See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/350365823

Rotor equivalent wind speed calculation method based on equivalent power

considering wind shear and tower shadow

Article in Renewable Energy · March 2021

DOI: 10.1016/j.renene.2021.03.089

CITATIONS READS

7 192

8 authors, including:

Liu Yongqian Yanhui Qiao

North China Electric Power University North China Electric Power University
199 PUBLICATIONS 3,915 CITATIONS 10 PUBLICATIONS 216 CITATIONS

SEE PROFILE SEE PROFILE

Shuang Han Tao Tao

81 PUBLICATIONS 1,350 CITATIONS

China Southern Power Grid Technology Co., Ltd.
14 PUBLICATIONS 85 CITATIONS
SEE PROFILE
SEE PROFILE

Some of the authors of this publication are also working on these related projects:

wind power forecasting View project

Монгол орны нөхцөлд BAPV ба BIPV технологийг ашиглах боломжийн судалгаа View project

All content following this page was uploaded by Jie Yan on 07 June 2021.

The user has requested enhancement of the downloaded file.

1 Rotor equivalent wind speed calculation method based on equivalent power
2 considering wind shear and tower shadow
3 Yongqian Liu a, Yanhui Qiao a, Shuang Han a, *, Tao Tao a, Jie Yan a, Li Li a, Galsan Bekhbat b,
4 Erdenebat Munkhtuya b
5
6 a
State Key Laboratory of Alternate Electrical Power System with Renewable Energy Sources (NCEPU), School of New Energy, North China
7 Electric Power University, Beijing 102206, China
8 b
School of Power Engineering, Mongolian University of Science and Technology, Ulaanbaatar, Mongolia
9 * E-mail address: hanshuang1008@sina.com
10
11 H I G H LI G HTS
12
13  Rotor equivalent wind speed calculation model based on equivalent power is proposed.
14  Wind shear, tower shadow and wind turbine operation characteristics are considered.
15  The proposed model matches the actual power control effect of wind turbine.
16  The proposed model contains the power loss caused by actual power control deviation.
17  The proposed method can improve theoretical power accuracy verified by real cases.
18
19 AB STRACT
20
21 Wind speed spatial variation in the whole wind wheel rotation plane caused by wind shear and tower shadow effect becomes
22 more obvious with the increase of rotor diameter and tower height, thus the hub height wind speed is not sufficient to represent
23 the actual wind speed effect and the rotor equivalent wind speed should be used. However, the existing rotor equivalent wind
24 speed calculation models do not consider wind turbine operation characteristics, which cannot match the actual power control
25 effect, thus ignoring the power loss caused by actual power control deviation. To address this problem, a rotor equivalent wind
26 speed calculation method based on equivalent power is proposed in this paper, which comprehensively considers wind shear,
27 tower shadow and wind turbine operation characteristics. The rationality of the proposed model was qualitatively evaluated
28 from different power control stages of wind turbine and the validity was quantitatively verified from the aspect of theoretical
29 power calculation. The results demonstrated that the proposed model can effectively reflect the variation law of rotor
30 equivalent wind speed under the actual control effect of different power control stages, which contains the power loss caused
31 by actual power control deviation, thus improving the theoretical power calculation accuracy.
32
33 Keywords: Wind turbine; equivalent power; rotor equivalent wind speed; wind shear; tower shadow
34
35 1. Introduction
36 Wind energy is one of the most significant and potentially useful renewable energy, which has attracted widespread
37 attention all over the world and has developed rapidly in recent years [1]. The global installed wind power capacity is
38 estimated to exceed 800 GW by the end of 2021 [2]. In the context of the rapid development of wind power industry and the
39 continuous innovation of wind power manufacturing technology, wind turbines present the trend of large-scale development
40 with increasing blade length and tower height [3]. Wind speed spatial variation in the whole wind wheel rotation plane caused
41 by the combined effect of wind shear and tower shadow increases with the increase of rotor diameter and tower height [4-6],
42 and its impact on the generation performance of wind turbines is more and more obvious and can not be ignored [7]. Therefore,
43 only the hub height wind speed is not sufficient to represent the actual wind speed effect in the whole wind wheel rotation
44 plane [8], and the rotor equivalent wind speed considering wind shear and tower shadow effects should be used, which is of
45 great significance to improve the accuracy of wind power prediction and theoretical power calculation of wind turbines.
46 Many scholars are dedicated to the study of wind shear and tower shadow effects. Shen et al. (2011) [9] used a lifting
47 surface method with time marching free wake model to investigate the periodic unsteady nature in the wind shear. Holtslag et
48 al. (2017) [10] extended the diabatic surface layer wind shear model for offshore wind energy purposes to higher altitudes
49 based on Gryning's wind profile and the resistance functions proposed by Byun, and analyzed the influence of adopting this
50 extended shear profile for wind energy in terms of the kinetic energy flux and blade root fatigue loads experienced by a wind
51 turbine. Sintra et al. (2014) [11] constructed wind shear and tower shadow models of three-blade variable speed wind turbines
52 in Matlab/Simulink, and studied the model construction methods of reducing mechanical stress and torque oscillation and
53 improving aerodynamic efficiency under rated wind speed. Zhang et al. (2014) [12] studied the aerophysical characteristics of
54 wind turbine intelligent load control caused by extreme wind shear. Gualtieri et al. (2016) [13] used the wind shear coefficient
55 of atmospheric stability to improve the precision of wind resource extrapolation. Pedersen et al. (2012) [14] studied the effect
56 of tower shadow on the fatigue life of wind turbine blades, and performed Monte Carlo simulation method for random
57 analysis.

1
 58 In addition, many literature have studied the combined effects of wind shear and tower shadow on the aerodynamic
59 characteristics and power quality of wind turbines by constructing the rotor equivalent wind speed calculation model that take
60 into account both wind shear and tower shadow effects. Rotor equivalent wind speed refers to the wind speed that can
61 represent the actual wind speed effect in the whole wind wheel rotation plane. The existing rotor equivalent wind speed
62 calculation methods are mainly divided into two categories: (1) Rotor equivalent wind speed calculation method based on
63 kinetic energy theory. Choukulkar et al. (2016) [15] constructed a rotor equivalent wind speed model considering wind shear
64 effect, wind direction fluctuation and turbulence intensity, and analyzed the influence of various factors on wind power. Clack
65 et al. (2016) [16] constructed a rotor equivalent wind speed model that takes into account vertical wind shear and directional
66 shear, and applied it to wind energy resource assessment. Redfern et al. (2019) [17] constructed a rotor equivalent wind speed
67 model based on kinetic energy theory, and applied the rotor equivalent wind speed to wind farm parametric weather prediction
68 model. Wagner et al. (2014) [18] used the calculation method of rotor equivalent wind speed based on kinetic energy theory in
69 IEC 61400-12-1 standard, and compared the difference between the hub height wind speed and the rotor equivalent wind speed
70 for wind power curve modeling under different conditions. Scheurich et al. (2016) [19] constructed a rotor equivalent wind
71 speed calculation model based on kinetic energy theory using the measured wind speed of different heights of laser radar, and
72 used it for wind turbine power curve test. Kavari et al. (2019) [20] studied the effects of wind shear on aerodynamic
73 performance and energy generation of horizontal-axis wind turbine based on improved momentum-blade element theory. Jeon
74 et al. (2017) [21] proposed a method for calculating the rotor equivalent wind speed based on the measured wind speed in the
75 cabin, and used to improve the performance prediction accuracy of wind turbine under wake conditions. (2) Rotor equivalent
76 wind speed calculation method based on equivalent torque. Dolan et al. (2006) [22] constructed an equivalent wind speed
77 calculation model considering both wind shear and tower shadow effects. The simulation results show that tower shadow effect
78 is the dominant factor causing wind turbine torque 3P oscillation when the dynamic torque is determined. Wan et al. (2015) [23]
79 constructed the wind turbine model, yaw error model and equivalent wind speed model including wind shear and tower
80 shadow effects, and studied the influence of yaw error on the operation characteristics of wind turbine under different wind
81 speeds and control stages. Fajardor et al. (2011) [24] constructed the per-blade equivalent wind speed model considering wind
82 shear and tower shadow effects, and used it to model and simulate the high-order transmission system. The simulation results
83 show that the wind speed determined by the per-blade equivalent wind speed model in wind turbine high-order transmission
84 system is less conservative than that determined by hub height. Wan et al. (2017) [25] constructed the equivalent wind speed
85 calculation model of large-scale n-blade wind turbine considering wind shear and tower shadow effects, and analyzed the
86 variation of equivalent wind speed with the parameters such as rotor radius, hub height, tower radius, distance between blade
87 and tower center, wind shear coefficient and number of blades.
88 In summary, many scholars have studied wind shear and tower shadow effects, and proposed the rotor equivalent wind
89 speed calculation methods based on kinetic energy theory and equivalent torque, but there are still some problems as follows.
90 (1) The rotor equivalent wind speed calculation method based on kinetic energy theory requires the measured wind speed and
91 wind direction of multiple heights in the wind wheel rotation plane, the commonly used anemometer tower data cannot
92 meet the calculation requirements, and the effect of tower shadow is not considered in this method.
93 (2) The rotor equivalent wind speed calculation methods based on kinetic energy theory and equivalent torque are ideal
94 methods, which ignore wind turbine operation characteristics. Therefore, the actual power control effect of wind turbine
95 and the power loss caused by actual control deviation cannot be reflected.
96 To solve the aforementioned problems, considering wind shear, tower shadow and wind turbine operation characteristics,
97 the rotor equivalent wind speed calculation method based on equivalent power is proposed in this paper. The main
98 contributions of this study are as follows:
99 (1) The rotor equivalent wind speed calculation method based on equivalent power was proposed. The proposed method
100 comprehensively takes into account wind shear, tower shadow and wind turbine operation characteristics, and the rotor
101 equivalent wind speed has the same output power with the actual wind speed in the whole wind wheel rotation plane.
102 (2) The rationality of the proposed model was qualitatively evaluated from different power control stages using the design
103 data of wind turbines with different capacities. Compared with the method based on equivalent torque, the proposed
104 model can effectively reflect the variation law of rotor equivalent wind speed under the actual control effect of different
105 control stages, which matches the actual power control effect of wind turbine and contains the power loss caused by
106 actual power control deviation.
107 (3) The validity of the proposed model was quantitatively verified from the perspective of wind turbine theoretical power
108 calculation using the design parameters and measured data of 2MW and 5MW wind turbines in different seasons.
109 Compared with the rotor equivalent wind speed calculation method based on equivalent torque and hub height wind speed,
110 the proposed model can effectively improve the theoretical power calculation accuracy.
111 This paper is organized as follows. Section 2 analyzes the influence of wind shear and tower shadow on the wind speed
112 spatial distribution. Section 3 introduces the rotor equivalent wind speed calculation method based on equivalent torque, which
113 is used as the comparative benchmark model. Section 4 describes the construction method of rotor equivalent wind speed
114 calculation model based on equivalent power. Section 5 verifies the rationality and validity of the proposed model from
115 different power control stages and theoretical power calculation. Section 6 concludes this paper.
116 2. Wind speed spatial distribution due to wind shear and tower shadow

2
117 2.1. Wind shear
118 Wind shear effect refers the property that wind speed changes obviously with the increase of vertical height above the
119 ground due to the influence of vegetation and buildings on the air motion in the atmospheric boundary layer. The higher the
120 vertical height from ground, the smaller the impact of the ground obstacles, and the greater the wind speed. Wind shear effect
121 is usually described by an exponential model [26] as follows.

Z
VZ  VH   (1)
H
122 Where VZ is the average wind speed at the height Z above the ground, VH is the average hub height wind speed, H is the
123 hub height of wind turbine,  is the wind shear coefficient, which is mainly affected by surface roughness.

124
125 Fig. 1 Wind turbine structural parameters.
126 The structural parameters of the three-blade horizontal axis wind turbine are defined as shown in Fig. 1. Taking into
127 account the influence of wind shear effect on the wind speed spatial distribution in the wind wheel rotation plane, the formula
128 (1) is converted to a function of r (the radial distance between blade element and rotor axis) and  (the blade azimuth angle),
129 and Wws is defined as the disturbance coefficient imposed by the wind shear effect on the hub height wind speed, as shown in
130 formula (2).

 r cos   H 
V  r ,   V H    VH 1  Wws r ,  (2)
 H 
131 Where V r ,  is the wind speed where the radial distance between blade element and rotor axis is r, and the blade
132 azimuth angle is  ; Wws is the wind shear disturbance coefficient, which characterizes the influence degree of wind shear
133 effect on the wind speed at each point in the wind wheel rotation plane.
134 The wind shear disturbance coefficient is the partial sum of the last  terms of the Taylor series.
   1  r 
2
 r 
Wws r ,      cos   2
  cos 
H 2 H
n
(3)
   1  2   r 
3    j  1  r 
j 1
n
3 n
   cos       cos 
6 H
  n! H
 
135 For the three-blade wind turbine, due to the symmetrical distribution of three blades, wind shear effect causes 3P
136 oscillation of aerodynamic torque. The third-order Taylor series expansion can well reflect the wind shear effect, and the
137 subsequent effects are relatively small and can be ignored [8]. Therefore, the approximate expression of wind shear disturbance
138 coefficients by the third-order Taylor series expansion can be easily calculated while retaining the non-linear characteristics of
139 model, which can be expressed as follows.
 r     1  r  2    1   2  r  3 3
Wws  r ,      cos     cos 2
    cos  (4)
H 2 H 6 H

3
   r     1  r  2    1   2  r  3 3 
V  r ,   VH 1     cos   2
  cos     cos   (5)
 H 2 H 6 H 
140 Assuming that the hub height wind speed is 11 m/s, the wind shear coefficient is 0.3, and the ratio of the radial distance
141 between blade element and rotor axis to hub height is 2/5, 1/2 and 2/3 respectively, the local wind speed of each blade element
142 under different blade azimuth angles is calculated. As shown in Fig. 2, during the process of blade rotation, the local wind
143 speed at each blade element varies periodically with the change of blade azimuth angle, and the greater the radial distance
144 between blade element and rotor axis, the greater the variation of wind speed. Taking blade element at r/H=2/3 as an example,
145 the local wind speed first decreases and then increases with the increase of blade azimuth. When blade element is located at the
top (   0 ), the maximum local wind speed is 12.88m/s, which is 1.88m/s different from the hub height wind speed. When

146
blade element is located at the bottom (   180 ), the minimum local wind speed is 8.09m/s, which is 2.91m/s different from

147
148 the hub height wind speed. The fluctuation of local wind speed in the lower half plane of wind wheel is larger than that in the
149 upper half plane, which corresponds to the exponential distribution of the average wind profile. The main reason is that the
150 closer the wind speed is to the ground, the greater the influence of surface roughness, the more obvious the spatial variation of
151 wind speed. The maximum fluctuation amplitude of local wind speed is 4.79m/s, which is about 43.5% of the hub height wind
152 speed, which indicates that the wind speed spatial distribution in the wind wheel rotation plane has obvious inhomogeneity,
153 which will produce large fluctuation of blade load, aggravate the fatigue loss of blade and cause periodic fluctuation of wind
154 turbine power output.

155
156 Fig. 2 The variation of wind speed with blade azimuth angle under different radial distances(Wind shear).
157 Assuming that the hub height wind speed is 11 m/s, the ratio of the radial distance between blade element and rotor axis to
158 hub height is 2/3, and the wind shear coefficient is 0.1, 0.2 and 0.3 respectively, the local wind speed of each blade element
159 under different blade azimuth angles is calculated. As shown in Fig. 3, during the process of blade rotation, the local wind
160 speed at each blade element varies periodically with the change of blade azimuth angle, and the greater the wind shear
161 coefficient, the greater the wind speed fluctuation. When   0.1 , the maximum fluctuation amplitude of local wind speed is
162 1.65m/s. When   0.2 , the maximum fluctuation amplitude of local wind speed is 3.25m/s. When   0.3 , the maximum
163 fluctuation amplitude of local wind speed is 4.79m/s. The maximum fluctuation amplitude of local wind speed increases
164 approximately linearly with the increase of wind shear coefficient, which indicates that wind shear effect is the key factor
165 causing the uneven spatial distribution of wind speed in the wind wheel rotation plane. For the wind farms in low wind speed
166 region, wind turbines are mostly located in the region of highland with complex terrain and harsh climate, which results in a
167 large wind shear coefficient. The wind speed spatial distribution is more uneven due to the large wind shear effect, which has a
168 more significant impact on wind turbine power output and can not be ignored.

4
169
170 Fig. 3 The variation of wind speed with blade azimuth angle under different wind shear coefficients(Wind shear).
171 2.2. Tower shadow
172 Tower shadow effect is a negative effect in the process of wind power generation. When the wind flows through tower,
173 the magnitude and direction of wind vector vary duo to the change of airflow distribution. It is generally believed that the
174 tower shadow effect only affects the wind speed spatial distribution in the lower half plane of wind wheel, that is, within the
175 blade azimuth angle    90, 270  . Taking into account the tower shadow effect, the wind speed at any point in the lower
176 half plane of wind wheel can be expressed as follows [27].
V  y , x   VH  Vts  y , x   VH 1  Wts  y , x  (6)
y 2  x2 r 2 sin2   x 2
Vts  y , x   V0a 2  V a 2
(7)
 y 2  x2  r 2 sin2   x 2 
2 0 2

177 Where Vts is the wind speed fluctuation value caused by tower shadow effect, y is the horizontal distance between blade
178 element and rotor axis ( y  r sin  ), x is the distance from the wind wheel rotation plane to the tower center, and a is the
179 tower radius, as shown in Fig. 1. Wts is the tower shadow disturbance coefficient, which characterizes the influence degree of
180 the tower shadow effect on the wind speed at each point in the lower half plane of wind wheel. V0 is the spatial average wind
181 speed in the wind wheel rotation plane, which is calculated by formula (8).
1 2 R     1  R  2 
V0  0 0 H 
V  1  W ws  r ,   
 rd rd   V H  1      m VH (8)
 R2  8  H  
182 It can be seen from formula (7). When calculating Vts , the reference wind speed chooses the spatial average wind
183 speed V0 in the wind wheel rotation plane instead of the hub height wind speed VH . This method will introduce the wind shear
184 effect into the analysis of tower shadow effect, but V0 and VH are extremely close, although there are certain defects in the
185 physical sense, it does not introduce too much deviation. Combining formulas (7) and (8) and substituting formulas (6), the
186 wind speed and tower shadow disturbance coefficient at each point in the lower half plane of wind wheel considering tower
187 shadow effect are obtained as follows.
 
r 2 sin 2   x 2 
V  r ,   VH 1  ma 2 (9)
  r 2 sin 2   x2  
2


r 2 sin 2   x 2
Wts  r ,   ma 2 (10)
r sin 2   x 2 
2 2

188 Based on the structural parameters of 7MW wind turbine, the hub height is 85.8m, the rotor radius is 77.8m, the distance
189 from the wind wheel rotation plane to the tower center is 5.5m, the tower radius is 1.7m. Assuming that the hub height wind
190 speed is 11 m/s, the wind shear coefficient is 0.3, and the radial distance between blade element and rotor axis is 5m, 30m, 55m
191 and 75m respectively, the local wind speed of each blade element under different blade azimuth angles is calculated, as shown
192 in Fig. 4.

5
193
194 Fig. 4 The variation of wind speed with blade azimuth angle under different radial distances(Tower shadow).
195 Fig. 4 shows that the same blade element is periodically affected by the tower shadow in the rotation process, which can
196 be divided into three stages according to the influence degree of tower shadow effect. When    0,90  and
197    270,360  , the tower shadow has almost no effect on local wind speed and can be ignored. When
198 180  arcsin( x / r)    180  arcsin( x / r) , which is called the center area of tower shadow, the local wind speed
199 decreases sharply and reaches the minimum at 180 ° due to the blocking effect of tower. When
200 90    180  arcsin( x / r) and 180  arcsin( x / r)    270 , which are called the acceleration area, due to the
201 narrow tube effect between tower and blade, the airflow in this area will be squeezed, and the local wind speed will increase
202 slightly, reaching the local maximum at the junction of the center area of tower shadow and the acceleration area. As the radial
203 distance between blade element and rotor axis increases, the blade azimuth range affected by the tower shadow effect gradually
204 decreases. When the radial distance r=5, the difference between local wind speed and hub height wind speed is large. Which is
205 most likely because the airflow is also blocked by the hub and the local wind speed is reduced.
206 2.3. Wind shear and tower shadow
207 Wind turbines are affected by both wind shear and tower shadow effects in the actual operation process. The wind speed
208 in the upper half plane of wind wheel, that is, the wind speed within the blade azimuth angle    0,90  and
209    270,360  , which is mainly affected by wind shear effect. The wind speed at any point can be expressed as:
  r     1  r  2    1   2  r  3 3 
V  r ,   VH 1     cos  
2
  cos     cos   (11)
 H  2 H 6 H 
210 The wind speed in in the lower half plane of wind wheel, that is, the wind speed within the blade azimuth angle
211    90, 270  , which is mainly affected by both wind shear and tower shadow effects. The wind speed at any point can be
212 expressed as:
V  r, , x  VH 1  Wws  r,   1  Wts  r, , x 
(12)
 VH 1  Wws  r,   Wts  r, , x  Wws r, Wts r, , x 
213 Since the value of item Wws  r , Wts  r , , x  is extremely small and can be ignored, it is simplified as:
V  r, , x  VH 1  Wws  r,   Wts  r, , x  (13)
214 By substituting formulas (4) and (10) into formulas (13), the wind speed at any point in the lower half plane of wind
215 wheel can be obtained:
 r    1  r 2 2 
1   cos    cos  
 H 2  H 
V  r, , x  VH   (14)
 
   1   2  r 3
r2 sin 2   x2
  cos 3
  ma 2 
 r 2 sin 2   x2 
2
 6 H  
 
216 Based on the above-mentioned structural parameters of 7MW wind turbine, assuming that the hub height wind speed is
217 11m/s and the wind shear coefficient is 0.1, and the radial distance between blade element and rotor axis is 30m, the local wind
218 under different blade azimuth angles is calculated when considering only the wind shear effect, only the tower shadow effect

6
219 and the combined effects of wind shear and tower shadow, as shown in Fig. 5.
220 Fig. 5 shows that the wind speed spatial distribution is only affected by wind shear effect when the blade element is
221 located in the upper half plane of wind wheel, i.e. the blade azimuth angle is within the range 0    90 and
222 270    360 , and the local wind speed are larger than the hub height wind speed. When the blade element is in the
223 acceleration area, i.e. the blade azimuth angle is within the range 90    176.87 or 183.13    270 , the wind speed
224 spatial distribution is affected by the combined effects of wind shear and tower shadow, which have opposite effects, but wind
225 shear effect is dominant, and the local wind speed at blade element is all less than the hub height wind speed. When the blade
226 element is in the center area of tower shadow, i.e. the blade azimuth angle is within the range 176.87    183.13 , the
227 wind speed spatial distribution is affected by the combined effects of wind shear and tower shadow, but the tower shadow
228 effect dominates and the local wind speed decreases sharply. When the blade azimuth angle is 180 ° , the local wind speed
229 reaches the minimum value of 9.49m/s, which decreases by about 14% compared to the hub height wind speed. The results
230 show that the combined effects of wind shear and tower shadow cause the obvious inhomogeneity of the wind speed spatial
231 distribution in the wind wheel rotation plane, and the hub height wind speed is not enough to represent the actual wind speed
232 effect in the wind wheel rotation plane.

233
234 Fig. 5 The variation of wind speed with blade azimuth angle.
235 3. Rotor equivalent wind speed calculation method based on equivalent torque
236 The above influence analysis of wind shear and tower shadow effect on wind speed spatial distribution show that the
237 spatial distribution of wind speed in the wind wheel rotation plane is uneven, and the non-uniformity becomes more obvious
238 with the increase of rotor diameter and tower height. Therefore, the hub height wind speed is not sufficient to represent the
239 actual wind speed effect in the whole wind wheel rotation plane, which could bring large errors in wind power prediction and
240 theoretical power calculation. To solve this problem, Dolan et al. [22] put forward the rotor equivalent wind speed calculation
241 method based on equivalent torque, which follows the momentum conservation law, that is, the rotor equivalent wind speed
242 has the same mechanical torque with the actual wind speed in the whole wind wheel rotation plane. The instantaneous rotor
243 equivalent wind speed based on equivalent torque is as follows.
    1  R 2   1  2  R 3 
VH 1      cos3  , others
  8 H 60 H 

    1  R    1  2  R  
2 3

 1       cos3 
8 H 60 H
Veq       (15)
    R sin i 
2 2
   3
VH  2 3  
ln 2
1  ,
2
 
2
   2 ma   x  R2
 
  3R2 
 2 2 2 
 2
2sin i 
x  R sin i 
 
i 1
 
    
244 Since the instantaneous rotor equivalent wind speed varies periodically with the change of blade azimuth angle, the
245 average rotor equivalent wind speed in the wind wheel rotation plane can be obtained.
2
V eq  d
    1  R  2 2ma 2 
Veq  0
 VH 1      (16)
2  8 H R 2 
246 The rotor equivalent wind speed calculation method based on equivalent torque is based on the equivalent mechanical
247 torque without considering wind turbine operation characteristics, which cannot match the actual power control effect of wind

7
248 turbine, thus ignoring the power loss caused by the actual control deviation. Because in the actual operation of wind turbine,
249 different power control strategies are used in different wind speed stages, and the actual control deviation will cause the
250 mechanical torque to deviate from the ideal value, thus the model based on equivalent torque cannot effectively reflect the
251 actual power control effect of different control stages. To address this problem, the rotor equivalent wind speed calculation
252 method based on equivalent power is proposed with comprehensively considering wind shear, tower shadow and wind turbine
253 operation characteristics in this paper, that is, the rotor equivalent wind speed has the same output power with the actual wind
254 speed in the whole wind wheel rotation plane.
255 4. Rotor equivalent wind speed calculation method based on equivalent power
256 4.1. Wind turbine operation characteristics
257 The important operation characteristics of variable-pitch variable-speed wind turbine are the switch features of operation
258 conditions vary with wind speed. With the change of wind speed, wind turbine switches between different operation conditions.
259 Due to the different control objectives and strategies under each operation condition, the variation of wind turbine power
260 output and rotor speed is also different [28]. Fig. 6 is the rotor speed-power curve corresponding to the active power control
261 strategy of variable-pitch variable-speed wind turbine [29-30], which is mainly divided into four control stages: the starting
262 and grid-connected control, the maximum power point tracking control, the constant rotor speed control and the constant wind
263 power control. (1) The starting and grid-connected control stage (segment AB). When wind speed reaches the cut-in wind
264 speed and maintains a certain time, the wind turbine is released from braking control and enters the stage of starting and
265 grid-connected control, and its main control objectives is to achieve the rotor speed-up and grid-connected of wind turbine. (2)
266 The maximum power point tracking control stage (segment BC). When wind speed is lower than the rated wind speed, the
267 wind turbine enters the maximum power point tracking control stage, and its main control objectives is to maximize the use of
268 wind energy and increase the power generation of wind turbine. The main control strategy is to keep the optimal pitch angle
269 constant (generally 2°~3 °), the rotor speed is controlled by the excitation adjustment to track maximum power point. (3)
270 The constant rotor speed control stage (segment CD). When rotor speed reaches the limit value and wind power does not reach
271 the rated power, the wind turbine enters the constant rotor speed control stage, and the rotor speed is no longer increased, but
272 the output power is still increased by the excitation control. (4) The constant wind power control stage (segment DE). When
273 wind speed exceeds the rated wind speed, the wind turbine enters the constant wind power control stage, and the output power
274 is maintained near the rated power mainly through pitch control without exceeding the power limit. In the actual operation
275 process of wind turbine, due to the different control algorithms, the actual control effect of each stage may deviate from the
276 ideal states, resulting in different torque and power output characteristics.

277
278 Fig. 6 The rotor speed-power curve of variable-pitch variable-speed wind turbine.
279 4.2. Rotor equivalent wind speed calculation model based on equivalent power
280 Wind wheel is one of the key components of wind turbine, and its function is to convert wind energy into mechanical
281 energy. According to the principle of aerodynamics, the output mechanical power and torque of wind turbine are deduced as
282 follows.
1
P AV 3C P   ,   (17)
2
P  AV C P   ,   1 C  ,  
3

T    ARV 2 P (18)
 2 2 
283 Where V is the hub height wind speed,  is the air density, A is the rotor swept area, R is the rotor radius,  is the
R
284 pitch angle,  is the blade tip speed ratio (   ),  is the rotor speed, and C P   ,   is the power coefficient
V
285 corresponding to the hub height wind speed, which does not exceed 0.593 according to the Betz momentum theory.

8
286 For three-blade horizontal-axis wind turbines, the aerodynamic torque generated by wind speed can be defined as [27]:
3 R
T    3M V0      r  V r ,   V0  dr (19)
r0
i 1

287 Where r0 is the hub radius,   r  is the aerodynamic torque influence coefficient, and its empirical formula is:
  r   kr (20)
288 Since the effect of rotor equivalent wind speed on wind turbine is consistent with the actual wind speed, based on the
289 principle of equivalent power, the actual wind speed V (r , ) of each point in the wind wheel rotation plane is replaced by the
290 equivalent wind speed Veq ( ) , and the substitution formula (19) can be obtained.
3 R
T    3M V0      r  V eq    V0  dr (21)
r0
i 1

291 Define V  Veq    V0 , since the hub radius r0 is much smaller than the rotor radius R, which can be ignored, and
292 the calculation can be obtained.
3kR 2
T    3M V0   V (22)
2
293 The partial differential linearization of the torque formula (18) at V0 , 0  is obtained as follows.
V  V0 T   V  V0 1 C p  0 ,  0  C p 0 ,  0 
T    T    V   ARV 0 2   ARV 0 V (23)
  0 V   0 2 0 0
294 Where 0 is the tip speed ratio corresponding to the spatial average wind speed V0 , and C P  0 ,  0  is the power
295 coefficient corresponding to the spatial average wind speed V0 .
296 Combining formulas (22) and (23) can be obtained:
1 C  ,  
3M V0    ARV0 2 p 0 0 (24)
2 0
2  AV0 C p  0 ,  0 
k (25)
3R 0
297 Substituting formula (24) and formula (25) into formula (19), the formula for calculating wind turbine aerodynamic
298 torque can be obtained as follows.
1 Cp 0 , 0  2 AV0 Cp 0 , 0  3 R
T     ARV02
2 0

3R 0
 r V r,  V0  rdr
i1 0
(26)

299 Substituting formula (11) into formula (26), the formula for calculating the aerodynamic torque of the upper half plane of
300 wind wheel can be obtained:
 r   1  r 2 2 
 1   cos     cos  
2 p  0 0  2 AV0
1 C  , Cp 0 , 0  3 R  H 2 H V dr
T     ARV0
2 0

3R 0
i 1
r0
r
   1  2  r  3
3  H (27)
   cos   m 
 6 H 
301 Substituting formula (14) into formula (26), the formula for calculating the aerodynamic torque of the lower half plane of
302 wind wheel can be obtained:

9
 1 Cp 0 , 0 
T     ARV02
2 0
 r   1  r 2 2 
1 cos     cos    (28)
H 2  H 
2 AV0 Cp 0 , 0  3 R  

3R 0
 r0   1  2  r 3
r  
r 2
sin2
  x 2 VH dr

i 1
 1
  cos   m a 2 2
3 2

 r sin   x  
6 H 2 2


2
303 For the three-blade wind turbine, the corresponding relations of blade azimuth angles are 1   ,  2  1  and
3
4
304 3   2  , and the calculation can be obtained:
3
3
 cos  i  0
 i 1
3 3
 cos  i 
2
(29)
 i 1 2
 3
3
 cos  i  cos3
3

 i 1 4
305 Since the hub radius r0 is much smaller than the rotor radius R, which can be ignored. Therefore, the aerodynamic
306 torque at each point in the wind wheel rotation plane can be calculated as:
2 p  0 0  2 AV0
1 C  , Cp 0 , 0    1  2  R5  
  ARV0  VH   3 cos3  , others
2 0 3R 0  40 H  

   1  2  R5  
   3 cos3  
 40 H  
T     (30)
C   ,   C  
2 AV0 p 0 0 ,      3
1 2 p 0 0    R sin i 
2 2

2  ARV0 0

3R 0
VH   ln 
x2
1 2  , 2    2
 3
R

ma 
2     2 2 2 
 i 1
 2sin2 i x  R sin i 
   
   
307 The instantaneous mechanical power absorbed by wind wheel can be expressed as:
P    T     (31)
308 According to formula (30), the instantaneous mechanical power absorbed by wind wheel varies periodically with the
309 change of blade azimuth angle, and the average mechanical power is obtained in the whole wind wheel rotation plane as
310 follows.
2
 P   d  AV0 2 C p  0 , 0  2
P 0
2

2R 0
 R  4 a2  (32)

311 Based on the principle of equivalent power, the rotor mechanical power can be expressed as:
1
Peq  C p  ,  AVeq 3  P (33)
2
312 Combining formulas (30), (31) and (33), the instantaneous rotor equivalent wind speed is obtained:

10
  RV 2 C  ,    4VH     1  2   R 5  
 0 P 0 0
1    cos 3  , others
 C P  ,   0  3R V0 
2
40  H3 
 



      1  2   R 5  
     cos 3  
  40  H3 
V eq        
   (34)
  C P 0 ,  0  2 4    R 2 sin 2  i    3
 mV H mR  
 ln 
 1  ,  
 C P  ,   0 3R 
 2 2
3 2 2
 x R
ma 2   
     2
2 sin  i 2 2 2 
x  R sin  i 
  
i 1  
    
    
313 Combining formulas (32) and (33), the average rotor equivalent wind speed is obtained:

Veq  3
R 
 4a 2  C P 0 ,  0  2     1 R 2 
2
VH 1  
2

(35)
C P  ,   R 0  8 H2
314 5. Case study
315 5.1. Rotor equivalent wind speed variation analysis with different parameters
316 5.1.1. Wind turbine design parameters
317 According to the rotor equivalent wind speed calculation model based on equivalent power, the rotor equivalent wind
318 speed for specific wind turbine varies with the wind shear coefficient  , blade azimuth angle  , and hub height wind
319 speed VH . Therefore, the variation law of rotor equivalent wind speed with different parameters in different control stages of
320 wind turbines is qualitatively analyzed using the design data of wind turbines with different capacities. The design parameters
321 of wind turbine are shown in Table 1. The static wind turbine power curve and the power coefficient corresponding to different
322 pitch angles are simulated and calculated by the Bladed software.
323 Table 1
324 Wind turbine design parameters.
Wind turbine #1 #2 #3 #4
Rated power 1.5MW 2MW 5MW 7MW
Rated wind speed 12m/s 11m/s 10m/s 11m/s
Rated rotor speed 17.6rpm 18rpm 11.4rpm 12rpm
Hub height 80m 61.5m 95m 85.8m
Rotor diameter 77.42m 86.656m 165.365m 155.6m
Tower radius 1.885m 0.95m 3.2m 1.7m
The distance from the wind wheel rotation plane to tower center 4.08m 3.7m 4.708m 5.5m
325
326 5.1.2. Rotor equivalent wind speed variation analysis
327 Assuming that the hub height wind speed is 7 m/s and wind shear coefficient is 0.3, the formula (8) is used to calculate the
328 spatial average wind speed, and its corresponding tip speed ratio is obtained. The power coefficient corresponding to the spatial
329 average wind speed, the theoretical power coefficient and rotor speed corresponding to the hub height wind speed are
330 determined using the look-up table method, and then the average rotor equivalent wind speed is calculated by formula (16) and
331 (35). In addition, the instantaneous rotor equivalent wind speed is calculated by formula (15) and (34), as shown in Fig. 7.

a) Wind turbine performance curve (1.5MW) b) Wind shear coefficient (1.5MW) c) Blade azimuth angle (1.5MW)

11
 d) Wind turbine performance curve (2MW) e) Wind shear coefficient (2MW) f) Blade azimuth angle (2MW)

g) Wind turbine performance curve (5MW) h) Wind shear coefficient (5MW) i) Blade azimuth angle (5MW)

j) Wind turbine performance curve (7MW) k) Wind shear coefficient (7MW) l) Blade azimuth angle (7MW)
Fig. 7 The variation of rotor equivalent wind speed in the rotor speed control stage.
332

333
334 Fig. 8 The simulation results of mechanical torque in the rotor speed control stage(1.5MW).
335 Fig. 7 shows the variation of rotor equivalent wind speed in the rotor speed control stage. As shown in Fig. 7 (b), the
336 average rotor equivalent wind speed of the two models is less than hub height wind speed, which decreases first and then
337 increases with the increase of wind shear coefficient, and reaches the minimum value when the wind shear coefficient is 0.5.
338 However, the wind shear coefficient of actual wind farm usually does not reach 0.5, so the average rotor equivalent wind speed
339 in the actual wind farm decreases gradually with the increase of wind shear coefficient. Comparing Fig. 7 (b) (e) (h) (k), the
340 average rotor equivalent wind speed of wind turbines with different capacities has similar changes with wind shear coefficient,
341 however, there are still some differences that the average rotor equivalent wind speed based on the equivalent power model of
342 the 1.5MW wind turbine is greater than that based on the equivalent torque model, which is exactly the opposite of the

12
343 calculation results of other wind turbines, and the wind speed difference between the two models of wind turbines with
344 different capacities is also different. The main reason is that wind turbines with different capacities may adopt different control
345 algorithms and have different control effects in the rotor speed control stage, as shown in Fig. 7 (a) (d) (g) (j), the tip speed
346 ratio tracked by the 1.5MW wind turbine during the rotor speed control stage is less than the design optimal tip speed ratio, the
347 simulation results in Fig. 8 show that the mechanical torque is greater than the optimal mechanical torque, and the average
348 rotor equivalent wind speed based on the equivalent power model is greater than that based on the equivalent torque model,
349 other wind turbines have the opposite results, and the greater the difference between the actual tip speed ratio and the design
350 optimal tip speed, the greater the difference between the average equivalent wind speed of the two models.
351 Fig. 7 (c) (f) (i) (l) shows the variation of instantaneous rotor equivalent wind speed with different blade azimuth angles of
352 wind turbines with different capacities respectively. As shown in Fig. 7 (c), the instantaneous rotor equivalent wind speed of
353 the two models is smaller than the hub height wind speed at different blade azimuth angles, and varies periodically with the
354 change of blade azimuth angle. The period is 120°, which is related to the installation of three blades of wind turbine with the
355 equal interval design of 120°. When the blade azimuth angle is 60°, 180° and 300°, the instantaneous rotor equivalent wind
356 speed reaches the minimum value, which is mainly because one of the blades is located in the center area of tower shadow, and
357 the local wind speed decreases sharply due to the tower shadow effect. When the blade azimuth angle is 0°, 120° and 240°, the
358 instantaneous rotor equivalent wind speed reaches the maximum value, the main reason is that one of the blades is located in
359 the upper half plane of wind wheel, which is not affected by tower shadow, and the other two blades are located in the
360 acceleration region of the lower half plane, the impact of tower shadow is the least. The variation amplitude of instantaneous
361 rotor equivalent wind speed based on equivalent power model is smaller, which is mainly because the rotor inertia and control
362 effect of wind turbines make the equivalent power model less sensitive to the change of aerodynamic torque than the
363 equivalent torque model. Comparing Fig. 7 (c) (f) (i) (l), the variation of instantaneous equivalent wind speed with blade
364 azimuth angle of wind turbines with different capacities is similar, but there are still some differences that the larger the
365 diameters of rotor and tower, the greater the influence of wind shear and tower shadow on rotor equivalent wind speed, and the
366 size of difference between the two models is determined by the actual control deviation of wind turbines.
367 Assuming that the hub height wind speed is 10 m/s, the parameters needed for rotor equivalent wind speed calculation are
368 obtained by the same method mentioned above. The average rotor equivalent wind speed is calculated by formula (16) and (35).
369 In addition, assuming that the hub height wind speed is 10 m/s and wind shear coefficient is 0.3, the instantaneous rotor
370 equivalent wind speed is calculated by formula (15) and (34), as shown in Fig. 9.

a) Wind shear coefficient (1.5MW) b) Blade azimuth angle (1.5MW)

13
 c) Wind shear coefficient (2MW) d) Blade azimuth angle (2MW)

e) Wind shear coefficient (5MW) f) Blade azimuth angle (5MW)

g) Wind shear coefficient (7MW) h) Blade azimuth angle (7MW)

Fig. 9 The variation of rotor equivalent wind speed in the torque control stage.
371 Fig. 9 shows the variation of rotor equivalent wind speed in the torque control stage. As shown in Fig. 9, the variation law
372 of equivalent wind speed with wind shear coefficient and blade azimuth angle in the torque control stage is similar to that in
373 the rotor speed control stage. As shown in Fig. 9 (a) (c) (e), the rotor equivalent wind speed based on equivalent power model
374 is greater than that based on equivalent torque model, the main reason is that when wind turbine is in the torque control stage,
375 the rotor speed remains the rated constant, while the output power is still increasing, which is equivalent to increasing the
376 mechanical torque. Therefore, the rotor equivalent wind speed based on the equivalent power model is greater than that based
377 on the equivalent torque model. As shown in Fig. 9 (g), the rotor equivalent wind speed calculation result of 7MW wind
378 turbine is opposite to other wind turbines, which is mainly because the actual tip speed ratio is far greater than the design
379 optimal tip speed ratio in the rotor speed control stage. When it is in the torque control stage, the actual tip speed ratio
380 decreases with the increase of wind speed, and the capture capacity of wind wheel increases. Therefore, the rotor equivalent
381 wind speed based on the equivalent torque model is greater than that base on the equivalent power model. It should be noted
382 that this situation is relatively rare under the normal control of wind turbines. As shown in Fig. 9 (b) (d) (f) (h), the difference
383 between the two models of 5MW wind turbine is different from other wind turbines, the main reason is that the wind speed at
384 this time is close to the rated wind speed, and the influence of wind speed change on output power decreases gradually, hence
385 the rotor equivalent wind speed is close to the hub height wind speed.
386 Assuming that the wind shear coefficient is 0.3 and the hub height wind speed ranges from 4 to 25 m/s, the variation of
387 average rotor equivalent wind speed of 2MW and 7MW wind turbines is calculated and analyzed by formula (16) and (35), and
388 the deviation between the average rotor equivalent wind speed and hub height wind speed is calculated, as shown in Fig. 10.

14
 a) 1.5MW wind turbine b) 2MW wind turbine

c) 5MW wind turbine d) 7MW wind turbine

Fig. 10 The variation of rotor equivalent wind speed with hub height wind speed.
389 Fig. 10 shows the variation of rotor equivalent wind speed with hub height wind speed. As shown in Fig. 10, wind speed
390 deviation is the difference between the rotor equivalent wind speed and the hub height wind speed. The wind speed deviation
391 based on the equivalent torque model increases linearly with the increase of hub height wind speed, however, the wind speed
392 deviation based on the equivalent power model presents different variation rules in different wind speed ranges, i.e. different
393 control stages of wind turbines. The variation law of rotor equivalent wind speed in the rotor speed and torque control stages
394 have been fully analyzed above. In the pitch control stage, wind turbine maintains constant power operation by adjusting the
395 pitch angle, the wind speed deviation is affected by the effect of pitch control and wind turbine performance characteristics,
396 which has little effect on wind power prediction and theoretical power calculation. In summary, the proposed model can
397 effectively reflect the variation law of rotor equivalent wind speed under the actual control effect of different control stages,
398 which matches the actual power control effect of wind turbine and contains the power loss caused by actual power control
399 deviation.
400 5.2. Wind turbine theoretical power calculation
401 5.2.1. Data
402 Since it is difficult to obtain the detailed wind turbine design parameters and its corresponding measured data at the same
403 time, the validity of the proposed model is verified only by using design parameters and measured data of the above-mentioned
404 2MW and 5MW wind turbines. The measured data includes the hub height wind speed, measured power and wind speed at
405 different heights of anemometer tower. The data time span of 2MW wind turbine is from 00:00 on January 1st 2017 to 23:45
406 on December 31th 2017, and 5MW wind turbine is from 00:00 on January 1st 2018 to December 31 23:45. The time resolution
407 of the data is 15min. According to formula (1), the wind shear coefficient of wind farm in different seasons is calculated by
408 using the wind speed at different heights of anemometer tower.
409 5.2.2. Performance criteria
410 In order to quantitatively evaluate the validity of the proposed model, Root Mean Square Error (RMSE) and Mean
411 Absolute Error (MAE) are used to calculate the model error. RMSE is used to evaluate the overall deviation of wind turbine
412 theoretical power, and MAE is used to evaluate the relative absolute deviation of wind turbine theoretical power. The
413 calculation formulas of the two indicators are as follows.

15
 n

1
 P  P 
i 1
i i
' 2
(36)
RMSE 
PN n
n

 P P
i 1
i i
'
(37)
MAE 
nPN
'
414 Where PN is the rated power of wind turbine, Pi and Pi are the actual power and theoretical power of wind turbine at time
415 i respectively, and n is the number of data.
416 5.2.3. Wind turbine theoretical power calculation results
417 The theoretical power of wind turbines are calculated based on the above-mentioned rotor equivalent wind speed
418 calculation models, and compared with hub height wind speed and the actual power of wind turbines under normal operation
419 condition, it should be pointed out that the theoretical power is generally considered to be close to the actual power under
420 normal operation condition of wind turbine. The error statistical results of wind turbine theoretical power calculation are shown
421 in Table 2, which only shows the calculation results of any day in different seasons.
422 Table 2
423 The error statistical results of wind turbine theoretical power calculation.
Rotor equivalent wind speed Rotor equivalent wind speed
Hub height wind speed
Wind turbines Seasons based on equivalent torque based on equivalent power
RMSE MAE RMSE MAE RMSE MAE
Spring 2.68% 2.40% 2.24% 1.94% 3.38% 3.11%
Summer 2.03% 1.72% 1.81% 1.49% 2.38% 2.07%
2MW
Autumn 2.18% 1.87% 1.91% 1.60% 2.59% 2.27%
Winter 2.16% 1.93% 1.92% 1.69% 2.46% 2.22%
Spring 8.03% 5.47% 7.23% 4.81% 8.71% 6.05%
Summer 7.61% 5.50% 7.31% 5.28% 8.13% 5.69%
5MW
Autumn 6.98% 5.80% 5.53% 6.61% 8.54% 7.43%
Winter 4.48% 3.17% 3.25% 2.27% 5.40% 4.03%
424
425 As shown in Table 2, the theoretical power calculation results based on rotor equivalent wind speed are better than those
426 based on hub height wind speed, among the theoretical power calculation models based on rotor equivalent wind speed, the
427 model based on equivalent power has higher accuracy, which mainly because that the proposed model contains the power loss
428 caused by actual power control deviation, and is more in line with the actual power control effect of wind turbines. However,
429 the theoretical power calculation accuracy is not significantly improved, the main reason is that the wind shear coefficient is
430 not large in the area where the two wind turbines are located, which are approximately 0.16 and 0.2 respectively, thus the
431 difference cannot be well highlighted in the calculation accuracy of wind turbine theoretical power.
432 In order to highlight the difference in wind turbine theoretical power calculation results, the annual power generation is
433 counted, as shown in Table 3.
434 Table 3
435 The statistics of wind turbine annual theoretical power generation.
Annual power generation (MWh)
2MW 5MW
Rotor equivalent wind speed based on equivalent torque 2162.7 14641
Rotor equivalent wind speed based on equivalent power 2105.3 14613
Hub height wind speed 2231.8 15198
Wind turbine actual power generation 1971.4 12560
436
437 As shown in Table 3, the deviation between the annual power generation based on the hub height wind speed and the
438 actual annual power generation is the most obvious, and the calculation results of the rotor equivalent wind speed based on
439 equivalent power are closer to the actual power generation. Compared with the rotor equivalent wind speed based on
440 equivalent torque, the difference of 2MW wind turbine is more obvious than that of 5MW wind turbine, the main reason is that
441 the control effect of 5MW wind turbine is better than that of 2MW wind turbine and more close to the ideal control effect,
442 which is just corresponding to the analysis in Section 5.1.2 above. In summary, the proposed rotor equivalent wind speed
443 calculation model can effectively reflect the power loss caused by the actual control deviation, and reduce the wind turbine
444 theoretical power generation calculation error, which can provide theoretical and data basis for the reduction coefficient setting
445 of wind turbine.
446 6. Conclusions
447 This paper proposed a novel rotor equivalent wind speed calculation method based on equivalent power considering wind

16
448 shear, tower shadow and wind turbine operation characteristics, and the rationality and validity of the proposed model were
449 evaluated qualitatively and quantitatively from different power control stages and theoretical power calculation. The
450 conclusions are as follows.
451 (1) The rotor equivalent wind speed calculation model based on equivalent power comprehensively takes into account wind
452 shear, tower shadow and wind turbines operation characteristics, and the rotor equivalent wind speed has the same output
453 power with the actual wind speed in the whole wind wheel rotation plane.
454 (2) Compared with the model based on equivalent torque, the proposed model can effectively reflect the variation law of
455 rotor equivalent wind speed under the actual control effect of different control stages, which matches the actual power
456 control effect of wind turbine and contains the power loss caused by actual power control deviation.
457 (3) The calculation accuracy of wind turbine theoretical power obtained by the rotor equivalent wind speed calculation model
458 based on equivalent power was effectively improved compared to the rotor equivalent wind speed calculation model
459 based on equivalent torque and hub height wind speed, which can provide theoretical and data basis for the reduction
460 coefficient setting of wind turbine.
461 (4) Future research plan: (i) The proposed rotor equivalent wind speed calculation model does not consider the influence of
462 blade airfoil, hence we plan to further research the rotor equivalent wind speed calculation method based on the blade
463 element-momentum theory. (ii) The wind shear coefficient used in the proposed model is the average wind shear
464 coefficient to simplify the calculation, but wind shear coefficient varies dynamically with the change of altitude in the
465 actual conditions, and under the condition of non-neutral atmospheric stability, the wind speed profile does not
466 completely obey the logarithmic or exponential distribution rate. Therefore, we plan to study the function of wind shear
467 coefficient varying with height and apply it to the rotor equivalent wind speed calculation model.
468 Acknowledgments
469 This work is supported by National Key Research and Development Program of China (No. 2017YFE0109000).
470 References
471 [1] Demolli Halil, Dokuz Ahmet Sakir, Ecemis Alper, Gokcek Murat. Wind power forecasting based on daily wind speed
472 data using machine learning algorithms[J]. Energy Conversion and Management, 2019, 198:111823.
473 https://doi.org/10.1016/j.enconman.2019.111823.
474 [2] Global Wind Energy Council. Global Wind Report 2016-Annual Market Update[R]. 2017.
475 [3] Ashuri Turaj, Zaaijer Michiel B, Martins Joaquim R R A, Zhang Jie. Multidisciplinary design optimization of large wind
476 turbines-Technical, economic, and design challenges[J]. Energy Conversion and Management, 2016, 123:56-70.
477 https://doi.org/10.1016/j.enconman.2016.06.004.
478 [4] Rehman Shafiqur, Al-Abbadi Naif M. Wind shear coefficients and their effect on energy production[J]. Energy
479 Conversion and Management, 2005, 46(15-16):2578-2591. https://doi.org/10.1016/j.enconman.2004.12.005.
480 [5] Mohammadi Ebrahim, Fadaeinedjad Roohollah, Naji Hamid Reza. Using a new wind turbine emulator to analyze tower
481 shadow and yaw error effects[J]. Energy Conversion and Management, 2018, 174:378-387.
482 https://doi.org/10.1016/j.enconman.2018.08.049.
483 [6] Emeksiz Cem, Tugce Cetin. In case study: Investigation of tower shadow disturbance and wind shear variations effects on
484 energy production, wind speed and power characteristics[J]. Sustainable Energy Technologies and Assessments, 2019,
485 35:148-159. https://doi.org/10.1016/j.seta.2019.07.004.
486 [7] Wen Binrong, Wei Sha, Wei Kexiang, Yang Wenxian, Peng Zhike, Chu Fulei. Influences of wind shear and tower shadow
487 on the power output of wind turbine [J]. Chinese Journal of Mechanical Engineering, 2018, 54(10):124-132.
488 https://doi.org/10.3901/JME.2018.10.124.
489 [8] Wharton Sonia, Lundquist Julie K. Atmospheric stability affects wind turbine power collection[J]. Environmental
490 Research Letters, 2012, 7(1):014005. https://doi.org/10.1088/1748-9326/7/1/014005.
491 [9] Shen Xin, Zhu Xiaocheng, Du Zhaohui. Wind turbine aerodynamics and loads control in wind shear flow[J]. Energy,
492 2011, 36(3):1424-1434. https://doi.org/10.1016/j.energy.2011.01.028.
493 [10] Holtslag M C, Bierbooms W A A M, Van Bussel G J W. Extending the diabatic surface layer wind shear profile for
494 offshore wind energy[J]. Renewable Energy, 2017, 101:96-110. https://doi.org/10.1016/j.renene.2016.08.031.
495 [11] Sintra H, Mendes V M F, Melício R. Modeling and simulation of wind shear and tower shadow on wind turbines[J].
496 Procedia Technology, 2014, 17:471-477. https://doi.org/10.1016/j.protcy.2014.10.255.
497 [12] Zhang Mingming, Yu Wei, Xu Jianzhong. Aerodynamic physics of smart load control for wind turbine due to extreme
498 wind shear[J]. Renewable Energy, 2014, 70(5):204-210. https://doi.org/10.1016/j.renene.2013.12.046.
499 [13] Gualtieri Giovanni. Atmospheric stability varying wind shear coefficients to improve wind resource extrapolation: A
500 temporal analysis[J]. Renewable Energy, 2016, 87:376-390. https://doi.org/10.1016/j.renene.2015.10.034.
501 [14] Pedersen R R, Nielsen S R K, Thoft-Christensen P. Stochastic analysis of the influence of tower shadow on fatigue life of
502 wind turbine blade[J]. Structural Safety, 2012, 35:63-71. https://doi.org/10.1016/j.strusafe.2011.12.005.
503 [15] Choukulkar Aditya, Pichugina Yelena, Clack Christopher T M, Calhoun Ronald. A new formulation for rotor equivalent
504 wind speed for wind resource assessment and wind power forecasting[J]. Wind Energy, 2016, 19(8):1439-1452.
505 https://doi.org/10.1002/we.1929.

17
506 [16] Clack Christopher TM , Alexander Anneliese, Choukulkar Aditya, MacDonald Alexander E. Demonstrating the effect of
507 vertical and directional shear for resource mapping of wind power[J]. Wind Energy, 2016, 19(9):1687-1697.
508 https://doi.org/10.1002/we.1944.
509 [17] Redfern Stephanie, Olson Joseph B, Lundquist Julie K, Clack Christopher TM. Incorporation of the rotor-equivalent wind
510 speed into the weather research and forecasting model’s wind farm parameterization[J]. Monthly Weather Review, 2019,
511 147(3):1029-1046. https://doi.org/10.1175/MWR-D-18-0194.1.
512 [18] Wagner Rozenn, Cañadillas Beatriz, Clifton Andy, Feeney Siobhan. Rotor equivalent wind speed for power curve
513 measurement comparative exercise for IEA Wind Annex 32[J]. Journal of Physics Conference Series, 2014, 524:012108.
514 https://doi.org/10.1088/1742-6596/524/1/012108.
515 [19] Scheurich Frank, Enevoldsen Peder B, Paulsen Henrik N, Dickow Kristoffer K, Fiedel Moritz, Loeven Ale-x, Antoniou
516 Ioannis. Improving the accuracy of wind turbine power curve validation by the rotor equivalent wind speed concept[J].
517 Journal of Physics: Conference Series, 2016, 753:072029. https://doi.org/10.1088/1742-6596/753/7/072029.
518 [20] Kavari Ghazale, Tahani Mojtaba, Mirhosseini Mojtaba. Wind shear effect on aerodynamic performance and energy
519 production of horizontal axis wind turbines with developing blade element momentum theory[J]. Journal of Cleaner
520 Production, 2019, 219:368-376. https://doi.org/10.1016/j.jclepro.2019.02.073.
521 [21] Jeon Sanghyeon, Kim Bumsuk, Huh Jongchul. Study on methods to determine rotor equivalent wind speed to increase
522 prediction accuracy of wind turbine performance under wake condition[J]. Energy for Sustainable Development, 2017,
523 40:41-49. https://doi.org/10.1016/j.esd.2017.06.001.
524 [22] Dolan DSL, Lehn PW. Simulation model of wind turbine 3p torque oscillations due to wind shear and tower shadow[J].
525 IEEE Transactions on Energy Conversion, 2006, 21(3):717-724. https://doi.org/10.1109/TEC.2006.874211.
526 [23] Wan Shuting, Cheng Lifeng, Sheng Xiaoling. Effects of yaw error on wind turbine running characteristics based on the
527 equivalent wind speed model[J]. Energies, 2015, 8(7):6286-6301. https://doi.org/10.3390/en8076286.
528 [24] FajardoR Luis Alberto, Medina Aurelio. Per-blade equivalent wind speed function for high-order drive-train modeling[C].
529 Power & Energy Society General Meeting, Detroit, USA, 2011:1-7. https://doi.org/10.1109/PES.2011.6039447.
530 [25] Wan Shuting, Cheng Lifeng, Sheng Xiaoling. Numerical analysis of the spatial distribution of equivalent wind speeds in
531 large-scale wind turbines[J]. Journal of Mechanical Science and Technology, 2017, 31(2):965-974.
532 https://doi.org/10.1007/s12206-017-0149-6.
533 [26] Jung Christopher, Schindler Dirk. Development of a statistical bivariate wind speed-wind shear model (WSWS) to
534 quantify the height-dependent wind resource[J]. Energy Conversion and Management, 2017, 149:303-317.
535 http://dx.doi.org/10.1016/j.enconman.2017.07.033.
536 [27] Sørensen Poul, Hansen Anca D , Pedro André Carvalho Rosas. Wind models for simulation of power fluctuations from
537 wind farms[J]. Journal of Wind Engineering and Industrial Aerodynamics, 2002, 90(12-15):1381-1402.
538 https://doi.org/10.1016/S0167-6105(02)00260-X.
539 [28] Asghar Aamer Bilal , Liu Xiaodong. Adaptive neuro-fuzzy algorithm to estimate effective wind speed and optimal rotor
540 speed for variable-speed wind turbine[J]. Neurocomputing, 2018, 272:495-504.
541 https://doi.org/10.1016/j.neucom.2017.07.022.
542 [29] Xu Daping. Principle of wind power generation [M]. China Machine Press, 2011.
543 [30] Bossanyi EA. GH Bladed: Theory manual[M]. Bristol: Garrad Hassan and Parters Limited, 2007.

18

View publication stats