1)Dr. Eng. Senior Researcher Engineer2)Dr. Eng. Associate Professor3)Dr. Eng. Senior Researcher Engineer

Wind Forces and Peak Wind Pressure Distributionson Wind Turbine Nacelle

Kenji Shimada 1) Takeshi Ishihara 2) Hiroshi Noda3)

1) Institute of Technology, Shimizu Corporation, 3-4-17, Etchujima, Koto-ku, Tokyo, Japan 2) Department of Civil Engineering, The University of Tokyo, 7-3-1, Hongo, Bunkyo-ku, Tokyo,Japan

3) Technical Research Institute, Sumitomo Mitsui Construction, 518-1 Komaki, Nagareyama, Chiba, Japan

1) shimaken@shimz.co.jp Phone : +81-3-3820-6897, Fax : +81-3-3820-5955

SUMMARY

In this paper, wind forces acting on a nacelle of wind turbine were measured in the wind tunnel experiments.In order to investigate the validity of application of surface mounted structure data on nacelle enclosure, two kinds ofexperiments, i.e. model set up on wind tunnel floor and the one apart from floor are carried out. Based on theexperimental results, empirical formula of wind force coefficients that are applicable in the design of tower and baseof the wind turbine were proposed. Furthermore a series of local peak pressure measurement was conducted. Basedon the results, empirical distributions of local peak pressure coefficients for nacelle enclosures were also proposed.

INTRODUCTION

Wind-generated power that is one of the recyclable energies is expected to beneficial energy for provisionagainst global warming well. In European countries in which numerous wind turbines were constructed,certification systems are maintained well and familiarized.

Although the wind turbines were certified in manufacturers, recently in Japan, a number of serious accidents onwind turbine structures such as collapses of tower and damages on nacelle covers have been reported. These arecaused by severe natural conditions, i.e. typhoon wind and highly gusty wind generated by complex terrains whichare inherent in Japan.　 However, aerodynamic data of nacelle which is applicable for the design of tower and baseof the wind turbine structure and nacelle enclosure is not well established.

In recently, Germanischer Lloyd , which has been well-known and the most popular certification authority,introduces pressure coefficients around the nacelle enclosure in its revised version of guideline (GL2003)[1] for windturbine certification.　Suction values in GL2003 range from –1.0 to –1.2. In GL2003, its design wind speed isdefined as a 50 year maximum peak wind speed which is 1.4 times of 50 year annual maximum wind speed, theabove values, therefore, should have been multiplied 1.96 to compare with AIJ recommendations [2] in which thecoefficients are normalized by mean wind speed. The values in GL2003 appear to be conservative estimationcompared with AIJ recommendation in which peak suction coefficient ranges from -2.0 to -5.4.

In the present work, wind forces acting on a nacelle were measured in the wind tunnel experiments. Based on theexperimental results, empirical formula of wind force coefficients which are applicable in the design of tower andbase of the wind turbine were proposed.

Furthermore a series of local peak pressure measurement was conducted. Based on the results, empiricaldistributions of local peak pressure coefficients for nacelle enclosures were also proposed.

EXPERIMENTAL SETUP

Experimental facilities and windExperiments were conducted in a large blower-type low-speed wind tunnel with the test section size of 2000mm

×2600mm. Boundary layer turbulent flow was used as experimental wind. Fig.1(a) shows mean wind speed andturbulent intensity profiles of experimental wind. The experiment wind was made for suburban and forest. The

power law exponent α is 0.2. Turbulent intensity at the height of nacelle is approximately 13%. Fig.1(b) showspower spectrum density. The power spectrum density form of experimental wind are seem to be very close to theKarman type spectrum.

Experimental models and casesConfigurations of wind turbine nacelles are classified into 4 types i.e. rectangular type, cylinder type, globe type

and disk type, generally. In this paper, rectangular type which is major configuration of wind turbine nacelle is subject.Experiment models are shown in Fig.2. Geometric scale of model is 1/50. The Experiment models of height H todepth D ratio of 1.0 and of length L to width D ratio of 2.0, 2.5 and 3.0 were tested. Aspect ratios of models weredecided according to results of survey on actual wind turbine configurations. Nacelle models and load-cell wereconnected by support rod. In order to measure the just wind force act on nacelle, support rod was covered by towermodel in case of wind force measurements. The tower model was separated from nacelle model, support rod andload-cell completely. Fig.3 shows wind force measurement set up. Definitions of wind forces are shown in Fig.4.

(a) Case1(L/D=2.5 on the ground) (b) Case4(L/D=2.5,at the top of tower)Fig.5 Set up Examples of Wind Force Measurement

D=80 L=160,200,240

H=80

φ6060

R=80

unit:mm

Fig.2 Experimental Model of Nacelle

SupportRod

Tower

Floor

Load-Cell

Nacelle

Fig.3 Setup of Model in Wind ForceMeasurements

Fig.4 Difinitions of Force and Wind Angle

FD

FL

θ

windangle

Table.1 Experimental Cases of Wind ForceMeasurements

Case No.

Case 1Case 2Case 3Case 4Case 5

NacellePosition

On the Ground

Top of Tower

Hub

w/o

w/

LengthRatio

L/D=2.5

L/D=2.5L/D=2.0

L/D=3.0

(a)Vertical Profile (b)Power SpectrumFig.1 Experimental Wind

101

102

103

100 101 102

Mean velocityα=0.2Turbulent Intensity

z[mm]

U[m/s],σ/U[%]

10-2

10-1

100

10-2 10-1 100 101

Experimental Wind

karman type spectrum

b4000u

n/UH

nSu(n)/σ2

z=400mm(at the height of nacelle)

The nacelle models also set up on wind tunnel floor in order to investigate ground effects and compare with otherwind force coefficients that were proposed for design of structure built on ground in some building code or standard.Measurement cases of wind forces were done for five cases summarized in Table 1. The set up examples are shownin Fig5. In order to assure large quantity of turbulent intensity, the tower height set to 400mm that is lower than realproportion of wind turbine tower, but it is enough to avoid the ground effects for wind force on nacelle. There aresome Influences of blades to wind forces on nacelle, but that effects were not considered in measurements. In windforces measurements, wind speed at the height of nacelles was 8m/s, sampling ratio( tΔ ) was 10ms, number of was6000 for a sample, 5 sampling were carried out and averaged 5 samples to calculate mean wind force coefficients.Wind angles are from o0 to o180 , as pitch is o5 .

Experimental model used in surface pressure measurement is only one that height to length ratio is 2.5 (Case4 inwind forces measurement). Arrangements of pressure orifices are shown in Fig.6. All pressure orifices wereplaced on one side of the model. Total number of pressure orifices was 187.

In surface pressure measurements, wind speed at the height of nacelles was 14m/s, sampling ratio ( tΔ ) was2.5ms, number of data was 8192 for a sample, 10 samples were carried out.

Evaluation of peak pressure coefficientSampling period of one sample in full scale was about 300sec based on design wind speed(50m/s). Sampling

period of peak pressure coefficients for using design of wind load must be longer than 600sec. in Japan. Peakpressure coefficients, therefore, were evaluate as follows,

1) A positive and a negative peak values were picked up from one sample. Moving average method wasconducted to each sample. Averaging time was decided by TVL method [3].

2) Larger peak value was adopted from 2 samples, that evaluating period regard as 600sec.3) Above procedures were carried out about at all wind angle, and largest coefficient among all wind

angle was defined as peak pressure coefficient.

Averaging time was calculated using TVL method as a time interval of peak pressure coefficients. Averagingtime defined by TVL method is given by

Hc U

lkT ⋅= (1)

where, Tc: averaging time, l: reference length, UH: reference wind velocity, k: decay constant The decay constant k is given by assuming root coherence of 2 points pressure fluctuation to take the formspecified by the Davenport type root coherence as

⋅−=

HU

dxnkncoh exp)(

(2)

where, n: frequency, dx: separation distance between 2 points Root coherence between 2 point pressure fluctuations act on nacelle in this experiment was covered by eq(2) inwhich setting to k=4. The evidences are described in experimental results of pressure measurements. Thereference length l is set to 1m considering the actual nacelle hatch size. Averaging time in full scale of peak pressureacting on 1m extents simultaneously comes to 0.08 sec. assuming that wind speed is 50m/s. Moving averagenumber, therefore, is 2 according to tTc Δ/ .

MEAN WIND FORCES

Influence of ground and hub to mean wind forces coefficientsDrag force coefficients and lift force coefficients defined as follows

AU

FC

H

DD

ρ

θθ

2

1)(

)( =,

AU

FC

H

LL

ρ

θθ

2

1)(

)( = (3a),(3b)

where, FD: mean drag force, FL: mean lift force, ρ : air density,θ : wind azimuth, A : reference area(=LH).

top surface

frontsurface

Fig.6 Pressure Measurement Points

rearsurface

side surface

s2

s1

s3

s4 s7 s10 s13

s6 s9 s12 s15

s5 s8 s11 s14

t2 t4 t6 t8

t1 t3 t5 t7 t9t10t11t12

The values of these coefficients are shown in Fig.7. It is seen that in Case2 (at the top of tower, without hub),maximum value of mean drag coefficient CD is 1.2 which occurred at o90=θ . In Case4 (at the top of tower, withhub), distributions of CD and CL are anti-symmetric about the o90=θ according to the effects of hub. Maximum CD

in Case4 is occurred at o80=θ and the value is 1.4. Increase of CD caused by influence of hub is almost 23% ato75=θ . Decreases in CD caused by hub effects in the wind angle from o0 to o45 were observed. In Case2 (at

the top of tower, without hub), CD were increased 50% than in Case1 (on the ground). The ground effects,therefore, were clearly seen in this study.

As to the value of mean lift coefficients CL, the values of CL with hub (Case4) have doubled in comparison withCL without hub (Case2) in the wind angle from o0 to o90 . The effects of hub are more influential on CL.

Effect of nacelle length on mean wind forces coefficients Fig.8 shows CD and CL defined by another reference area A* that face area of hub is comprehended as follow,

4* HRAA

π+= (4)

Using A* as reference face area, difference of CD and CL caused by nacelle length comes to decrease. Oncondition using A* as reference face area, CD and CL are represented by proposal equations well, which could be usedfor design load for tower and base of wind turbine as follow, respectively

( )oo 180180,68.0)2.3cos(06.0)95.1cos(4.0)( ≤≤−+−−= θθθθDC (5a)

( )( )( ) ( )oo 180180,)43.0cos()4cos(05.09.0)3sin(08.02sin9.0)( ≤≤−++−= θθθθθθLC (5b)

In order to compare with CD or CL provided in BSLJ or GL2003, We re-define the CD and CL using a furtheranother face area A(θ)** as follow,

θπ

θθ sin4

cos)(** HLR

HRA

++= (6)

CD and CL defined by A** as reference face area are compared with values provided in BSLJ (CD =1.2) andGL2003 (CD =1.3) in Fig.9 together with eq.(5a,5b). It is seen that the values of CD in BSLJ and GL2003seem to be conservative rather than proposal values.

(a) CD (b) CLFig.9 Effects of Hub Length on Nacelle Load(A** by eq.(4))

0.0

0.5

1.0

1.5

0 30 60 90 120 150 180

L/D=2.0L/D=2.5L/D=3.0Approx.

CD

θ [deg.]

GL2003 : CD=1.3

BSLJ : CD=1.2

-0.2cos(2.7a-45)+0.23a/180+0.75

-1.0

-0.5

0.0

0.5

1.0

0 30 60 90 120 150 180

L/D=2.0L/D=2.5L/D=3.0Approx.

CL

θ [deg.]

(-1.15sin2θ-0.14sin4θ)×(0.77+0.115cos4θ)×(180-0.4θ)/180

(a) CD (b) CLFig.8 Effects of Hub Length on Nacelle Load(A* by eq.(2))

0.0

0.5

1.0

1.5

0 30 60 90 120 150 180

L/D=2.0L/D=2.5L/D=3.0Approx.

CD

θ [deg.]

-0.4cos(1.95θ)-0.06cos(3.2θ)+0.68

-1.0

-0.5

0.0

0.5

1.0

0 30 60 90 120 150 180

L/D=2.0L/D=2.5L/D=3.0Approx.

CL

θ [deg.]

(-0.9sin(2θ)+0.08sin(3θ))×(0.9+0.05cos(4θ))cos(0.43θ)

PEAK PRESSURE COEFFICIENTS

Root Coherences of surface pressuresExamples of root coherence of pressure fluctuation on surface of nacelle are shown in Fig.10 together with

eq.(2) with k =8 and k=4. These wind angle are θ =55°and θ =215°in which negative peak pressures wereoccurred, respectively. The root coherences between 2 points pressure fluctuations have varied from point to pointand from wind angles. Eq.(2) with k=8 is center of variations in root coherences. The root coherences of t9-t12and s1-s2 are smaller than eq. (2) with k=8. The points of s1 and t9 were at the corner of side surface and at theedge of top surface where largest negative peak pressures were occurred, respectively. If the value of k is set to 4,almost root coherences of pressure fluctuation on nacelle are covered by eq.(2). Therefore, value of k is set to 4 inthis study and estimate number of moving average of peak pressures.

Distributions of mean and peak wind pressure coefficientsLargest and smallest mean pressure coefficients in all of wind angles are defined as maximum mean pressure

coefficients and minimum mean pressure coefficients, respectively. Distributions of maximum mean pressurecoefficients and minimum mean pressure coefficients are shown in Fig.11. First, maximum mean pressurecoefficients are described. It is seen that, negative values of mean pressure on the top surface of nacelle do notoccur. On the side and rear surface, values are 0.8 in almost region, and slightly decrease near upper and lower edge.But these decrease are not seen near vertical edges. On the front surface, values on the region shaded by hub areabout 0.6, and about 0.8 on other regions. Distributions of minimum mean pressure on each surface are morecomplex rather than that of maximum except on rear surface. Positive and negative peak pressure coefficients distributions on nacelle surfaces are shown in Fig.12. Thevalues of positive peak pressure coefficients are from 0.4 to 0.6 on top surface, from 1.6 to 2.2 on the other surfaces.It is seen that the locally particularly large values are not

Fig.10 Root Coherence of Fluctuating Pressure Coefficients(a) top surface, wind angle:215° (b) side surface, wind angle:55°

0.0

0.2

0.4

0.6

0.8

1.0

0.0 0.1 0.2 0.3 0.4

t1-t2t3-t4t5-t6t7-t8t9-t12eq.(2),k=8eq.(2),k=4

rtcoh

n|dx|/UH

√coh(n)

θ=215°,top

0.0

0.2

0.4

0.6

0.8

1.0

0.0 0.1 0.2 0.3 0.4

s1-s2s3-s2s4-s5s6-s5s6-s8s9-s8

s10-s11s12-s11s13-s14s15-s14eq.(2),k=8eq.(2),k=4

rtcoh

n|dx|/UH

√coh(n)

θ=55°,side

toppitch:0.2

0.8

0.80.8

0.8

-0.2

0.6

0.6

0.6

0.6

0.8

0.8

-1.0

-2.4

-1.2

-1.0

-1.4

-0.6

-1.0

-0.6

-0.8-0.8

-1.6

-1.2

-1.2

-1.8

pitch:0.2

side rearfront

top

side rearfront

Fig.11 Maximum and Minimum Mean Pressure Coefficient Distributions(a) Maximum mean pressure (b) Minimum mean pressure

0.4 0.4 0.6

2.2

2.0

2.0

1.6

1.61.8

1.8

2.0

1.6

1.6

2.0

2.0

1.8

1.8

-1.8

-3.2

-1.6

-1.8 -1.8

-2.0

-3.4

-3.4

-3.4

-4.0

-2.8

-2.8

-3.0

-2.6

-3.0 -5.8

-3.0-2.8

-4.4

-2.6

-4.0

toppitch:0.2

side rearfront

toppitch:0.2

side rearfront

(a) Positive peak pressure (a) Negative peak pressureFig.12 Peak Pressure Coefficient Distributions

appeared on any surfaces. As to the negative peak pressure coefficients local large peak pressures, on the other hand,are occurred near the corner on side surface, on rear surface and on top surface. The values of local peak pressurecoefficients are from –3.4 to –4.0 on side surface, –3.0 on rear surface, respectively. The local peak pressure at sidesurface corner near front are occurred at wind angle 55°(see Fig.13), that at the corner nearby rear are occurred atwind angle 195°,the local peak pressures on side surface occurred in leeward slightly. The local peak pressures areoccurred on top surface. Particularly at the edge near rear, where influence of hub is seem to disappear, largest peakpressure appears. The value of local peak pressure at the edge on the top surface is –5.8. It seems that this localpeak pressure caused by conical vortices, which appear on upwind edges at diagonal wind angle[4]. In point of fact, itwas occurred at wind angle 215°(see Fig.13).

-6.0

-5.0

-4.0

-3.0

-2.0

-1.0

0.0

0 90 180 270 360

t9

t10

t11

t12

129

θ

Cp

-5.0

-4.0

-3.0

-2.0

-1.0

0.0

1.0

0 90 180 270 360

s-1

s-2

s-3

s-4

s-5

129

θ

Cp

Fig.13 Minus Peak Pressure Coefficients(a) top surface (b) side surface

Comparison with standards and recommendations GL2003 has recommended pressure coefficients for design of nacelle enclosure. These pressure coefficients aremean pressure coefficients based on 3sec. gust wind speed. Pressure coefficients for nacelle enclosure in GL2003are shown in Fig.14. The values in fig.14 are converted into peak pressure coefficients based on 10 min. mean windspeed using the square of conversion rate (=1.4) between gust wind speed and mean wind speed determined inGL2003. The values of pressure coefficient recommended in GL2003 are 1.6 for positive pressure and –1.2 fornegative pressure, which appear to be considerably underestimated compared with results in this study. Positivepeak pressure coefficients for cladding on walls provided by BSLJ and AIJ recommendations are 2.3 and 2.0 (heightis 60m, category III), respectively. Negative peak pressure coefficients provided in BSLJ and AIJ recommendationsare shown in Fig.15. As to the positive peak pressure coefficients, values in BSLJ and in AIJ recommendations areclose to results in this experiment results well. Because main cause of positive peak pressure is turbulence inoncoming flow and its almost unrelated to configrations or locations of nacelle. Thought it seems that applying thepositive peak pressure coefficients in BSLJ and AIJ for wind resistant design of nacelle enclosure are reasonable.As to the negative peak pressure coefficients, distributions obtained in this experiments are kind of similar to that onroof on which peak pressure comes to large near the corner. Though the values could not be compared directlybecause those are based on mean hourly wind speed, distributions of pressure coefficients on roof for design providedin ASCE standard[5] are similar to that in this study, in BSLJ and in AIJ recommendations.

-2.4

-2.5

-4.3

-3.2

-3.0

-2.5

-3.2-4.3

-2.4

-3.0

(a)BSLJ,roof

(b)BSLJ,walls

(c)AIJ,roof

(d)AIJ,walls

Fig.15 Minimun Peak Pressure Coefficients inBSLJand AIJ Recomendations　

Fig.14 Pressure Coefficients forNacelles in GL2003

wind direction

1.6 -1.2 -1.0

-1.2

-5.0 -5.0

-3.0-4.4 -4.4

-5.8-4.0

-3.5

-1.8-2.6

-3.0-2.0

-3.5 -3.5

Fig.16 Schematic Distribusions of MinimunPeak Pressure Coefficients for Panel Design

Loads of Wind Tubine Nacelles 　

top

side rearfront

On the basis of above results, schematic distributions of positive peak pressure could be drawn with consideringabove reasons are shown in Fig.16.

CONCLUDING REMARKS

Total mean wind forces and local peak pressures on nacelles were measured in order to investigate applicabletotal load on nacelle mounted on tower and distributions of peak pressure coefficients for nacelle enclosure. Meandrag coefficients and mean lift coefficients for estimate total design load and peak pressure coefficients on account ofwind resistant design for nacelle enclosure are proposed. The values of mean drag coefficients in Building StandardLow of Japan and GL2003 seem to be conservative rather than proposal values.

However, the values of pressure coefficients recommended in GL2003 appear to be considerably underestimatedcompared with results in this study, the positive peak pressure coefficients obtained in this study are close to that inBSLJ and in AIJ recommendations. Thought, distributions of negative peak pressure coefficients are kind of similarto that on roof in BSLJ and in AIJ recommendations, the values are larger than that in BSLJ and AIJrecommendations on some regions.

REFERENCES1.Germanischer Lloyd (2003), Rules and Guidelines IV Industrial Services 1 Guideline for the Certification of Wind

Turbines, Chapter62.Architectural Institute of Japan (2003), AIJ Recommendations for loads on buildings3.Lawson, T. V.,(1980), Wind Effects on Buildings, Vol.2, Applied Science Publishers, p.1924.Cook, N.J. ( 1985), The designer’s guide to wind loading of building structures, Part1, p.371,5.American Society of Civil Engineers (2002), Minimum design loads for buildings and other structures

Wind Forces and Peak Wind Pressure DistributionOn Wind Turbine Nacelle

Kenji SHIMADA* shimaken@shimz.co.jpTakeshi ISHIHARA** ishihara@bridge.t.u-tokyo.ac.jp

Hiroshi NODA*** hnoda@smcon.co.jp

*Institute of Technology, Shimizu Corporation, 3-4-17 Etchujima Koto TOKYO 135-8350 JAPAN**Institute of Engineering Innovation, School of Engineering, The University of Tokyo, 2-11-6 Yayoi Bunkyo TOKYO 113-8656 JAPAN

***Technical Research Institute, Sumitomo Mitsui Construction, 518-1 Komaki Nagareyama CHIBA 270-0132 JAPAN

Object

• Ground effects are apparent in CD. Drag force acting on a nacelle with tower is almost twice as large as those acting on a hub on the ground.

• The effects of hub are found to be more influential on CL.

Mean Wind Force Coefficients Local Peak Pressure Coefficients

Wind turbine certification system are maintained well, and familiarized, however, serious damages of nacelle are reported. Aerodynamic data on nacelle does not seem to be well established.

In this paper, wind forces acting on a nacelle were measured in the wind tunnel experiments. At first, based on the experimental results, empirical formula of wind force coefficients which is applicable in the design of tower and base of the wind turbine were proposed. Next, a series of local peak pressure measurement was conducted. Based on the results, empirical distributions of local peak pressure coefficients for a nacelle enclosures were proposed.

Influence of Ground and Hub

FD

FL

θ

windangle

Drag Coefficient Lift Coefficient

Wind Tunnel Experiments

Wind Force Test Model

SupportRod

Tower

Floor

Load-Cell

Nacelle

• For the evaluation of local peak pressure, appropriate evaluation time should be determined. In this study, it was estimated by TVL method.

• In TVL method, pressure fluctuation which is below the critical frequency of L/TcU can be regarded as acting on nacelle enclosure simultaneously.

-3.0-4.4 -4.4-5.8-4.0

-3.5

-2.6-3.0-2.0

-3.5 -3.5

top

side rearfront

-0.8

-3.2

Pitch:0.2

-3.4-4.0

-3.0 -5.8

-4.4

-3.0-2.8

-2.6

-4.0

-1.8-2.0

-3.4

-3.4

-2.0

-2.4

-1.6

-2.0

-1.6-1.8

-2.8

-3.0

-2.6-2.8

-1.6-2.4-3.3-4.1-5.0-5.8

Front Side Rear

Top

• Total mean wind forces and local peak pressures on nacelles were measured in order to investigate reasonable wind load on nacelles

• For mean wind force coefficients, empirical formula were proposed based on the experimental results.

• Distributions of peak pressure coefficients which are applicable in the design of nacelle cover were also proposed.

• The drag coefficients recommended in GL2003 were found to be conservative as total wind forces.

• However, the values in GL2003* appear to be underestimated as local pressure coefficients.

Conclusion

-5.0

-4.0

-3.0

-2.0

-1.0

0.0

1.0

0 90 180 270 360

s-1s-2s-3

s-4s-5

θ

Cp

-6.0

-5.0

-4.0

-3.0

-2.0

-1.0

0.0

0 90 180 270 360

t9t10t11t12

θ

Cp

0.0

0.2

0.4

0.6

0.8

1.0

0.0 0.1 0.2 0.3 0.4

t1-t2t3-t4t5-t6t7-t8t9-t12eq.(2),k=8eq.(2),k=4

n|dx|/UH

√coh(n)

α=215°,top

• Pressure time series at each measuring point was smoothed by moovingaverage by evaluation time Tc.

• The above procedure was applied for all the points at all angles.

Coherence of Pressure fluctuation

• 2 point root coherence of pressure fluctuations was covered by k=4.

• The reference length L is set to 1m considering the actual nacelle hatch size.

• Averaging time in full scale comes to 0.08 sec. assuming that wind speed is 50m/s.

• The largest coefficient amoung all wind angles was defined as a local peak pressure.

topsurface

t2 t4 t6 t8

t1 t3 t5 t7 t9t10t11t12

• Based on the measurements, local peak pressure distribution was proposed.

• For the negative peak pressure, GL2003* appears to be underestimated compared with the proposed distribution.

wind direction

1.6 -1.2 -1.0

-1.2

*Coefficients was redefined based on mean wind speed.

0.0

0.5

1.0

1.5

0 30 60 90 120 150 180

L/D=2.0L/D=2.5L/D=3.0Approx.

CD

θ [deg.]

GL2003 : CD=1.3

-0.2cos(2.7a-45)+0.23a/180+0.75

-1.0

-0.5

0.0

0.5

1.0

0 30 60 90 120 150 180

L/D=2.0L/D=2.5L/D=3.0Approx.

CL

θ [deg.]

(-1.15sin2θ-0.14sin4θ)×(0.77+0.115cos4θ)×(180-0.4θ)/180

θπθθ sin4

cos)( HLRHRA ⎟⎠⎞

⎜⎝⎛ ++=

topsurface

t9t10t11t12

α

sidesurface

s2

s1

s3

s4

s5

• Drag and lift coefficients are fitted by empirical formula below.• The values of mean drag coefficient 1.3 in GL2003 appears to be

conservative rather than proposed values.

AU

FCAU

FCH

LL

H

DD

22

21,

21 ρρ

==

0.0

0.5

1.0

1.5

2.0

0 30 60 90 120 150 180

CD

θ [deg.]

On the Ground(w/o Hub)

w/ tower (w/o Hub)w/ tower (w/Hub)

Ground effects

-1.0

-0.5

0.0

0.5

1.0

0 30 60 90 120 150 180

CL

θ [deg.]

On the Ground(w/o Hub)

w/ tower (w/o Hub)

w/ tower (w/Hub)

Hub effects

Comparison with GL2003

FD

FL

θ

windangle

LHAAU

FCAU

FCH

LL

H

DD === ,

21

,

21 22 ρρ

D=80 L=2D,2.5D,3D

H=

80

φ60 60

R=80

unit:mm

Case2 (w/o hub)

Case1

Case 4

Distribution of MeasuredNegative Peak Pressure

Proposed Negative LocalPeak Pressure Distribution

GL2003*

Negative PeakPressure

• Wind force measurement testand pressure measurement test are conducted.

• 1/50 scaled models were used.

• In order to measure only wind forces acting on nacelle, support rod was covered by tower model which was separated from nacelle model

• In the force tests, effects of ground and hub were investigated.

• Also effect of length of the nacelle was also investigated.

• In order to assure the large turbulence intensity, height of the model was set to 400mm that is lower than real proportion, however, it is high enough to avoid the ground effects for wind force on nacelle.

101

102

103

100 101 102

z[mm]

U[m/s],σ/U[%]

TurbulenceIntensity

Mean Velocityα=0.2

Experimental Wind

Pressure Test Modeltop surface

frontsurface

rearsurface

side surface

s2

s1

s3

s4 s7 s10 s13

s6 s9 s12 s15

s5 s8 s11 s14

t2 t4 t6 t8

t1 t3 t5 t7 t9t10t11t12• Total number of pressure orifices was187

• L/D=2.5 model was used.

• Turbulent boundary layer flow of α=0.2 and I=0.13 at the hub height was used.

Case No.

Case 1Case 2Case 3Case 4Case 5

NacellePosition

On the Ground

Top of Tower

Hub

w/o

w/

LengthRatio

L/D=2.5

L/D=2.5L/D=2.0

L/D=3.0

Collapse of nacelle cover

Damage at a man hatch

Damage of nacelle cover