- Email: [email protected]
Preliminary evaluation of monopile foundation dimensions for an offshore wind turbine by analyzing hydrodynamic load in the frequency domain
Renewable Energy 54 (2013) 211e218
Contents lists available at SciVerse ScienceDirect
Renewable Energy journal homepage: www.elsevier.com/locate/renene
Preliminary evaluation of monopile foundation dimensions for an offshore wind turbine by analyzing hydrodynamic load in the frequency domain Ki-Yong Oh*, Ji-Young Kim, Jun-Shin Lee Technology Commercialization Office, KEPCO Research Institute, Daejeon 305-760, Republic of Korea
a r t i c l e i n f o
a b s t r a c t
Article history: Received 27 January 2012 Accepted 1 August 2012 Available online 28 August 2012
Although design of offshore wind turbines has many similarities to that of onshore turbines, a lot of considerations should be made for the additional substructure imposed on hydrodynamic loads. The additional substructure prolongs the total tower length, increasing the tower bending moment and lowering the natural bending frequencies of the tower. Accordingly, system dynamic analyses associated with hydrodynamic load should be performed in the frequency domain in order to avoid bending modes of tower from the operation frequency ranges. In this paper, a method to generate hydrodynamic load for a finite element analysis is introduced, considering the characteristics of sea conditions for a candidate site of demonstration offshore wind farm in the west sea of Korea. In addition, a wind energy conversion system with a monopile foundation is fully modeled using the finite element method to simulate the various conditions based on IEC standard. Based on the FEM analyses of tower bending modes, optimal dimensions of the monopile for the candidate site are proposed. Crown Copyright Ó 2012 Published by Elsevier Ltd. All rights reserved.
Keywords: Design of monopile Frequency domain analysis Monopile foundation Offshore wind turbine Offshore wind farm
1. Introduction Offshore wind farms provide higher energy density, have lower spatial limitations, and are less likely to generate complaints from the public than onshore wind farms. Despite such advantages, because the installation of offshore wind turbines is difficult and construction costs are far higher than those of onshore wind turbines [1], a stringent feasibility study must be conducted in advance of construction. Various researches regarding foundation design have been carried out. R. S. Nehal et al. [2] designed a monopile foundation for 3.6 and 6.0 MW wind turbines to investigate the possibility of placing wind turbines in the North Sea, some kilometers from the Dutch west coast. B. W. Byrne et al. [3] evaluated foundation modeling methods from a civil engineering point of view to support and accelerate the large-scale commercial development of offshore wind farms in the UK. H. J. T. Kooijman et al. [4] described the predesign and analyzed the stationary aerodynamic performance and natural frequency of a 6 MW wind turbine with PHATAS in the framework of the DOWEC project. While intensive research, as aforementioned, has been carried out on the foundations of wind turbines, systematic studies on the foundation of offshore wind turbine that consider the characteristics of sea conditions for candidate offshore wind farm site have not been performed to date. * Corresponding author. Tel.: þ82 42 865 5376; fax: þ82 42 865 5202. E-mail addresses: [email protected], [email protected] (K.-Y. Oh).
In this paper, a method to generate the hydrodynamic load for a finite element analysis considering the characteristics of a candidate site is introduced. A full model of the wind turbine and foundation using the finite element method is then designed and an analysis technique involving a multi body dynamic system is introduced to simulate the various conditions based on IEC code [5]. A wind turbine for low wind speed is introduced in order to enhance the economic feasibility of the offshore wind farm. Optimal dimensions of the monopile for the candidate site are finally proposed based on an analysis of the hydrodynamic load and operation range in the frequency domain. 2. Hydrodynamic load 2.1. Sea conditions To select the candidate site for a wind farm, not only wind resources but also diverse factors including the distance from substations, water depth, and the use of the sea must be taken into consideration [6]. Through an analysis of these factors, the Korea Electric Power Corporation Research Institute (KEPCO Research Institute) has determined that the area near the island Wi-do in the West Sea (Yellow Sea) is the optimal site for demonstration offshore wind farm. The IEC stipulates the evaluation of hydrodynamic loads in the normal sea state and extreme sea state in a recurrence period of 50 years to estimate the integrity of wind turbines and foundations [5].
0960-1481/$ e see front matter Crown Copyright Ó 2012 Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.renene.2012.08.007
212
K.-Y. Oh et al. / Renewable Energy 54 (2013) 211e218
Nomenclature a Ai cd cm D f fi fp F Ftotal h H
significant wave height shape function PiersoneMoskowitz spectrum spectrum density of wave acceleration spectrum density of wave elevation spectrum density of wave velocity wave velocity velocity of the flow resolved normal to the member acceleration of the flow resolved normal to the member mean water depth of site density of water random phase in the [p, p] range at spectral density
Hs N SPM SPMa SPMd SPMu u U U_
wave acceleration i-th magnitude at spectral density drag coefficient inertia coefficient diameter of foundation frequency i-th frequency at spectral density peak frequency wave force per unit length of foundation total wave load depth of water height of foundation
z
r
ɸi
For reliability assessments of the wind turbine and foundation, significant wave height and the peak frequency of waves with a recurrence period of 50 years and waves for the normal sea state should be measured in order to describe the wave conditions of the candidate site. In this paper, hindcasted wave data estimated for the past 24 years from January 1979 to December 2002 by the KORDI (Korea Ocean Research and Development Institute) [7] has been analyzed, as there is no long-term measurement data for the candidate site. Fig. 1 denotes the calculation lattices of the hindcasted wave data for the candidate site. Extreme wave conditions for a recurrence period of 50 years per wave direction at locations 055120 and 056120 between the islands Wi-do and Anma-do, where a test bed is to be constructed, were analyzed and are shown in Table 1. Through the height and period of the waves thus calculated retrogressively, the candidate site was observed to have the strongest wave height and period at NW for a recurrence period of 50 years. This is identical to the results of investigations of wind resources on the west coast, where NW was the main wind direction [8]. When the correlation between the wind and the waves is taken into consideration, it is possible to infer that the derived results are reliable. Because both lattice points were being considered as candidate sites for an offshore wind farm, in the present study, it was decided that the wave height and the wave period of lattice point 055120 NW, which had the greatest wave height and the longest wave period from among the two lattice points, would be selected as the extreme waves for a recurrence period of 50 years and the hydrodynamics would be reproduced. In addition, it could be observed that the average and maximum wave heights were 0.6 m and 1.8 m, respectively, and the average and maximum peak periods were 3.6 s and 6.7 s, respectively, in the normal sea state.
4 fp ; SPM ðf Þ ¼ 0:3125$Hs2 $fp4 $f 5 $exp 1:25 f
(2)
where Hs denotes the significant wave height (m), fP is the peak 1 (Hz), and f is the frequency (Hz). The spectrum frequency ¼ Tp of velocity and acceleration can be converted by substituting Eq. (2) into Eqs. (3) and (4) [9]:
SPMu ðf Þ ¼ 2pf $SPMd ðf Þ;
(3)
SPMa ðf Þ ¼ 2pf $SPMu ðf Þ;
(4)
where SPMd represents the spectral density of wave elevation, SPMu the spectral density of wave velocity, and SPMa the spectral density of wave acceleration. Calculated spectrums are shown in Fig. 2. The velocity and acceleration spectrums generated using Eqs. (3) and (4) can be transformed into non-stationary time series data by using Eqs. (5) and (6) [10]:
uðtÞ ¼
N X
Aui sinð2pfi t þ fi Þ;
(5)
Aai sinð2pfi t þ fi Þ;
(6)
i¼1
aðtÞ ¼
N X i¼1
where Ai and fi represents the i-th magnitude and frequency at the spectral density. ɸi is generated randomly in the [p, p] range. Ai can be calculated using Eq. (7).
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi SPM ðfi Þ$Df :
2.2. Non-stationary random hydrodynamic load
Ai ¼
Morison’s equation, given as Eq. (1), is commonly used to depict hydrodynamic loads [5]:
By substituting Eqs. (5) and (6) into Eq. (1), F(t) can be calculated. Total wave load can be estimated by using Eq. (8).
F ¼
pD2 _ 1 U; cd rDjUjU þ cm r 2 4
(1)
where F represents the wave force per unit length of the foundation, r is the density of water, D is the diameter of the foundation, U denotes the velocity of the flow resolved normal to the member, U_ denotes the acceleration of the flow resolved normal to the member. cd and cm are the drag coefficient and the inertia coefficient. When Eq. (1) is examined, to calculate the hydrodynamic load, the velocity and acceleration of the waves that act on the structures must be calculated. To simulate the waves that act on the structures, the PiersoneMoskowitz spectrum delineated in Eq. (2) was used [5]:
(7)
Zz Ftotal ðtÞ ¼
NðhÞ$FðtÞdh;
(8)
0
where h denotes depth of water, z is the mean water depth of site and N(z) denotes the shape function as given in Eq. (9):
NðzÞ ¼
3 z 2 1 z 3 ; 2 H 2 H
(9)
where H is the height of the foundation. The non-stationary random hydrodynamic load regenerated using the above procedure is described in Fig. 3 (Fig. 4).
K.-Y. Oh et al. / Renewable Energy 54 (2013) 211e218
213
Fig. 1. Calculation lattices of the hindcasted wave data for candidate site [7].
3. Design 3.1. Wind turbine The wind class of the sea around the Korean Peninsula is class 3e5, which makes the region a feasible for construction of wind farm. However, when compared to the wind class of Horns Rev offshore wind farm and Nysted offshore wind farm in Europe, which are class 7 or above, it is relatively inferior in terms of economic efficiency [11]. In the present study, to improve the economic efficiency of the proposed wind farm, the wind turbine Table 1 Extreme wave conditions with a recurrence period of 50 years. Point
055120
Direction
Height [m]
Period [s]
Height [m]
056120 Period [s]
N NNE NE ENE E ESE SE SSE S SSW SW WSW W WNW NW NNW
5.52 3.85 3.45 4.03 4.89 5.43 6.09 6.13 6.53 7.06 6.48 4.90 6.01 6.85 7.22 5.98
9.4 7.7 7.1 7.5 8.2 8.7 9.5 10.1 10.9 11.9 11.8 10.1 11.0 11.7 11.9 10.4
5.50 3.46 2.98 3.13 3.65 4.48 4.90 4.84 5.30 5.84 5.70 4.61 5.56 6.91 7.00 5.16
9.4 7.2 6.6 6.6 7.2 8.2 9.1 9.5 10.2 11.2 11.4 9.9 10.7 11.9 11.8 9.6
blade has been designed as a low wind speed-type. Because wind turbines with low wind speed blades have identical rated power but are relatively longer in blade length, the load imposed on the tower and the foundation increases. It is possible to infer that the foundation designed in this paper can withstand greater load than wind turbine foundations of the same capacity that are commercially operated at present. Specification of wind turbine in this paper is given in Table 2. The rotor was designed as a fixed axle type where only the torque is transmitted to the rotor shaft and the horizontal load and axial load are supported by the nacelle, thus decreasing the load on the rotor shaft [12]. The total diameter of the rotor is 100 m, making it 10 m longer than that of Vestas V90, the most widely installed type among 3 MW wind turbines. Table 2 Specification of wind turbine. Rating Rotor orientation, configuration Control Gearbox type Gearbox ratio Rotor diameter Hub diameter Cut-in wind speed Rated wind speed Cut-out wind speed Rated rotor velocity Tower height Tower inner diameter Tower outer diameter Foundation height Axle type
3 MW Upwind, 3 blades Variable velocity, collective pitch 2 Planetary, 1 helical 104.8672 100.4 m 3m 3 m/s 13 m/s 25 m/s 15 rpm 80 m 3.94 m 4.30 m 30 m Fixed axle
214
K.-Y. Oh et al. / Renewable Energy 54 (2013) 211e218
Fig. 4. Finite element model of wind turbine components.
Fig. 2. Power spectrum of wave elevation, velocity and acceleration.(a) Normal sea state (Hs ¼ 0.6 m, Ts ¼ 3.6 s).(b) Extreme sea state (Hs ¼ 7.22 m, Ts ¼ 11.9 s).
The gearbox is assumed to be a typical multiple-stage gearbox but with no frictional losses. The generator is assumed to be a PMG (Permanent Magnetic Generator), as it shows high efficiency in low wind speed. The electrical efficiency of the generator was taken as 94.4%. This was chosen to be roughly the same as the total mechanical-to-electrical conversion loss of the DOWEC turbine at rated power [4]. With rated generator speed of 1575 rpm, rated electric power of 3 MW, and a generator efficiency of 94.4%, the rated mechanical power is 3.178 MW and the rated generator torque is 19,260 Nm. The tower height is set to 80 m, the typical hub height of 3 MW class wind turbines. The tower is composed of five flanges and four segments. Because the water depth of the candidate site is approximately 15 m, a monopole-type foundation, which is known to be the most affordable in such shallow depths, was selected. Further, in consideration of tide differences and the extreme wave height, the total height was designed to be 30 m. The mode is to be analyzed through a multi body dynamic analysis of the turbine and the foundation; therefore, based on this, the optimal design is to be derived, considerations of the ground and other aspects were omitted, and research was conducted on the premise that the ground was fixed. FEM model for analysis is shown in Fig. 4. SAMCEF Wind Turbines, which has a generic non-linear finite elements solver including multi body simulation features, was used for the integrated load analyses of wind turbine. It can make 3D turbulence wind profile interlinked with TurbSim [13]. The aerodynamic torque imposed on the blade can be calculated using the blade element method (BEM) [14]. Based on this load, a finite element analysis was conducted on the components of each part, and a multi body simulation was simultaneously carried out for full model. The super element method was used for the components of parts in the wind turbine [15]. 3.2. Foundation
Fig. 3. Extreme hydrodynamic load of candidate site with a recurrence period of 50 years.
For the foundation design, aerodynamic load and hydrodynamic load must be taken into consideration. As can be observed in Eq. (1), wave velocity and acceleration should be considered in frequency domain in that the hydrodynamics, load stemming from velocity and acceleration are both in operation. The foundation receives period fatigue load from the waves in the normal condition and,
K.-Y. Oh et al. / Renewable Energy 54 (2013) 211e218 Table 3 Exclusive conditions for design of the foundation.
Force Force Force Force Force Force 1P 2P 3P 4P
by by by by by by
Vel. (NSS, mean) Acc. (NSS, mean) Vel. (NSS, max) Acc. (NSS, max) Vel. (ESS) Acc. (ESS)
Min. freq.
Max. freq.
0.244 0.256 0.127 0.133 0.074 0.078 0.225 0.450 0.638 0.900
0.374 0.420 0.195 0.219 0.113 0.127 0.275 0.550 0.863 1.100
though infrequent, is subjected to great strength under extreme load. Accordingly, the average and maximum figures under normal conditions and extreme load for a recurrence pffiffiffi period of 50 years are all taken into consideration. Because 1= 2 or less of the amplitude of the significant frequency is the value where the strength is reduced by half, the exclusive condition for the waves is to be designed such that the bending mode of the foundation and the pffiffiffi tower exist at locations 1= 2 or less of the velocity and acceleration power spectrum values. In a turbine with a three-bladed rotor, the aerodynamic frequency of excitation occurs at three times the rotational frequency of the rotor (3P). The foundation’s first natural bending frequency coupled with the tower must not under any circumstances coincide with the critical exciting forces. Moreover, care must be taken to ensure that a certain distance from the remaining multiples of the rotor frequency is maintained. The distances between frequencies at which excessive vibrations will occur are determined by system damping, i.e., both structural as well as aerodynamic damping. Experience from existing turbines indicates that a safety distance of 0.25P from the dominant frequency of excitation (3P) and of 0.20P from the less critical frequency (1P, 3P, 4P) are good guide values [16]. Table 3 lists the exclusive conditions for the design of the foundation. The allowable region is 0.219e0.225 Hz, 0.420e 0.450 Hz, 0.550e0.638 Hz, and 0.863e0.9000 Hz, as shown in Fig. 5. To select the optimal foundation thickness in consideration of the exclusive condition, a modal analysis was conducted, and the
215
results are shown in Table 4. The main mode of the tower consists of the two modes of fore-after and sideeside. When the above exclusive condition is taken into consideration, it can be observed that, as the foundation thickness, values from 0.036 m to 0.039 m are possible. Because the construction costs decrease as the thickness decreases, it can be inferred that 0.036 m is the optimal thickness. When the design is such that the thickness is increased so that the natural frequency falls in a range of 0.420e0.450 Hz, the structures’ integrity increases further but the costs for producing the foundation increase geometrically. Currently, at commercially operated Horns Rev offshore wind farms, the rotor’s diameter is 80 m yet the thickness is 0.05 m [17]. Considering that foundationrelated costs generally take up approximately 20% of the total costs in constructing offshore wind farms, when foundations are designed in consideration of the aerodynamics and hydraulics, as suggested in this paper, the thickness should be decreased dramatically in comparison with existing design methods, and thus it can be anticipated that costs will be reduced dramatically.
4. Assessment of reliability In order for a manufacturer to obtain the certificate of its wind turbines by the certification authority, many design load cases must be analyzed and their reliability must be verified [18]. With offshore wind turbines, because hydrodynamic load acts on the foundation, the design evaluation must be conducted on more Design Load Cases (DLCs) than for onshore wind turbines and the reliability must accordingly be verified. For this, the IEC stipulates the categories necessary for offshore wind turbines and foundation design evaluation in 61400-3 and demands that a reliability evaluation be conducted. To validate the propose method for the design of offshore wind turbine, reliability of designed wind energy conversion system should be evaluated based on this standard. In this paper, the wind turbine and the foundation were evaluated for their reliability in terms of gusts and the extreme wind speed, to which wind turbines are considered to be especially vulnerable as well as DLC 1.1 among the many design load cases. Considering that offshore wind farms are inferior to onshore wind farms in terms of the surrounding environment, integrity was evaluated with wind
Fig. 5. Exclusive conditions for design of the foundation.
216
K.-Y. Oh et al. / Renewable Energy 54 (2013) 211e218
Table 4 Bending mode variation of foundation coupled with tower. Thickness (m)
Fore-after bending mode (Hz)
Sideeside bending mode (Hz)
0.027 0.030 0.033 0.036 0.039 0.042 0.045 0.048 0.051 0.054 0.057 0.060
0.1982 0.2083 0.2142 0.2196 0.2244 0.2288 0.2328 0.2364 0.2398 0.2429 0.2458 0.2485
0.1974 0.2079 0.2138 0.2192 0.2241 0.2285 0.2325 0.2361 0.2395 0.2426 0.2455 0.2482
turbine class IA, which is the strongest design basis, as the standard. Details of design load cases for simulations are depicted in Table 5. Fig. 6. DLC 1.1 (V ¼ 25 m/s, I15 ¼ 0.18).
4.1. DLC 1.1 In this design load case, the wind turbine is connected to electric load and producing electric power. As the load condition to confirm that the wind turbine operates without any problems when 3D isotropic turbulence occurs, this can be used as the basis to calculate a wind turbine performance curve. Fig. 6 shows the pitch angle and the generation power in terms of the wind speed when the average wind speed is 25 m/s and the turbulence intensity is 0.18. It can be observed that, even though the wind constantly changes and blows, control is satisfactory and electric power is generated. To evaluate the performance of wind turbines for low wind speed, the power curve of the wind turbine, calculated by sequentially increasing the average velocity from 3 m/s to 25 m/s, is shown in Fig. 7. Because the suggested wind turbine is 10 m longer than the Vestas V90 in the rotor diameter, it could generate more electric power at low wind speed. Also its performance was outstanding at roughly the rated output, because the generator of wind turbine was designed as a PMG-type. In consideration of the fact that the average wind speed of the West Sea is 7e8 m/s [19], the annual energy production is calculated with an average wind speed of 7 m/ s; the figure is 6123 MWh for the KEPCOeRI wind turbine and 5784 MWh for Vestas V90, thus showing that the annual energy production is increased by 5.86%. In case that the average wind speed is 8 m/s, the annual energy production is 8162 MWh for the KEPCOeRI wind turbine and 7752 MWh for Vestas V90, respectively, thus constituting an increase of 5.29%. In calculating the annual energy production above, loss was disregarded. Considering that the wind turbine operation period is 20 years, it is evident that the introduction of wind turbines for low wind speed will improve the economic efficiency of wind farms. In particular, considering that development costs for offshore wind farms are higher than those for onshore wind farms, it is imperative that wind turbines for low wind speed be introduced in constructing the former.
during power generation. Although an integrity evaluation must be conducted on a recurrence period of 1 year or 50 years, in the present study, the integrity evaluation was conducted on the condition of a recurrence period of 50 years, which is a relatively extreme condition. Fig. 8 shows the results of the evaluation when the average wind speed is 25 m/s. The statues of wind turbine were observed in case that gust with a recurrence period of 50 years occurs from 150 s to 164 s and wind turbine is disconnected from grid during 150 s to 180 s as shown in Fig. 8(a). As the blade pitch angle is controlled during gust, the aerodynamic torque is decreased. In addition, it can be observed that the blade pitch angle is adjusted also during system loss, because no electric power is transmitted to grid. Fig. 8(b) shows the distribution of stress. The maximum stress amounts to 72 MPa on the ground and at the contact point, which are subjected to the greatest bending moment. The wind turbine stably operate even under load conditions identical to this condition. 4.3. DLC 3.2 This case is the category for evaluating the integrity of the load that acts on the wind turbine during the transient region from a suspended or idling state to power generation. The integrity evaluation was conducted for a recurrence period of 50 years,
4.2. DLC 2.3 This load condition is the category for evaluating the reliability of the wind turbine when gust and grid loss occur simultaneously Table 5 Design load cases [5]. Load case
Wind condition
Wave condition
1.1 2.3 3.2 6.1c
NTM, Vin < Vhub < Vout EOG, Vhub ¼ Vr 2 m/s and Vout EOG, Vhub ¼ Vin, Vr 2 m/s and Vout Vhub ¼ Ve50
NSS NSS NSS EWH(H ¼ H50)
Fig. 7. Power curves (kW). (a) Pitch angle and generated power via gust and grid loss (b) Stress distribution.
K.-Y. Oh et al. / Renewable Energy 54 (2013) 211e218
217
Fig. 9. DLC 3.2 (EOG50, V ¼ 25 m/s).
Fig. 8. DLC 2.3 (EOG50, V ¼ 25 m/s). (a) Pitch angle and generated power via gust (b) Stress distribution.
which is a relatively extreme condition. Gusts began at 70 s, which was before the turbine entered the normal condition, and controller is operated to regulate the power and reduce the load. As shown in Fig. 9(a), even when gusts are occurred, the wind turbine operates without any problems and withstands the transient load in the transient region. Fig. 9(b) shows the distribution of stress in this case. The maximum stress amounts to 57 MPa on the ground and at the contact point, which are subjected to the greatest bending moment. Maximum stress of this design load case is smaller than one of DLC 2.3 because there is no grid loss, which could be considered as additional load from the wind turbine point of view. Consequently, the wind turbine reliably operate even under load conditions identical to this condition.
load act on the wind turbine and the foundation is shown in Fig.10. It can be observed that the maximum stress amounts to 257 MPa on the ground and at the contact point, which are subjected to the greatest bending moment. Considering that the ultimate strength of iron is 400 MPa, the safety factor is 1.56 or above, which satisfies the
4.4. DLC 6.1c This case is the category for evaluating the reliability of the wind turbine and the foundation when they are parked in a suspended or idling state and subjected to extreme wind speed and extreme waves. An integrity evaluation is accomplished for a recurrence period of 50 years because it is more severe case than extreme load case with a recurrence period of 1 year. The maximum load distribution in cases where extreme aerodynamic load and hydrodynamic
Fig. 10. DLC 6.1c (EWM50, Vhub ¼ 50 m/s).
218
K.-Y. Oh et al. / Renewable Energy 54 (2013) 211e218
stipulation in the regulations (i.e. 1.35 or above). Consequently, the wind turbine and the foundation proposed in the present study are equivalent to wind turbine class I in terms of the ultimate load. Through an evaluation of the above conditions, which are seen as especially weak from among the design conditions of wind turbines, it could be observed that, even when the thickness of the foundation was set at 0.036 m, it was strong enough to withstand the external load imposed on the wind turbine. If the wave conditions of candidate sites for offshore wind farms are analyzed in the frequency domain and, based on this, the foundation is designed, it will then be possible to improve the economic efficiency of offshore wind farms, as has been mentioned above.
5. Conclusions This paper presents optimal foundation design methods that take into consideration aerodynamic load and hydrodynamic load on the basis of the environmental conditions of a candidate site for an offshore wind farm. Each load was analyzed in the frequency domain to select possible locations for designing the main modes of the tower and the foundation. Based on results of analysis, the optimal dimensions for a monopile foundation were presented. An integrity evaluation was conducted on the designed turbine and foundation in accordance with IEC standards to verify the reliability of the optimized turbine and foundation. In addition, methods to improve economic efficiency when constructing offshore wind farms such as the introduction of wind turbine for low speed and construction cost reduction through the use of optimally designed foundations were presented.
Acknowledgment This work was supported by the New & Renewable Energy of the Korea Institute of Energy Technology Evaluation and Planning (KETEP), grant funded by the Korea government Ministry of Knowledge Economy. (20113040020010)
References [1] Krohn Soren, Morthorst Poul-Erik, Awerbuch Shimon. The economics of wind energy. The European Wind Energy Association; 2009. [2] Nehal RS, Nedam Ballast. Foundation design monopile 3.6 & 6.0 mW wind turbines. DOWEC, Dutch; 2001. [3] Byrne BW, Houlsby GT. Foundations for offshore wind turbines. Philosophical Transactions of the Royal Society, A: Mathematical, Physical and Engineering Sciences 2003;361:2909e30. [4] Kooijman HJT, Lindenburg C, Winkelaar D, van der Hooft EL. DOWEC 6 MW pre-design: aero-elastic modeling of the DOWEC 6 MW pre-design in PHATAS. ECN-CXe01e135. Energy Research Centre of the Netherlands; 2003. [5] IEC 61400-3 wind turbinesdPart 3: design requirements for offshore wind turbines. IEC; 2009. [6] Kim JY, Kang KS, Oh KY, Lee JS and Ryu MS. A study on the site selection of offshore wind farm around Korean Peninsula. In. 3rd International conference on ocean energy, Bilbao, Spain; 2010. [7] Lee GS and Gwon OS. Analysis of ocean characteristics and design support for construction of offshore wind farm. 1st year interim report, KORDI; 2009. [8] Oh Ki-Yong, Kim Ji-Young, Lee Jun-Shin, Ryu Ki-Wahn. Wind resource assessment around Korean Peninsula for feasibility study on 100 MW class offshore wind farm. Renewable Energy 2012;42:217e26. [9] Choi YH, Kwon SH, Kim DW, Park SK. A reliability study of coastal structures under the influence of wave and currentsdrandom analysis of fixed structures. Journal of Ocean Engineering and Technology 1995;9(2):186e92. [10] Ki-Yong Oh, Ji-Young Kim, JaeKyung Lee, Joon-Young Park, Moo-Sung Ryu, Keum-Seok Kang and Jun-Shin Lee. Design of offshore wind turbine with foundation based on modal analysis. In. The 17th International congress on sound & vibration, Cairo, Egypt; 2010. [11] Archer Cristina L, Jacobson Mark Z. Evaluation of global wind power. Journal of Geophysical Research 2005;110. [12] Heege A, Viladomiu P, Betran J, Radovcic Y, Latorre M, Cantons JM, et al. Impact of wind turbine drive train concepts on dynamic gearbox loads. Milan, Italy. In: Proceedings of European wind energy conference 2007; 2007. [13] Jonkman BJ. TurbSim user’s guide: version 1.50. National Renewable Energy Laboratory; 2009. [14] Munteanu Lulian, Bratcu Antoneta luliana, Cutululis Nicolaos-Antonio, Ceanga Emil. Optimal control of wind energy systems. Springer; 2008. 30e57. [15] Kohno Kazukiyo, Tsunada Tomohiro, Seto Hideyuki, Tanaka Michio. Hybrid stress analysis of boundary and finite elements by a super-element method. Finite Elements in Analysis and Design 1991;7(4):279e90. [16] Hau Erich. Wind turbines: fundamentals, technologies, application, economics. 2nd ed. Springer; 2006. p. 408e11. [17] http://www.hornsrev.dk/Engelsk/default_ie.htm. [18] IEC 61400-1 wind turbinesdpart 1: design requirements. 3rd ed. IEC; 2005. [19] Jang Jea-Kyung, Yu Byoung-Min, Ryu Ki-Wahn, Offshore Jun-Shin Lee. Wind resource assessment around Korean Peninsula by using QuikSCAT satellite data. Journal of the Korean Society for Aeronautical & Space Sciences 2009; 37(11):1121e30.
Contents lists available at SciVerse ScienceDirect
Renewable Energy journal homepage: www.elsevier.com/locate/renene
Preliminary evaluation of monopile foundation dimensions for an offshore wind turbine by analyzing hydrodynamic load in the frequency domain Ki-Yong Oh*, Ji-Young Kim, Jun-Shin Lee Technology Commercialization Office, KEPCO Research Institute, Daejeon 305-760, Republic of Korea
a r t i c l e i n f o
a b s t r a c t
Article history: Received 27 January 2012 Accepted 1 August 2012 Available online 28 August 2012
Although design of offshore wind turbines has many similarities to that of onshore turbines, a lot of considerations should be made for the additional substructure imposed on hydrodynamic loads. The additional substructure prolongs the total tower length, increasing the tower bending moment and lowering the natural bending frequencies of the tower. Accordingly, system dynamic analyses associated with hydrodynamic load should be performed in the frequency domain in order to avoid bending modes of tower from the operation frequency ranges. In this paper, a method to generate hydrodynamic load for a finite element analysis is introduced, considering the characteristics of sea conditions for a candidate site of demonstration offshore wind farm in the west sea of Korea. In addition, a wind energy conversion system with a monopile foundation is fully modeled using the finite element method to simulate the various conditions based on IEC standard. Based on the FEM analyses of tower bending modes, optimal dimensions of the monopile for the candidate site are proposed. Crown Copyright Ó 2012 Published by Elsevier Ltd. All rights reserved.
Keywords: Design of monopile Frequency domain analysis Monopile foundation Offshore wind turbine Offshore wind farm
1. Introduction Offshore wind farms provide higher energy density, have lower spatial limitations, and are less likely to generate complaints from the public than onshore wind farms. Despite such advantages, because the installation of offshore wind turbines is difficult and construction costs are far higher than those of onshore wind turbines [1], a stringent feasibility study must be conducted in advance of construction. Various researches regarding foundation design have been carried out. R. S. Nehal et al. [2] designed a monopile foundation for 3.6 and 6.0 MW wind turbines to investigate the possibility of placing wind turbines in the North Sea, some kilometers from the Dutch west coast. B. W. Byrne et al. [3] evaluated foundation modeling methods from a civil engineering point of view to support and accelerate the large-scale commercial development of offshore wind farms in the UK. H. J. T. Kooijman et al. [4] described the predesign and analyzed the stationary aerodynamic performance and natural frequency of a 6 MW wind turbine with PHATAS in the framework of the DOWEC project. While intensive research, as aforementioned, has been carried out on the foundations of wind turbines, systematic studies on the foundation of offshore wind turbine that consider the characteristics of sea conditions for candidate offshore wind farm site have not been performed to date. * Corresponding author. Tel.: þ82 42 865 5376; fax: þ82 42 865 5202. E-mail addresses: [email protected], [email protected] (K.-Y. Oh).
In this paper, a method to generate the hydrodynamic load for a finite element analysis considering the characteristics of a candidate site is introduced. A full model of the wind turbine and foundation using the finite element method is then designed and an analysis technique involving a multi body dynamic system is introduced to simulate the various conditions based on IEC code [5]. A wind turbine for low wind speed is introduced in order to enhance the economic feasibility of the offshore wind farm. Optimal dimensions of the monopile for the candidate site are finally proposed based on an analysis of the hydrodynamic load and operation range in the frequency domain. 2. Hydrodynamic load 2.1. Sea conditions To select the candidate site for a wind farm, not only wind resources but also diverse factors including the distance from substations, water depth, and the use of the sea must be taken into consideration [6]. Through an analysis of these factors, the Korea Electric Power Corporation Research Institute (KEPCO Research Institute) has determined that the area near the island Wi-do in the West Sea (Yellow Sea) is the optimal site for demonstration offshore wind farm. The IEC stipulates the evaluation of hydrodynamic loads in the normal sea state and extreme sea state in a recurrence period of 50 years to estimate the integrity of wind turbines and foundations [5].
0960-1481/$ e see front matter Crown Copyright Ó 2012 Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.renene.2012.08.007
212
K.-Y. Oh et al. / Renewable Energy 54 (2013) 211e218
Nomenclature a Ai cd cm D f fi fp F Ftotal h H
significant wave height shape function PiersoneMoskowitz spectrum spectrum density of wave acceleration spectrum density of wave elevation spectrum density of wave velocity wave velocity velocity of the flow resolved normal to the member acceleration of the flow resolved normal to the member mean water depth of site density of water random phase in the [p, p] range at spectral density
Hs N SPM SPMa SPMd SPMu u U U_
wave acceleration i-th magnitude at spectral density drag coefficient inertia coefficient diameter of foundation frequency i-th frequency at spectral density peak frequency wave force per unit length of foundation total wave load depth of water height of foundation
z
r
ɸi
For reliability assessments of the wind turbine and foundation, significant wave height and the peak frequency of waves with a recurrence period of 50 years and waves for the normal sea state should be measured in order to describe the wave conditions of the candidate site. In this paper, hindcasted wave data estimated for the past 24 years from January 1979 to December 2002 by the KORDI (Korea Ocean Research and Development Institute) [7] has been analyzed, as there is no long-term measurement data for the candidate site. Fig. 1 denotes the calculation lattices of the hindcasted wave data for the candidate site. Extreme wave conditions for a recurrence period of 50 years per wave direction at locations 055120 and 056120 between the islands Wi-do and Anma-do, where a test bed is to be constructed, were analyzed and are shown in Table 1. Through the height and period of the waves thus calculated retrogressively, the candidate site was observed to have the strongest wave height and period at NW for a recurrence period of 50 years. This is identical to the results of investigations of wind resources on the west coast, where NW was the main wind direction [8]. When the correlation between the wind and the waves is taken into consideration, it is possible to infer that the derived results are reliable. Because both lattice points were being considered as candidate sites for an offshore wind farm, in the present study, it was decided that the wave height and the wave period of lattice point 055120 NW, which had the greatest wave height and the longest wave period from among the two lattice points, would be selected as the extreme waves for a recurrence period of 50 years and the hydrodynamics would be reproduced. In addition, it could be observed that the average and maximum wave heights were 0.6 m and 1.8 m, respectively, and the average and maximum peak periods were 3.6 s and 6.7 s, respectively, in the normal sea state.
4 fp ; SPM ðf Þ ¼ 0:3125$Hs2 $fp4 $f 5 $exp 1:25 f
(2)
where Hs denotes the significant wave height (m), fP is the peak 1 (Hz), and f is the frequency (Hz). The spectrum frequency ¼ Tp of velocity and acceleration can be converted by substituting Eq. (2) into Eqs. (3) and (4) [9]:
SPMu ðf Þ ¼ 2pf $SPMd ðf Þ;
(3)
SPMa ðf Þ ¼ 2pf $SPMu ðf Þ;
(4)
where SPMd represents the spectral density of wave elevation, SPMu the spectral density of wave velocity, and SPMa the spectral density of wave acceleration. Calculated spectrums are shown in Fig. 2. The velocity and acceleration spectrums generated using Eqs. (3) and (4) can be transformed into non-stationary time series data by using Eqs. (5) and (6) [10]:
uðtÞ ¼
N X
Aui sinð2pfi t þ fi Þ;
(5)
Aai sinð2pfi t þ fi Þ;
(6)
i¼1
aðtÞ ¼
N X i¼1
where Ai and fi represents the i-th magnitude and frequency at the spectral density. ɸi is generated randomly in the [p, p] range. Ai can be calculated using Eq. (7).
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi SPM ðfi Þ$Df :
2.2. Non-stationary random hydrodynamic load
Ai ¼
Morison’s equation, given as Eq. (1), is commonly used to depict hydrodynamic loads [5]:
By substituting Eqs. (5) and (6) into Eq. (1), F(t) can be calculated. Total wave load can be estimated by using Eq. (8).
F ¼
pD2 _ 1 U; cd rDjUjU þ cm r 2 4
(1)
where F represents the wave force per unit length of the foundation, r is the density of water, D is the diameter of the foundation, U denotes the velocity of the flow resolved normal to the member, U_ denotes the acceleration of the flow resolved normal to the member. cd and cm are the drag coefficient and the inertia coefficient. When Eq. (1) is examined, to calculate the hydrodynamic load, the velocity and acceleration of the waves that act on the structures must be calculated. To simulate the waves that act on the structures, the PiersoneMoskowitz spectrum delineated in Eq. (2) was used [5]:
(7)
Zz Ftotal ðtÞ ¼
NðhÞ$FðtÞdh;
(8)
0
where h denotes depth of water, z is the mean water depth of site and N(z) denotes the shape function as given in Eq. (9):
NðzÞ ¼
3 z 2 1 z 3 ; 2 H 2 H
(9)
where H is the height of the foundation. The non-stationary random hydrodynamic load regenerated using the above procedure is described in Fig. 3 (Fig. 4).
K.-Y. Oh et al. / Renewable Energy 54 (2013) 211e218
213
Fig. 1. Calculation lattices of the hindcasted wave data for candidate site [7].
3. Design 3.1. Wind turbine The wind class of the sea around the Korean Peninsula is class 3e5, which makes the region a feasible for construction of wind farm. However, when compared to the wind class of Horns Rev offshore wind farm and Nysted offshore wind farm in Europe, which are class 7 or above, it is relatively inferior in terms of economic efficiency [11]. In the present study, to improve the economic efficiency of the proposed wind farm, the wind turbine Table 1 Extreme wave conditions with a recurrence period of 50 years. Point
055120
Direction
Height [m]
Period [s]
Height [m]
056120 Period [s]
N NNE NE ENE E ESE SE SSE S SSW SW WSW W WNW NW NNW
5.52 3.85 3.45 4.03 4.89 5.43 6.09 6.13 6.53 7.06 6.48 4.90 6.01 6.85 7.22 5.98
9.4 7.7 7.1 7.5 8.2 8.7 9.5 10.1 10.9 11.9 11.8 10.1 11.0 11.7 11.9 10.4
5.50 3.46 2.98 3.13 3.65 4.48 4.90 4.84 5.30 5.84 5.70 4.61 5.56 6.91 7.00 5.16
9.4 7.2 6.6 6.6 7.2 8.2 9.1 9.5 10.2 11.2 11.4 9.9 10.7 11.9 11.8 9.6
blade has been designed as a low wind speed-type. Because wind turbines with low wind speed blades have identical rated power but are relatively longer in blade length, the load imposed on the tower and the foundation increases. It is possible to infer that the foundation designed in this paper can withstand greater load than wind turbine foundations of the same capacity that are commercially operated at present. Specification of wind turbine in this paper is given in Table 2. The rotor was designed as a fixed axle type where only the torque is transmitted to the rotor shaft and the horizontal load and axial load are supported by the nacelle, thus decreasing the load on the rotor shaft [12]. The total diameter of the rotor is 100 m, making it 10 m longer than that of Vestas V90, the most widely installed type among 3 MW wind turbines. Table 2 Specification of wind turbine. Rating Rotor orientation, configuration Control Gearbox type Gearbox ratio Rotor diameter Hub diameter Cut-in wind speed Rated wind speed Cut-out wind speed Rated rotor velocity Tower height Tower inner diameter Tower outer diameter Foundation height Axle type
3 MW Upwind, 3 blades Variable velocity, collective pitch 2 Planetary, 1 helical 104.8672 100.4 m 3m 3 m/s 13 m/s 25 m/s 15 rpm 80 m 3.94 m 4.30 m 30 m Fixed axle
214
K.-Y. Oh et al. / Renewable Energy 54 (2013) 211e218
Fig. 4. Finite element model of wind turbine components.
Fig. 2. Power spectrum of wave elevation, velocity and acceleration.(a) Normal sea state (Hs ¼ 0.6 m, Ts ¼ 3.6 s).(b) Extreme sea state (Hs ¼ 7.22 m, Ts ¼ 11.9 s).
The gearbox is assumed to be a typical multiple-stage gearbox but with no frictional losses. The generator is assumed to be a PMG (Permanent Magnetic Generator), as it shows high efficiency in low wind speed. The electrical efficiency of the generator was taken as 94.4%. This was chosen to be roughly the same as the total mechanical-to-electrical conversion loss of the DOWEC turbine at rated power [4]. With rated generator speed of 1575 rpm, rated electric power of 3 MW, and a generator efficiency of 94.4%, the rated mechanical power is 3.178 MW and the rated generator torque is 19,260 Nm. The tower height is set to 80 m, the typical hub height of 3 MW class wind turbines. The tower is composed of five flanges and four segments. Because the water depth of the candidate site is approximately 15 m, a monopole-type foundation, which is known to be the most affordable in such shallow depths, was selected. Further, in consideration of tide differences and the extreme wave height, the total height was designed to be 30 m. The mode is to be analyzed through a multi body dynamic analysis of the turbine and the foundation; therefore, based on this, the optimal design is to be derived, considerations of the ground and other aspects were omitted, and research was conducted on the premise that the ground was fixed. FEM model for analysis is shown in Fig. 4. SAMCEF Wind Turbines, which has a generic non-linear finite elements solver including multi body simulation features, was used for the integrated load analyses of wind turbine. It can make 3D turbulence wind profile interlinked with TurbSim [13]. The aerodynamic torque imposed on the blade can be calculated using the blade element method (BEM) [14]. Based on this load, a finite element analysis was conducted on the components of each part, and a multi body simulation was simultaneously carried out for full model. The super element method was used for the components of parts in the wind turbine [15]. 3.2. Foundation
Fig. 3. Extreme hydrodynamic load of candidate site with a recurrence period of 50 years.
For the foundation design, aerodynamic load and hydrodynamic load must be taken into consideration. As can be observed in Eq. (1), wave velocity and acceleration should be considered in frequency domain in that the hydrodynamics, load stemming from velocity and acceleration are both in operation. The foundation receives period fatigue load from the waves in the normal condition and,
K.-Y. Oh et al. / Renewable Energy 54 (2013) 211e218 Table 3 Exclusive conditions for design of the foundation.
Force Force Force Force Force Force 1P 2P 3P 4P
by by by by by by
Vel. (NSS, mean) Acc. (NSS, mean) Vel. (NSS, max) Acc. (NSS, max) Vel. (ESS) Acc. (ESS)
Min. freq.
Max. freq.
0.244 0.256 0.127 0.133 0.074 0.078 0.225 0.450 0.638 0.900
0.374 0.420 0.195 0.219 0.113 0.127 0.275 0.550 0.863 1.100
though infrequent, is subjected to great strength under extreme load. Accordingly, the average and maximum figures under normal conditions and extreme load for a recurrence pffiffiffi period of 50 years are all taken into consideration. Because 1= 2 or less of the amplitude of the significant frequency is the value where the strength is reduced by half, the exclusive condition for the waves is to be designed such that the bending mode of the foundation and the pffiffiffi tower exist at locations 1= 2 or less of the velocity and acceleration power spectrum values. In a turbine with a three-bladed rotor, the aerodynamic frequency of excitation occurs at three times the rotational frequency of the rotor (3P). The foundation’s first natural bending frequency coupled with the tower must not under any circumstances coincide with the critical exciting forces. Moreover, care must be taken to ensure that a certain distance from the remaining multiples of the rotor frequency is maintained. The distances between frequencies at which excessive vibrations will occur are determined by system damping, i.e., both structural as well as aerodynamic damping. Experience from existing turbines indicates that a safety distance of 0.25P from the dominant frequency of excitation (3P) and of 0.20P from the less critical frequency (1P, 3P, 4P) are good guide values [16]. Table 3 lists the exclusive conditions for the design of the foundation. The allowable region is 0.219e0.225 Hz, 0.420e 0.450 Hz, 0.550e0.638 Hz, and 0.863e0.9000 Hz, as shown in Fig. 5. To select the optimal foundation thickness in consideration of the exclusive condition, a modal analysis was conducted, and the
215
results are shown in Table 4. The main mode of the tower consists of the two modes of fore-after and sideeside. When the above exclusive condition is taken into consideration, it can be observed that, as the foundation thickness, values from 0.036 m to 0.039 m are possible. Because the construction costs decrease as the thickness decreases, it can be inferred that 0.036 m is the optimal thickness. When the design is such that the thickness is increased so that the natural frequency falls in a range of 0.420e0.450 Hz, the structures’ integrity increases further but the costs for producing the foundation increase geometrically. Currently, at commercially operated Horns Rev offshore wind farms, the rotor’s diameter is 80 m yet the thickness is 0.05 m [17]. Considering that foundationrelated costs generally take up approximately 20% of the total costs in constructing offshore wind farms, when foundations are designed in consideration of the aerodynamics and hydraulics, as suggested in this paper, the thickness should be decreased dramatically in comparison with existing design methods, and thus it can be anticipated that costs will be reduced dramatically.
4. Assessment of reliability In order for a manufacturer to obtain the certificate of its wind turbines by the certification authority, many design load cases must be analyzed and their reliability must be verified [18]. With offshore wind turbines, because hydrodynamic load acts on the foundation, the design evaluation must be conducted on more Design Load Cases (DLCs) than for onshore wind turbines and the reliability must accordingly be verified. For this, the IEC stipulates the categories necessary for offshore wind turbines and foundation design evaluation in 61400-3 and demands that a reliability evaluation be conducted. To validate the propose method for the design of offshore wind turbine, reliability of designed wind energy conversion system should be evaluated based on this standard. In this paper, the wind turbine and the foundation were evaluated for their reliability in terms of gusts and the extreme wind speed, to which wind turbines are considered to be especially vulnerable as well as DLC 1.1 among the many design load cases. Considering that offshore wind farms are inferior to onshore wind farms in terms of the surrounding environment, integrity was evaluated with wind
Fig. 5. Exclusive conditions for design of the foundation.
216
K.-Y. Oh et al. / Renewable Energy 54 (2013) 211e218
Table 4 Bending mode variation of foundation coupled with tower. Thickness (m)
Fore-after bending mode (Hz)
Sideeside bending mode (Hz)
0.027 0.030 0.033 0.036 0.039 0.042 0.045 0.048 0.051 0.054 0.057 0.060
0.1982 0.2083 0.2142 0.2196 0.2244 0.2288 0.2328 0.2364 0.2398 0.2429 0.2458 0.2485
0.1974 0.2079 0.2138 0.2192 0.2241 0.2285 0.2325 0.2361 0.2395 0.2426 0.2455 0.2482
turbine class IA, which is the strongest design basis, as the standard. Details of design load cases for simulations are depicted in Table 5. Fig. 6. DLC 1.1 (V ¼ 25 m/s, I15 ¼ 0.18).
4.1. DLC 1.1 In this design load case, the wind turbine is connected to electric load and producing electric power. As the load condition to confirm that the wind turbine operates without any problems when 3D isotropic turbulence occurs, this can be used as the basis to calculate a wind turbine performance curve. Fig. 6 shows the pitch angle and the generation power in terms of the wind speed when the average wind speed is 25 m/s and the turbulence intensity is 0.18. It can be observed that, even though the wind constantly changes and blows, control is satisfactory and electric power is generated. To evaluate the performance of wind turbines for low wind speed, the power curve of the wind turbine, calculated by sequentially increasing the average velocity from 3 m/s to 25 m/s, is shown in Fig. 7. Because the suggested wind turbine is 10 m longer than the Vestas V90 in the rotor diameter, it could generate more electric power at low wind speed. Also its performance was outstanding at roughly the rated output, because the generator of wind turbine was designed as a PMG-type. In consideration of the fact that the average wind speed of the West Sea is 7e8 m/s [19], the annual energy production is calculated with an average wind speed of 7 m/ s; the figure is 6123 MWh for the KEPCOeRI wind turbine and 5784 MWh for Vestas V90, thus showing that the annual energy production is increased by 5.86%. In case that the average wind speed is 8 m/s, the annual energy production is 8162 MWh for the KEPCOeRI wind turbine and 7752 MWh for Vestas V90, respectively, thus constituting an increase of 5.29%. In calculating the annual energy production above, loss was disregarded. Considering that the wind turbine operation period is 20 years, it is evident that the introduction of wind turbines for low wind speed will improve the economic efficiency of wind farms. In particular, considering that development costs for offshore wind farms are higher than those for onshore wind farms, it is imperative that wind turbines for low wind speed be introduced in constructing the former.
during power generation. Although an integrity evaluation must be conducted on a recurrence period of 1 year or 50 years, in the present study, the integrity evaluation was conducted on the condition of a recurrence period of 50 years, which is a relatively extreme condition. Fig. 8 shows the results of the evaluation when the average wind speed is 25 m/s. The statues of wind turbine were observed in case that gust with a recurrence period of 50 years occurs from 150 s to 164 s and wind turbine is disconnected from grid during 150 s to 180 s as shown in Fig. 8(a). As the blade pitch angle is controlled during gust, the aerodynamic torque is decreased. In addition, it can be observed that the blade pitch angle is adjusted also during system loss, because no electric power is transmitted to grid. Fig. 8(b) shows the distribution of stress. The maximum stress amounts to 72 MPa on the ground and at the contact point, which are subjected to the greatest bending moment. The wind turbine stably operate even under load conditions identical to this condition. 4.3. DLC 3.2 This case is the category for evaluating the integrity of the load that acts on the wind turbine during the transient region from a suspended or idling state to power generation. The integrity evaluation was conducted for a recurrence period of 50 years,
4.2. DLC 2.3 This load condition is the category for evaluating the reliability of the wind turbine when gust and grid loss occur simultaneously Table 5 Design load cases [5]. Load case
Wind condition
Wave condition
1.1 2.3 3.2 6.1c
NTM, Vin < Vhub < Vout EOG, Vhub ¼ Vr 2 m/s and Vout EOG, Vhub ¼ Vin, Vr 2 m/s and Vout Vhub ¼ Ve50
NSS NSS NSS EWH(H ¼ H50)
Fig. 7. Power curves (kW). (a) Pitch angle and generated power via gust and grid loss (b) Stress distribution.
K.-Y. Oh et al. / Renewable Energy 54 (2013) 211e218
217
Fig. 9. DLC 3.2 (EOG50, V ¼ 25 m/s).
Fig. 8. DLC 2.3 (EOG50, V ¼ 25 m/s). (a) Pitch angle and generated power via gust (b) Stress distribution.
which is a relatively extreme condition. Gusts began at 70 s, which was before the turbine entered the normal condition, and controller is operated to regulate the power and reduce the load. As shown in Fig. 9(a), even when gusts are occurred, the wind turbine operates without any problems and withstands the transient load in the transient region. Fig. 9(b) shows the distribution of stress in this case. The maximum stress amounts to 57 MPa on the ground and at the contact point, which are subjected to the greatest bending moment. Maximum stress of this design load case is smaller than one of DLC 2.3 because there is no grid loss, which could be considered as additional load from the wind turbine point of view. Consequently, the wind turbine reliably operate even under load conditions identical to this condition.
load act on the wind turbine and the foundation is shown in Fig.10. It can be observed that the maximum stress amounts to 257 MPa on the ground and at the contact point, which are subjected to the greatest bending moment. Considering that the ultimate strength of iron is 400 MPa, the safety factor is 1.56 or above, which satisfies the
4.4. DLC 6.1c This case is the category for evaluating the reliability of the wind turbine and the foundation when they are parked in a suspended or idling state and subjected to extreme wind speed and extreme waves. An integrity evaluation is accomplished for a recurrence period of 50 years because it is more severe case than extreme load case with a recurrence period of 1 year. The maximum load distribution in cases where extreme aerodynamic load and hydrodynamic
Fig. 10. DLC 6.1c (EWM50, Vhub ¼ 50 m/s).
218
K.-Y. Oh et al. / Renewable Energy 54 (2013) 211e218
stipulation in the regulations (i.e. 1.35 or above). Consequently, the wind turbine and the foundation proposed in the present study are equivalent to wind turbine class I in terms of the ultimate load. Through an evaluation of the above conditions, which are seen as especially weak from among the design conditions of wind turbines, it could be observed that, even when the thickness of the foundation was set at 0.036 m, it was strong enough to withstand the external load imposed on the wind turbine. If the wave conditions of candidate sites for offshore wind farms are analyzed in the frequency domain and, based on this, the foundation is designed, it will then be possible to improve the economic efficiency of offshore wind farms, as has been mentioned above.
5. Conclusions This paper presents optimal foundation design methods that take into consideration aerodynamic load and hydrodynamic load on the basis of the environmental conditions of a candidate site for an offshore wind farm. Each load was analyzed in the frequency domain to select possible locations for designing the main modes of the tower and the foundation. Based on results of analysis, the optimal dimensions for a monopile foundation were presented. An integrity evaluation was conducted on the designed turbine and foundation in accordance with IEC standards to verify the reliability of the optimized turbine and foundation. In addition, methods to improve economic efficiency when constructing offshore wind farms such as the introduction of wind turbine for low speed and construction cost reduction through the use of optimally designed foundations were presented.
Acknowledgment This work was supported by the New & Renewable Energy of the Korea Institute of Energy Technology Evaluation and Planning (KETEP), grant funded by the Korea government Ministry of Knowledge Economy. (20113040020010)
References [1] Krohn Soren, Morthorst Poul-Erik, Awerbuch Shimon. The economics of wind energy. The European Wind Energy Association; 2009. [2] Nehal RS, Nedam Ballast. Foundation design monopile 3.6 & 6.0 mW wind turbines. DOWEC, Dutch; 2001. [3] Byrne BW, Houlsby GT. Foundations for offshore wind turbines. Philosophical Transactions of the Royal Society, A: Mathematical, Physical and Engineering Sciences 2003;361:2909e30. [4] Kooijman HJT, Lindenburg C, Winkelaar D, van der Hooft EL. DOWEC 6 MW pre-design: aero-elastic modeling of the DOWEC 6 MW pre-design in PHATAS. ECN-CXe01e135. Energy Research Centre of the Netherlands; 2003. [5] IEC 61400-3 wind turbinesdPart 3: design requirements for offshore wind turbines. IEC; 2009. [6] Kim JY, Kang KS, Oh KY, Lee JS and Ryu MS. A study on the site selection of offshore wind farm around Korean Peninsula. In. 3rd International conference on ocean energy, Bilbao, Spain; 2010. [7] Lee GS and Gwon OS. Analysis of ocean characteristics and design support for construction of offshore wind farm. 1st year interim report, KORDI; 2009. [8] Oh Ki-Yong, Kim Ji-Young, Lee Jun-Shin, Ryu Ki-Wahn. Wind resource assessment around Korean Peninsula for feasibility study on 100 MW class offshore wind farm. Renewable Energy 2012;42:217e26. [9] Choi YH, Kwon SH, Kim DW, Park SK. A reliability study of coastal structures under the influence of wave and currentsdrandom analysis of fixed structures. Journal of Ocean Engineering and Technology 1995;9(2):186e92. [10] Ki-Yong Oh, Ji-Young Kim, JaeKyung Lee, Joon-Young Park, Moo-Sung Ryu, Keum-Seok Kang and Jun-Shin Lee. Design of offshore wind turbine with foundation based on modal analysis. In. The 17th International congress on sound & vibration, Cairo, Egypt; 2010. [11] Archer Cristina L, Jacobson Mark Z. Evaluation of global wind power. Journal of Geophysical Research 2005;110. [12] Heege A, Viladomiu P, Betran J, Radovcic Y, Latorre M, Cantons JM, et al. Impact of wind turbine drive train concepts on dynamic gearbox loads. Milan, Italy. In: Proceedings of European wind energy conference 2007; 2007. [13] Jonkman BJ. TurbSim user’s guide: version 1.50. National Renewable Energy Laboratory; 2009. [14] Munteanu Lulian, Bratcu Antoneta luliana, Cutululis Nicolaos-Antonio, Ceanga Emil. Optimal control of wind energy systems. Springer; 2008. 30e57. [15] Kohno Kazukiyo, Tsunada Tomohiro, Seto Hideyuki, Tanaka Michio. Hybrid stress analysis of boundary and finite elements by a super-element method. Finite Elements in Analysis and Design 1991;7(4):279e90. [16] Hau Erich. Wind turbines: fundamentals, technologies, application, economics. 2nd ed. Springer; 2006. p. 408e11. [17] http://www.hornsrev.dk/Engelsk/default_ie.htm. [18] IEC 61400-1 wind turbinesdpart 1: design requirements. 3rd ed. IEC; 2005. [19] Jang Jea-Kyung, Yu Byoung-Min, Ryu Ki-Wahn, Offshore Jun-Shin Lee. Wind resource assessment around Korean Peninsula by using QuikSCAT satellite data. Journal of the Korean Society for Aeronautical & Space Sciences 2009; 37(11):1121e30.