Home

GH WindFarmer 4.0

1. 3 5 5 Model parameters The parameters in the calculation are e Base roughness that represents the surface without wind turbines a default value for the Base roughness of 0 0002 m offshore is used For onshore applications a single value of Base roughness which is representative of the site and surrounding area is appropriate e Increased roughness representing the effect of the wind turbines An increase from the base roughness between 0 02 and 0 03m has been found to provide a good match A higher value of 0 05 can be used when looking at specific wind speed cases with deep wakes The default setting for the increased roughness is 0 03m e Geometric width This parameter determines the sector for which an upstream turbine has an affect on the downwind turbines The default setting is 1 turbine diameter e Maximum row spacing This parameter is used to detect situations when the wind is blowing along a row of turbines In this case the offshore effect is not visible in the wind farm data The row spacing represents the distance to the next row of turbines in the projection plane perpendicular to the wind direction If this distance is greater than the maximum row spacing then the offshore correction is not applied The default value is 5 turbine diameters e Start and stop parameters for recovery It is expected that the total disturbance declines as the distance behind the wind turbine increases A linear recovery to the unaffected free
2. V05403 d Vine where dj is the distance to the neighbouring turbine that causes the wake in direction bin j normalised by the rotor diameter is the mean ambient wind speed standard deviation in wind direction bin j and Gi wind speed bin i For wind directions without upstream turbines or turbine distances greater than 10 rotor diameters Orij 0 L The estimate of the Effective turbulence intensity Zey is calculated in GH WindFarmer for all wind speeds and directions and is available in the Flow and Performance Matrix GH WINDFARMER THEORY MANUAL 21 The representative value of wind speed standard deviation o ep is defined as the 90 quantile of the turbulence standard deviation O rep Oo Yu 1 280 4 2 8 where O is the standard deviation of wind speed standard deviation in wind speed bin Oi i given as a mean over all wind directions If o is not available from measurements it can be assumed that it is 20 of the mean wind speed standard deviation 4 2 so that le inp 0 2567 4 2 9 where O is the mean of standard deviation of ambient wind speed in wind speed bin i given as a mean over all wind directions O re j The corresponding Representative turbulence intensity i is available in the Flow and V hub i Performance Matrix of GH WindFarmer To get a measure of the suitability of a wind turbine for the site a wind turbine should satisfy the condition that the estimate of the representative
3. 4 4 2 the load increases as soon as a part of the rotor is within the wake The model given by equation 4 4 3 should only be used for calculation of wake turbulence aggregated over a complete wind rose since it can produce non conservative turbulence estimates for individual wind directions at close spacing 4 4 2 Other model assumptions In the simple model effects from an upwind turbine causing a wake on a downwind turbine are neglected The free wind speed is assumed on the upwind turbine In the advanced model the wind speed at the upwind turbine is reduced if there are any wakes on the upwind turbine The differences in the free wind speed due to the terrain are taken into account in the advanced model while the same free wind speed at all turbine locations is assumed in the simple model GH WINDFARMER THEORY MANUAL 25 In the simple model it is assumed that the thrust coefficient of the turbine causing the wake is the same as for the turbine receiving the wake In the advanced model the actual trust coefficient of the upstream turbine is used in the calculation In both models the hub heights of the turbines are taken into account when determining if the turbine rotor is completely below or above the wake of an upstream turbine so that there is no wake effect As soon as the tip of the downwind turbine gets into the wake it is assumed that the turbine has the same hub height as the upstream turbine and it is determined how much of
4. The points A P and S are represented by their vectors a p and s a Asb The vector b is the unit vector pointing from the receptor to the middle of the sun It is given by cos e sin az b cos el cos az sin e For vector AS to be perpendicular to the vector PS we require be s p 0 us DO a eS This leads to the vector SP perpendicular to the vector AS sp a b p GH WindFarmer compares the norm of the vector PS with the radius R of the turbine This is repeated in time intervals of for example 1 minute through a year to detect if shadow is produced at the point of analysis at this time The program counts the minutes per day and the hours per year of shadow flicker caused by that wind turbine GH WindFarmer always considers the topography in the calculation using the height ASL specified in the DTM file GH WindFarmer optionally allows you to detect if the direct line of sight between receptor and turbine or between turbine and sun is blocked by terrain features 9 5 Modelling the rotor as a disc If the rotor is modelled as a sphere then this represents the worst possible geometric scenario This is the method we recommend Alternative options in GH WindFarmer allow the rotor to be modelled in a vertical plane of defined orientation When the distance between the rotor and the turbine tower centre is taken into account and the rotor disc model is being used the worst case scenario occurs whe
5. _ E Em F e 20 1 f E or with Em the central estimate of the predicted energy prediction calculated net energy and o the overall absolute standard deviation of the predicted energy production The equation is shown graphically in the following plot The plot shows the central most likely energy yield predicted as the Net Yield of GH WindFarmer this is also called the P50 value or central estimate Normal distribution Probability P30 Central Energy yield GH WINDFARMER THEORY MANUAL 34 The probability P that E exceeds a certain value e is P E gt e 1 fdt 1 F e F te oo In order to be able to use tabulated values for F e it is necessary to convert F e into a Normalised Normal Gaussian Distribution Babess Es 0 P E gt eP P gt 2 F tz As we want to know the energy production that is exceeded with a specific probability we need F z This can not be derived analytically but there are look up tables for the normal distribution for specific probabilities of exceedance and the corresponding values for z This is the reason why GH WindFarmer provides confidence limits only for specific probabilities of exceedance These are 1 10 25 50 75 80 84 85 90 95 and 99 The energy production that is exceeded with a certain probability is then derived from 26 gt Ei Sg This results in e E ZO Z The central estimate is exceeded with a probability of 50 So for t
6. analysis e The low wind speed cut off set in the Measure Correlate Predict window refers to the wind speed at the reference station only This option is implemented by activating calculation of the site cut off All points x y having x smaller than the cut off wind speed are excluded from the correlation analysis The calculated speed up is then applied to the cut off wind speed to derive the corresponding wind speed at the site All points x y with y smaller than the cut off wind speed at the site are then excluded from the second correlation iteration and a new speed up determined The new speed up is then applied to obtain a new site cut off and this process is repeated until the slope of the linear fit no longer changes Note that in the prediction phase of the MCP process the correlation relationships are applied to data of all wind speeds to be sure that a complete wind speed distribution is derived 5 2 Calculation of direction shifts Systematic direction shifts between the reference and site measurements can be applied if required when the long term site wind speed and direction distribution is predicted Otherwise the long term wind direction distributions at the reference station and site are assumed to be the same The direction shift between the time series is calculated for each 30 degree direction sector in the following way First the reference direction time series is binned into 12 direction sectors For each t
7. e For each downstream turbine calculate the reactive power Q produced or consumed by the turbine for the given hub height wind speed from the user defined table of Q or PF versus P e For each downstream PFCD calculate the reactive power Q produced by this element as shown in Section 8 3 GH WINDFARMER THEORY MANUAL 45 e Summate to find total P and total Q passing through the transformer from all downstream turbines e Calculate apparent power S yP Q kVA e Calculate current I for transformer secondary side o 1 S 1000 V3 U Amps o where U is the nominal voltage of the secondary side 8 4 3 Calculation of the resistance of transformer For each transformer the user will specify a nominal rating S kVA a nominal secondary voltage U volts and a load loss P at nominal rating W lom S 1000 V 3 u R Pi 3 lhom Ohms The result is the transformer series resistance in ohms referred to the secondary side 8 4 4 Calculation of the transformer series reactance For each transformer the transformer series reactance X is obtained by means of the following expressions X Z2 R Z Zvaro U J S 1000 100 Ohms where X Is transformer series reactance per phase Ohms R is transformer resistance as calculated above U is transformer nominal secondary voltage Volts Ztrafo IS transformer series impedance S is transformer nominal rating kVA Note the first equation will cause errors if
8. when the projected wind farm will be operational The equation below defines the conversion from an annual wind speed variability uncertainty to an uncertainty over a reference number of years U swsy JN U ws 6 1 With Unwsv uncertainty wind speed variability over N years Uawsv uncertainty annual wind speed variability N reference number of years It is assumed that the annual wind speed variability is the same for the past and the future periods In some regions there are sufficient data to describe a region specific value for annual wind speed variability In the absence of this an annual wind speed variability of 6 is recommended The Future Uncertainty describes the uncertainty for the target period Future Uncertainty are usually 1 year to assess uncertainty over a single operational year of the projected wind farm 10 years which is a typical period for simple payback period of the investment and 20 years being standard life span of a wind farm project 6 4 Sensitivity The type of uncertainty can be user defined or one of four predefined types wind speed energy yield topographic effect or wake losses Before combining them to a global uncertainty level all uncertainties are first converted to a uncertainty on energy yield The conversion factor of a particluar uncertainty to an energy yield uncertainty percentage is defined called a sensitivity The four predefined types are e Wind speed th
9. IEC 61400 1 Edition 2 In this Standard 4 2 the design turbulence is defined as a Characteristic value The maximum allowed Characteristic turbulence is defined according to the turbine class The Characteristic turbulence and the longitudinal wind velocity standard deviation o are linked through the following formula o Ving Lis 15 4 Ving a 1 4 2 1 where O1 is in this case called the Characteristic value of the standard deviation of the longitudinal wind speed component Iis is the characteristic value of the turbulence intensity at 15m s a is the slope parameter Vhub is the wind speed at hub height in m s The parameters and a are specified in the standard according to the turbulence level of the turbine class For low turbulence a 0 16 3 and for high turbulence 7 5 a 0 18 2 The standard does not state how wake effects at the site may be taken into account in the calculation of the site specific characteristic turbulence One method to do this is to use a method published in 4 3 whereby the wake effects from neighbouring turbines are taken into account by an Effective turbulence The Effective standard deviation is defined as 1 m Cog Vrbi By eu 4 2 2 J where Pj is the probability of the wind direction bin j In GH WindFarmer the predicted turbine specific probability is used m is the W hler coefficient representative for the turbine and component receiving turbulence Th
10. THEORY MANUAL 57 11 SUMMARY GH WindFarmer is a sophisticated and versatile package not just for the prediction of the energy yield of a wind farm but also for Long term correlation of measured wind data Optimisation of wind farm layout Wind farm noise calculation Noise constrained turbine placement Electrical infrastructure design and loss calculation Shadow flicker calculation Prediction of wind farm induced turbulence intensity Financial modelling Visualisation and Photomontage Because of the large number and complexity of the topics it is unfortunately not possible to describe all of them in as much detail as we would like to Please consult the quoted literature for more in depth reading and the GH WindFarmer Validation document for comparisons with measured data Useful information regarding the application of the models of GH WindFarmer that are presented in this manual can also be found in the User Manual Last but not least please do not hesitate to ask the GH WindFarmer user support team if you need any further information or have suggestions to further improve this document GH WINDFARMER THEORY MANUAL 58 REFERENCES 1 1 1 2 1 3 3 1 3 2 3 3 3 4 3 5 3 6 3 7 3 8 3 9 Ainslie J F Development of an eddy viscosity model for wind turbine wakes Proceedings of 7 BWEA Wind Energy Conference Oxford 1985 Ainslie J F Development of an Eddy Viscosity model of a Wind Turbine Wake C
11. are also different turbine types in the wind farm Some assumptions have been made to take this into account On the other hand a simple approach has also been made which is called the Simple Model in GH WindFarmer The differences of the advanced and simple models are explained below Depending on the circumstances either the simple or the advanced model may give more conservative results 4 4 1 Direction sectors with increased turbulence To implement the above it is necessary to identify the wind direction sectors in which a turbine will experience increased loads due to the wakes of upstream turbines These sectors are determined by applying the approach proposed in 4 3 The effective turbulence GH WINDFARMER THEORY MANUAL 23 is calculated using equations 4 2 3 4 2 7 or 4 2 10 as soon as a part of the rotor is within the wake of an upstream turbine as represented by a view angle D2 a 2 ae oa p X er j Pa Wake producing turbine Wake affected turbine Definition of the view angle Dp is the rotor diameter of the turbines and D is the diameter of the wake at the downstream distance s The view angle can be described as 0 arctan 1 2 i 4 4 1 z arcan arctan k for l lt lt l s s where S is the downstream distance in rotor diameters k is the wake decay constant which following recommendations of 4 3 is derived on the basis of the linear wake expansion equati
12. dependent turbine sound power levels are available The Simple noise model in GH WindFarmer calculates the noise propagation at a fixed reference frequency of 500 Hz The continuous octave band sound pressure level at a receiver location L is calculated using the equation ISO 3 La f D A 6 2 1 where Lu is the sound power level in dB A produced by each turbine taking the turbine as a point source D is the directivity correction in decibels For the case of an assumed omni directional point sound source Wind Turbine D 0 dB The directivity of the wind turbine noise is considered when measuring the sound power level A is the attenuation that takes place during the propagation from the point sound source to the receiver in decibels Note GH WindFarmer does not perform an A weighting Please enter A weighted sound power levels if you need your result in dB A GH WINDFARMER THEORY MANUAL 36 Equation 6 2 1 relates the sound power level of a turbine to the sound pressure level at a reference distance from an omni directional sound source The attenuation A in equation 6 2 1 is defined by ISO 4 A Aaiv ae Aatm Agr Abar Amisc Amet 6 2 2 where Agiv is the attenuation due to geometrical divergence Aatm is the attenuation due to atmospheric absorption Agr is the attenuation due to ground effects Abar is the attenuation due to barriers Amis is the attenuation due to other effects like foliage and a
13. distance The correction is applied only if the adjacent lines of turbines are closer than a threshold e Downstream recovery From a certain distance the wind speed recovers to ambient levels and no correction is applied The component models are presented in the following sections 3 5 1 Boundary layer modification Momentum is continuously generated on top of the boundary layer and transferred downward to the ground surface in a dynamic equilibrium Wind turbines that take out some of this momentum are part of the dynamic equilibrium similar to trees or other roughness elements GH WINDFARMER THEORY MANUAL 12 However as the impact of the wind turbines on the boundary layer profile is not as well researched we take some guidance from forest canopy and roughness change models height D distance from 1st turbine D Internal boundary layer IBL development after disturbance caused by first turbine The ambient wind speed for downstream turbines is reduced An internal boundary layer IBL develops from each turbine location The height h of the IBL for a roughness change as function of the fetch x and z o the larger of the two roughness values Z91 and Zo2 can be determined from m4 1 05 amp Zo Zo Zo To take into account that the momentum is not extracted at ground level an offset of 2 3 hub height z is used leading to a new height h It is further assumed that the disturbance to the ambient wind
14. equations given in Table 3 ISO 9613 2 7 3 4 Alternative method to calculate the Ground Attenuation Agr This alternative method is suggested in ISO 9613 2 as an option for specific situations where only the A weighted sound power level is of interest The calculation then depends on two terms ISO 10 and ISO 11 that are not dependent on frequency or on ground properties Agr 4 8 2hm d 17 300 d dB 6 3 3 Do 101g 1 d 7 hg h d h h dB 6 3 4 Where m is the mean height of the propagation path above ground is the distance of the receiver from the wind turbine is the hub height of the wind turbine is the ground height of the receiver a 7p 2 DS ne is the distance from the wind turbine base to the receiver base projected onto the ground plane The first term is for the ground attenuation the second term is a correction for the reflections of sound from the ground near the source Ag is limited to a number equal or greater than OdB GH WINDFARMER THEORY MANUAL 39 7 3 5 Meteorological correction Amet The meteorological correction allows the user to correct for statistically changing meteorological conditions This correction may be applied if a long term average sound pressure level is required and sufficient information about the local meteorological statistics is available to establish a value for the site specific factor Co 7 3 6 Miscellaneous types of attenuation Amisc This parameter c
15. estimates are given below as guidance GH WINDFARMER THEORY MANUAL 31 Correlation coefficient Wind speed correlation uncertainty gt 0 9 lt 1 0 9 0 8 1 2 0 7 0 6 3 5 It is recommended not to use correlation results with a correlation coefficient less than 0 6 In this case it may be worthwhile looking for other reference stations or to carry out an energy calculation with only the wind data measured on site This will then increase the historical uncertainty due to the annual variability but this may be balanced by the absence of a correlation uncertainty 6 2 Modeling uncertainties Topographic model uncertainties The wind flow model is not always capable to predict the site wind regime generated from an initiation mast on site to the turbine location The complexity of terrain the site roughness the presence of obstacles and the distance between each turbine and the initiation mast are factors determining the magnitude of this uncertainty The difference in height between the mast and the predicted hub height is also a factor computed in the assessment of this uncertainty Wake model uncertainty Wake model uncertainty relates to the wake loss calculations with respect to the calculated wake loss The area of uncertainty includes wake model uncertainties but also uncertainties related to the thrust characteristic of the turbine Vertical extrapolation uncertainty Uncertainty lies in the wind sh
16. information In this case the user must use a long term reference tab that is not seasonally biased In the first step monthly seasonal wind speed and direction distributions are created using all data in the time series in a specific month e g all data in a January The annual tab file is then calculated by summing these monthly tables using a weighting according to the number of days in each month To remove the seasonal bias the time series must contain data for all months It should be noted that seasonal bias is not removed in wind roses and turbulence wti files GH WINDFARMER THEORY MANUAL 30 6 UNCERTAINTY ANALYSIS The energy production calculation of a wind farm is always subject to uncertainties that should be accounted for assessing the degree of accuracy of the calculation model The three main areas of uncertainty are outlined below e Measurement uncertainties due to anemometry characteristics of reference meteorological stations and on site masts Furthermore correlation between masts and wind index contribute to the measurement uncertainty e Modelling uncertainties due to complexity of topography and site roughness layout induced wakes wind shear turbine power curve meter accuracy utility downtime etc e Wind speed variability uncertainties due to the statistical fluctuations of the wind resource over the historical period when measurements have been carried out and the future period when the wind farm will b
17. phase Alternatively this calculation could be further simplified for those expected cases up to 3 elements in parallel For N 2 the equivalent impedance is calculated as 1 1 Zeq 27 being the currents flowing across each branch l1 Z2 Z4 Z2 l2 Z4 Z4 Z2 l h l For N 3 the equivalent impedance is calculated as 1 1 Zeq Z being the currents flowing across each branch l Z2 Za Z2 ZZ Z2 Z3 GH WINDFARMER THEORY MANUAL 47 l2 Z4 Z3 Z1 Z3 Z1 Z2 Z3 l3 Z Zal Z1 Z2 Z4 Z2 Zs l lo l3 GH WindFarmer electrical module provides only the results for the equivalent impedance If the user is particularly interested in knowing an electrical parameter of one of the branches the user will have to determine this on a separate calculation 8 5 Caveats Users attention is drawn to the caveats given in the Electrical Chapter of the User Manual GH WINDFARMER THEORY MANUAL 48 9 SHADOW FLICKER 9 1 Introduction Shadow flicker is the occurrence of periodic changes in light intensity due to the shadow of a wind turbine blade passing over a point of interest The shadow flicker module simulates the path of the sun over the year and assesses at each time interval the possible shadow flicker at one or multiple receptor position s The output of the module can be used to design a wind farm to fulfill planning requirements The results of the mod
18. stream conditions is assumed between the start and stop parameters Default is 60 respectively 80 diameters Considerable caution is required with regards to the application of the large wind farm correction The model has not yet been validated against a substantial number of wind farms As soon as operational data from additional wind farms become available an update of the model is likely and therefore the current model results may be subject to change as the model is further refined 3 6 Terrain modification of the wake and turbulence The wind flow and wake around a wind turbine are accelerated when passing over terrain This acceleration is calculated by the wind flow model and affects both mean wind speeds and turbulence GH WindFarmer allows optional consideration of this effect on the mean wind speed incident on the next turbine by scaling the wake deficit with the terrain speed up derived from the flow model Using the assumption that the standard deviation stays constant while the mean wind speed is modified over the terrain GH WindFarmer also varies the turbulence intensity over a given terrain Three dimensional modifications of the turbulence and terrain induced modifications of the turbulence spectrum are neglected GH WINDFARMER THEORY MANUAL 16 4 ESTIMATIONS OF DESIGN TURBULENCE 4 1 Introduction From the definition of the mean wind speed turbulence is defined as all wind fluctuations with periods less than the averagi
19. to reference data Removing seasonal bias in tab files 6 Uncertainty Analysis Measurement uncertainties Modeling uncertainties Wind speed variability uncertainties Sensitivity Combination of the uncertainties Exceedance levels 7 Noise Model 7 1 7 2 7 3 7 4 Introduction Simple Noise Model Complex Noise Models Turbine noise vs background noise 8 Electrical Calculations 8 1 8 2 8 3 8 4 8 5 Input data checks Electrical losses Reactive Power Formulae Caveats 9 Shadow Flicker 9 1 Introduction GH WINDFARMER THEORY MANUAL Program inputs Methodology of the shadow flicker calculation 9 4 Occurrence of shadow flicker 9 5 Modelling the rotor as a disc 9 6 Modelling the sun as a disc 9 7 Distance between rotor and turbine centre 9 8 Notes 10 Visual Influence 10 1 The terrain model 10 2 Line of sight algorithm 10 3 Standard ZVI for hub and for tip visibility 10 4 Vertical subtended angle 10 5 Horizontal subtended angle 10 6 Visibility of site 10 7 Field of view 10 8 Horizontal field of view 11 Summary References GH WINDFARMER THEORY MANUAL 1 INTRODUCTION The GH WindFarmer software package allows the user to design a wind farm to achieve maximum energy production within the geometric and environmental constraints of the site Through continuous development GH WindFarmer is at the forefront of wind farm technical analysis while meeting the changing needs of the industry GH WindFarmer allows the u
20. value of the turbulence standard deviation o Equation 4 2 5 shall be greater than or equal to oer at hub height wind speeds Vrw between 0 2 Ver and 0 4 Ver Of the turbine with Ver the reference wind speed of the turbine class defined in the standard When the turbine properties are known it is adequate to verify this for hub height wind speeds between 0 6 rated wind speed and cut out wind speed 4 2 3 DIBt The calculation method defined in 4 3 must be read in conjunction with the DIBt Standard 4 4 In the Standard 4 4 the design turbulence is defined as a Characteristic value In general the maximum allowed turbulence that is defined according to the turbine class is equivalent to the definition in IEC 61400 1 Edition 2 The calculation method for the site specific design equivalent turbulence intensity is defined in 4 4 similarly to the method used in the GH WindFarmer calculation according to IEC 61400 1 Edition 2 But contrary to the IEC 61400 1 Edition 2 calculation the thrust coefficient is not replaced by the approximation 7m s Vpub as proposed in 4 3 The site specific effective turbulence is defined by v hub i 2 Off ij zZ Di tO char ij 4 2 10 where Ct Vhup i is the turbine thrust coefficient of the upwind turbine at wind speed Vnu GH WINDFARMER THEORY MANUAL 22 The general calculation methodology is equivalent to the methodology described for IEC 61400 1 Edition 2 4 3 Large wind farm t
21. wake of another turbine the initial wake velocity deficit is corrected from the incident rotor wind speed to the free stream wind speed This correction is necessary in order to ensure that at distances far downstream the wake wind speed will recover to the free stream value rather than that incident on the rotor Therefore the initial centre line velocity Uwi is scaled by the ratio of the average influx velocity U and the free upstream wind velocity according to the following formula Uw U0 U Uwi 3 1 3 To combine the wakes of two wind turbines onto a third turbine the overall wake effect is taken as the largest wind speed deficit and other smaller wake effects are neglected This methodology is based on the results of the assessment of measured data from a number of wind farms 3 2 Eddy Viscosity model The Eddy Viscosity wake model is a CFD calculation representing the development of the velocity deficit field using a finite difference solution of the thin shear layer equation of the Navier Stokes equations in axi symmetric co ordinates The Eddy Viscosity model automatically observes the conservation of mass and momentum in the wake An eddy viscosity averaged across each downstream wake section is used to relate the shear stress to gradients of velocity deficit The mean field can be obtained by a linear superposition of the wake deficit field and the incident wind flow Below is an explanation of how this theory has been adapted
22. wind speed component Leet is the mean value of the turbulence intensity at 15 m s 5 6 m s Vhub is the wind speed at hub height in m s The parameter J is specified in the standard according to the turbulence level of the turbine class For low turbulence is 0 12 for medium turbulence it is 0 14 and for high turbulence 0 16 The wake effects from neighbouring turbines can be taken into account by an effective turbulence which according to the standard may be derived using a Frandsen method The Effective turbulence estimate is defined as l m Oeog Vrbi og 4 2 6 7 where P is the probability of the wind direction bin j In GH WindFarmer the predicted turbine specific probability is used m is the W hler coefficient The W hler exponent is specific to component material and geometry and is derived from the slope of the log log plot of the S N curve magnitude of a cyclical stress S against the cycles to failure N where the relation S Nis assumed Typical values for wind turbines range between 3 and 15 where 4 is appropriate for simple steel components and values 10 to 15 are suitable for simple composite components The turbine manufacturer should be consulted for advice on an acceptable range of W hler indices for these calculations Within GH WindFarmer the site specific Effective turbulence estimate derived from Frandsen is expressed as 0 9 Vabi 2 4 2 7 oe 0 2 ff i j 2 i j
23. Angle is the angular displacement of the sun west or east of the local meridian due to the rotation of the earth on its axis at 15 per hour In order to obtain the hour angle the program first calculates GH WINDFARMER THEORY MANUAL 49 9 3 1 Julian Date Julian Date JD is defined as the difference in days between the current Julian day and the Julian day at noon 1 January 2000 hour JD 2432916 5 365 delta leap day where delta year 1949 EI leap int Pa where int is defined as the integer portion of the argument 9 3 2 Elliptic coordinates The elliptic coordinates are the mean longitude L mean anomaly g ecliptic longitude I and the obliquity of the ecliptic ep These parameters are calculated using the equations n JD 2451545 0 L 280 460 0 9856474 n 0 lt L lt 360 g 357 528 0 9856003 n 0 lt g lt 360 l L 1 915 sin g 0 020 sin 2g 0 lt 1 lt 360 ep 23 439 0 0000004 n 9 3 3 Celestial coordinates For the calculation of the celestial coordinates right ascension ra and declination dec GH WindFarmer uses the following equations sin cos tan ra cos ep sin dec sin ep sin The next step is to calculate the azimuth angle az and elevation el To calculate az and el it is necessary to calculate the hour angle for which we must first calculate the following terms 9 3 4 Greenwich mean sidereal tim
24. ERL Memorandum TPRD L AP 0081 M83 1983 Sanderhoff P PARK User s Guide RIS 1993 Jensen N A note on wind generator interaction RIS M 2411 1983 Katic I Hostrup J and Jensen N A simple model for cluster efficiency EWEC 86 Tennekes H and Lumley J A first course in turbulence MIT Press 1980 Prandtl L Bemerkungen zur Theorie der freien Turbulenz ZAMM 22 5 1942 Ainslie J F Calculating the flowfield in the wake of wind turbines Journal of Wind Engineering and Industrial Aerodynamics Vol 27 1988 Taylor G J Wake Measurements on the Nibe Wind Turbines in Denmark National Power ETSU WN 5020 1990 Press W H et al Numerical Recipes in FORTRAN 2 Edition Cambridge University Press 1992 Quarton D C and Ainslie J F Turbulence in Wind Turbine Wakes J Wind Eng Vol 14 No 1 1990 Vermeulen P and Builtjes P Mathematical Modelling of Wake Interaction in Wind Turbine Arrays Part 1 report TNO 81 01473 1981 Vermeulen P and Vijge J Mathematical Modelling of Wake Interaction in Wind Turbine Arrays Part2 report TNO 81 02834 1981 Hassan U A Wind Tunnel Investigation of the Wake Structure within Small Wind Turbine Farms Department of Energy E 5A CON 5113 1890 1992 Smith G et al Advanced wake model for very closely spaced turbines EWEC conference Athens 2006 Barthelmie R et al Efficient development of offshore wind fa
25. GH WindFarmer Wind Farm Design Software a Zz lt gt be LLI 4 _ ig Wind Energy Experts worldwide DISCLAIMER Garrad Hassan and Partners Ltd accepts no liability for any loss or consequential damage arising directly or indirectly from the use of its products This document is subject to change without notice COPYRIGHT All rights reserved Duplications of this document in any form are not allowed unless agreed in writing by Garrad Hassan and Partners Ltd 2009 Garrad Hassan and Partners Ltd Garrad Hassan and Partners Ltd St Vincent s Works Silverthorne Lane Bristol BS2 0QD England www garradhassan com windfarmer garradhassan com GH WindFarmer Theory Manual May 2009 Contents 1 Introduction 2 Energy Calculation Program inputs Association Method Net Yield without losses Net Yield including losses Methodology of the energy calculation Directional correction 3 Wake Models Modified PARK model Eddy Viscosity model Turbulence intensity Eddy Viscosity wake model for closely spaced turbines Correction for large wind farms Terrain modification of the wake and turbulence 4 Estimations of Design Turbulence Introduction Design Standards Large wind farm turbulence correction Simple and Advanced Approach Wind speed and sector management Evaluation 5 MCP Methodology Correlation methods Calculation of direction shifts Application of speed ups and direction shifts
26. Yield including losses is the product of the Ideal Energy and all efficiencies GH WINDFARMER THEORY MANUAL 4 2 5 Methodology of the energy calculation Before performing an energy calculation the program determines the topographic speed ups over the site Using the Association Method this is the ratio of the wind speed at each grid point in the WAsP wind resource grid file to the wind speed at the reference location The topographic speed up is determined separately for each of the direction sectors in the wind resource grid file The speed up factors are then applied to the measured wind speed and direction frequency distribution table that is assumed to be representative for the site This method avoids errors due to fitting a Weibull distribution to measured data The program considers each wind direction sector in turn and each wind speed bin individually For each wind direction the program determines the topographic speed up of wind speed for the grid points nearest each turbine location This speed up factor is assumed constant for all wind speeds considered The wake effect of each turbine on the others is calculated for each wind speed step as the wake effect varies with wind speed The methodology for modelling the wake of each turbine is detailed in Section 3 on Wake Models The first step in this process is to calculate the wind speed and turbulence intensity incident upon the turbine If a turbine is in more than one wake the over
27. a tick box or other device for each point to allow the user to specify whether reactive power is imported or exported default is import Section 8 4 presents the expression to calculate the reactive power from the active power and power factor GH WINDFARMER THEORY MANUAL 43 8 3 2 Reactive power cables and lines The reactive power produced and consumed by each cable and line is calculated as follows e Calculate the series reactance per phase X ohms for each cable and line from specified data and calculated length e Calculate the capacitance per phase C microfarads for each cable and line from specified data and calculated length o Asa function of site wind speed and direction o calculate the current in the cable or line See Section 8 4 1 o calculate the reactive power consumption Q due to the series reactance X in the cable or line Q 3 1 X VAr this will be a negative number o calculate the reactive power generation Q due to the capacitance C in the cable or line Qe U 211 f C 10 VAr this will be a positive number and is constant for each cable or line o Add Q and Q to get total reactive power produced or consumed by the cable or line 8 3 3 Reactive power transformers The reactive power consumed by each transformer is calculated as follows e Calculate the transformer series reactance per phase X ohms for each transformer See Section 8 4 4 e Asa function of site wind speed and dire
28. all wake effect is taken as the largest wind speed deficit and other smaller wake effects are neglected This methodology is based on the results of the assessment of measured data from a number of wind farms The wind speed incident on each turbine is therefore the combination of the topographic speed up and the wake loss This incident wind speed can then be used to determine the power output A power look up table is created where the power output of each turbine is stored for each reference wind direction and wind speed considered The energy output is calculated as the sum product of the reference wind speed and direction frequency distribution table and the power output look up table 2 6 Directional correction The direction of the wind flow is changed when passing over terrain The flow modification is usually calculated by the wind flow model and represented in the wind resource file GH WindFarmer allows optional use of this shift in direction Directional correction factors are obtained with the same methodology as the non directional speed up factors By using this function the energy results can be improved in areas where the measured distribution is no longer an adequate representation of the site conditions Users are however advised to use this function with care as sparsely occupied directions sectors may lead to unrealistic results GH WINDFARMER THEORY MANUAL 5 3 WAKE MODELS The calculation of wake effects employs a systemat
29. being completely in the wake The ambient turbulence intensity is best derived from measurements Alternatively GH WindFarmer can predict the turbulence intensity from an input surface roughness length which is representative of the site using 3 3 1 I gt 3 3 3 P In h z The turbulence intensity is defined here as the quotient of standard deviation and mean wind speed at high wind speeds 3 4 Eddy Viscosity wake model for closely spaced turbines It is known that where turbines are closely spaced within a row the wake losses of the subsequent downwind turbines are significantly underpredicted by industry standard wake models including the Modified PARK and standard Eddy Viscosity models in GH WindFarmer This underprediction occurs when the turbine spacing within row is around 2D or less and where the turbines are placed in rows perpendicular to the prevailing wind directions This type of layout is usually only chosen where there is a uni or bi directional wind regime Garrad Hassan has developed a modification of the Eddy Viscosity model specifically for turbine arrays with these characteristics which has been validated with data from several wind farms 3 12 For closely spaced turbines the following alterations are made to the Eddy Viscosity model e The velocity deficit is allowed to add up cumulatively e The added turbulence is reduced in the wake e The Gaussian profile is replaced by a blunter profile If the
30. ction o calculate the current in the transformer See Section 8 4 2 o calculate the reactive power consumption Q due to the series reactance X Q 3 1 X VAr this shall be a negative number o calculate the reactive power consumption Qm due to the magnetising reactance Xm Qn U X m VAr this will be a negative number and is constant for each transformer type Note that U is the nominal secondary voltage of the transformer See Section 8 4 5 to see how to obtain Xm from lec o Add Q and Q to get total reactive power consumed by the transformer 8 3 4 Total reactive power import and export As a function of site wind speed and direction the net production or consumption at each metering point is calculated This calculation takes into account the reactive power produced and consumed by e each turbine e each cable and line e each transformer Then by summating the above amounts for the whole year the outputs specified in the User Manual are given GH WINDFARMER THEORY MANUAL 44 8 3 5 Power factor correction requirements If this option is chosen for each PFCD the following steps are carried out As a function of site wind speed and direction e Calculate the active P and reactive Q power produced and consumed by all the downstream components at the metering node e Calculate the PFCD size Qpfcep Ppp tan arccos cos Opr o positive number for capacitor bank exporting kVAr o negative
31. ction 8 4 3 e Asa function of site wind speed and direction o calculate the current in the transformer See Section 8 4 2 o calculate the loss L in the transformer L 3 1 R Watts e Check for transformer overloading i e compare if the highest value of current in the transformer exceeds the nominal current rating of the transformer The nominal current rating of the transformer Inom is as calculated in Section 8 4 3 If there are N transformers in parallel at the substation current and losses are split into these elements Section 8 4 6 shows the methodology to obtain the currents for each transformer l4 l2 In from the total current and the impedances Z Zo Zy 8 2 4 Total active power import and export As a function of site wind speed and direction the wind farm net production or consumption at each metering point is calculated This calculation takes into account e production of each turbine e cable and line losses e transformer losses Then by summating the above amounts for the whole year the outputs specified in the User Manual are given 8 3 Reactive Power 8 3 1 Reactive power turbines The reactive power produced and consumed by each turbine as a function of wind speed and direction is calculated The relationship between active and reactive power can ideally be entered by the user as either of e a table of reactive versus active power e a table of power factor versus active power with
32. ds Values used in the standards are e Characteristic value mean turbulence intensity plus one standard deviation This represents the 84 quantile of a normal distribution which means that 84 of the data are smaller than or equal to the characteristic value e Representative value mean turbulence intensity plus 1 28 times the standard deviation This is the 90 quantile of a normal distribution which means that 90 of the data are smaller than or equal to the representative value The Mean Characteristic and Representative turbulence intensities are marked in the figure below for specific wind speed bins D7 aS tet aa nn ET av AT ee u e se 7 t wap E SR S eer al t a aeann 50 ae 5 See tee E En u EEE un eit tate 50 s amp Sate tS e e i sj om 5 2 1HE 431 2 N Sat ne A 5 ap mna a EES ee Si VA 0 Res a 8 RENTE Ten 2 e i i e Fr 4 s NCADP ee eater El te Si BEE pS OF a OR CREA a 3 ERBEN ae Stee rear see we on pet at 0 0 5 10 15 2 5 30 wind speed mis Mean red Characteristic green and Representative blue turbulence intensities of turbulence measurements 4 2 Design Standards The assessment of structural integrity of wind turbines is done by comparison of the site specific wind parameters at the wind turbine site with those used for the design There are different design standards that define in different ways the wind conditi
33. e a neighbouring wind farm included in the analysis in order to represent wake effects correctly In this case the user could leave those turbines without connection or use another root node for that wind farm if required 8 2 8 2 1 Electrical losses Cable and line losses The annual loss in each cable and line is calculated as follows Calculate the resistance R ohms per phase for each cable and line from specified data and calculated length As a function of site wind speed and direction o calculate the current in the cable or line See Section 8 4 1 o calculate the loss L in the cable or line L 3 R Watts Check for cable and line overloading i e compare if the highest value of current exceeds the maximum continuous current rating of the cable or line if specified by the user If there are N parallel cables lines between two nodes current and losses are split into these elements Section 8 4 6 shows the calculation to obtain the currents for each cable line l4 l2 In from the total current and the impedances Z Zo Zy GH WINDFARMER THEORY MANUAL 42 8 2 2 Transformer no load losses For each transformer the annual no load loss is the stated no load loss figure Watts multiplied by the number of hours per year 8 2 3 Transformer load losses The annual loss in each transformer is calculated as follows e Calculate the resistance R ohms per phase referred to the secondary side See Se
34. e gmst The formula to approximate the Greenwich mean sidereal time to an arbitrary time is given by gmst 6 697375 0 0657098242 n hour UT 0 lt gmst lt 24 h GH WINDFARMER THEORY MANUAL 50 9 3 5 Local mean sidereal time Imst To calculate the local mean sidereal time Imst from a given gmst we just have to add the east longitude to the gmst east longitude Imst gmst 15 9 3 6 Hour angle Then the Hour Angle ha can be calculated using ha Imst ra 12 lt ha lt 12h The Hour Angle is defined as negative before the sun reaches the meridian and positive when has already reached the postmeridian hemisphere 9 3 7 Azimuth and Elevation Finally to calculate the parameters that define the solar position which are the azimuth az and the elevation el GH WindFarmer use the following equations sin el sin dec sin lat cos dec cos lat cos ha and then the calculation for azimuth angle which is measured from North 0 sin ha sin az cos dec 0 lt az lt 360 cos el 9 4 Occurrence of shadow flicker The occurrence of the shadow flicker is determined by the wind turbine position point P and sun position elevation angle and azimuth angle The program calculates from these the minimum distance from the wind turbine hub to any point S on the line between the sun and the point of interest A GH WINDFARMER THEORY MANUAL 51 Shadow flicker calculation
35. e W hler coefficient is specific to component material and geometry and is derived from the slope of the log log plot of the S N curve magnitude of a cyclical stress S against the cycles to failure N where the relation S Nis assumed Typical values for wind GH WINDFARMER THEORY MANUAL 19 turbines range between 3 and 15 where 4 is appropriate for simple steel components and values 10 to 15 are suitable for simple composite components The turbine manufacturer should be consulted for advice on an acceptable range of Wohler indices for these calculations The estimate of a site specific effective turbulence can be expressed following 4 3 as 2 Oy N Fo 4 2 3 Pii YA5 0 3 d PWV where dj is the distance to the neighbouring turbine that causes the wake in direction bin j normalised by the rotor diameter Ocharij is the characteristic ambient wind speed standard deviation in wind direction bin j and wind speed bin i For wind directions where there are no upstream turbines or turbine distances greater than 10 rotor diameters Ogrij Ocharij With these definitions Oo Vis i represents the Characteristic value of the wind speed standard deviation The estimate of the Effective turbulence is calculated in GH WindFarmer for all wind speeds and directions and is available as output in the Flow and Performance Matrix The characteristic wind speed standard deviation is defined as O hari j T Oj 04 j 4 2 4 where
36. e hub height without any losses B Gross Yield All turbines at their true locations with topographic speedups relative to the mast and without any allowance for wake losses C Net Yield without losses All turbines with the topographic speed ups and calculation of wake losses The Net Yield without losses Calculation C is equivalent to the ideal output Calculation A multiplied by the two following efficiencies e Topographic efficiency Calculation B Calculation A e Array efficiency Calculation C Calculation B The program calculates the net energy output array and topographic efficiency for each individual turbine and the wind farm as a whole To calculate the net energy production of each wind turbine only calculation C is required Calculations A and B are used to estimate the wake and topographic effects experienced by each turbine 2 4 Net Yield including losses The following additional losses are calculated by GH WindFarmer and expressed in the form of efficiencies Sector management includes losses due to the shut down of the turbines and its impact on the wake effects in the wind farm Electrical efficiency requires the Electrical Module Additional efficiencies can be entered manually by the user For each new loss the efficiency is derived by carrying out a yield calculation taking the loss factor into account and dividing the original energy yield by the new energy yield Following this procedure the final Net
37. e operational These uncertainties can be either calculated or estimated Ideally the wind farm development should aim to minimise the overall uncertainty 6 1 Measurement uncertainties Anemometry uncertainties e Steady state calibration uncertainty is associated to the assessment of the transfer function frequency Hz to wind speed in wind tunnel This uncertainty is usually expressed in the calibration certificate as of wind speed or as standard deviation m s per each speed bin e Variability of calibration uncertainty is associated to the variation of the transfer function with time Weather conditions affect the transfer function of the sensor during the measurement campaign e Flow inclination uncertainty is expected when the axis of rotation of the anemometer is not perfectly vertical e Flow distortion uncertainty is caused by the proximity of the anemometer to the supporting mast and horizontal vertical booms The closer the anemometer is to these elements and higher is the uncertainty caused by flow distortion The anemometry uncertainties above are usually combined into one uncertainty value The correlation uncertainty is determined by the scatter of the correlation between reference mast and site mast The lower the scatter the lower is the correlation uncertainty Note that there is no simple relationship between the correlation coefficient see chapter on MCP methodology and the correlation uncertainty Nevertheless some
38. e wind speed is perturbed by a percentage 3 and the resulting perturbed energy calculated The normalised relative increase is the wind speed sensitivity in percent e Topographic the sensitivity is 100 Topographic Efficiency e Wake Loss the sensitivity is 100 Array Efficiency e Energy Yield Uncertainties defined with respect to the energy yield have a sensitivity of 100 in net yield e User Defined Sensitivity is a manual input derived outside WindFarmer GH WINDFARMER THEORY MANUAL 33 6 5 Combination of the uncertainties Once the uncertainties have been converted with appropriate sensitivity to energy yield uncertainties they are combined assuming statistical independence to form the Total Uncertainty U ror le gA where 1 2 2 2 u2 Uror Total Uncertainty Uup Historical Period uncertainties measurements modelling and wind variability over the reference period of the measurement campaign Urp Future Period uncertainty over reference periods when the projected wind farm is operational 6 6 Exceedance levels The estimated annual Net Energy Production and the Total Uncertainty determine respectively the mean and the standard deviation of a Normal Gaussian Distribution The absolute standard deviation is obtained by multiplying the Total Uncertainty by the Net Energy Production In general the probability distribution of the energy production E when assuming a Gaussian Distribution is
39. ear model used to extrapolate the frequency distribution from mast height to hub height at the mast position This should be accounted for if it is not already included in the topographic model uncertainty Horizontal extrapolation uncertainty This is usually an alternative to the topographic uncertainty if the latter is split between vertical and horizontal extrapolation uncertainty It relates to the extrapolation of the wind regime at turbine position from the wind regime predicted at mast position at hub height Power curve uncertainty The uncertainty of the power curves used in the energy prediction Uncertainties can be measured uncertainties uncertainties due to the measurement setup model uncertainties if a calculated power curve is used or can represent tolerances between different turbines of the same type Further to the list above other uncertainties can be identified for a specific project For instance wind vane misalignments may lead to wind rose uncertainties wind indexes also carry uncertainties when used in extrapolating long term site wind regime In this regard for each project the modeling uncertainty list and the relevant uncertainty values have to be individually assessed to reflect the specific issues GH WINDFARMER THEORY MANUAL 32 6 3 Wind speed variability uncertainties Wind speed is stochastically variable over the historical period when the measurement campaign is carried out and over the future period
40. eating the wake The velocity profile across the turbine rotor at the hub height is integrated to produce a mean wind speed incident across the rotor at the hub height This is assumed to represent the incident wind speed across the whole rotor disc 3 3 Turbulence intensity The Eddy Viscosity model relies on a value of incident ambient turbulence intensity for equations 3 2 6 and 3 2 7 For a turbine in the free wind stream the calculation must be initiated using the ambient turbulence level For a turbine within a wind farm it is necessary to calculate the increased turbulence level which results from the presence of upstream turbines Wind farm turbulence levels are calculated using an empirical characterisation developed by Quarton and Ainslie 3 8 This characterisation enables the added turbulence in the wake to GH WINDFARMER THEORY MANUAL 10 be defined as a function of ambient turbulence the turbine thrust coefficient the distance downstream from the rotor plane and the length of the near wake eA a else 3 3 1 where x is the calculated length of the near wake using the method proposed in 3 9 3 10 The characterisation was subsequently amended slightly to improve the prediction as shown below 3 11 Laa SS Le ele 3 3 2 Using the value of added turbulence and the incident ambient turbulence the turbulence intensity at any turbine position in the wake can be calculated The model also accounts for the turbine not
41. eed and direction frequency distributions at planned wind farm sites from short time measurements at the wind farm site and long term measurements at a nearby reference station In these processes concurrent time series at both locations are compared and then relationships between the two are determined These will comprise wind speed relationships and possibly direction shifts These relationships are then applied to the long term reference measurement to obtain a long term frequency distribution at the site This derived long term frequency distribution will be representative of the period of the reference measurements 5 1 Correlation methods In the MCP Module of GH WindFarmer linear correlation methods are used to derive the relationships between reference and site wind speed measurements It is assumed that for each direction sector the pairs of wind speed data can be fitted to a straight line that is characterised by its slope This slope describes the speed up between the two locations To calculate the slope either a least squares method or a PCA method can be selected in GH WindFarmer It is assumed that winds at the site and the reference station are being driven by the same meteorological systems This implies that when it is calm at the site it is also calm at the reference station Because of this assumption the linear fit is forced through the origin with both fitting methods The general linear relationship between site and ref
42. en Einwirkungen und Standsicherheitsnachweise f r Turm und Gr ndung Schriften des Deutschen Instituts f r Bautechnik Reihe B Heft 8 2004 International Standard ISO International Organisation for Standardization 9613 2 Acoustics Attenuation of sound during propagation outdoors Part 2 General method of calculation International Standard ISO International Organisation for Standardization 9613 1 Acoustics Attenuation of sound during propagation outdoors Part 1 Calculation of the absorption of sound by the atmosphere Klug H Mellert V Radeke U Zuverl ssigkeit von Ger uschimmissionsprognosen bei gewerblichen Anlagen Forschungsbericht 89 105 02 702 Umweltforschungsplan des Bundesministeriums f r Umwelt Naturschutz und Reaktorsicherheit 1989 Seidelmann K ed Explanatory supplement to the Astronomical Almanac 1992 Michalsky JJ The Astronomical Almanac s Algorithm for Approximate Solar Position 1950 2050 Solar Energy 40 3 pp 227 235 1988 GH WINDFARMER THEORY MANUAL 60
43. erence wind speed is therefore defined as y b x where x is the wind speed at the reference station b is the slope of the linear fit also called speed up and y is the concurrent wind speed at the site The correlation coefficient indicates the strength and direction of a linear relationship between two random variables It ranges from 1 to 1 A correlation coefficient of 1 means that there is a perfect positive linear relationship between variables if the reference variable increases also the dependent variable increases A correlation coefficient of 1 means that there is a perfect negative linear relationship between variables if the reference variable increases also the dependent variable decreases A correlation of 0 means there is no linear relationship between the two variables The directional and overall correlation coefficients are given in the speed ups table in the Correlation window R The linear correlation coefficient is defined as ye x y 7 R i EGF L0H 5 1 1 Least Squares method Assume at a specific time the measured wind speed is x at the reference station and y at the site For a number n of such points the least square method derives the linear fit that minimises the sum of the squared distances in y from the fit as illustrated in the figure below GH WINDFARMER THEORY MANUAL 27 This gives dyr Eo bx OEE en I b x 0 5 1 1 b b ob L L L Least Squares method This lead
44. for use within GH WindFarmer A list summarising the notation used in this explanation can be found on at the end of this document An illustration of the wake profile used in the Eddy Viscosity model is shown in the figure below Centre Line Turbine Incident rotor ind d a Wake wind speed profile Wake profile used in the Eddy Viscosity model GH WINDFARMER THEORY MANUAL 8 3 2 1 Model equations The Navier Stokes equations with Reynolds stresses and the viscous terms dropped gives 3 3 U oU V oU 2 1 ruv Ox Or r of 3 2 1 The turbulent viscosity concept is used to describe the shear stresses with an eddy viscosity defined by 3 4 ex La x U x 3 2 2 oU and uv 8 3 2 3 or Lm and Um are suitable length and velocity scales of the turbulence as a function of the downstream distance x but independent of r The length scale is taken as proportional to the wake width Bw and the velocity scale is proportional to the difference U U across the shear layer Equation 3 2 3 permits the shear stress terms uv to be expressed in terms of the eddy viscosity The governing differential equation to be solved becomes py ee 3 2 4 Ox o r or When considering the ambient wind flow for a wind farm it must be considered as turbulent Therefore the eddy viscosity in the wake cannot be wholly described by the shear contribution alone but an ambient term is included Hence the
45. he central estimate of the predicted energy production z is equal to 0 P z 99 2 326 95 1 645 90 1 282 85 1 036 84 1 000 80 0 842 75 0 674 50 0 25 0 674 10 1 282 1 2 326 GH WINDFARMER THEORY MANUAL 35 7 NOISE MODEL 7 1 Introduction The noise functions in the Base Module allow the user to input the sound power levels for the turbines and calculate the noise distribution The aim of the functions is to design wind farms within legal limits based on giving the information about sound power levels of the turbines There are three noise models available within GH WindFarmer All three are based on ISO 9613 2 5 1 e Simple noise model e Complex ISO9613 General e Complex ISO9613 Alternative The Simple noise model calculates the attenuation for a single representative frequency and assumes hard ground surfaces The model Complex ISO9613 General considers noise attenuation for separate octave bands and includes the effect of ground attenuation as well as directional meteorological effects The model Complex ISO9613 Alternative does not assume frequency dependent ground attenuation Additionally a Custom model has been implemented where all parameters are freely selectable Equations from the ISO standard are referred to below as ISO x where x refers to the equation number in the standard 7 2 Simple Noise Model The Simple noise model is recommended for a first fast assessment or if no frequency
46. he rotor The figure below outlines the flow field used by the model U Turbine D 2kx rotor X Flow field used in the Modified PARK model to calculate the wind turbine output 3 1 2 Wake development The downstream wind speed is calculated using the following formula 3 1 D 2 U uf iG SP 3 1 1 Here k is the wake decay constant that is defined by the following expression 3 2 A kie 3 1 2 In h z where A is a constant which is equal to 0 5 The nature of the program means that k is set at the same value for all wind directions This assumes that there is no significant variation in surface roughness length over the site and the surrounding area 3 1 3 Wake superposition For each turbine downstream of the turbine under consideration the program determines the axial displacement assuming rotational symmetry of the wake The wake width and the wind speed at this displacement are then calculated The turbines affected by the wake may not be totally in the wake so the percentage cover of the turbine s rotor in the wake is determined If the whole rotor is within the wake then the turbine wind speed is set as Uw If some of the GH WINDFARMER THEORY MANUAL 7 rotor is outside the wake the wind speed at the turbine is the sum of Uy and the upstream velocity of the turbine creating the wake multiplied by the relative percentages of rotor cover If the turbine under consideration is in the
47. he standard ZVI answering the question How large are the turbines in my field of view 10 5 Horizontal subtended angle Tip Hub horizontal subtended angle A Horizontal subtended angle GH WINDFARMER THEORY MANUAL 55 The horizontal subtended angle is used in a similar way to the vertical subtended angle It expresses the angle of horizontal view which includes turbines 10 6 Visibility of site Tip Hub Visibility of site The site visibility is an expression for the percentage of the whole wind farm which is visible from a location 10 7 Field of view The Visualisation Module in GH WindFarmer allows you to create visualisations and photomontages of wind farms The view in the Visualisation Window can be presented as it would be on a photograph when using the Fit to Film option This requires the correct definition of width and height of the film as well as the focal length of the camera From these data the field of view and the horizontal field of view are calculated fim width film helqlir fiald of view The field of view is defined as film width film height Field of view degrees 2 arctan 2 focal length GH WINDFARMER THEORY MANUAL 56 10 8 Horizontal field of view film width film horizontal field of view height The field of view is defined as film width film height Field of view degrees 2 acta z 3 2 focal length GH WINDFARMER
48. ic absorption is calculated for each octave the same way as described above for the simple noise model The attenuation coefficients can be set by the user and are a function of frequency humidity and temperature This dependency is described in detail in ISO 961 3 1 6 2 The set default attenuation coefficients in GH WindFarmer are valid for 10 Celsius and 70 humidity and represent a conservative choice GH WINDFARMER THEORY MANUAL 38 7 3 3 Ground Attenuation Agr The ground attenuation considers the sound reflected or absorbed by the ground surface ISO 9613 2 divides the path where the sound propagation is affected into three regions the source region the receiver region and a middle region The acoustic properties of each of these regions are taken into account separately through ground factors G The table below defines the ground factors G for three different kinds of surfaces Hard Ground Low porosity surfaces Paving water ice concrete Porous Ground Porous surfaces suitable for 1 growth of vegetation ground covered with grass trees and vegetation Mixed Ground Both hard and porous ground Between 0 and 1 The total ground attenuation for the octave band is calculated as the sum of individual absorption coefficients for the source region A the receiver region A and the middle region An Age Ast Ar Am 6 3 2 where As Ar and Am are calculated as a function of G by GH WindFarmer using the
49. ic approach where each turbine is considered in turn in order of increasing axial displacement downstream By this method the first turbine considered is not subjected to wake effects The first turbine s incident wind speed the thrust coefficient and the tip speed ratio are calculated Its wake is then modelled as described below and the parameters which describe its wake are stored The effect of this wake on all the turbines downstream can then be modelled If any of the downstream turbines fall within the wake of this turbine the velocity and turbulence incident on these turbines solely due to this upstream turbine being considered can be determined As the calculation progresses through the turbines the incident wind speed on the turbine is the sum of the wake and topographic effects as described in Section 2 There are two wake models available within GH WindFarmer e Modified PARK model based on the method presented by Jensen and Katic 1 3 e Eddy Viscosity model based on work conducted by Ainslie 1 1 1 2 Due to the complexity of the wake directly behind the rotor all models are initiated from two diameters downstream This is assumed to be the distance where pressure gradients no longer dominate the flow If a turbine is within this limit the program resets the axial distance offset to a value of two rotor diameters The input parameters to the two models are listed in the table below For a comparison of the model results w
50. ime step the difference is calculated between the direction measured at the site and the middle of the direction bin the reference direction belongs to For example if the reference wind direction is 32 degrees it will be binned in the direction bin centred at 30 degrees If the concurrent site wind direction is 35 degrees the resulting direction shift is therefore calculated to be 35 30 5 degrees For all reference wind directions within a GH WINDFARMER THEORY MANUAL 29 direction sector the mean of these direction differences is calculated and is output in the Correlation window as the Calculated Site Offset 5 3 Application of speed ups and direction shifts to reference data In the prediction phase of the MCP method the correlation relationships can be applied to either the long term time series or the long term wind speed and direction frequency distribution tab file measured at the reference station 5 3 1 Applying speed ups and direction shifts to long term reference time series If a long term time series at the reference mast is used the directional speed ups and optional direction shifts are applied directly to each line of the reference time series according to the measured wind direction at the reference mast The adjusted time series is then used for the creation of the long term representative tab file This tab file can then be used in flow models to calculate the long term representative wind flow over the wi
51. in the source region As the middle region Am and the receiver region A In the source and receiver regions the ground attenuation is 1 5dB The ground attenuation in the middle region is given by A 3q dB 30 h h i when d gt 30 h h 6 2 6 q p 0 when d lt 30 h h where hs is the hub height of the wind turbine h is the height of the receiver above ground dp is the distance from the wind turbine base to the receiver base projected onto the ground plane 7 3 Complex Noise Models The complex noise model in GH WindFarmer considers the noise attenuation as a function of the frequency distribution of the noise The noise emission of the turbine s needs to be defined in octave bands Frequency specific attenuation coefficients are then used to calculate the attenuation of the noise 7 3 1 Noise in Octave Bands GH WindFarmer calculates the noise propagation by summing the contributing sound pressures for each source and for each octave band The next equation replaces Equation 6 2 3 and shows the method of adding octave bands Dron Dal 6 3 1 j l i l where n is the number of sources i indicates the eight standard octave band frequencies 53 Hz to 8KHz Let is the octave band sound pressure level Note GH WindFarmer does not perform any A weighting If you need your result in dB A you need to enter the A weighted sound pressure levels 7 3 2 Atmospheric Attenuation Azim The atmospher
52. ith that of real wind farms please refer to the Validation Manual MEERE Default Description Wake expansion factor k Surface roughness length Zo Eddy Viscosity Describes the rate of the assumed linear expansion of the wake Increase this value for high turbulence situation decrease for offshore Formula 3 1 2 is used to obtain the wake expansion factor from a surface roughness length Maximum length of wake Maximum allowable turbulence Wake Recovery Ambient turbulence intensity Maximum length of the wake to limit the calculation Turbines further away should be treated as sources of roughness Calculation will return a warning if a set incident turbulence level is exceeded the turbulence is calculated as average over all wind speed conditions A parameter used to speed up the calculation of the wake The calculation stops and assumes the wake has recovered at a percentage deviation of the incident wind speed The ambient turbulence intensity can be entered as function of wind speed and or direction GH WINDFARMER THEORY MANUAL 6 Summary of wake calculation inputs 3 1 Modified PARK model This is a simple model of the wake that is based on the same algorithm used in PARK 3 1 3 2 3 1 1 Model initialisation This two dimensional model uses the momentum theory to predict the initial profile assuming a rectangular wind speed profile and that the wake expands linearly behind t
53. ly direction shift Not ticked Shifts the directional probabilities following the to sector probability predictions of the flow model used Use with care Apply topography Ticked The wake of a wind turbine is modified by the correction to wake topography if ticked Older models calculate the wake model losses as they would be in flat terrain Site roughness This parameter can be used to calculate turbulence length m levels at hub height The equations used are only valid for flat smooth uniform terrain Maximum wind Maximum wind speed for which energy is calculated speed m s The mean wind speed calculation from a Weibull distribution requires a high value Number of direction Complex geometry of wind farms needs to be captured steps Do not use less than 72 steps on some sites more than 120 steps can improve accuracy of calculation Summary of energy calculation inputs 2 2 Association Method For the association method it is recommended that the following optional input files are used in addition e A single point WAsP wind resource grid file for the reference mast at measurement height and when using an RSF file an additional single point WRG file at hub height at the mast location e A joint wind speed and direction frequency distribution table for the reference mast location in the format of the WAsP table file TAB If these two files are not specified GH WindFarmer will use the information from the WAsP wind resou
54. model is used to determine the changes to the incident wind speeds at each turbine within a wind farm due to the effects of other turbines The accuracy of wake prediction has become increasingly important as larger wind farms are being developed and turbines are being placed closer together 2 1 Program inputs The following inputs to the GH WindFarmer model are required to produce an estimate of the wind farm energy production e A WASP wind resource grid WRG format file at the turbine hub height with extents covering all the intended turbine locations or a WAsP discrete resource RSF format file with wind speed results at individual turbine locations e Turbine locations as grid co ordinates e Turbine performance data which includes power thrust and rotor speed characteristics e Turbine dimensions specifically hub height and diameter The resource file is generated within WAsP and is a grid representation of the directional wind speed distribution at each point over the site Program parameters used in the energy yield calculation are listed in the table below GH WINDFARMER THEORY MANUAL 2 ENERGY Default Description CALCULATION value Site reference air The air density for each turbine is calculated from the density kg m3 site air density according to height above sea level ASL and the power curve used is adjusted Lapse rate 0 113 The Lapse rate describes the variation of the air density kg m3 km with height App
55. n the rotor is facing 180 degrees away from the sun s azimuth By taking into account the wind speed and direction frequency distribution a more likely occurrence of shadow flicker can be derived as compared with this geometric worst case GH WINDFARMER THEORY MANUAL 52 9 6 Modelling the sun as a disc If shadow flicker is calculated modelling the sun as a disc the vector b is pointing to the edge of the sun disc closest to the rotor disc The sun s diameter is assumed to be 0 351 degrees Please note that current standards and limits assume a point source and this calculation option while more accurately representing theoretical shadow flicker should not be selected if the result is to be compared with such a limit 9 7 Distance between rotor and turbine centre Usually the turbine positions are defined at the centre of the turbine tower For the shadow flicker calculation the position of the rotor is important as the rotor is the source of the shadow flicker The calculation is very sensitive against changes in the relative position of source and receptor the distance between the turbine rotor and the tower can become crucially important GH WindFarmer calculates the rotor offset using 1 O rotor offset O top diameter tower position disc depth Tower top diameter tower position and disc depth are defined in the 3D Designer in the Turbine Studio 9 8 Notes The shadow flicker calculation makes the following simplifica
56. nd farm site 5 3 2 Applying speed ups and direction shifts to long term tab files If a long term wind speed and direction distribution tab file is used the directional speed ups and optional direction shifts are applied to the reference tab file according to the wind direction in the tab file First the directional speed ups are applied according to the wind direction sector in the tab file For each sector this is done by applying the speed up to the wind speed bin boundaries and then re binning the probabilities to obtain new wind speed bins in steps of 1 m s After this has been done for all 12 direction sectors the direction shifts are applied by adding them to the respective direction bin boundaries and then re binning the probabilities to obtain 12 direction sectors in equal steps of 30 degrees 5 4 Removing seasonal bias in tab files When a tab file is created that is based on a time series which does not cover an integer number of complete years or when there are data missing in the time series the tab file will be biased according to the seasonal variation of wind speed This bias can be removed using the Remove Seasonal Bias function in GH WindFarmer This can be applied when creating a tab file directly from a time series or when applying the MCP method to the short term time series This function is not available when a long term reference tab file is used because tab files do not include seasonal
57. ng period One common measure of turbulence is the turbulence intensity defined as Oo 4 1 1 V where Oy standard deviation of wind speed vin the averaging period v mean wind speed in the averaging period The same notation is used here as in the relevant IEC Standards 4 1 4 2 Both values have to be determined from the same set of measured data samples of wind speed The typical averaging period is 10 min When a set of turbulence intensity measurements is taken over a period it is possible to plot all the single values of turbulence intensity as a function of the single measured mean wind speeds v The plot may look like this Point Data a 8 8 turbulence intensity 8 8 5 10 15 20 25 30 Measured turbulence intensities as a function of the mean wind speed over the averaging period For every wind speed bin a mean turbulence intensity is calculated from the single values With regard to turbine loads those single turbulence intensities greater than the mean value are more relevant To analyse these it is assumed that all the single values within a wind GH WINDFARMER THEORY MANUAL 17 speed bin have a normal distribution with a mean value and a standard deviation that can be calculated from the sample data in the specific bin In terms of turbine loads specific quantiles of the turbulence intensity distribution are used depending on the definition in the design standar
58. number for reactor bank importing kVAr The program should keep a record of the lowest and highest Qprcp values which will be part of the output as shown in the User Manual 8 4 Formulae 8 4 1 Calculation of the current in each cable or line e Identify each downstream turbine and PFCD i e each turbine and each PFCD whose output must flow through the cable or line in question before it reaches a root node e For each downstream turbine determine the power P produced or consumed by the turbine for the given hub height wind speed e For each downstream turbine calculate the reactive power Q produced or consumed by the turbine for the given hub height wind speed from the user defined table of Q or PF versus P e For each downstream PFCD calculate the reactive power Q produced by this element as shown in Section 8 3 e Summate to find total P and total Q passing through the cable or line from all downstream turbines e Calculate apparent power S P Q _ kVA e Calculate current for cable or line o 1 1000 3 U Amps o where U is the operating voltage 8 4 2 Calculation of the current in each transformer e Identify each downstream turbine and PFCD i e each turbine and each PFCD whose output must flow through the transformer in question before it reaches a metering point e For each downstream turbine determine the power P produced or consumed by the turbine for the given hub height wind speed
59. o is the mean of wind speed standard deviation in wind direction bin j and wind speed bin i O s the standard deviation of wind speed standard deviation in wind direction bin j and wind speed bin i sigma sigma fo is not available from measurements it can be assumed 4 3 that it is 20 of the oi j mean wind speed standard deviation so that 1 2 5 O char i j i j nn u ep aie The corresponding estimate of the Characteristic turbulence intensity as a V hub i function of wind speed is available in the Flow and Performance Matrix of GH WindFarmer To get a measure of the suitability of a wind turbine for the site a wind turbine should satisfy the condition that the estimated Characteristic value of the turbulence standard deviation o Equation 4 2 1 shall be greater than or equal to Ge at hub height wind speeds v7 in the operational range of the turbine 4 2 2 IEC 61400 1 Edition 3 In this Standard 4 1 the design turbulence is defined as a Representative value GH WINDFARMER THEORY MANUAL 20 The maximum allowed turbulence is defined according to the turbine class The expected value for the hub height turbulence intensity at a 10 min average wind speed of 15m s and the turbulence standard deviation c are linked through the following formula o Ving Le 0 75 Ving b 4 2 5 where O14 is in this case called the representative value of the standard deviation of the longitudinal
60. o the lower roughness offshore Onshore the effect of a large wind farm is less pronounced and to some degree masked by the higher roughness of the local terrain A large wind farm correction has been developed by GH 3 19 and 3 20 based on modelling the disturbance caused by each individual turbine This allows numerous wind farm layouts to be considered during the design phase of a wind farm layout The model approach can be described in three steps 1 Use the wind flow model and data of choice that best describe the ambient wind flow over the proposed wind farm site 2 Place the turbines in the wind flow and calculate the large wind farm correction to the ambient flow due to the presence of the turbines 3 Use a standard wake model with the corrected ambient wind speeds as boundary conditions to describe the inter turbine wake deficits The above second step details an empirical correction to describe the disturbance of the atmospheric flow caused by the wind farm This breaks with the traditional assumption that the wind flow can be treated as independent from the wind farm The large wind farm correction consists of the following components e Boundary layer modification Establishes the magnitude of correction to the ambient the wind profile behind each individual turbine e Wind turbine density Considers relative turbine positions and decides if a correction should be applied in a particular direction sector e Row to row
61. ons of the PARK model With k 0 087 equation 4 4 1 becomes 0 H2 arctan 10 4 4 2 2 S Also in 4 3 equation 3 18 a characteristic view angle is defined as 1 1 6 0 arctan 10 4 4 3 2 S Equation 4 4 2 is used in the advanced model while equation 4 4 3 is used in the simple model Both definitions of the view angle comply with the IEC61400 1 Ed 3 and are compared in the figure below k 0 087 GH WINDFARMER THEORY MANUAL 24 3 5 3 0 25 2 0 WSs NAN WY distance perpendicular to wind direction in rotor diameters 0 1 2 3 4 5 6 7 8 9 10 downwind distance in rotor diameters The area that a downstream turbine can be placed without being exposed to increased load due to the wake of an upstream turbine when using Equation 4 4 3 red and Equation 4 4 2 green The area shaded in grey represents the wake behind the upstream turbine The bold red and green lines represent the rotor edge In both cases it is assumed that the wake angle is measured downstream from the rotor centre of the turbine that generates the wake The figure illustrates the area where a downstream turbine can be placed without being exposed to increased load due to the wake of an upstream turbine It can be seen that when using equation 4 4 3 the downstream turbines can be placed so that half the rotor is within the wake without getting any increase in load whereas when using equation
62. ons that a wind turbine shall be designed to withstand In GH WindFarmer design turbulence estimates can be produced with respect to the following design standards e IEC 61400 1 1998 Edition 2 e IEC 61400 1 2005 Edition 3 e DIBt GH WINDFARMER THEORY MANUAL 18 The design turbulence definitions in these standards and the methods used in GH WindFarmer to calculate the site specific design turbulence are given in the following sections For calculation of design turbulence estimates the turbulence intensity values measured at the site masts are input in the Turbulence Intensity tab in the Project Properties window It is important not to input Characteristic or Representative turbulence values here Values for the standard deviation of the wind speed standard deviation known as sigma sigma can be calculated with the MCP Module of GH WindFarmer when time series of wind speed and wind speed standard deviation data are available The calculations by GH WindFarmer of Effective turbulence estimates are designed for use with the Modified PARK wake model The turbulence intensity outputs from the Eddy Viscosity wake model of GH WindFarmer are true meteorological turbulences They should not be compared directly with the IEC or DIBt turbulence models but it is recommended that they are used in conjunction with the GH WindFarmer Bladed Link and GH Bladed software tools to obtain the most accurate description of the design loads 4 2 1
63. onsiders two factors that attenuate sound e Attenuation due to foliage Ata e Attenuation due to areas of buildings or industrial sites Asie GH WindFarmer allows the user to input site specific attenuation Ayisc The attenuation due to the foliage of trees and bushes is usually small but in some cases when the foliage is dense close to the source of sound to the receiver or to both the attenuation could increase considerably It is suggested the values of Table A1 ISO 9613 2 are used to calculate the attenuation due to foliage To calculate the attenuation due to industrial sites use Table A2 ISO9613 2 This table is an approximation as this parameter has a strong dependency on the site conditions 7 3 7 Accuracy of the method The propagation of sound and its attenuation depends on the meteorological and geographical conditions along the propagation path The typical error associated with an assessment using the above method is 3dB The ISO 9613 2 standard that is used in the noise propagation models is strictly only applicable when the terrain is almost flat or with a constant slope The application of this standard to wind turbine noise is with large heights above ground and large propagation distances outside the usual scope of the standard 6 3 We recommend the complex model with G 0 hard ground is used for all surfaces to account for possible differences in behaviour of wind turbine noise as opposed to typical indust
64. orrected from the incident rotor wind speed to the free stream wind speed This correction is necessary in order to ensure that at distances far downstream the wake wind speed will recover to the free stream value rather than that incident on the rotor Therefore the initial centre line velocity Dmi is scaled by the ratio of the average influx velocity U and the free upstream wind velocity according to the following formula D i U CD 3 2 9 U U 3 2 3 Wake development Using the Crank Nicholson method 3 7 a tri diagonal matrix can be produced and solved at the first grid node of the wake giving the centre line velocity at the next grid point downstream The process can then be repeated for each grid node in turn From these solutions only the wake centreline wind speed and wake width are stored for each axial centreline node for computational efficiency Once the wake width and centreline wind speed for the whole length of the wake have been calculated the effect of the wake on the downstream turbines can be determined For each downstream turbine that falls inside the wake the incident wind speed needs to be calculated The velocity profile across the turbine affected by wake is calculated by assuming a Gaussian profile based on the centre line velocity at that distance downstream If some of the rotor is outside the wake then the wind speed for that portion of the rotor is set as the incident wind speed of the turbine cr
65. overall eddy viscosity is given by 3 5 e FKB U U Emb 3 2 5 where the filter function F is a factor applied for near wake conditions This filter can be introduced to allow for the build up of turbulence on wake mixing The dimensionless constant K is a constant value over the whole flow field and the default value is 0 015 The ambient eddy viscosity term is calculated by the following equation proposed by Ainslie 3 5 Em F Kr Im 1100 3 2 6 amb K is the von Karman constant with a value of 0 4 As a result of comparisons between the model and measurements reported by Taylor in 3 6 the filter function F is fixed at unity 3 2 2 Initialisation of the model The centre line velocity deficit Dmi can be calculated at the start of the wake model two diameters downstream using the following empirical equation proposed by Ainslie 3 5 GH WINDFARMER THEORY MANUAL 9 D 1 72 0 05 16C 0 5 I 1000 3 2 7 mi amb L Assuming a Gaussian wind speed profile and momentum conservation an expression for the wake width is obtained 3 56C B 3 2 8 8D 1 0 5D The wake width B used is defined as 1 89 times the half width of the Gaussian profile Using the above equations the average eddy viscosity at a distance 2D downstream of the turbine can be calculated However if the turbine under consideration is in the wake of another turbine this initial wake velocity deficit is c
66. rce grid file which is based on a fit of a Weibull distribution to the measured data For certain wind regimes the Weibull distribution may not give a good representation of the wind climate at the site Using the Association Method allows the user to e Use the measured wind speed and directional distribution instead of a Weibul distribution e Model the variation of the turbulence intensity over a wind farm e Model the influence of the topography on turbine wakes e Assess the flow models modification to the directional probability GH WINDFARMER THEORY MANUAL 3 The additional parameter settings used with the Association Method are listed in the table below In the following Section it is assumed that the two optional input files are available ENERGY Default Description CALCULATION value Apply direction shift Not ticked Shifts the directional probabilities following the to sector probability predictions of the flow model used Use with care Apply topography Ticked The wake of a wind turbine is modified by the correction to wake topography if ticked Older models calculate the wake model losses as they would be in flat terrain Summary of optional energy calculation inputs 2 3 Net Yield without losses In analysing the wind farm energy production the program performs three energy calculations for the following situations A Ideal Energy All turbines experiencing the same wind regime as at the reference location at th
67. reas of buildings Amet is the attenuation due to meteorological effects In the case of a wind farm with multiple turbines T4 To Th it is necessary to find the combined sound pressure level for all the turbines at a point GH WindFarmer calculates the effective sound pressure level using Lin Li Las Lan Liora 10log 10 1 10 10 10 4 10 6 2 3 7 2 1 Geometrical Divergence Agiv This attenuation accounts for spherical spreading in the free field from a point sound source over hard ground The next equation is used in GH WindFarmer to calculate the attenuation due to geometrical divergence ISO 7 An 20log d 11 dB 6 2 4 where d is the 3 dimensional distance between the source and the receiver The combination of this spherical spreading and a hard ground plane is sometimes called a hemispherical model 7 2 2 Atmospheric Attenuation Aam The attenuation due to atmospheric absorption is calculated with the next equation ISO 8 ad Am 6 2 5 1000 where a is the atmospheric attenuation coefficient in decibels per kilometer for each octave band see Table 2 ISO 9613 2 GH WINDFARMER THEORY MANUAL 37 7 2 3 Ground Attenuation Agr The ground attenuation is calculated using the equations given in Table 3 ISO 9613 2 using a ground factor of zero hard ground and the fixed reference frequency of 500 Hz The total ground attenuation is the sum of the ground attenuation
68. rial noise 7 4 Turbine noise vs background noise Depending on the national limits and guidelines on wind turbine noise it can become necessary to calculate the wind farm noise relative to the background noise at sensitive locations Both wind turbine noise and background noise are a function of their respective different local wind speeds The turbine sound power level is provided by the turbine manufacturer as a function of wind speed at a reference height typically 10m and reference roughness usually 0 05m Within the wind farm the wind speeds will at any given time vary from turbine to turbine as will the resulting turbine sound power level Each turbine sound power level is calculated using the turbine sound power levels at different wind speeds as provided by the turbine manufacturer The background noise level is usually measured as a function of the local wind speed in a location representative for the dwelling Also required is a measure of the wind speed differences between turbines and between turbines and dwellings For this the mean wind speed defined at the background noise reference is compared with the mean free wind speed at the turbine hub calculated using WRGI RSF information Using this ratio the hub height wind speed at the turbine for each background noise reference wind speed is internally calculated GH WINDFARMER THEORY MANUAL 40 The process is illustrated in the flow chart below wind speed at mean wind
69. rms ENDOW Riso National Laboratory Final report Riso R 1407 EN 2003 Rathman O Barthelmie R and Frandsen S Turbine Wake Model for Wind Resource Software EWEC2006 Athens Bossanyi E A Maclean C Whittle G E Dunn P D Lipman N H Musgrove P J The Efficiency of Wind Turbine Clusters International Symposium on Wind Energy Systems Lyngby Denmark 1980 Crafoord C An Estimate of the Interaction of a Limited Array of Windmills Department of Meteorology University of Stockholm Sweden Report DM 16 1975 D J Milborrow The Performance of Arrays of Wind Turbines Journal of Industrial Aerodynamics 5 pp 403 430 1980 GH WINDFARMER THEORY MANUAL 59 3 20 4 1 4 2 4 3 4 4 6 1 6 2 6 3 8 1 8 2 Newman B G The Spacing of Wind Turbines in Large Arrays Energy Conversion Vol 16 pp 169 171 1977 Schlez W Neubert A and Smith G M New Developments in Precision Wake Modelling DEWEK 2006 Bremen Schlez W Neubert A New Developments in Large Wind Farm Modelling EWEC2009 Marseille France 2009 IEC 61400 1 2005 Wind turbines Part 1 Design requirements Third edition IEC 61400 1 1998 Wind turbines Part 1 Design requirements Second edition Sten Tron s Frandsen Turbulence and turbulence generated structural loading in wind turbine clusters Ris R 1188 EN 2005 DIBt Deutsches Institut f r Bautechnik Richtlinie f r Windenergieanlag
70. s to the estimation of the slope of the best fit of Ye De 5 1 2 SG i 5 1 2 PCA Principal Components Analysis method The PCA fitting method involves a mathematical procedure that transforms a number of correlated variables into a smaller number of uncorrelated parameters called principle components The approach is equivalent to a least squares fit that minimises the squared orthogonal i e perpendicular distance of the measured points from a linear function as illustrated The perpendicular distance from the fit is Y Ox b 1 d I 5 1 3 Minimising the sum of the squared distance results in the estimation of the slope of the best fit being De y b B vVB 1 with B 5 1 4 ar l GH WINDFARMER THEORY MANUAL 28 PCA method 5 1 3 Handling of low wind speed cut offs In the MCP Module it is possible to exclude data at low wind speeds when determining the correlation relationships Usually for the low wind speeds the scatter is relatively high and from the energy point of view low wind speeds below turbine cut in wind speeds are not relevant There are two options available in the MCP Module for handling the correlation cut offs e The low wind speed cut off set in the Measure Correlate Predict window refers to the wind speeds both at the reference and the site location All points x y having x and or y smaller than the cut off wind speed are excluded from the correlation
71. ser to efficiently design a wind farm to achieve maximum energy production within the geometric and environmental constraints of the site This manual provides background information on some of the core features e Wind data analysis and correlation e Energy and wake loss calculation e Site conditions assessment e Noise propagation model e ZVI analysis e Electrical loss calculation e Shadow Flicker calculation GH WINDFARMER THEORY MANUAL 1 2 ENERGY CALCULATION An energy calculation combines the incident wind speeds at each turbine with the power curve of the turbine to give the power output for the whole wind farm applying the frequency distribution results in the expected energy yield The energy production of the wind farm is calculated using GH WindFarmer in conjunction with WAsP or a wind flow model with compatible output The wind flow model is used to determine the ambient wind speeds at each turbine location The output of a wind flow model consists of the directional Weibull A and k parameters which represent the wind speed probability distribution and a directional probability for every point on a grid GH WindFarmer enables you to use the probability distribution of wind speed and direction measured at an on site mast by associating this file with the predictions of the wind flow model Through this Association Method the measured data are scaled to the turbine locations using the predictions of the wind flow model A wake
72. simplification of the problem GH WindFarmer uses the third method checking the line of sight at regular intervals against the terrain height at this point This method takes little longer than alternative methods but provides highest accuracy 10 3 Standard ZVI for hub and for tip visibility Tip Hub height ZWl height Calculation of hub and tip visibility GH WINDFARMER THEORY MANUAL 54 A ZVI answers the question How many turbines can see from a location A ZVI calculation can be carried out for a single point or an area Each ZVI point is calculated for a specific ZVI height usually 2m above ground level A standard ZVI shows how many wind turbines are visible at a observer point A where visibility is defined as either e atleast the hub being visible or more sensitively e atleast the tip of a blade being visible The accuracy of a ZVI can be assessed by comparing the result of the ZVI with a visualisation from the same point of interest In the example above the blade tip is visible but the hub is not 10 4 Vertical subtended angle Tip Hub Hub height f u angle gt ZW height Vertical subtended angle In flat offshore areas or on top of hills wind turbine visibility presented in a ZVI will show turbine visibility for large distances The severity of the impact will however decline with distance The vertical subtended angle is used to add a quantitative measure of severity to t
73. speed u4 felt at the lower edge of the rotor z is decisive for the turbine performance The wind speed u is then expressed as u z for 22 0 3h ulz 2 5 ade 03 vif x for 0 094 lt z lt 0 3H Zii Z 09 0 09 H A rn GH WINDFARMER THEORY MANUAL 13 3 5 2 Wind Turbine Density The momentum extracted per given area is increasing with the number of wind turbines in that area Changing the area roughness to achieve this is an option but it is impractical for wind farm design purposes as the distribution of wind turbines may be irregular and is subject to iterative change and no fixed relationship with roughness can be established Instead a geometric measure of turbine density is used 45 40 35 30 25 Distance D Distance D Geometric model to consider the turbine density for a 30 degree sector The large wind farm correction to ambient wind speed is applied for wind arriving from the red sectors For each small direction sector the horizon is scanned and the presence of upstream turbines detected The ambient wind speed correction is applied if and only if such a turbine is present in the sector The consequences of the geometric model is that the overall impact of the correction is reduced with distance from turbines due to a smaller aspect ratio and increased for a fixed distance with increased turbine density due to more turbines contributing The geometric model considers ho
74. speed at background noise reference mean wind speed at turbine background noise reference speed up betw een dw elling and turbine wind speed at turbine logarithmic w ind profile z 0 05m wind speed of turbine noise measurement noise at turbine corresponding to w ind speed at background noise reference GH WINDFARMER THEORY MANUAL 41 8 ELECTRICAL CALCULATIONS 8 1 Input data checks The following data checks are required before attempting any calculation Check node connections each node must be connected eventually to one and only one root node Check cable and line connections each cable and line must be connected between two different nodes Check for substations cables lines and transformers are not beyond root nodes i e between a root node and a turbine or PFCDs if installed Check for loops check that there is only one route from each turbine to a metering point Check for correct voltage allocation check that each cable or line has both ends eventually connected to turbines transformers and substations with the same nominal voltage Note assume the primary side of a transformer is the side closest to the root node Check for correct cable or line selection check that each cable or line has a nominal voltage the same as or greater than its operating voltage Note that it is not necessary for all turbines to be connected to the electrical system for exampl
75. the rotor is in the wake Then the analysis described in chapter 4 4 1 is carried out 4 5 Wind speed and sector management During the calculation of the design equivalent turbulence the program checks whether the turbine under consideration is operational or not taking into account the cut in and cut out wind speeds and sector management and using the incident wind speed that includes any wake effects If the turbine is not in operation the design equivalent turbulence is set to zero 4 6 Evaluation After deriving the estimates of Characteristic or Representative turbulence intensity for a particular turbine they can be plotted as a function of hub height wind speed and compared with the allowable values over the operational wind speed range of the turbines It should be noted that the allowed wind speed standard deviations O Vin defined in the design standards must be divided by the hub height wind speed for direct comparison with the Flow and Performance Matrix output of GH WindFarmer This is because GH WindFarmer provides outputs of turbulence intensity and not wind speed standard deviation If there are any doubts concerning the suitability of the turbine a site specific load calculation should be considered For this purpose the use of the GH WindFarmer Bladed Link is recommended GH WINDFARMER THEORY MANUAL 26 5 MCP METHODOLOGY Measure Correlate Predict MCP methods are used to derive long term representative wind sp
76. the transformer series impedance Zia has been left at the default value of zero by the user In this case set X to zero 8 4 5 Transformation of excitation current into magnetising reactance The first thing is to calculate excitation current referred to secondary as follows lec lec Inom 100 Amps lhom S 1000 v3 U same calculation to obtain Series Resistance nom where lec is the excitation current as percent on the nominal current laom iS the nominal current of the transformer S is the nominal rating kVA U is the nominal secondary voltage U Volts GH WINDFARMER THEORY MANUAL 46 Then the magnetising resistance is calculated as Rm U Puuu Ohms where Part is the no load losses W And finally the magnetising reactance is calculated from the expression Xm _ Ohms lec yA U V3 R Note If excitation current is not stated this should be equal to 0 so that the magnetising reactance should be assigned infinite 8 4 6 Calculate equivalent impedance of N elements in parallel When N elements in parallel are to be connected in an electrical circuit the following set of equations must be solved to obtain the equivalent circuit l l2 l3 Has Inet In Z4 l4 Zo lo 0 Zl Z3 l3 0 Zil Zy 1 lyN 1 0 Zo lo 23 13 0 Zo lo Zus In 0 23 13 Zu lyr 0 It must be said that all the matrix calculations have to be conducted using complex numbers where Z R jX Ohms
77. tions e that there are no clouds e thatthe turbines are always rotating e alimit of human perception of the shadow flicker is not considered GH WINDFARMER THEORY MANUAL 53 10 VISUAL INFLUENCE 10 1 The terrain model The Digital Terrain Model used to model the visual influence consists of a regular grid of spot heights These spot values are provided in a three column ASCII file x y z with the file extension DTM Other digital terrain file formats can easily converted to the DTM file format The precision that can be achieved will be limited to the precision of the digital terrain model Typically a grid resolution of 50 m is used It is worth to keep in mind in the following that locations e g of turbines or houses as well as heights in the DTM have typically an associated error of 5 10m 10 2 Line of sight algorithm The algorithm checks either for a viewpoint or for each point of an area if the line of sight to the turbine is interrupted by a terrain feature Different methods for checking this are used in the field e Checking if the line of sight falls below the connecting lines between pairs of adjacent DTM grid points e Checking if the line of sight crosses any of the two triangular areas formed by sets of three out of four corner points of a DTM cell e Checking if the line of sight crosses the interpolated area between the four corner points Methods like first two give an advantage in speed and performance through
78. ule can also be used to reduce the shadow flicker annoyance at the receptors by providing the turbine controller or SCADA system with the time and date of shadow flicker occurrences so that the turbines can be switched off at these times 9 2 Program inputs The following inputs to the GH WindFarmer model are required to produce an estimate of the wind farm s shadow flicker effect Latitude where the wind farm s is located y Longitude where the wind farm s is located A Zone longitude where the wind farm s is located Azone Minimum elevation angle of the sun Calculation time interval Maximum distance from turbine for calculation Resolution of calculation points Turbine and shadow receptor locations Turbine dimensions hub height rotor diameter distance between rotor and turbine tower centre It is possible to use the following optional input data Time Zone The zone longitude and the time zone are equivalent sets of information because for each hour deviation from the GMT there is 15 degree offset in the zone longitude The time zone is used in GH WindFarmer 9 3 Methodology of the shadow flicker calculation Before performing the actual shadow flicker calculation the program determines the position of the sun at any time of the year The following definitions and equations are used to determine the elevation and azimuth angle that determine the position of the sun For more details and definitions see 8 1 and 8 2 The Hour
79. urbulence correction The ambient turbulence level increases within large wind farms An adjustment has been suggested when the number of wind turbines from the receiving wind turbine to the edge of the wind farm is more than 5 or the spacing in the rows perpendicular to the dominant wind direction is less than 3 rotor diameters 4 1 In this case is replaced by gj which is 4 2 2 Oj 5 No 0 0 with 4 3 1 0 36 Vps S d d 1 0 2 io a C V hubi where Ct Vhub i is the the turbine thrust coefficient of the upwind turbine at wind speed Vrun d is the turbine separation within rows in rotor diameters d is the turbine separation between rows in rotor diameters The definitions of d and d demand a regular turbine layout GH WindFarmer approximates these parameters for all layouts to be the distance to the closest neighbouring turbine and the distance to the upstream turbine causing the wake If this correction is selected by the user the correction is applied to all wind directions whether or not there are 5 or more upstream turbines 4 4 Simple and Advanced Approach The calculations models given in the Design Standards are especially for wind farms with a regular turbine layout The model described in 4 3 has been derived for wind farms in simple terrain and consisting of only one turbine type In many cases the wind turbines are not in a regular layout they are located in complex terrain and there
80. w much of the horizon is filled with turbines Wake expansion is not considered in this step since the model is to be used in conjunction with standard wake models that already consider the effect of wake expansion and consequential wake recovery GH WINDFARMER THEORY MANUAL 14 3 5 3 Wind blows along wind farm geometry axis When for any wind direction the cross wind spacing between two rows of wind turbines is more than a fixed number of turbine diameters then the wind turbine wakes in each row are considered to develop independent from other rows Experimental evidence suggests that no large wind farm correction should be applied for this narrow wind direction sector xi XE xX x X ix i i gt turbine under m Se A a a NHEesaskderatian n n 7 Ya wind direction The crosswind distance is determined by projection of upstream turbines onto a plane perpendicular to the flow direction 3 5 4 Downstream recovery The disturbance of the wind profile caused by a wind turbine is expected to subside after a certain distance Investigation is ongoing to establish details of the processes and relevant scales However as a first approximation this recovery can be modelled as linear 1 01 7 1 00 0 99 0 98 0 97 0 96 0 95 0 94 0 93 0 92 0 91 0 90 T T T T T T T relative wind speed downwind distance D Corrected ambient wind speed u and linear recovery to ambient level GH WINDFARMER THEORY MANUAL 15
81. within row spacing is greater than approximately 2D then this modified model will give the same results as the standard Eddy Viscosity model Please note if you use the closely spaced wake model the turbine array in question must be similar to that described above for it to fall within the realms of validity of the model The validation for this model has been undertaken using very specific uni and bi directional wind regimes It is not recommended at present that the results are used for wind regimes and layouts which differ from this GH WINDFARMER THEORY MANUAL 11 3 5 Correction for large wind farms The modelling of wind farms is traditionally a two step process In the first step the ambient wind flow without the presence of a wind farm is established The wind turbines are placed within this wind flow in the second step The ambient wind flow is assumed to be independent from the wakes generated by the wind turbines It is well recognised that wind turbines do not only react passively to the wind regime but at the same time they are part of it 3 15 3 18 Qualitatively the wind farm can be thought of an area of higher roughness It is considered that modern wind farm developments do not affect weather systems significantly However locally the boundary layer profile is disturbed by the extraction of momentum by the wind turbines In particular offshore wind farms cause a more pronounced local effect similar to an onshore forest due t

GH WindFarmer 4.0

Contents

Download Pdf Manuals

Related Search

Related Contents