See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/290730448

The wind energy system performance overview: Capacity factor vs. technical
efficiency

Article · January 2011

CITATIONS READS

28 1,434

2 authors, including:

Ciprian Nemes
Gheorghe Asachi Technical University of Iasi
70 PUBLICATIONS   201 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Demand-Side Management of residential prosumers with photovoltaic systems. Project PN-III-2017-0823. View project

Power Quality and Reliability of Supply View project

All content following this page was uploaded by Ciprian Nemes on 21 June 2016.

The user has requested enhancement of the downloaded file.

 INTERNATIONAL JOURNAL OF MATHEMATICAL MODELS AND METHODS IN APPLIED SCIENCES

The wind energy system performance overview:

capacity factor vs. technical efficiency
Ciprian Nemes 1, Florin Munteanu

suitable wind turbine for a specific wind profile.

Abstract—The main objective of the paper is to develop a A measure of the suitability of wind turbine to a specific
probabilistic model for capacity factor and technical efficiency location is given by the capacity factor and efficiency values.
estimation for a wind turbine located in a specific area, model based The capacity factor is defined as the ratio of the expected
on the output power distribution of wind turbine. This model was
applied for a wind turbine located to a region in the North-East of
output power over a period of time to the rated power of wind
Romania, the model results being validated by results from Monte- turbine generator. All power plants have capacity factors, and
Carlo simulation. Finally, the model was used to evaluate the effects they vary depending on resource, technology and purpose.
of wind turbine generator parameters, for a given wind profile, on the Typical wind power capacity factors are 20-40%, with values
capacity factor and technical efficiency values. at the upper end of the range in particularly favourable areas
[4]. The capacity factor is not an indicator of efficiency. A
Keywords— wind energy, Weibull distribution, output power measure of turbine efficiency is the power coefficient. This
distribution, capacity factor, technical efficiency.
coefficient indicates how efficiently a turbine converts the
wind energy into electricity. This coefficient varies with the
I. INTRODUCTION
wind speed [2]. Efficiency is the expected power coefficient,

E NVIRONMENTAL factors such as global warming and

pollution have heightened the need to introduce into the
generation mix a greater percentage of renewable sources. In
over a period of time, and is defined as ratio of the useful
output energy to the input wind energy.
In literature are presented various approaches for capacity
the last time, wind power has drawn much attention as a factor and efficiency estimations, mostly obtained from
promising renewable energy resource, which has shown some simulations techniques based on wind speed data [5,6] and
prospects in curtailing fuel consumption and reducing the sometimes from computational models [1,7], that need
emission of pollutants into the atmosphere. Unlike other numerical integration techniques or some approximations.
renewable energy sources, wind energy has become Having in view the stochastic nature of the primary energy,
competitive with conventional power generation sources and the probabilistic methods can be proper solutions for capacity
therefore application of wind turbine generators has the most factor and efficiency evaluations. In the paper, a probabilistic
growth among other sources. Wind is one of fastest growing model is developed to evaluate these values and to analyze the
energy source and is considered as an important alternative to dependence of the wind turbine generator characteristics. The
conventional power generating sources. proposed model is based on probability density function of
The energy production from a wind turbine or a wind park, output power generated by the wind turbine. In order to
in a specific location, depends by many factors. The main validate the model, the results model were compared with the
factors include the wind speed conditions from the area, and results of the other model, namely with Monte Carlo
most importantly, the characteristics of the wind turbine technique. The model has the advantage that can be easily
generator itself, particularly the cut-in, rated and cut-off wind implemented in computer programs and require a computing
speed parameters. The output power of a wind turbine time considerably less than in the case of simulation or
generator does not vary linearly with the wind speed. The numerical methods.
output power increases with the wind speed between the cut-in Selection of the optimal wind turbine was discussed in
speed and the rated wind speed, after that the power output different manner in various papers, among which the
remains constant at the rated power level, until the cut-out maximization of capacity factor and/or efficiency [5,6,7]. The
speed, when the turbine is stopped for safety reasons [1,2]. choice of turbine involves choosing parameters that lead to
Different types of wind turbines are commercially available on maximizing these factors. The turbines must be chosen with
the market. It is therefore desirable to select a wind turbine the parameters that match those of wind profile area. Based on
which is best suited for a particular location in order to obtain these issues, in the paper, a numerical analysis is realized to
a maximum power from power transposed by the wind. These have a comparison between effects of different parameters of
important aspects bring suitability concerns regarded by the the wind turbine generator on capacity factor and efficiency
energy potential of a specific location and the selection of the values, analyzing the importance and weight of each parameter
of those values.
1
Corresponding author: Tel +40232278683, email [email protected]

Issue 1, Volume 5, 2011 159

 INTERNATIONAL JOURNAL OF MATHEMATICAL MODELS AND METHODS IN APPLIED SCIENCES

II. WIND AND OUTPUT POWER WIND TURBINE PROBABILISTIC To establish the parameters of probability density
CHARACTERISTICS distribution is necessary an accurate dataset of wind speed.
The output power from a wind turbine depends by the The wind speed database can be obtained from meteorological
availability of the energy source, namely the wind speed and station, where, usually, the measurement point (anemometer)
height above the surface may be to 10 m or 50 m. Depending
the power-wind characteristics of the wind turbine generator.
by the wind measurement level, the speed data must be
A. Probabilistic model of wind speed adjusted for the change in height desired according to a
Wind is a turbulent movement mass of air resulting from the logarithmic profile previously mentioned.
differential pressure at different locations on the earth surface. B. Wind speed-power relationship
One of the main characteristics of wind is that it is highly
The power transported by an air stream flowing with a given
variable in time and space, the variation of wind exists from
speed, v, can be calculated according to [1] using the following
instantaneous, hourly, daily to seasonal, and its properties vary
simple expression:
from one location to another. The wind property of interest in
the power generation problems is the wind speed probabilistic PV = 1 2 A ⋅ ρ ⋅ v 3 (3)
model. where ρ is the air density and A the area of the air stream,
The speed of the wind is continuously changing, making it measured in a perpendicular plane to the direction of the wind
desirable to be described by the probabilistic models. The speed.
probability density function of wind speed is important in The calculation of the mechanical power that can be
numerous wind energy applications. A large number of studies extracted by the rotor of a wind turbine, requires Betz’ law to
have been published in scientific literature related to wind be taken into account. This law specifies that only a maximum
energy, which propose the use of a variety of functions to 16/27 of the wind energy can be converted into mechanical
describe wind speed frequency distributions [9],[10]. The power. This value is known as the Betz limit. In practice, the
conclusion of these studies is that the Weibull distribution of collection efficiency of a rotor is not as high as 59%. A more
two parameters may be successfully utilized to describe the typical efficiency is 35% to 45%.
principle wind speed variation. The Weibull probability The mechanical power is converted in electrical power by
density and cumulative distribution function are given by: generator, so, the output electric power of a wind turbine is a
β −1 function of the wind speed. The power curve gives a relation
β v   v β    v β 
f W (v) =   exp  −    ; FW ( v ) = 1 − exp  −    (1) between the wind speed and the output electric power, a
α α   α     α   typical curve of the wind turbine generator is nonlinear related
The scale parameter α (m/s) and a shape parameter β to the wind speed. However, the assumption of the linear
(dimensionless) of the Weibull distribution can be found using characteristic of power with the wind speed, brings a
different estimation methods [2],[4]. Each method has a significantly simplifies of calculations, without roughly errors.
criterion, which yields estimates that are best in some The power output characteristic can be assumed in such way:
situations. Different results are produced based on that − it starts generating power when the speed wind exceeds
criterion. The most commonly methods are Maximum the minimum wind speed, namely cut-in speed (vcut-in);
Likelihood Estimator, Method of Moments, Least Squares − the power output increases with the wind speed, when
Method or Regression Method. The Maximum Likelihood wind varies between cut-in and rated speed wind (vrated),
Estimator is so commonly applied in engineering and value for that the power achieves the rated power (Prated).
mathematics problems [6], so, this method is used in this paper − the rated power of a wind turbine, generally the maximum
to establish the parameters of wind speed distribution. power output of a generator at highest efficiency, is
In many studies, the shape parameter is often chose to 2 produced when the speed lies between rated and cut-off
and therefore a Rayleigh distribution can be used, with a same wind speed (vcut-off).
accuracy and with a simpler model. − cut-off wind speed is the maximum wind speed at which
The wind blows faster at higher altitudes because of the the turbine is allowed to produce power, usually limited
reduced influence of drag of the surface and lower air by engineering design and safety constraints.
viscosity. The effect of the altitudes in the wind speed is most Thus, the electric power PE may be calculated from the wind
dramatic near the surface and is affected by topography, speed as follows:
surface roughness, and wind obstacles such as trees or
 Prated ⋅ (v − v cut − in )
buildings. The most common expression for the variation of  (v for v cut − in < v < v rated
wind speed with hub height is the power law having the  rated − v cut − in ) (4)
PE (v ) =  Prated for v rated < v < v cut − off
following logarithmic profile model [8],[9].  0 other else
v ( z ) = v ( z r ) ⋅ (ln ( z z 0 ) ln ( z r z 0 )) (2) 

where v(z) and v(zr) are the wind speeds at a desired z and This curve comes available from the wind turbine
registered zr height, and z0 is the surface roughness length, a manufacturer or plotted using recorded wind speed and
characterization of a ground terrain. corresponding output power data, a typical curve of the wind

Issue 1, Volume 5, 2011 160

 INTERNATIONAL JOURNAL OF MATHEMATICAL MODELS AND METHODS IN APPLIED SCIENCES

turbine generator is shown in the figure 1. β

−1
β
 2 PV 3   (6)
Electric power f PV ( PV ) =
2β
⋅  ⋅ exp  − PV( A ρα 3 ) 3

(MW) 3 A ρα 3  A ρα 3 
  
Prated
Therefore the power transported by the wind can be
represented by a Weibull distribution, with α’ and β’
parameters given by:
α' = 1 2 ⋅ A ⋅ ρ ⋅ α 3 , β' = β 3 (7)

Wind speed where ρ is the air density, A the area of the wind turbine rotor,
(m/s) and α,β are the wind distribution parameters.
vcut-in vrated vcut-off The output power probabilistic model for a wind turbine and
Fig. 1 The power curve of a wind turbine its practical evaluations were developed and evaluated by
authors in [11],[12]. Similar results has also been obtained by
Equations (3) and (4) express the instantaneous values of others authors in [ 13],[14].
wind power and electric output power, as a function of the The possible values of FWT(P) may be roughly classified in
instantaneous wind speed. However, the wind speed may vary 0, Prated and in the interval that lies between mentioned values,
during a period of time. To consider this effect, we are going respectively. Each possible value has been evaluated, having
to work with the wind speed probability distribution function. in view the probability to achieve that value. The cumulative
C. Probabilistic model of wind power and wind turbine distribution function of the output power of the wind turbine
output power is:
The probability distribution functions of wind power and of [
 1 − FW (vcut − off ) − FW (vcut − in ) ] for PE = 0

the wind turbine output power can be obtained using the FPE ( PE ) =  FWT (0) + FW (W ) − FW (vcut − in ) for 0 < PE < Prated (8)
analytical dependence between wind speed, wind power, and  1 for PE = Prated

the output power respectively, operating a change of variables. The probability density function results from differential of
Lets assume that v is a continuous random variable with cumulative distribution function:
cumulative distribution function, FV(v) and that P=J(v) defines
 ℜ1 for PE = 0
a one-to-one transformation from a region of the wind-space,   (v
rated − v cut − in )
 (9)
to a region of the power-space, with inverse transformation f PE ( PE ) =    ⋅ fW (W ) for 0 < PE < Prated
v=J-1(P), the cumulative distribution function of power can be  Prated 
calculated according [16],[17]:  ℜ2 for PE = Prated

F ( P ) = Pr( P < p ) = Pr( J ( X ) < p ) = where:

− ℜ1 = 1 − [FW ( v cut − off ) − FW ( v cut − in ) ] = F PE ( 0 ) , represents
(5)
−1 −1
= Pr( X < J ( p )) = FW ( J ( p ))
In order to obtain the probability density function, a the value of output power cumulative distribution function
differential of cumulative distribution function must be in the 0 point,
operated. In figure 2 is presented the intuitive process of the − ℜ 2 = FW ( v cut − off ) − FW ( v rated − 0 ) = 1 − F PE ( Prated − 0 ) ,
random variables transformation, the wind speed variables represents the increase value of the cumulative
being transformed in wind power variable PW, on the right, distribution function in the Prated value, and
respectively in the output electrical power variable PE, on the  P .
− W =  (v rated − v cut − in ) ⋅ E + v cut − in 
left. P 
 rated 
PE,PV PV
PE
PE=J1 (v) PV=J2 (v)
III. CAPACITY FACTOR AND EFFICIENCY EVALUATIONS
Capacity factor and efficiency depend both on the wind
speed distributions in the area and the turbine parameters. To
illustrate the effect of wind turbine parameters on capacity
fPE(PE) v fPV(PV)
fV(v) factor and efficiency values, a probabilistic model for these
indicators is developed. The capacity factor (CF) of wind
v turbine is the ratio of expected output power over a period of
Fig. 2 Transformation of the wind speed variable
time to rated power. The efficiency (EF) of wind turbine is the
ratio of useful output energy to the input wind energy, or
If (5) is applied for the wind power relationship (3), having expected power output from wind machine to expected power
in view the Weibull distribution, the probability density available in wind over a period of time. The expected output
function of the wind power may be expressed as a function of power is used in both on the capacity factor and efficiency
the variable PV: evaluations, so, it will be firstly evaluated.

Issue 1, Volume 5, 2011 161

 INTERNATIONAL JOURNAL OF MATHEMATICAL MODELS AND METHODS IN APPLIED SCIENCES

A. The expected output power of a wind turbine of wind turbine generator.

The expected output power of a wind turbine depends on α 1
Γ 
β  β    v  1 
β β
the output power values and the probability to achieve that E ( PE )  1  v
CF = = ⋅  P   rated  ,  − P   cut − in  ,  
power, described by the probability distribution functions of Prated v rated − vcut − in    α  β    α  β 
 

output power. The expected output power (E(PE)) from a wind
− exp( − ( vcut − off / α )β ) (15)
turbine generator can be estimated from its power-wind and
wind characteristics, being represented by the probability The efficiency is a measurement of how much energy from
distribution of wind speed. This is given by [3] as: the wind is converted in electrical power energy. For that, the
P rated expected output power must be divided to the expected wind
E(PE ) = ∫0
PE ⋅ f PE ( PE )dPE (10) power input to measure how technically efficient is the wind
where PE is the output power function variable and fPE(PE) is turbine.
probability density function of output power from wind Prated α 1
⋅ Γ 
E(PE ) vrated − vcut −in β  β     vrated  1    vcut−in  1 
β β
turbine. EF = = ⋅ P   , −P  , 
Having in view the expression of output power distribution E( PV )  1    α  β    α  β 
α'⋅Γ1 + 
function, from (9), the expected output power can be obtained  β' 
by: Prated (16)
Pr ated − ⋅ exp( − ( v cut − off / α ) β )
 1
E ( PE ) = 0 ⋅ ℜ1 + ∫ 0
PE ⋅ f PE ( PE ) dP E + Prated ⋅ ℜ 2 =
(11) α '⋅Γ  1 + 
β' 

v rat v − v cut − in
= Prated ⋅ ∫ ⋅ f W (v ) ⋅ dv + Prated ⋅ ℜ . The equations (15) and (16) show the results of authors’
v cut − in v rated − v cut − in
research based on laborious calculations and detailed analysis
where fW(v) is a probability density function of wind speed.
of statistical distributions. This relationship represents an
Taking into account that dF W (v ) = f W (v ) ⋅ dv , the equation equation which shows the effects of cut-in, rated, and cut-off
(11) may be written as: speeds parameters on the capacity factor value. For a given
 1 − exp( − ( v / α ) β )  wind regime, with known α and β parameters, we can select
( ) ∫
v rated
E ( PE ) = Prated ⋅  FW v cut − off − dv  =
 vcut − in v rated − v cut − in  that values of vcut-in, vrated and vcut-off that maximize the
vrated exp( −(v / α ) β ) expected output power, and thereby maximize the capacity
= Prated ⋅ ∫
vcut −in vrated − vcut −in
dv − Prated ⋅ exp( −(vcut −off / α ) β ) (12) factor.
The integration can be accomplished by making the change
IV. MODEL VALIDATION AND NUMERICAL EXAMPLE
in variable y = (v / α ) β , and therefore dy = β ( v / α ) β −1 d ( v / α ) .
For validation of probabilistic model, its results have been
After substitution of integration limits and their reduction to
compared with results from other model. The Monte Carlo
the minimum number of terms, the result is:
simulation has been used to provide information related to the
α 1
Prated ⋅ Γ  average values of the capacity factor. A Matlab program has
β  β     vrated  1   1 
β β
 v
E ( PE ) = ⋅ P   , − P   cut − in  ,   (13) been developed to validate the probabilistic model.
vrated − vcut − in    α  β    α  β 
 
 The program has been structured by two main functions.
− Prated ⋅ exp( − (vcut − off / α )β ) First function is based on a probabilistic model previously
developed and modelled with (15) and (16), respectively.
where Γ ( ) and P( ) are the gamma and the lower incomplete
Second function has been developed based on Monte Carlo
gamma functions, respectively [15]. simulations technique. This technique generates different
B. The expected output power model of a wind turbine values of wind speed, in accordance with their Weibull
distribution (with the shape and scale parameters estimated
As it has been presented in (6), considering the wind power
from the real data base) and these wind values are used to
having a Weibull distribution, the expected value of the wind
generate the output power, having in view the characteristics
power can be expressed as a function of the parameters α’, β’ of the wind turbine generator. The expected output power from
and the Gamma function: a wind turbine is the power produced at each wind speed
 1 sample, integrated over all possible wind speeds. The required
E[PV ] = α'⋅Γ1+  (14)
 β'  capacity factor values may be observed from the average of all
output power values, over a long number of samples. The
C. The probabilistic model for capacity factor and efficiency is observed from average of all output power values,
technical efficiency and also from all power transported by any values of wind
The capacity factor of a wind turbine means its energy speed. The simulation can be stopped when a specified degree
output divided by the theoretical maximum output, if the wind of confidence has been achieved.
turbine generator were running at its rated (maximum) power The methodology presented in this paper was applied to a
during all the time. So, the capacity factor is the ratio of the real wind turbine and for a real wind speed database, to
expected output power over a period of time to the rated power validate the probabilistic model and to evaluate the influence

Issue 1, Volume 5, 2011 162

 INTERNATIONAL JOURNAL OF MATHEMATICAL MODELS AND METHODS IN APPLIED SCIENCES

of main parameters of wind turbine generator to capacity The probability density function (PDF) and cumulative
factor and efficiency value. The wind speed database was distribution function (CDF) of the output power for a 1.5 XLE
collected from the north-east area of Romania, for a wind turbine, considering a Weibull distribution with
measurement interval to one hour for the year 2008. The mentioned parameters, are presented in the figure 5.a,b.
figures 3.a,b present the wind speed collected from wind
station height (10m) and adjusted to the hub wind turbine 0.9
Probability density function for 1.5 XLE Wind Turbine (GE)

height (80m).
0.8
Wind speed to measured height (10m)
15
0.7
wind speed (m/s)

10 0.6

0.5

pdf(Pel)
5
0.4
0
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 0.3
time (hr)
Wind speed to 80 m height 0.2
30
0.1
wind speed (m/s)

20
0
0 0.5 1 1.5
10 Pel[MW]

Cumulative distribution function fot 1.5 XLE Wind Turbine (GE)

0 1
0 1000 2000 3000 4000 5000 6000 7000 8000 9000
time (hr) 0.9

Fig. 3.a,b Wind speed data base from the north-east area of 0.8

Romania, for 10m, respectively 80 m height 0.7

0.6
The parameters of the Weibull distribution have been
cdf(Pel)

0.5
estimated using the hourly wind data base. The probability
0.4
density and the density function fitted for different wind speed
0.3
values in 1 m/s steps are presented in figure 4. The probability
distribution function used to fit is a Weibull distribution with 0.2

scale parameter α=4.82253 m/s and a shape parameter 0.1

β=1.8656. 0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6
Pel[MW]

Wind speed data Fig.5.a,b The PDF and CDF of output power wind turbine
0.16
Weibull fit

0.14
Considering the ρ=1.225 kg/m3 for the air density (at 15°C)
0.12
and the rotor are A=5346 m2, in the previous equation (7), the
0.1 scale parameter and shape parameter of Weibull distribution
Dens ity

0.08 parameters are obtained as α’=3.6725×105m/s and β’=0.6219,

0.06 respectively. The probability density function and cumulative
0.04 distribution function of the wind power are in accordance with
0.02
Weibull distribution with mentioned parameters.
-4
x 10
0
0 5 10 15 20
Wind speed data (m/s)

Fig. 4 The wind data base associated distribution

The wind turbine chose for analyze is an active blade pitch 2

control wind turbine, namely 1.5 XLE GE-Energy,
pdf(Pw)

manufactured by GE Energy [18], with their technical

specifications presented in table 1:
1
Table 1 Technical specifications of wind turbine GE Energy
Rated Cut-in Rated Cut-off Rotor Hub
Turbine
power speed speed speed diameter Heights
Model
(MW) (m/s) (m/s) (m/s) (m) (m)
1.5xle - GE 0
1.5 3.5 11.5 20 82.5 80 0 0.5 1 1.5
Energy Wind Power Pw [MW]

Issue 1, Volume 5, 2011 163

 INTERNATIONAL JOURNAL OF MATHEMATICAL MODELS AND METHODS IN APPLIED SCIENCES

1 0.4

0.9
0.35

0.8
0.3
0.7
0.25

Efficiency (%)
0.6
cdf(Pw)

0.5 0.2

0.4
0.15
WTG 1.5xle - GE Energy
0.3 Vin=3.5 m/s
0.1 Vrat=11.5 m/s
0.2 Voff=20 m/s
0.05
0.1 Wbl(a=4.82253;b=1.8656)

0 0
0 0.5 1 1.5 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Wind Power Pw [MW] Number of samples

Fig.6.a,b The PDF and CDF of wind power Fig. 8 Efficiency result from Monte Carlo simulation

Using the output power and wind power distribution The capacity factor values provided by probabilistic model
functions in the probabilistic model previously developed, the and sequential Monte Carlo simulation are shown in Table 2.
capacity factor and efficiency values were evaluated and For a better comparison between models, three ranges of speed
presented in table 2 and 3. For validation, in the following is parameters of wind turbine were considered in capacity factor
presented an example of capacity factor and efficiency evaluation. Commercial wind turbines typically have cut-in
evaluation, for a 1.5 XLE GE-Energy wind turbine, using the speeds between 2.5 and 4.5m/s, a rated wind speeds between
Monte Carlo Simulation technique (MCS). This technique 10 and 15 m/s and a cut-off speeds between 20 and 25 m/s.
creates a fluctuating convergence coefficient of variation range
for various numbers of samples for capacity factor (in figure 7) Table 2. Capacity factor values from probabilistic model (PM) and
and for efficiency (in figure 8). The coefficient of variation of Monte Carlo simulation (MCS)
Capacity Factor Capacity Factor Capacity Factor
the wind speed generated range can be used to improve the vrat=11.5m/s vcut-in=3.5m/s vcut-in=3.5m/s
effectiveness of MCS, this being often used as the convergence vcut-off=20m/s vcut-off=20m/s vrat=11.5m/s
criterion in simulation techniques. vcutin PM MCS vrat PM MCS vcutoff PM MCS
Number of simulations results from condition that the 2.5 22.3301 22.2012 10 20.4575 20.3350 20 16.8492 16.7547
3 19.5020 19.6152 11 17.9203 17.9318 21 16.8492 16.9726
deviation of the coefficient of variation of CF and EF ranges 3.5 16.8492 16.6911 12 15.8886 15.6188 22 16.8493 16.7824
to expected value to be under a settled value. Using 4 14.4048 14.8282 13 14.2455 14.1276 23 16.8493 16.7034
simulations techniques, for a settled value (0,01%) is obtained 4.5 12.1901 11.9975 14 12.8995 12.8572 24 16.8493 16.8065
about 10.000 necessary samples, the convergence process of 5 10.2157 10.3042 15 11.7815 11.7896 25 16.8493 16.7920
CF being presented in the figure 7, and for EF in figure 8.
As described above, table 3 shows the efficiency values for
24 different values of cut-in, rated and cut-off turbine speeds.
WTG 1.5xle - GE Energy
Vin=3.5 m/s
22 Vrat=11.5 m/s
Table 3. Efficiency values from probabilistic model (PM) and
Voff=20 m/s Monte Carlo simulation (MCS)
Wbl(a=4.82253;b=1.8656)
Efficiency Efficiency Efficiency
20
Capacity Factor (%)

vrat=11.5m/s vcut-in=3.5m/s vcut-in=3.5m/s

vcut-off=20m/s vcut-off=20m/s vrat=11.5m/s
18 vcutin PM MCS vrat PM MCS vcutoff PM MCS
2.5 37.5769 37.3589 10 32.1705 32.2894 20 28.3533 28.4966
3 32.8178 32.6990 11 30.1561 29.7850 21 28.3535 27.9834
16 3.5 28.3536 27.7772 12 28.3536 27.7068 22 28.3536 29.1203
4 24.2402 24.2985 13 26.7372 26.9143 23 28.3537 28.5333
14
4.5 20.5134 20.8663 14 25.2835 25.4662 24 28.3537 27.9599
5 17.1909 16.3235 15 23.9722 23.5458 25 28.3537 28.2550

12
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 It can be seen that the results obtained from both methods
Number of samples
are very close. The probabilistic method provides comparative
Fig. 7 Capacity factor result from Monte Carlo simulation results with Monte Carlo simulation, these proving the
accuracy of probabilistic model, developed in equations (15)
and (16).
The analytical expressions developed in (15) and (16) were

Issue 1, Volume 5, 2011 164

 INTERNATIONAL JOURNAL OF MATHEMATICAL MODELS AND METHODS IN APPLIED SCIENCES

used to study the effect of mean wind speed on both

30
coefficients. Graphic representation of dependence is shown in Vcut-in (3.5 m/s) affected by wind speed steps

figure 9, with solid line for efficiency and dashed line for Vrated (11.5 m/s) affected by wind speed steps
Vcut-off (20 m/s) affected by wind speed steps
capacity factor. 25

0.7

Capacity Factor (%)

Capaciy factor 20
0.6 Efficiency
Capacity factor and efficiency

0.5 15

0.4
10

0.3

0.2 5
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2
wind speed step (m/s)
0.1
Fig. 10 Effect of wind turbine parameters on capacity factor
0 0.5
0 5 10 15 20
Mean wind spead (m/s) Vcut-in (3.5m/s)
0.45 Vrated (11.5 m/s)
Fig. 9 Effect of mean wind speed on CF and EF Vcut-off (20.5 m/s)

0.4
As it can be seen, the wind turbine efficiency is largest (in
Efficiency (%)

this case 50%) at a relatively low wind speeds, around some 4 0.35
m/s. But, at low wind speeds, efficiency is not so important,
because there is not much energy to be converted. At higher 0.3

wind speeds, the turbine can not convert the excess energy that
0.25
exceeds the limits of the generator. So, the efficiency is not the
best indicator for evaluating the suitability of wind turbine to a 0.2
specific location. For suitability evaluation, the capacity factor
is a better indicator. If the wind turbine is located in an area -1.5 -1 -0.5 0 0.5 1 1.5
with an average speed around some 10m/s, the maximum wind speed step (m/s)
output power will be generated, even if the efficiency is low. Fig. 11 Effect of wind turbine parameters on efficiency values
So, it is not an aim in itself to have a high technical efficiency
of a wind turbine. Since the fuel is free, the technical As it can be seen from the figure 9, a certain value of
efficiency is not important, the wind energy can be used or will capacity factor or efficiency can be achieved by changing the
be lost. What really matters is the amount of generated energy, two parameters of wind turbine generator. Most important
even with a lower efficiency parameter and providing the greatest degree of freedom is cut-
Also, the proposed model may be used to analyze the effects in wind speed. It has been shown that the cut-in wind speed
of different cut-in, rated and cut-off wind speeds on the has a significant effect on the capacity factor and efficiency
capacity factor value. Using a 1.5-XLE GE Energy wind values. Their values decrease approximately linearly as the
turbine, placed in Iasi location, with previously wind profile, cut-in wind speed increases.
the capacity factor will be 16.85% and the efficiency, 28.35%. The second parameter of wind turbine generator with effect
for this wind profile, from table 2, it can see, a wind turbine on the capacity factor and efficiency is the rated wind speed. It
generator can be expected to operate with a maximum capacity has been shown that the rated wind speed has a relatively small
factor of 22.33% and a maximum efficiency of 37.57% for a effect on these values. The rated wind speed growth leads to
wind turbine generator characterised by a wind speed the capacity factor and efficiency values decrease, but this
parameters set to vcut-in=2.5m/s, vrated=11.5m/s and vcut- effect is less significant than that of the cut-in wind speed
off=20m/s, respectively. A capacity factor of 22.33% from a It has been shown that the cut-off wind speed has no effect
1.5kW wind generator means a mean output power of 0.335 on either the capacity factor nor efficiency values. The cut-off
kW or an annual power output of 2934.6 kWh. wind speed is a safety parameter and is usually large. For
The effects of the wind turbine generator parameters on the relatively few times the instantaneous wind speed at a
capacity factor and efficiency are shown in figure 10, and particular area will be greater than the cut-off speed. The
figure 11, respectively. In the same system coordinates is selection of the cut-off speed parameter is therefore less
shown the dependence of capacity factor and efficiency for important than that of the cut-in and the rated wind speed
various wind speeds values around of wind turbine generator parameters.
parameters (cut-in, rated, cut-off speeds).

Issue 1, Volume 5, 2011 165

 INTERNATIONAL JOURNAL OF MATHEMATICAL MODELS AND METHODS IN APPLIED SCIENCES

V. CONCLUSION [14] D. Villanueva, A. Feijoo, “Wind power distributions: A review of their

applications”, Renewable and Sustainable Energy Reviews 14 (2010)
Integration of wind energy is an important activity in the 1490–1495
developing process of the electric power system. Knowing the [15] MathWorks Products. Statistics Toolbox. Function Reference, Gamma
Distribution.
capacity factor values is a key factor when examining wind
[16] L. Bain, M. Engelhardt, Introduction to Probability and Mathematical
energy potential for a wind turbine located in a specific area. Statistics, Duxbury Press, California, 1992.
The probabilistic methods are the recommended solution for [17] A. Papoulis, Probability, Random Variables and Stochastic Processes,
wind integration analysis, since they can take into account the McGraw-Hill, New York, 1984.
[18] http://www.ge-energy.com/prod_serv/products /wind_turbines/
wind power uncertainty.
This paper presents a probabilistic model to evaluate the
capacity factor and technical efficiency of a wind turbine Ciprian Nemes was born in Turda, Romania, on
based on the output power distribution. The results were May, 1975. He graduated from “Gh. Asachi”
Technical University of Iasi and received the MSc
validated using the Monte Carlo simulations, and the analysis degree in electrical engineering and PhD degree in
demonstrates that the probabilistic model results are very reliability engineering.
accurate. The model has the advantage that can be easily He is currently a Senior Lecturer and research
interests are in the area of power equipment
implemented in computer programs and require a computing reliability, power system planning based on risk
time considerably less than in the case of simulation methods. assessment, renewable energy sources operation and
The electric energy output of a wind turbine for a specific planning.
area depends on many factors. These factors include the wind
Florin Munteanu was born in Campina, Romania,
speed conditions at the area, and the characteristics of the wind on April 10, 1954. He graduated from “Gh. Asachi”
turbine generator. Technical University of Iasi and he received the MSc
The case studies show that turbine cut-in wind speed has a degree in Power Engineering in 1979 and PhD
degree in Reliability Engineering in 1995. Starting
significant effect on the both capacity factor and technical with 1984 he is with “Gh. Asachi” Technical
efficiency values while the cut-off wind speed has almost no University of Iasi where, from 1999, he is holding a
effect. Significant benefits can be obtained by selecting full time professor position and from 2008 he is also
the head of Power Engineering Department. The
suitable wind turbine parameters for the specific wind profile. main fields of interest included transients of power
systems, power quality and reliability.
REFERENCES
[1] Tony Burton, David Sharpe, Nick Jenkins, Ervin Bossanyi, Wind
Energy Handbook, John Wiley & Sons 2001.
[2] Thomas Ackermann, Wind Power in Power Systems, Ed. John Wiley &
Sons 2005.
[3] R. Billinton, W. Y. Li, Reliability assessment of electrical power
systems using Monte Carlo method, Plenum Press, New York, 1994.
[4] Nicolas Boccard, Capacity Factor of Wind Power, Realized Values vs.
Estimates, Raport project SEJ2007-60671. Depart. d’Economia,
Universitat de Girona, Spain. October 2008.
[5] Z. M. Salameh, I. Safari, Optimum windmill –site matching, IEEE
Trans. Energy Convers., vol. 4, no. 7, pp. 669-675, 1992.
[6] Tsang-Jung Changa, Yi-Long Tua, Evaluation of monthly capacity
factor of WECS using chronological and probabilistic wind speed data:
A case study of Taiwan, Renewable Energy 32, 2007, pp.1999–2010.
[7] E. Kavak Akpinar, S. Akpinar, “An assessment on seasonal analysis of
wind energy characteristics and wind turbine characteristics” Energy
Conversion and Management 46, 2005, pp. 1848–1867.
[8] Isaac Y. F. Lun, Joseph C. Lam, “A study of Weibull parameters using
long-term wind observations” Renewable Energy Journal, Volume 20,
Issue 2, June 2000, pages 145-153.
[9] J.A. Carta, P. Ramırez, “A review of wind speed probability
distributions used in wind energy analysis. Case studies in the Canary
Islands” Renewable and Sustainable Energy Reviews, volume 13, Issue
5, June 2009, pp. 933-955.
[10] Mohammad A. Al-Fawzan, “Algorithms for Estimating the Parameters
of the Weibull Distribution” Interstat Journal, October 2000.
http://interstat.statjournals.net/YEAR/2000/articles/0010001.pdf
[11] Nemes C., Munteanu F., “Optimal Selection of Wind Turbine for a
Specific Area”, OPTIM 2010, Brasov, Romania. 978-1-4244-7020-
4/10/26.00 '2010 IEEE, pp. 1224-1229.
[12] Nemes, C., Munteanu, F., “Development of Reliability Model for Wind
Farm Power Generation”, Advances in Electrical and Computer
Engineering, ISSN 1582-7445, e-ISSN 1844-7600, vol. 10, no. 2,
2010pp. 24-29.
[13] M.H. Albadia, E.F. El-Saadanyb “New method for estimating CF of
pitch-regulated wind turbines”, Electric Power Systems Research 80
(2010) 1182–1188.

Issue 1, Volume 5, 2011 166

View publication stats