Proceedings of the Institution of Mechanical Engineers, Part M: Journal of Engineering for the Maritime Environment http://pim.sagepub.

com/

Wave-current interactions in marine current turbines

N Barltrop, K S Varyani, A Grant, D Clelland and Xuan Pham Proceedings of the Institution of Mechanical Engineers, Part M: Journal of Engineering for the Maritime Environment 2006 220: 195 DOI: 10.1243/14750902JEME45 The online version of this article can be found at: http://pim.sagepub.com/content/220/4/195

Published by:
http://www.sagepublications.com

On behalf of:

Institution of Mechanical Engineers

Additional services and information for Proceedings of the Institution of Mechanical Engineers, Part M: Journal of Engineering for the Maritime Environment can be found at: Email Alerts: http://pim.sagepub.com/cgi/alerts Subscriptions: http://pim.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav Citations: http://pim.sagepub.com/content/220/4/195.refs.html

Downloaded from pim.sagepub.com at NANYANG TECHNOLOGICAL UNIV on September 4, 2011

SPECIAL ISSUE PAPER

195

Wave-current interactions in marine current turbines

N Barltrop1, K S Varyani1, A Grant2, D Clelland1, and Xuan Pham1* 1Department of Naval Architecture and Marine Engineering, Universities of Glasgow and Strathclyde, Glasgow, UK 2Department of Mechanical Engineering, University of Strathclyde, UK The manuscript was received on 7 October 2005 and was accepted after revision for publication on 9 June 2006. DOI: 10.1243/14750902JEME45

Abstract: The inuence of waves on the dynamic properties of bending moments at the root of blades of tidal stream vertical-axis rotors is reported. Blade element-momentum theory for wind turbines is combined with linear wave theory and used to analyze this inuence. Experiments were carried out with a 350 mm diameter rotor to validate the simulation and the comparison shows the ability of the theoretical approach to predict the blade root bending moments. It can be concluded that, in steep waves, linear theory underestimates the dynamic behaviour of bending moments. However, in long waves, linear theory works well. Bending moments at roots of rotor blades uctuate with signicant amplitudes (as much as 50 per cent of mean value for out-of-plane bending moment and 100 per cent of mean value for in-plane bending moment), which will be important for design of tidal stream rotors. Keywords: tidal stream, current turbine, wave-current interaction, linear blade element-momentum theory

1 INTRODUCTION Horizontal-axis wind turbines have achieved a dominant position in the market and for tidal streams the arguments in their favour are even more compelling [1, 2]. Vertical-axis rotors are subject to cyclic loads even in uniform ow [3, 4], and in tidal streams these will make the design more dicult in all but small machines. Reversing gravity loads, which are very important for horizontal-axis wind turbines, should be much less important for tidal stream turbines because buoyancy can be used to balance the weight. For horizontal-axis machines, relatively low levels of turbulence minimize stochastic structural loads, but deterministic cyclic loading (from velocity shear and yaw error eects) may be comparatively large. The eect of waves is likely to be important. It is desirable to locate rotors near the free surface to make maximum use of the available crosssectional area and to intercept the highest stream velocities. The eect of waves will therefore be an
* Corresponding author: Naval Architecture and Marine

Engineering, University of Strathclyde, 100 Montrose Street, Glasgow G4 0LZ, UK. email: xuan.pham@strath.ac.uk

important consideration in determining limits for device location and rotor operational envelopes. A substantial body of knowledge on wave loading already exists, but not in this context. There is an urgent need for further research to determine the response and practical limitations posed by the associated phenomena. The main objectives of this study were, rstly, to assess the limits imposed by waves on the performance of tidal stream rotors, to investigate ways in which these limits might be relaxed by renements of rotor design and the control of operating conditions, and to conrm that the concept is structurally and mechanically possible. In this paper, the authors specically look into the dynamic property of bending moments acting about the roots of tidal stream rotors. Experiments are carried out and data are used to validate simulation results based on a mathematical model, i.e. linear wave and blade element-momentum theory. It is noticed that, in long waves, the theory predicts very well the dynamic response of these bending moments. However, in steep waves, probably due to the presence of high non-linearities, linear wave theory underestimates these responses.

JEME45 IMechE 2006

Proc. IMechE Vol. 220 Part M: J. Engineering for the Maritime Environment

Downloaded from pim.sagepub.com at NANYANG TECHNOLOGICAL UNIV on September 4, 2011

196

N Barltrop, K S Varyani, A Grant, D Clelland, and Xuan Pham

2 THEORY The torque, thrust and bending moments induced by stream ow are calculated based on blade element-momentum theory [5]. Each blade of the wind turbine is divided into a number of sections (Fig. 1). Even though chord and pitch or twist angle vary along the blade, within each section, these values are presumed constant to simplify the calculations. Lift and drag coecients as functions of incident angles are then calculated and adjusted appropriately to account for 3D eects. Velocities of water ow passing the blade section are next calculated taking into consideration the eects from current and waves. As well as adding to the wave particle velocities, current changes the encountered wave frequency as follows.
1 1 U f = = + c cos(h ) e T w T L e w w
(1)

will increase incident ow velocity if they are in the same direction, or decrease this velocity if they are not in the same direction. Vertical wave particle velocity: pH sinh[k ((zrds)+d )] w sinh(w ) (3) w T cosh(k d ) w w will change the ow incidence angle and hence modify the lift and drag forces. The wave eects on the rotor are periodical if the waves are regular waves. The wave eects on the rotor are also reduced when the rotor is positioned at greater water depth because the wave kinematics decrease with depth. The wave eects on the rotor can be ignored if wave ow is small compared with tidal currents. The waves are important if the waves ow speeds are a signicant proportion of, or are greater than, those of the tidal current. This generally happens in extreme weather. To illustrate this comment, Table 1 shows the dimension and a typical operating condition of a full-scale marine current turbine, and Table 2 shows the horizontal and vertical particle velocities at the 0.7R position of the blade as it passes three locations, i.e. upright, down right, and horizontal, respectively. As seen, the gures in this illustrative example reects well the remark noted. v=

where U is the current velocity, T and L are wave c period and length, respectively. Subscript e and w represent encountered and wave, respectively. h is the incident angle of current. The possible problems waves can cause to tidal stream water turbine are mainly embodied in change of ow velocity eld and pressure eld around the rotor. The former change will aect the uid forces on the rotor. The latter and former may trigger or depress the occurrence of cavitation. The eect of waves on the rotor depends on magnitude and direction of ow velocity that is induced by waves. Horizontal wave particle velocity:
u= pH cosh[k ((zrds)+d )] w cosh(w ) w T sinh(k d ) w w
(2)

Table 1 Dimensions and operating conditions of rotor at full scale

Diameter (m) Number of blades Rotation rate (rpm) Immersion of shaft (m) Water depth (m) Current speed (kts) Maximum wave height (m) 20.0 3 12.0 20.0 40.0 5.0 (or 2.5 m/s) 9.0

Fig. 1 Co-ordinate system and velocity diagram in blade theory

Proc. IMechE Vol. 220 Part M: J. Engineering for the Maritime Environment JEME45 IMechE 2006

Downloaded from pim.sagepub.com at NANYANG TECHNOLOGICAL UNIV on September 4, 2011

Wave-current interactions in marine current turbines

197

Table 2 Horizontal and vertical particle velocities at 0.7R position

Wave period set at 12 seconds
Upright position
u (m/s)
H =8 m wave % current velocity H =2 m wave % current velocity 1.99 79.65% 0.50 19.91%

Down right
u (m/s)
1.64 65.44% 0.41 16.36%

Horizontal
u (m/s)
1.78 71.18% 0.44 17.79%

v (m/s)
1.02 40.99% 0.26 10.25%

v (m/s)
0.46 18.38% 0.11 4.59%

v (m/s)
0.73 29.13% 0.18 7.28%

Apart from velocity components due to waves and current, there are also velocity components due to the rotation of the blade. All together, velocity of water ow around each blade segment is calculated. Directional components of this velocity consist of out-of-rotor-plane velocity, which is in the positive x-axis direction, and in-rotor-plane velocity, which is in the direction tangent to the positive y-z rotation (Fig. 1). From these ndings, the out-of-plane (short for out-of-rotor-plane) liftdrag force and in-plane tangential liftdrag force on the blade section are calculated. Note that tip loss from the blade is also introduced into momentum equations. The tip loss coecient F is a function of distance r from root and angle of attack w and is expressed as F(w, r)=
1 (Rr) 2 acos exp N p 2 blades (R|sin w|)

3 ROTOR DESIGN AND MANUFACTURE For the project, three rotor designs were produced, with progressive renement based on results of testing. The largest practical rotor diameter was used to minimize scale eects. The rst model rotor [6] was essentially a wind turbine conguration, with a slight increase in blade chord and thickness for structural strength. The aerofoil prole (NREL S814) was chosen for its good performance at low Reynolds numbers, and its tolerance of surface imperfections. The rotor has three blades, and an overall diameter of 350 mm. At the root, the blades taper down to a circular section to increase the local strain and to facilitate the tting of strain gauges. The second rotor [7] also uses three blades. Its diameter is increased to 400 mm to increase torque and thrust, and blade chord is increased in an attempt to combat the cavitation observed on the rst rotor. The same aerofoil prole is used, but the conguration at the blade root is changed: the blades merge into the hub without taper, and strain gauges are not tted. However, blade angles can be adjusted over a range of about 15 degrees. In the third rotor [8], an attempt is made to produce a lightly loaded, high-solidity turbine with good resistance to cavitation and the eects of marine growth. The blade number, diameter, and hub design are the same as for the second rotor. However, a higher-lift aerofoil is used (NREL WA3-02) and blade chord is further increased. In this paper, analysis focuses on the rst model rotor. The S814 is one of the series developed by the National Renewable Energy Laboratory (NREL), USA, for wind turbines. One particularly important characteristic of the S814 is the minimal sensitivity of its maximum lift coecient to roughness eects, a critical property for stall-regulated wind turbines. The aerofoil has a very low drag coecient and is also not sensitive to change of angle of attack at around stalling angle. The prole shape is illustrated

G C

DH

(4)

The torque acting on the rotor is then the sum of torque acting on individual blade, which in turn is the sum of products of in-plane liftdrag forces and the arm lengths from centres of blade sections to the axis of the rotor. Thrust on the rotor, similarly, is the sum of out-of-plane liftdrag forces on blade sections of all blades on the rotor. Bending moments acting around the root of each blade are calculated in the similar fashion. The outof-plane moment is the sum of products of out-ofplane liftdrag forces and the arm lengths from centres of blade sections to the root of the blade. The in-plane moment is the sum of products of in-plane liftdrag forces and the arm lengths from centres of blade sections to the root of the blade. The calculations are carried out for every time step. The simulation modelling uses assumption of quasisteady ow and was coded in MathCAD by the authors. The time step was selected based on the rotational speed of the rotor and the frequency of the incident waves such that the amount of data obtained is sucient to reect the eects of both the rotational speed of the rotor and the encountered frequency with the waves.
JEME45 IMechE 2006

Proc. IMechE Vol. 220 Part M: J. Engineering for the Maritime Environment

Downloaded from pim.sagepub.com at NANYANG TECHNOLOGICAL UNIV on September 4, 2011

198

N Barltrop, K S Varyani, A Grant, D Clelland, and Xuan Pham

in Fig. 2a; specic technical features of the NREL S814 are detailed in [9]. The length of the blade used for rotor 1 was 150 mm with maximum chord of 44 mm. Nominal chord length at the tip was 20 mm, but the tip was rounded to reduce shed vortices. Chord and pitch distributions are shown in Fig. 2b; an increase of about 5 degrees from the nominal pitch angle was found to give best performance. Figure 2c shows the two-dimensional characteristics of lift and drag coecients based on tunnel

tests at Reynolds number of 3106. In the simulation, the three-dimensional eects on the lift coecient are considered and three-dimensional lift coecient is calculated based on two-dimensional lift coecient using the following equation. C L3D =C +DC 3 L2D L

A B

c(r) 2 r

(5)

where DC is the dierence between C obtained L L from linear potential ow theory and the actual twodimensional lift coecient C c(r) is the chord L2D length as the function of radial distance r from root. The drag coecient, on the other hand, is assumed to be equal to the two-dimensional drag coecient. Curve-t functions were developed based on these lift and drag coecients as function of angle of attack. Intermediate coecients can therefore be estimated using these generic fuctions.

4 EXPERIMENT Rotor performance (shaft torque, axial thrust, and blade bending moments) was investigated in the Universities of Glasgow and Strathclydes 774.6 2.4 m-deep wave/towing tank. A two-dimensional aerofoil section boat made of glass bre (Fig. 3a) was used to house the motor, gearing, and torque and thrust transducers. The rotor was supported in front of the boat which itself was cantilevered down from the towing tank carriage as in Fig. 3b. Strain gauges were used for the measurement of bending moments about the root of the blade in the case of rotor 1. The strain gauges were made watertight using M-Line J-Coat. The shaft to which the rotor was attached was hollow. The cables of the strain gauges were accommodated through the inside of the shaft and the signals were transferred through slip rings. Provision was made for amplifying the signals before passing them through the slip rings, but slip ring tests and the nal results demonstrated remarkably good performance without the need for ampliers on the rotating shaft. The shaft was supported by linear bearings to allow rotation and axial movement. To allow precise speed control, a motor/generator with gearbox and tachometer was used with closed loop control. The motor/generator could provide power (at low current speeds) or absorb power (at the higher current speeds). The shaft extended in front of the boat bow suciently for minimal interference between the boat and the rotor. A grease box prevented the entry of water around the shaft. The boat was clamped and bolted by steel beams to a rail-mounted moving carriage
JEME45 IMechE 2006

Fig. 2 (a) Sectional prole of NREL S814 blade. (b) Chord and pitch distribution along blade. (c) Lift and drag coecients based on tunnel tests at Rn=3106

Proc. IMechE Vol. 220 Part M: J. Engineering for the Maritime Environment

Downloaded from pim.sagepub.com at NANYANG TECHNOLOGICAL UNIV on September 4, 2011

Wave-current interactions in marine current turbines

199

Fig. 3 (a) Arrangement of rotor and devices. (b) Prole and cross-section views of the experiments

over the tank. The carriage moved the boat along the tank at steady speed to simulate tidal currents. A ap type wave generator produced waves of a known wavelength and period. These are of course encountered at a higher frequency than the generated wave frequency when the rotor is moving towards the waves. (Note that, within the limitations of potential ow theory, the model was a precise Froudescaled representation of the real case of a xed rotor encountering waves of the generated length on the current.) Experiments were carried out to test the performance of the tidal stream rotors for dierent conditions. The most important parameters were the dynamic properties of the rotor under dierent wave heights, wave frequencies, and current speeds. Only the rst rotor was strain gauged because the requirements for high stresses in the root under the towing test conditions were not compatible with the need to load the rotor more highly to obtain a high Reynolds number in the cavitation tests. The experiments covered a range of towing speeds
JEME45 IMechE 2006

(0.01.6 m/s), wave heights and frequencies, and depths of immersion whilst the rotational speed is set at about 200 rpm. The aims were to investigate the magnitude and variation of thrust, torque, and bending moment at the root of the blade in the case of rotor 1 and thrust, torque, and rpm in the case of rotor 2. The results were then compared with calculations. This paper presents the dynamic property of bending moments acting about the roots of blades of rotor 1 in waves of 150 mm height (model scale), two frequencies, i.e. 0.5 Hz and 1.0 Hz, and towing speed 0.01.6 m/s. For a 10-m diameter, full scale turbine, these frequencies would correspond to 10.69 and 5.35 second periods, and to towing speed 08.55 m/s. The upper currents are clearly much higher than would be found in practice, but are useful for testing the theory. Reection from the tank side walls are assumed to be insignicant as the waves travel straight down the tank and there is a clearance of 2.1 m on either side of the turbine from the tank wall. For tests at low velocities, the starting position of the testing is set at approximately one

Proc. IMechE Vol. 220 Part M: J. Engineering for the Maritime Environment

Downloaded from pim.sagepub.com at NANYANG TECHNOLOGICAL UNIV on September 4, 2011

200

N Barltrop, K S Varyani, A Grant, D Clelland, and Xuan Pham

third of the tank length from the beach so that reected waves from the beach would not reach the turbine by the end of the test run. It should also be noted that, in these tests, the Reynolds number varies in the range of 4.32103 to 6.0104, which is lower than the range of 0.75106 to 1.5106 for S814 data. A low Reynolds number can degrade the dynamical properties of the airfoil and can be a source of discrepancy between the test data and the simulation data.

(at model scale), respectively. The maximum wave slope is approximately 8 degrees. The corresponding tipspeed ratios are 12.22 and 3.77, respectively. Sim_Mx and Sim_My represent the simulation results of out-of-plane and in-plane bending moments, respectively, whilst Exp_Mx and Exp_My refer to their corresponding experimental results. As can be seen, in long waves, blade element with linear wave theory can predict the dynamic response of both out-of-plane and in-plane bending moments. 5.2 Steep waves of 150 mm height and 1.0 Hz frequency Figures 5a and 5b, and 5c and 5d, show the dynamic properties of bending moments about roots of rotor blades in steep waves of 1.0 Hz frequency and 150 mm height at current speeds of 0.3 m/s and 1.0 m/s, respectively. The corresponding tipspeed ratios are therefore 12.22 and 3.77, respectively. The maximum wave slope is approximately 29.5 degrees. It can be

5 VALIDATION 5.1 Long waves of 150 mm height and 0.5 Hz frequency Figures 4a and 4b, and 4c and 4d, show dynamic properties of bending moments about roots of rotor blades in waves of 0.5 Hz frequency and 150 mm height at current speeds of 0.3 m/s and 1.0 m/s

Fig. 4 (a) Out-of-plane bending moment in waves of 150 mm height and 0.5 Hz frequency and current speed of 0.3 m/s at model scale. (b) In-plane bending moment in waves of 150 mm height and 0.5 Hz frequency and current speed of 0.3 m/s at model scale. (c) Out-of-plane bending moment in waves of 150 mm height and 0.5 Hz frequency and current speed of 1.0 m/s at model scale. (d) In-plane bending moment in waves of 150 mm height and 0.5 Hz frequency and current speed of 1.0 m/s at model scale
Proc. IMechE Vol. 220 Part M: J. Engineering for the Maritime Environment JEME45 IMechE 2006

Downloaded from pim.sagepub.com at NANYANG TECHNOLOGICAL UNIV on September 4, 2011

Wave-current interactions in marine current turbines

201

Fig. 5 (a) Out-of-plane bending moment in waves of 150 mm height and 1.0 Hz frequency and current speed of 0.3 m/s at model scale (b) In-plane bending moment in waves of 150 mm height and 1.0 Hz frequency and current speed of 0.3 m/s at model scale (c) Out-of-plane bending moment in waves of 150 mm height and 1.0 Hz frequency and current speed of 1.0 m/s at model scale (d) In-plane bending moment in waves of 150 mm height and 1.0 Hz frequency and current speed of 1.0 m/s at model scale

seen that blade element with linear wave theory can predict relatively well the uctuation of in-plane bending moment (My) both in low and high current speed ranges. However, for out-of-plane bending moment, the uctuation predicted is less than that measured. This highlights the signicance of high non-linearities associated with steep waves. Figures 4a and 4c show that the amplitude of bending moment variation recorded during the test is of the order of two times that predicted. 5.3 General discussion A general observation on the dynamic characteristics of bending moments at the root of a rotor blade is that it sees that the main oscillation at the encountered wave frequency. The rotor frequency shows in the smaller superimposed oscillation in the in-plane bending moment and is mainly the selfweight eect. The numerical model is clearly giving reasonable results, especially in the middle of the short segment of time history where the numerical
JEME45 IMechE 2006

wave has been synchronized with the measured wave (Fig. 5b). (It is interesting that in these relatively steep waves the dierence between the linear wavelength and the non-linear wavelength results in a loss of synchronization between rotor position and wave phase in just a few wave cycles. This is signicant when comparing predictions with measurements, but is not important for design.)

6 CASE STUDY USING LINEAR BLADE SIMULATION MODEL 6.1 Eects of ow incidence angle Of theoretical interest is the eect of the vertical component of wave particle velocity, which mainly aects the ow incidence angle onto the aerofoil (Fig. 6). The numerical calculations show a small eect at the main peaks of the wave loads (because these correspond to peaks in the horizontal velocity where the vertical wave velocity is zero). Also

Proc. IMechE Vol. 220 Part M: J. Engineering for the Maritime Environment

Downloaded from pim.sagepub.com at NANYANG TECHNOLOGICAL UNIV on September 4, 2011

202

N Barltrop, K S Varyani, A Grant, D Clelland, and Xuan Pham

Fig. 6 Out-of-plane bending moment with and without vertical wave velocity

numerically it has been possible to consider cases where wave and current are in dierent directions. Figure 7 shows rstly a case with the current and wave normal to the rotor plane and secondly a case with the wave direction at 45 degrees to the rotor plane. The highest total and uctuating wave loads occur with the wave propagating normal to the rotor plane. 6.2 Eects of waves on bending moments Figure 8a and 8b show the inuence of waves on outof-plane and in-plane bending moments, respectively. It appears that within the practical wave frequency range, steeper waves impose less bending moments in both directions about the roots of the rotor blade. In steady water condition, i.e. without waves, the out-of-plane bending moment is steady but the in-plane bending moment varies harmonically with the frequency of rotor due to the gravity bending moment component. The amplitude of the out-ofplane bending moment uctuation is more than halved as wave frequency increases from 0.5 Hz to 1.0 Hz. In steep waves, Fig. 8b shows that the uctuation of bending moment (My) is mainly due to the gravity component as the in-plane bending moment in the steep waves is more or less equivalent

Fig. 8 (a) Out-of-plane bending moment in dierent conditions (b) In-plane bending moment in dierent conditions

to that in steady water condition. The uctuation of My is signicant in long waves, as seen in Fig. 8b. 6.3 Full-scale case study Figure 9 presents simulation results of bending moments acting on a hypothetical full-scale rotor of 20 m diameter rotating at 7 rpm in waves of 3.0 m height and 10.0 s period on top of current speed of

Fig. 7 Out-of-plane bending moment corresponding to wave normal to rotor plane and wave at 45 degrees to rotor plane

Fig. 9 Bending moments about root of blades of a hypothetical full-scale rotor

JEME45 IMechE 2006

Proc. IMechE Vol. 220 Part M: J. Engineering for the Maritime Environment

Downloaded from pim.sagepub.com at NANYANG TECHNOLOGICAL UNIV on September 4, 2011

Wave-current interactions in marine current turbines

203

2.0 m/s. The tipspeed ratio (ratio between tangential speed of tip of the blade by the current speed) is then equivalent to 3.6, which corresponds to the current speed of 1.0 m/s at model scale. The water depth is set at 40.0 m and the clearance between still water surface and highest point of blade tip is 10.0 m. The hub radius of the rotor is 1.4 m and root radius is 0.3 m. As a result, the mean thrust acting on the rotor is 0.124 MN and mean torque is 0.142 MNm. The corresponding mean power generated is 0.104 MW. As can be seen in Fig. 9, the variation of bending moments acting about the root of the rotor blades is very signicant. The range of uctuation of out-ofplane bending moment is approximately 0.78 MNm about the mean value of 0.38 MNm. The in-plane bending moment is smaller, but its dynamic property is equally signicant, with the range of uctuation approximately 0.24 MNm about the mean of about 0.1 MNm.

REFERENCES
1 Pham, X., Liu, Y., Varyani, K., and Barltrop, N. Report on tidal stream rotor tests. Universities of Glasgow and Strathclyde, 2004. 2 Myers, L. and Bahaj, A. S. Basic operational parameters of a horizontal axis marine current turbine. In Proceedings of Eighth World Renewable Energy Congress, Denver, USA, 2004. 3 Sutherland, H. J. Fatigue case study and loading spectra for wind turbines, IEA Fatigue Experts Meeting, pp. 7787, April 1994. 4 Smith, K., Randall, G., and Malcolm, D. Evaluation of wind shear patterns at Midwest Wind Energy Facilities, American Wind Energy Association (AWEA) WINDPOWER 2002, Portland, Oregon, June 2002. 5 Eggleston, D. M. and Stoddard, F. S. Wind turbine engineering design, 1978 (Van Nostrand Reinhold, New York). 6 Wang, D., Atlar, M., and Paterson, I. Performance tests of the rst tidal stream rotor. Report No. MT2003-016, School of Marine Science and Technology, University of Newcastle upon Tyne, October 2003. 7 Wang, D., Atlar, M., and Paterson, I. Performance tests of the second tidal stream rotor. Report No. MT2004-027, School of Marine Science and Technology, University of Newcastle upon Tyne, December 2004. 8 Wang, D., Atlar, M., and Paterson, I. Performance tests of the third tidal stream rotor. Report No. MT2005-002, School of Marine Science and Technology, University of Newcastle upon Tyne, January 2005. 9 Tangler, J. L. and Somers, D. N. NREL airfoil families for HAWTs. http://wind.nrel.gov/, 1995.

7 CONCLUSION The paper has briey introduced blade elementmomentum theory, which includes wave eects in analyzing dynamic properties of bending moments acting at roots of rotor blades of a tidal stream rotor. Experiments were also described in the validation of simulation results. Based on what has been obtained, the following conclusions are drawn.
$ In-plane and out-of-plane bending moments both

uctuate signicantly and this is worse in steep waves. In long wave condition, the uctuation of these bending moments can be predicted well by linear wave theory with blade element-momentum theory. However, in steep waves, due possibly to high non-linearities involved, the out-of-plane bending moment uctuates signicantly more than that predicted (by about two times in this case). $ Waves approaching along the normal to the rotor plane lead to the largest bending moments on the blades.

APPENDIX Notation d ds H k w T w zr water depth rotor axis depth wave height wave number wave period instantaneous vertical position of centroid of blade section with respect to rotor axis angle of attack

w w

JEME45 IMechE 2006

Proc. IMechE Vol. 220 Part M: J. Engineering for the Maritime Environment

Downloaded from pim.sagepub.com at NANYANG TECHNOLOGICAL UNIV on September 4, 2011