Methodology and formulas used to calculate Wind Power Density

Pakistan Meteorological Department

WIND SPEED VARIATION WITH HEIGHT

It has been observed that wind speed increases with height. The variation of wind speed with height is dependent on the atmospheric stability.

Log law.
Log law is slightly dependent and used for the calculations of wind speed at a given height. It is given by

V=V*/K ln [Z-D/Z0]
Where V* is friction notify K is von Kerman constant and is Z0 is roughness length D is displacement height, usually taken as zero because turbines are rarely sited in forests and towns. Since V* is difficult to calculate, the above formula is usually rewritten in terms of a reference wind speed VR at a reference height ZR as given by,

V/VR= Ln (Z/Z0)/Ln (ZR/Z0)

Where VR is the wind speed at reference height ZR. The above equation is valid up to heights of about 100 m., which is sufficient for installation of all wind turbines currently in use. ZO can be estimated using

ZO = (H^VR/HR^V)^(1/(VR-V))
Power law
Engineers often prefer to use a power law to describe the wind variation with height. This formula is used for neutral conditions, which simple estimates of the distribution of the mean wind speed with height are required is given as

V/VR =(Z-D/ZR)
Where is power law exponent VR is Wind Speed at reference heights ZR D is displacement heights usually taken as zero. The power law exponent typically varies between 0.1 and 0.32 depending upon the landscape type. The exponent can be calculated from the roughness length.

= 1/Ln (Z.ZR/Zo)1/2 Or =LOG(V/VR)/LOG(Z/ZR)

Where Z is measurement height ZR is reference height. Z0 is roughness length

Prepared by Abdul Hameed (Director), Yaqub Hussain Khan (Meteorologist) and Asif Hussain (Asst. Meteorologist)

15

Methodology and formulas used to calculate Wind Power Density

Pakistan Meteorological Department

WIND SPEED FREQUENCY DISTRIBUTION.

Wind speed frequency distribution can simply be obtained by plotting the differences wind speed against their frequencies/related frequencies. For obtaining frequency distribution the following two procedures are necessary

BINNIG OF DATA
The sorting of the data into narrow wind speed bands is called binning of the data. In our case a bin width of 1 m/sec has been used e.g.; a measured wind speed of 3.5 m/sec would be placed in 3<X<=4 m/sec bin. The central value of each bin i.e. 0.5 m/sec, 1.5 m/sec etc has been used in calculations and frequency distribution groups.

RELATIVE FREQUENCY.
It is the proportion wind speed in each bin. It can be reviewed as the estimate of probability of given wind speed in the bin. Relative frequency is defined as R.F=probability P (Vi)=Frequency of given wind speed/Total period

WIND SPEED STATISTIC

The statistical Mean It is the average of a set of n numbers. Mathematically, we can write

Mean = [ n xi] /N
i=1

The Mean Wind Speed V can be calculated by a formula.

Prepared by Abdul Hameed (Director), Yaqub Hussain Khan (Meteorologist) and Asif Hussain (Asst. Meteorologist)

16

Methodology and formulas used to calculate Wind Power Density

Pakistan Meteorological Department

V = n Vi P(Vi)
i=1 I and P(V ) is the probability/relative frequency that Where Vi is the central wind speed of bin i

the wind speed has in bin i. VARIANCE: It is one of the several indices of variability that statistician, use to characterize the dispersion among the measures in a given set of data. Mathematically, variance is written as

Variance = 2 = (Xi-V)2
Where V is mean of data set. In case of wind speed data, we can write it, as

2 = Vi2 P (Vi) (V)2

STANDARD DEVIATION: It is the square root of the variance, denoted by

= ( )1/2 = ( Vi2 P(Vi) (V)2)1/2 Power of wind Energy

A parcel of Wind possesses kinetic energy

E= mV2
From this, power density is calculated as

P= e/t = dm/dt V2
Where dm/dt is the mass of air following time. From fluid dynamics, it can be proved that

dm/dt = AV
Volume of cylindrical cross section can be written as

V = r2 L ---- (1)
Prepared by Abdul Hameed (Director), Yaqub Hussain Khan (Meteorologist) and Asif Hussain (Asst. Meteorologist)

17

Methodology and formulas used to calculate Wind Power Density

Pakistan Meteorological Department

Where r is radius of cylinder and L is length of it The wind moving with velocity V travels this distance L in time t so

S = L= Vt,
So equation L takes the form

V = r2 Vt
Now mass of wind can be written as

M= Avt
Differentiating

dm/dt= AV d/dt(t) = AV
Where is density of wind and others parameters have been defined in diagram

So the power is then,

P = dm/dt V2 = AVT/t V2 = AV3

And power density

P/A = V3
Density of wind at mean sea level is 1.225 kg/m3
Prepared by Abdul Hameed (Director), Yaqub Hussain Khan (Meteorologist) and Asif Hussain (Asst. Meteorologist)

18

Methodology and formulas used to calculate Wind Power Density

Pakistan Meteorological Department

At 15C. The area depends upon the size of the rotor. Therefore it is clear that power density chiefly depends on wind velocity and goes up as a cube of it.

Wind power Calculation using Mean wind Speed

Wind power Calculated form Mean wind speed is not true representative of wind power .In real world, the wind varies constantly. Actual wind power density at most sites can rang form 1.0 to3 times greater then that calculated. For example we take wind speed of 5,7 and 8 m/sec respectively the respective power densities are 76 wat/m2, 210 watt/m2, and 313 watt/m2. The average of which is 200 watt/m2. On the other hand the average wind speed is 6.7 m/sec and power density of this av wind is 181 watt/m2. So the power of the wind calculated by Mean wind speed is less than the actual power present in wind. I.e. mean wind speed is not true representative for the wind power calculations. To overcome this drawback we find some alternative arrangement, which reduces the deficit. The weibull distribution is the best fit of wind data to calculate wind power based on mean wind speed and variance/standard deviation.

WEIBULL DISTRIBUTION
The weibull distribution is one of the most widely used lifetime distribution in reliability engineering. It has been found that frequency distribution at most sites can be conveniently and adequately represented by the weibull distribution with the help of weibull distribution probability of the wind speed V can be bound as

P(V) = (k/c) (V/C)k-1 EXP (-(V/C)k)

Where K is shape parameter the weibull distribution gets more packed as k gets larger value with peak moving in the direction of higher wind speeds. The parameter C scales the xaxis to fit different wind regimes. For K greater than Zero, the probability of calm wind becomes zero. So it is clearly a shortcoming in this distribution.

Prepared by Abdul Hameed (Director), Yaqub Hussain Khan (Meteorologist) and Asif Hussain (Asst. Meteorologist)

19