guideline for wind turbines in japan: measurements of wind conditions ...
2008年10月 再生可能エネルギー2008国際会議 RENEWABLE ENERGY 2008（韓国釜山） 発表論文
GUIDELINE FOR WIND TURBINES IN JAPAN: MEASUREMENTS OF WIND CONDITIONS AND TURBINE LOADS KARIKOMI Kai, HONDA Akihiro and HIRAI Shigeto Nagasaki R&D Center of Mitsubishi Heavy Industries, Ltd. 5-717-1, Fukahori-machi, Nagasaki, 851-0392, Japan The Guideline for Wind Turbines in Japan has been developed, which describes the procedure of selecting wind turbines in terms of wind conditions and clarifying problems to be solved. In order to collect data as evidence for the development of the Guideline, measurement campaigns on wind conditions and loads on turbines at seven existing turbine sites were carried out during about one and a half years. The seven sites were selected as target sites where typhoons frequently attack and/or high turbulences are generated. This paper describes the main measurement results of wind conditions and loads as well as the comparison of wind turbine loads between the measurement data and simulated results by aeroelastic simulations
Keywords: Wind Turbine, Wind Condition, Turbulence, Typhoon, Load 1. INTRODUCTION
2. MEASURING CAMPAGIN
Guideline for Wind Turbines in Japan has been developed, which is aimed to be used by wind turbine business companies installing wind power plants in Japan sites. The content of the Guideline includes the following items: • Methods of estimating extreme wind speeds and turbulence intensity at objective sites; • Procedures of selecting suitable wind turbines; • Technical information to be referred in the selection of turbines; • Results of measurement and simulation performed during the development of the Guideline.
Sites and Measuring Equipment The seven selected measuring sites are shown in Fig.1. The sites cover most of Japan from Hokkaido to Okinawa. Most of all sites are considered not flat but complex terrain. In this paper, the results for four sites A, B, C, D are focused on. Specifications for the turbines installed for the sites are summarized in Table 1. All the turbines are conventional typed turbines; upwind, three-bladed, fixedspeed and pitch regulated turbines. At each of the sites a meteorological mast (MET mast) with the same height as the turbine was installed within one or two rotor diameter distance from the turbine. An example picture is shown in Fig.2. The met masts were equipped with wind speed and direction sensors at three heights. The top heights are same level as turbine hub heights. Three kinds of sensors were used such as cup/vane, propeller and ultrasonic anemometers. The measurements of loads on the turbine structural components were performed with strain gazes. The measured quantities included flapwise and edgewise root bending moments, tower bending moments in two perpendicular directions. The turbine operational statuses were recorded by means of power output, pitch angles, rotor speed and nacelle yaw position.
Before installing wind power plants, it is important to find sites where annual wind speeds are high enough to yield high energy productions as well as to select wind turbines which can withstand wind conditions such as extreme wind speeds and turbulences expected at the sites. Generally in Japan, the external conditions are considered to be severer than in European countries because many landfalls and approaches of typhoons occur in every year and most of the land is covered with complex terrains. Furthermore, the number of damage cases of wind turbines by typhoons is increasing with growth in installation of wind turbines, and harmful influences from the regional peculiarity and the terrain conditions on turbines has been found to be outstanding. The above circumstance has increased the needs to select turbines carefully for objective sites performing evaluation of extreme wind speeds and turbulence intensity to be expected at the sites. The Guideline for Wind Turbines in Japan has been developed, which describes the procedure of selecting wind turbines in terms of wind conditions and clarifying problems to be solved. In order to collect data as evidence for the development of the Guideline, measurement campaigns on wind conditions and loads on turbines at four existing turbine sites were carried out during about one and a half years [1]. The seven sites were selected as target sites where typhoons frequently attack and/or high turbulences are generated. This paper describes the main measurement results of wind conditions and loads as well as the comparison of wind turbine loads between the measurement data and simulated results by aeroelastic simulations.
C D B A
Fig.1 Measuring Sites
Fig.2 MET and Turbine (Site B)
Table 1 Turbines Specifications Rated Power Rotor Diameter Hub Height
Site A
Site B
Site C
Site D
1 MW
300 kW
1 MW
600 kW
57 m
29 m
61.4 m
45 m
60 m
30.5 m
68 m
37 m
3. RESULTS Wind climates: Maximum wind speed Table 2 describes maximum wind speeds measured at each sites during the measurement campaigns from June 2006 to December 2007. The highest maximum speed of the all sites was 73 m/s that was measured at Site A in Okinawa at 13 July 2007. This wind speed was caused by the typhoon No. 0704 and has a level beyond the design gust speed of 70 m/s defined as the IEC Class1[2]. The second one is 58 m/s found at Site D, which is also brought by the typhoon No.0720. This is comparable to IEC Class 2 59.5 m/s. Although these IEC extreme wind speeds are defined as the value that might be occurred in 50 years, the comparable gust speed were measured at Japan sites in just 1 and a half year.
make a comparison of site TI to design TI defined by IEC Class A / B / C for the site assessment. The characteristic TI approximately corresponds to 90 percentile value of TI in each wind speed bin according to IEC 61400-1. In Fig.4 the IEC Class A line is drawn for comparison. At Site A, CharTI is approximately equal to Class A for wind speeds below 20 m/s. For above 20 m/s it becomes above Class A probably because of typhoons. At Site B, Char TI is slightly larger than Class A in high wind speed region. For Site C and D, the tendency of Char TI behaviours are similar such that for the lower wind speeds than 15 m/s the CharTI is below Class A and for the above 15 m/s wind speeds it is higher than Class A.
Site A
Site B
Site C
Site D
Table 2 Maximum wind speed Date yy/mm/dd 10 min mean speed Maximum wind speed
Site A
Site B
Site C
Site D
07/07/13
07/01/06
06/09/18
07/10/27
52 m/s
26 m/s
25 m/s
40 m/s
73 m/s
47 m/s
48 m/s
58 m/s
Wind climates: Annual wind conditions Annual wind conditions measured by the METs were analysed in terms of the wind speed distribution, the turbulence intensity and the ratio of u-, v-, w-turbulence components. For all analysis, the wind directions where the METs are behind the turbines so that the measured wind conditions are disturbed by the turbines were excluded. Figure 3 depicts the wind speed distributions for the four sites together with the average wind speed. The IEC wind speed distribution of Class 2 (8.5 m/s) is also depicted for comparison. It should be noted that the sum of the probability of all wind speeds is not equal to 100% because the wake directions were excluded as mentioned above. All the mean wind speeds are below IEC Class2 of 8.5 m/s. The highest average speed of the all sites is 7.68 m/s obtained at Site A, which is comparable to IEC Class3 of 7.5 m/s. Turbulence intensity (TI) distributions are shown in Fig. 4. A scatter point “Raw” means 10-min turbulence intensity. “AveTI” is the average of TI in each wind speed bins with width of 1 m/s. “CharTI” means the characteristic TI obtained according to the below equation:
Fig. 3 Wind Speed Distribution
Site A
Site B
Site C
Site D
Char.TI = Mean (TI) + 1.28 * Std.dev. (TI) . Fig. 4 Turbulence Intensity The equation is from the IEC 61400-1 Ed.3 and is used to
Site A
Site B
estimated loads according to Guidelines for Design of Wind Turbine Support Structures and Foundations where the tower loads in the extreme wind condition are simply estimated from the assumption of aerodynamic characteristics and geometry for blades, nacelle and tower [3]. From the comparison it can be deduced that the simple estimation shows good validation and it is useful for calculation of extreme tower loads. 4 Tower Base Lateral Moment Mxtb [-]
In order to resolve the turbulence intensity into three components, the standard deviations for three components of wind speed u-, v-, w-components were calculated as σu, σv, σw and then the ratio of σv/σu and σw/σu were estimated. The longitudinal direction of u was taken from 10-min averaged wind direction. According to IEC 61400-1, the values of σv/σu and σw/σu should be chosen as 0.8 and 0.5 used for the Kaimal spectrum, respectively. Figure 5 shows the bin-averaged ratios distributions against the wind speeds. The IEC values of 0.8 and 0.5 are also present. At Site A, σv/σu is below IEC for above 5 m/s speed and σw/σu shows similar values to IEC. For Site B, C and D, the both of σv/σu and σw/σu are beyond IEC values. Especially at Site D, the largest values can be found that σv/σu and σw/σu are averagely 0.9 and 0.6 respectively.
3 2 1 0 -1 -2 -3 -4 0
5
10
15
20
25
30
35
40
45
40
45
10 min Average Wind Speed [m/s]
Site C
Site D
Tower Base Tilt Moment Mytb [-]
4 3 2 1 0 -1 -2 -3 -4 0
5
10
15
20
25
30
35
10 min Average Wind Speed [m/s]
Fig. 6 Tower Base Loads (Site D) Fig. 5 Turbulence Component Ratio Tower loads on wind turbines In order to understand the behaviour of tower base loads against wind speeds, the 10 min statistical values for tower base loads are shown in Fig.6. The measured tower base loads are two components that are the tilt moment Mytb and the lateral moment Mxtb. The definition is shown in Fig.7.The tilt moment acts on the tower such that the tower is moved in the fore-aft direction. The lateral moment exert the side-side tower motion. The amplitude of Mxtb for wind speeds above cut-out is larger than that for wind speed below cut-out. Mytb shows the opposite behaviour that Mytb amplitude becomes smaller after the cut-out speed. The reason for it is that feathering the pitch to keep power output to the rated power makes the lift force direction change from the fore-aft direction to the side-side direction. The design loads for tower are determined by the resultant moment Mxytb of Mxtb and Mytb. For wind speeds below cut-out Mytb dominates Mxytb and for larger wind speeds Mxytb is largely dependent of Mxtb. Therefore the extreme loads on tower are determined by Mxtb. In Fig.8 the measured tower loads were compared to
Mytb
Fig. 7 Tower Moment Definition
Mxtb
タワー基部荷重 Mxytb
•
Comparing to the IEC turbulence classes, the measured turbulence intensity were found to be beyond the IEC Class A. On the contrary, the annual average wind speeds were below Class II. The turbine load responses in the extreme wind speed conditions were obtained during the measurement period. The calculation results by the aeroelastic code shows good agreement with the measured loads.
• Max Ave Ref.3
1
10
•
100
Mxtb Max Mxtb Ave Mytb Max Mytb Ave
Fig. 8 Comparison of Tower Loads with Ref.3 (Site A) Aeroelastic simulation results Aeroelastic load calculations were performed using a commercial code for wind turbine design and compared with the measured tower base loads in Fig.9. The data set measured during 2 or 3 days around the day when the maximum wind speeds were found were used. The target load components are 10 min max and average of Mxtb, Mytb and Mxytb. The measured values were binned in terms of wind speeds and averaged in each bins. For aeroelastic calculations, GH BLADED[4], in which the Blade Elementary Momentum Method for the aerodynamics modeling and the finite element model of mechanical structures are employed, was used that is made use of for the certification by Germanicher Lloyd. The inputs of the external conditions to the calculations were the air density, the upflow angle, the wind shear and the turbulence intensity. The normal operation and idling status were simulated. In Fig.9, the results for all the four sites are plotted together with the lines that mean 1, 1.2, 0.8, and 1.35 x calculated values. It is noted that 1.35 corresponds to the partial safety factor of loads defined by IEC 61400-1. The measured and calculated values agree well and most of values are in the range of 20% although in the calculations the detailed tuning has not been done such as the three components ratio of turbulence intensity, the atmospheric turbulence spectrum and the blade and tower eigenfrequency, etc. In order to improve the calculation accuracy the model tuning has to be done, however, the comparison shows that the calculation is valid without the detailed tuning. 4. CONCLUSIONS Main results are listed below: • The highest instantaneous wind speed during the measurement campaign was 73 m/s beyond Class I.
Measurement [-]
MET wind speed [m/s]
Mxytb Max Mxytb Ave x1 x 1.2 x 0.8 x 1.35
Calc. [-]
Fig. 9 Comparison of Measured and Calculated Tower Base Loads ACKNOWLEDGEMENT This work was carried out under the project of guideline for wind turbines in Japan, funded by New Energy and Industrial Technology Development Organization which are gratefully acknowledged. The authors also thank to Toyo sekkei Co. Ltd. and the Guideline committee members. REFERENCES [1] A.Honda, et al., “Overview of Japanese national project “Guideline for Wind Turbines in Japan,” Renewable Energy 2006. [2] IEC 61400-1 (Ed.3) Wind turbines Part1: Design Requirements, Aug. 2005. [3] T.Ishihara (ed.), “Guideline for Desgin of Wind Turbine Support Structures and Foundations (in Japanese),” Japan Society of Civil Engineers, June 2007. [4] Garrad Hassan & Partners Ltd., Web http://www.garradhassan.com/
GUIDELINE FOR WIND TURBINES IN JAPAN: MEASUREMENTS OF WIND CONDITIONS AND TURBINE LOADS KARIKOMI Kai, HONDA Akihiro and HIRAI Shigeto Nagasaki R&D Center of Mitsubishi Heavy Industries, Ltd. 5-717-1, Fukahori-machi, Nagasaki, 851-0392, Japan The Guideline for Wind Turbines in Japan has been developed, which describes the procedure of selecting wind turbines in terms of wind conditions and clarifying problems to be solved. In order to collect data as evidence for the development of the Guideline, measurement campaigns on wind conditions and loads on turbines at seven existing turbine sites were carried out during about one and a half years. The seven sites were selected as target sites where typhoons frequently attack and/or high turbulences are generated. This paper describes the main measurement results of wind conditions and loads as well as the comparison of wind turbine loads between the measurement data and simulated results by aeroelastic simulations
Keywords: Wind Turbine, Wind Condition, Turbulence, Typhoon, Load 1. INTRODUCTION
2. MEASURING CAMPAGIN
Guideline for Wind Turbines in Japan has been developed, which is aimed to be used by wind turbine business companies installing wind power plants in Japan sites. The content of the Guideline includes the following items: • Methods of estimating extreme wind speeds and turbulence intensity at objective sites; • Procedures of selecting suitable wind turbines; • Technical information to be referred in the selection of turbines; • Results of measurement and simulation performed during the development of the Guideline.
Sites and Measuring Equipment The seven selected measuring sites are shown in Fig.1. The sites cover most of Japan from Hokkaido to Okinawa. Most of all sites are considered not flat but complex terrain. In this paper, the results for four sites A, B, C, D are focused on. Specifications for the turbines installed for the sites are summarized in Table 1. All the turbines are conventional typed turbines; upwind, three-bladed, fixedspeed and pitch regulated turbines. At each of the sites a meteorological mast (MET mast) with the same height as the turbine was installed within one or two rotor diameter distance from the turbine. An example picture is shown in Fig.2. The met masts were equipped with wind speed and direction sensors at three heights. The top heights are same level as turbine hub heights. Three kinds of sensors were used such as cup/vane, propeller and ultrasonic anemometers. The measurements of loads on the turbine structural components were performed with strain gazes. The measured quantities included flapwise and edgewise root bending moments, tower bending moments in two perpendicular directions. The turbine operational statuses were recorded by means of power output, pitch angles, rotor speed and nacelle yaw position.
Before installing wind power plants, it is important to find sites where annual wind speeds are high enough to yield high energy productions as well as to select wind turbines which can withstand wind conditions such as extreme wind speeds and turbulences expected at the sites. Generally in Japan, the external conditions are considered to be severer than in European countries because many landfalls and approaches of typhoons occur in every year and most of the land is covered with complex terrains. Furthermore, the number of damage cases of wind turbines by typhoons is increasing with growth in installation of wind turbines, and harmful influences from the regional peculiarity and the terrain conditions on turbines has been found to be outstanding. The above circumstance has increased the needs to select turbines carefully for objective sites performing evaluation of extreme wind speeds and turbulence intensity to be expected at the sites. The Guideline for Wind Turbines in Japan has been developed, which describes the procedure of selecting wind turbines in terms of wind conditions and clarifying problems to be solved. In order to collect data as evidence for the development of the Guideline, measurement campaigns on wind conditions and loads on turbines at four existing turbine sites were carried out during about one and a half years [1]. The seven sites were selected as target sites where typhoons frequently attack and/or high turbulences are generated. This paper describes the main measurement results of wind conditions and loads as well as the comparison of wind turbine loads between the measurement data and simulated results by aeroelastic simulations.
C D B A
Fig.1 Measuring Sites
Fig.2 MET and Turbine (Site B)
Table 1 Turbines Specifications Rated Power Rotor Diameter Hub Height
Site A
Site B
Site C
Site D
1 MW
300 kW
1 MW
600 kW
57 m
29 m
61.4 m
45 m
60 m
30.5 m
68 m
37 m
3. RESULTS Wind climates: Maximum wind speed Table 2 describes maximum wind speeds measured at each sites during the measurement campaigns from June 2006 to December 2007. The highest maximum speed of the all sites was 73 m/s that was measured at Site A in Okinawa at 13 July 2007. This wind speed was caused by the typhoon No. 0704 and has a level beyond the design gust speed of 70 m/s defined as the IEC Class1[2]. The second one is 58 m/s found at Site D, which is also brought by the typhoon No.0720. This is comparable to IEC Class 2 59.5 m/s. Although these IEC extreme wind speeds are defined as the value that might be occurred in 50 years, the comparable gust speed were measured at Japan sites in just 1 and a half year.
make a comparison of site TI to design TI defined by IEC Class A / B / C for the site assessment. The characteristic TI approximately corresponds to 90 percentile value of TI in each wind speed bin according to IEC 61400-1. In Fig.4 the IEC Class A line is drawn for comparison. At Site A, CharTI is approximately equal to Class A for wind speeds below 20 m/s. For above 20 m/s it becomes above Class A probably because of typhoons. At Site B, Char TI is slightly larger than Class A in high wind speed region. For Site C and D, the tendency of Char TI behaviours are similar such that for the lower wind speeds than 15 m/s the CharTI is below Class A and for the above 15 m/s wind speeds it is higher than Class A.
Site A
Site B
Site C
Site D
Table 2 Maximum wind speed Date yy/mm/dd 10 min mean speed Maximum wind speed
Site A
Site B
Site C
Site D
07/07/13
07/01/06
06/09/18
07/10/27
52 m/s
26 m/s
25 m/s
40 m/s
73 m/s
47 m/s
48 m/s
58 m/s
Wind climates: Annual wind conditions Annual wind conditions measured by the METs were analysed in terms of the wind speed distribution, the turbulence intensity and the ratio of u-, v-, w-turbulence components. For all analysis, the wind directions where the METs are behind the turbines so that the measured wind conditions are disturbed by the turbines were excluded. Figure 3 depicts the wind speed distributions for the four sites together with the average wind speed. The IEC wind speed distribution of Class 2 (8.5 m/s) is also depicted for comparison. It should be noted that the sum of the probability of all wind speeds is not equal to 100% because the wake directions were excluded as mentioned above. All the mean wind speeds are below IEC Class2 of 8.5 m/s. The highest average speed of the all sites is 7.68 m/s obtained at Site A, which is comparable to IEC Class3 of 7.5 m/s. Turbulence intensity (TI) distributions are shown in Fig. 4. A scatter point “Raw” means 10-min turbulence intensity. “AveTI” is the average of TI in each wind speed bins with width of 1 m/s. “CharTI” means the characteristic TI obtained according to the below equation:
Fig. 3 Wind Speed Distribution
Site A
Site B
Site C
Site D
Char.TI = Mean (TI) + 1.28 * Std.dev. (TI) . Fig. 4 Turbulence Intensity The equation is from the IEC 61400-1 Ed.3 and is used to
Site A
Site B
estimated loads according to Guidelines for Design of Wind Turbine Support Structures and Foundations where the tower loads in the extreme wind condition are simply estimated from the assumption of aerodynamic characteristics and geometry for blades, nacelle and tower [3]. From the comparison it can be deduced that the simple estimation shows good validation and it is useful for calculation of extreme tower loads. 4 Tower Base Lateral Moment Mxtb [-]
In order to resolve the turbulence intensity into three components, the standard deviations for three components of wind speed u-, v-, w-components were calculated as σu, σv, σw and then the ratio of σv/σu and σw/σu were estimated. The longitudinal direction of u was taken from 10-min averaged wind direction. According to IEC 61400-1, the values of σv/σu and σw/σu should be chosen as 0.8 and 0.5 used for the Kaimal spectrum, respectively. Figure 5 shows the bin-averaged ratios distributions against the wind speeds. The IEC values of 0.8 and 0.5 are also present. At Site A, σv/σu is below IEC for above 5 m/s speed and σw/σu shows similar values to IEC. For Site B, C and D, the both of σv/σu and σw/σu are beyond IEC values. Especially at Site D, the largest values can be found that σv/σu and σw/σu are averagely 0.9 and 0.6 respectively.
3 2 1 0 -1 -2 -3 -4 0
5
10
15
20
25
30
35
40
45
40
45
10 min Average Wind Speed [m/s]
Site C
Site D
Tower Base Tilt Moment Mytb [-]
4 3 2 1 0 -1 -2 -3 -4 0
5
10
15
20
25
30
35
10 min Average Wind Speed [m/s]
Fig. 6 Tower Base Loads (Site D) Fig. 5 Turbulence Component Ratio Tower loads on wind turbines In order to understand the behaviour of tower base loads against wind speeds, the 10 min statistical values for tower base loads are shown in Fig.6. The measured tower base loads are two components that are the tilt moment Mytb and the lateral moment Mxtb. The definition is shown in Fig.7.The tilt moment acts on the tower such that the tower is moved in the fore-aft direction. The lateral moment exert the side-side tower motion. The amplitude of Mxtb for wind speeds above cut-out is larger than that for wind speed below cut-out. Mytb shows the opposite behaviour that Mytb amplitude becomes smaller after the cut-out speed. The reason for it is that feathering the pitch to keep power output to the rated power makes the lift force direction change from the fore-aft direction to the side-side direction. The design loads for tower are determined by the resultant moment Mxytb of Mxtb and Mytb. For wind speeds below cut-out Mytb dominates Mxytb and for larger wind speeds Mxytb is largely dependent of Mxtb. Therefore the extreme loads on tower are determined by Mxtb. In Fig.8 the measured tower loads were compared to
Mytb
Fig. 7 Tower Moment Definition
Mxtb
タワー基部荷重 Mxytb
•
Comparing to the IEC turbulence classes, the measured turbulence intensity were found to be beyond the IEC Class A. On the contrary, the annual average wind speeds were below Class II. The turbine load responses in the extreme wind speed conditions were obtained during the measurement period. The calculation results by the aeroelastic code shows good agreement with the measured loads.
• Max Ave Ref.3
1
10
•
100
Mxtb Max Mxtb Ave Mytb Max Mytb Ave
Fig. 8 Comparison of Tower Loads with Ref.3 (Site A) Aeroelastic simulation results Aeroelastic load calculations were performed using a commercial code for wind turbine design and compared with the measured tower base loads in Fig.9. The data set measured during 2 or 3 days around the day when the maximum wind speeds were found were used. The target load components are 10 min max and average of Mxtb, Mytb and Mxytb. The measured values were binned in terms of wind speeds and averaged in each bins. For aeroelastic calculations, GH BLADED[4], in which the Blade Elementary Momentum Method for the aerodynamics modeling and the finite element model of mechanical structures are employed, was used that is made use of for the certification by Germanicher Lloyd. The inputs of the external conditions to the calculations were the air density, the upflow angle, the wind shear and the turbulence intensity. The normal operation and idling status were simulated. In Fig.9, the results for all the four sites are plotted together with the lines that mean 1, 1.2, 0.8, and 1.35 x calculated values. It is noted that 1.35 corresponds to the partial safety factor of loads defined by IEC 61400-1. The measured and calculated values agree well and most of values are in the range of 20% although in the calculations the detailed tuning has not been done such as the three components ratio of turbulence intensity, the atmospheric turbulence spectrum and the blade and tower eigenfrequency, etc. In order to improve the calculation accuracy the model tuning has to be done, however, the comparison shows that the calculation is valid without the detailed tuning. 4. CONCLUSIONS Main results are listed below: • The highest instantaneous wind speed during the measurement campaign was 73 m/s beyond Class I.
Measurement [-]
MET wind speed [m/s]
Mxytb Max Mxytb Ave x1 x 1.2 x 0.8 x 1.35
Calc. [-]
Fig. 9 Comparison of Measured and Calculated Tower Base Loads ACKNOWLEDGEMENT This work was carried out under the project of guideline for wind turbines in Japan, funded by New Energy and Industrial Technology Development Organization which are gratefully acknowledged. The authors also thank to Toyo sekkei Co. Ltd. and the Guideline committee members. REFERENCES [1] A.Honda, et al., “Overview of Japanese national project “Guideline for Wind Turbines in Japan,” Renewable Energy 2006. [2] IEC 61400-1 (Ed.3) Wind turbines Part1: Design Requirements, Aug. 2005. [3] T.Ishihara (ed.), “Guideline for Desgin of Wind Turbine Support Structures and Foundations (in Japanese),” Japan Society of Civil Engineers, June 2007. [4] Garrad Hassan & Partners Ltd., Web http://www.garradhassan.com/
