, "A comparison of smart rotor control approaches using trailing edge flaps and individual pitch control," Wind Energy, vol. 13, no. 2‐3, pp. 117-134, 2010. I. Houtzager, J.-W. van Wingerden, and M. Verhaegen, "Rejection of periodic wind disturbances on a smart rotor test section using lifted repetitive control," IEEE Transactions on Control Systems Technology, vol. 21, no. 2, pp. 347-359, 2013. C. Plumley, W. Leithead, P. Jamieson, E. Bossanyi, and M. Graham, "Comparison of individual pitch and smart rotor control strategies for load reduction," in Journal of Physics: Conference Series, 2014, vol. 524, no. 1, p. 012054: IOP Publishing. B. F. Ng, R. Palacios, E. C. Kerrigan, J. M. R. Graham, and H. Hesse, "Aerodynamic load control in horizontal axis wind turbines with combined aeroelastic tailoring and trailing‐edge flaps," Wind Energy, vol. 19, no. 2, pp. 243-263, 2016. M. Zhang, B. Tan, and J. Xu, "Smart load control of the large-scale offshore wind turbine blades subject to wake effect," Science Bulletin, vol. 60, no. 19, pp. 1680-1687, 2015. M. Zhang, H. Yang, and J. Xu, "Numerical investigation of azimuth dependent smart rotor control on a large-scale offshore wind turbine," Renewable Energy, vol. 105, pp. 248-256, 2017. K.-S. Tang, K. F. Man, G. Chen, and S. Kwong, "An optimal fuzzy PID controller," IEEE Transactions on Industrial Electronics, vol. 48, no. 4, pp. 757-765, 2001. M. Mizumoto, "Realization of PID controls by fuzzy control methods," Fuzzy sets and systems, vol. 70, no. 2-3, pp. 171-182, 1995. J. Carvajal, G. Chen, and H. Ogmen, "Fuzzy PID controller: Design, performance evaluation, and stability analysis," Information sciences, vol. 123, no. 3, pp. 249-270, 2000. Z.-W. Woo, H.-Y. Chung, and J.-J. Lin, "A PID type fuzzy controller with self-tuning scaling factors," Fuzzy sets and systems, vol. 115, no. 2, pp. 321-326, 2000. H. Mamur, "Design, application, and power performance analyses of a micro wind turbine," Turkish Journal of Electrical Engineering & Computer Sciences, vol. 23, no. 6, pp. 16191637, 2015. W. Tong, Wind power generation and wind turbine design. WIT press, 2010. A. Hemami, Wind turbine technology. Cengage Learning, 2011. Z. Wenjing and X. Hongze, "Active disturbance rejection based pitch control of variable speed wind turbine," in Control Conference (CCC), 2011 30th Chinese, 2011, pp. 5094-5098: IEEE. C. H. Chen, C.-M. Hong, and T.-C. Ou, "Hybrid fuzzy control of wind turbine generator by pitch control using RNN," International Journal of Ambient Energy, vol. 33, no. 2, pp. 56-64, 2012. A. M. S. Hwas and R. Katebi, "Wind turbine control using PI pitch angle controller," in IFAC Conference on Advances in PID Control PID'12, 2012.

EP

[16]

AC C

487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539

18

ACCEPTED MANUSCRIPT

TE D

M AN U

SC

RI PT

Z. Civelek, E. Çam, M. Lüy, and H. Mamur, "Proportional–integral–derivative parameter optimisation of blade pitch controller in wind turbines by a new intelligent genetic algorithm," IET Renewable Power Generation, vol. 10, no. 8, pp. 1220-1228, 2016.

EP

[38]

AC C

540 541 542 543 544 545 546

19

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

Research Highligths 1-) A new FL-P-C to be implemented in order to optimize the balance between the output power and the mechanical load via the individual pitch angle control of WTs was suggested. 2-) For stability of electrical power of the WT, an extra FL-PID controller also was designed since it is suitable for nonlinear systems in terms of fast response. 3-) The controllers have been suggested for operations of mechanical loads reduction and stabilizing electrical outputs.

ACCEPTED MANUSCRIPT Reviewers' comments: Reviewer #1:

RI PT

In this paper, the authors have investigated the method to design a new fuzzy logic proportional controller (FL-P-C) to decrease moment loads on the blades and tower while keeping the output power of wind turbine systems at a constant value. The controller has been designated and simulated in MATLAB/Simulink programme. The paper is well organized. Overall, the paper can be accepted after minor revision.

SC

1. The main result is presented by integrating PID controller and fuzzy logic control. The method of fuzzy PID has been well investigated in the existing literature. Hence, the authors should clarify the motivation of designing a fuzzy PID controller. Moreover, the novelty and the main contribution of this paper should be stressed.

M AN U

The main purpose of this study is the individual control of the wind turbine blades. As the referee has said, the Fuzzy PID controller has been used in the literature in many areas. However, the designed controller was applied for the first time in this area. In addition, the parameters selected as the controller input were also used as input for the first time. Furthermore, no study has been conducted on the contribution and control of the Myc moments to the individual pitch angle. This part is also merged with Fuzzy PID for the first time. All the above sections show the innovations of the controller.

TE D

2. The literature review in this paper seems to miss the recent developments in terms of the investigated topic. The authors should better position the current paper among existing works and reassess the contributions. In line with the referee's suggestion, the literature has been re-examined and updated works have been added.

EP

3. The paper is too long, please present the paper concisely.

AC C

It is desired to shorten some places in the direction of suggestions. However, due to additional comments the other referee wanted, the article remained the same size.

4. When designing controller, the reviewer would say that almost all the systems are subject to uncertainty and external disturbance, so the wind turbine systems. Therefore, the authors are suggested to take these two issues into consideration in future works with the works "Tracking control of robotic manipulators with uncertain kinematics and dynamics", "Disturbance Observer based Integral Sliding Mode Control for Systems with Mismatched Disturbances", and "Velocity-free Fault tolerant and uncertainty attenuation control for a class of nonlinear systems" cited.

The proposal will be studied in the future. We thank the referee for his valuable comment

1

ACCEPTED MANUSCRIPT 5. Please improve the quality of all figures. Figures were re-examined and improved.

RI PT

Reviewer #2:

SC

Dear Editor, This paper presented a numerical study on the development of fuzzy controller for IPC fatigue control on a wind turbine blade. However, the development process is very vague. Plus, the physics behind the load mitigation is also not clear. In addition, the creative points in terms of fuzzy controller is not indicated. These are the main weaknesses of the paper. The authors need to go ahead to fix them. Additionally, the authors also needed to clarify the following points

AC C

EP

TE D

M AN U

（1）In the literature summary part, some other papers, concerned with advanced control schemes to control the fluctuating fatigue load on single turbine blade and even mutli turbine blades, should be mentioned and evaluated for completeness. For example, [1] Lackner MA, van Kuik GAM. A comparison of smart rotor control approaches using trailing edge flaps and individual pitch control. Wind Energy 2010; 13:117-34. [2] Barlas TK, van der Veen GJ, van Kuik GAM. Model predictive control for wind turbines with distributed active flaps: incorporating inflow signals and actuator constraints. Wind Energy 2012; 15:757-71. [3] Houtzager I, van Wingerden JW, Verhaegen M. Rejection of periodic wind disturbances on a smart rotor test section using lifted repetitive control. IEEE T Contr Sys T 2013; 21:347-59. [4] Plumley C, Leithead W, Jamieson P, Bossanyi E, Graham M. Comparison of individual pitch and smart rotor control strategies for load reduction. J Phys. 2014; 524:012054. [5] Ng BF, Palacios R, Kerrigan EC, Graham JMR, Hesse H. Aerodynamic load control in horizontal axis wind turbines with combined aeroelastic tailoring and trailing-edge flaps. Wind Energy 2016; 19:243-63. [6] Zhang M.M., Tan B., Xu J.Z. 2015 Smart Load Control of the Large-scale Offshore Turbine Blades subject to Wake Effect. Science Bulletin, 60(19), 1680-1687. [7] Zhang M.M., Yang H.L., Xu J.Z. 2017 Numerical Investigation of Azimuth Dependent Smart Rotor Control on a Large-scale Offshore Wind Turbine. Renewable Energy, 105, 248-256. The literature has been reviewed and some suggested sources have been added. （2）Please give a detailed deduction of Eqs. (4) and (5). "λi" in equations 4 and 5 is a transitional variable used to simplify the formulas. There is no interest in any physical parameter in the wind turbines.

（3）Please give the relationships among different dynamic loads that the authors mentioned. 2

ACCEPTED MANUSCRIPT We tried to solve the referee's wish by referring to the literature since the first referee wanted to shorten the work and also the number of pages of the work increased.

（4）The detailed information on the development of the fuzzy controller should be given, e.g. control target, adjustment of controller coefficients and control law. Plus, the control process should be given in terms of block diagram to clearly clarify the description.

RI PT

The developed fuzzy control block diagram was given in Figure 6. Membership functions of power and moment inputs were also given in Figures 7 (a) and (b). The output membership function was also given in Figure 7 (c). Additionally, the relationship between the inputs and the output was given in Figure 7 (d). All of these are available on the study.

SC

（5）What is the creative idea of the developed fuzzy controller compared with the previous ones in the past? Please add more description on this.

M AN U

When reviewing the literature, it has not been found that a fuzzy controller designed to take into account the contribution rate of output power quality and momentum amount to Myc moments for individual control in any of the past studies. Therefore this part is the difference of the study.

（6）From Fig. 8(b). why the fluctuation in the output power was enhanced within the period of 100-500s? What has happened?

EP

TE D

As the FL-P controller reaches the target value as indicated in line 360 of 13rd page of the study (as seen in Figure 9(b), this value reaches 500th seconds), it makes the output power more stable. This is because the fuzzy PID controller, which controls the output power, uses fine tuning capability over the coefficients. Therefore, up to 500 seconds, both the systems have the same oscillations on their outputs.

AC C

（7）How to obtain a optimal control Kp, Ki and Kd using Fuzzy controller? What is the physics behind their variations in Fig. 10? Due to the page limitations of the work, the relevant explanations were cited with reference.

（8）Why the edgewise and flapwise moments can be simultaneously reduced by the IPC control? Please give a detailed analysis on the control physics behind it. Due to the page limitations of the work, the relevant explanations were cited with reference.

（9）From Figs. 14(b)-(d), the fluctuations in the yaw moments of tower base, tap and bearing evidently increased, suggesting the fault controller. Please explain why? In Figure 14 (b) and (d), although the fluctuations in the individual control were greater than the fluctuations in the collective control, the average values were less in the individual control both at 3

ACCEPTED MANUSCRIPT the tower base yaw moment and tower bearing yaw moment. The controller in this study was designed to reduce average values. Therefore, the controller worked according to purpose.

The work was conducted and re-organized in English.

RI PT

（10）The English writing in this paper was poor and a native English speaker was required to improve the quality.

Finally, we wish we were able to improve the quality of the article in this second submission; thanks to the Reviewers’ and Editors’ comments. We look forward to having our article published in your journal at the soonest. Please do not hesitate to contact me for any clarifications or inquiries.

SC

Sincerely yours,

AC C

EP

TE D

M AN U

Zafer Civelek

4