Acoustic measurement for 12%scaled model of NREL Pha VI wind turbine by using beamforming
Taehwan Cho *,Cheolwan Kim,Duckjoo Lee
Korea Aerospace Rearch Institute,Daejeon 305-333,Republic of Korea
a r t i c l e i n f o Article history:available
Received 1December 2008Accepted 9June 2009
rigAvailable online 12November 2009Keywords:Wind turbine Blade
Beamforming Acoustics Wind tunnel
a b s t r a c t
Wind tunnel test for the 12%scaled model of NREL Pha VI wind turbine was conducted at Korea Aero-space Rearch Institute (KARI)low speed wind tunnel.Test condition for the scaled model was decided to match the blade tip mach number with real scale model test which was conducted at NASA Ames (800Â1200)Tunnel.Aerodynamic performance reprented by torque of the blades was measure
d by using the torque nsor installed in rotating shaft and compared with real scale model test results.Acoustic noi for scaled model was also measured at clod type test ction with acoustic array of 144microphones.Time bad beamforming method to identify the rotating noi source position was applied to analyze the test results.1/3octave band was ud in post processing for various wind speeds.Test results shows that the main acoustic noi source position moves toward the blade tip as frequency increas and the noi level at low frequency below 2kHz has much higher when the blade is in stall condition.
Ó2009Elvier B.V.All rights rerved.
1.Introduction
Wind turbine noi,mostly aerodynamic noi,is one of the major issues in wind energy and has been paid attention to the aerodynamic characteristics of turbine blade to reduce noi.Many rearchers,first of all,have tried to identify the noi source and clarified that turbulent boundary layer noi is the dominant source of the wind turbine noi in broadband.Wind turbine noi prediction method bad on airfoil noi prediction method by Brooks et al.[1]was rearched by Moriarty and Migliore [2].In experimental field wind turbine noi test was conducted by Sijts-ma and
solisOerlemans by using microphone array [3].They developed the time bad beamforming method for rotating source and adapted the methodology to wind turbine and helicopter blades to identify noi sources.
Wind tunnel test,a major tool for aerodynamic configuration development,is also indispensable for aerodynamic noi study.Mostly scaled model of aerodynamic configuration is ud at wind tunnel test for the aerodynamic configuration development be-cau the test ction does not fit to the full scale.By considering the cost and schedule in development,it is inevitable to conduct
the test with scaled model and correct the scale effect for the aero-dynamic characteristics of the full scale model.
For the study of the scaled effect of wind turbine model,a ries of wind tunnel tests were conducted in KARI (Korea Aerospace Re-arch Institute)low speed wind tunnel with the 12%scaled model of NREL (National Renewable Energy Laboratory)Pha VI wind turbine model which was tested in NASA 80in.Â120in.tunnel with the real scale model and who blade configuration and aero-dynamic performance results are available in public domain.KARI also measured the acoustic noi with microphone array with 144channels and the results are summarized in this paper.
2.Experiments今天是几月几号
The experiment was performed at KARI low speed wind tunnel of which test ction is 4m Â3m and its wall is solid structure without any acoustic treatment.Aerodynamic performance and acoustic test were conducted for the 12%scaled model.To match the blade tip Mach number with the real scale model test,the wind turbine model was operated at constant rotating speed,600RPM.The turbulence level is less than 0.1%and the boundary layer tran-sition dot was attached on the blade surface for all tests.The chord-extended blade which was ud in helicopter rotor test to correct the scaled effect [4]was also tested to find out the scaled effect in wind turbine.The 15%chord-extended blade shows
1567-1739/$-e front matter Ó2009Elvier B.V.All rights rerved.doi:10.1016/j.cap.2009.11.035
recommend*Corresponding author.Fax:+82428602628.E-mail address:kr (T.Cho).
Current Applied Physics 10(2010)
什么是pk
S320–S325
Contents lists available at ScienceDirect
Current Applied Physics永远在一起的英文>永远在一起的英文
j o u r n a l h o m e p a g e :w w w.e l s e v i e r.c o m /l o c a t e /c a
p
similar torque slope with full scale model until the wind speed reaches the stall speed ($9m/s)in Fig.1.
Wind speed was incread from 0to 13m/s while two-bladed turbine was rotated at constant rotating speed.The 12%scaled blade shown in Fig.2is simply denoted as ‘B3’and 15%chord-ex-tended blade of ‘B3’is ‘B4’.The 0.18mm transition dots were at-tached at 10%chord line on both of blade surfaces.The array of 1m diameter with 144microphones (RTI 1207A)was installed on the bottom wall of the wind tunnel test ction (Fig.3).The ar-ray center is 1.88m away from the turbine rotation center in the wind direction and 1.49m to the tunnel floor.The nsitivity and pha of each microphone were calibrated by using B&K 4295speaker and the effect of test ction wall reflection was also checked.
The signals of microphones were acquired by the VIPER (Gbmh)system with phad lock mode.The 500Hz high pass filter was ud to protect the electric power noi for all tests.The sampling frequency for the microphone signal was 2000samples per revolu-tion and 10revolutions data was procesd for data analysis.Back-ground noi with turbine stand,motor and motor housing was measured and compared with test data of rotating blades.PC with ‘core2quad’CPU was ud for data analysis.
Time bad beamforming methodology for rotating source which was developed by Sijtsma was applied to analyze the test re-sults [2].The relation between measurement time (t )and emission time(s e )is given by
c ½t Às e ¼k ~x À~n ðs e ÞÀM ðt Às e Þ~e x k
ð1Þ
The acoustic pressure field at ~x generated by a monopole source at ~n is given by
p ð~x ;t Þ¼T ð~x ;~n ðs e Þ;t ;s e Þr ðs e Þ
where T À1¼4p c ðt Às e þQ ð~x ;~n ðs e Þ;t ;s e ÞÞwhere c 2Q ¼½À~n 0ðs e ÞþM ~e x ½~x À~n ðs e ÞÀM ðt Às e Þ~e x
ð2Þ
The reconstructed signal bad on emission time can be derived from Eqs.(1)and (2)and the power spectra at ~n can be directly cal-culated from source signal r ðs e Þ.The transfer function (T and Q )in Eq.(2)for all analysis points with same radius has periodic form with polar angle of analysis positi
on.Therefore,the adoption of cylindrical coordinate describing the source positions with the cri-teria between the grid angle step (D h )and time step (D s ),enables to reduce the calculation time from 7h per 1test ca in rectangu-lar coordinate to 1.5h
D h ¼
2p ‘RPM ’
60
D s ÂN
ð3Þ
The area around the blade was divided by five zones with blade radius like Fig.4and the average source strength for each zone was calculated for given octave bands.The noi level is prented as source power level with reference power,10À12W.3.Results and discussion
The Background noi with only turbine stand and motor was measured and compared with the ‘B3’test results in Fig.5.The mo-
Fig.2.NREL Pha VI blade (12%scale model [mm]).
Fig.3.Test tup for 12%scaled model.
T.Cho et al./Current Applied Physics 10(2010)S320–S325S321
tor noi is dominant below 1kHz and the blade noi is not iden-tified in that frequency.Fig.5confirms that the blade noi is sig-nificantly higher than the background noi between 1kHz and 7kHz in the tests.The repeatability of the test by using averaged source strength was less than 1dB.
The noi source contours are shown in Figs.6and 7.The contour level is the relative value with the maximum power in the analysis domain.0to À6dB was plotted in the figures.The effective angle of attack is about 3°for all radius at wind speed 5.4m/s and larger than 11°at wind speed 13.3m/s.Noi source position moves from inboard to blade tip as frequency increas for both cas.The noi source at 2kHz for the wind speed higher than the stall speed in Fig.7is located inner region than one for the wind speed lower than the stall speed in Fig.6.
The average source strength for A1and A2zone are prented in Fig.8.The noi spectra for the wind speed lower than the stall speed shows similar value for all frequencies,but the noi spectra between 1kHz and 2kHz increas as the wind speed increas above the stall speed.The noi sp
ectra for each area zone are shown in Fig.9.Blade tip area (A1)is dominant source for the wind speed 5.4m/s,but inboard area (A2,A3)is also dominant when the wind speed is 13.3m/s.The noi level between A1and A4at 4kHz is 5.3dB at wind speed,5.4m/s and 2.5dB at wind speed,13.3m/s.The noi spectra for extended blade (B4)are shown in Fig.10.The noi spectra of the wind speed,5.4m/s depends on the fre-quency normalized by tip chord length (right)but the spectra de-pends on the real frequency (left)when the wind speed is 10.0m/s.For various blade pitch angles,the noi spectra and blade performance reprented by torque are prented in Fig.11.At low wind speed below 8m/s,the blade performance of Pitch (0)which means that pitch angle is 0°,is higher than oth-ers,but the noi level is lower than Pitch (3)and slightly higher than Pitch (6).
S322T.Cho et al./Current Applied Physics 10(2010)S320–S325
4.Conclusions
Acoustic measurement was conducted for the12%scaled model of NREL Pha VI wind turbine.Microphone array was ud to identify the noi source position of the blade.The noi spectra for each wind speed and blade zone was calculated.The noi spectra have similar trends fovanessa hudgens
r the wind speed below the stall speed but the level of low frequency increas for the wind speed above the stall speed.The position of dominant noi source moves toward the blade tip as frequency increas.The noi spectra of the chord-extended blade collap with ba blade in normalized
T.Cho et al./Current Applied Physics10(2010)S320–S325S323
sinceS324T.Cho et al./Current Applied Physics10(2010)S320–S325
