Renewable Energy 46 (2012) 289e 289 e296
Contents lists available at SciVerse ScienceDirect
Renewable Energy j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m/ m/ l o c a t e / r e n e n e
Technical note
Impact of wind speed variations on wind farm economy in the open market conditions Zeljko Ðurisic , Jovan Mikulovi c, Iva Babic *
University of Belgrade, Faculty of Electrical Engineering, Bulevar kralja Aleksandra 73, Belgrade, Serbia
a r t i c l e
i n f o
Article history: Received 15 October 2011 Accepted 11 March 2012 Available online 5 April 2012 Keywords: Wind farm Electricity market Correlation index
a b s t r a c t
Stimulating measures for the production of electrical energy from renewable energy sources (RES) are guaranteed over certain period of operation. Their aim is that the wind power plants (WPP) and other RES are brought to a concurrent position with the conventional power plants at the open market of electrical electrical energy. energy. In perspective, perspective, power market conditions should be made equal for all electricity producers. In such conditions, the economy aspects of a WPP are affected not only by the amount of produced electricity but also by the daily and yearly production pro �les. The paper de�nes correlation indices between the price pro�le of electricity at the open market and seasonal and daily pro �les of WPP production. These indices could serve as a quantitative measure of the in �uence of the seasonal and daily pro�les of wind speed on the economy of the project of a WPP. 2012 Elsevier Ltd. All rights reserved.
1. Introduction
Product Production ion costs of electric electrical al energy energy in WPP in the existing existing market market condit conditio ions ns are still still higher higher compa compared red to those those of the conventional power plants. On the other hand, production costs of electr electrica icall energy energy still still do not take take into into accoun accountt the exter external nal costs costs [1] [1],, which which will signi�cantly cantly affect affect the costs costs of conventio conventional nal power power plants [1,2] plants [1,2].. Inclusion of the external costs will create conditions where RES have become competitive in the open market of electrical energy. In the existing conditions, for the purpose of motivating electrical energy production from RES, the governments of many countries have undertaken various stimulating measures for production of electrical energy from RES. The literature contains [3,4] a [3,4] a survey of different instruments for stimulation of electrical energy production from WPP and other RES. Over Over the the last last ten ten years years,, Feed-In Tariff (FIT) system systemss have have Tariff (FIT) become become an effective effective instrument instrument of generating generating electricity electricity from RES, especially through WPP production. The central principle of FIT policies is to offer guaranteed prices for � xed periods of time for electricity electricity produced produced from WPP [5,6]. [5,6]. Throug Through h this system system,, (TSO) and Distributio Transmiss Transmission ion System Operator Operator (TSO) Distribution n System (DSO) are required to buy the energy generated by WPP. Operator (DSO) This This mechani mechanism, sm, in variou variouss forms, forms, is widely widely accepted accepted by the majority of European countries [5] countries [5],, including Germany, Spain, and Denmark [7] [7],, the leading leading countri countries es in Europ Europe e as regard regardss wind wind *
Corresponding author. E-mail address:
[email protected] ( Z. Ðurisic).
0960-1481/$ e see front matter 2012 Elsevier Ltd. All rights reserved. doi:10.1016/j.renene.2012.03.015 doi:10.1016/j.renene.2012.03.015
power penetration in electricity production [8] [8].. About 86% of all wind onshore capacity installed by the end of 2009 in Europe was initiated initiated by FIT systems [9] systems [9].. Fig. 1 shows the remuneration levels and the period of guaranteed support for onshore WPP for � xed option of FIT in several European countries. The level and duration of FIT for offshore WPP is equal or higher compared to onshore WPP. Detailed data concerning FIT for all EU countries are available in the literature [5] [5].. In some countries the level of FIT is not �xed. The tariff range in Greece shows the difference between the tariffs on mainland and autonomous islands. In Germany a system where the tariff level varies varies accord according ing to the wind wind yield yield is applie applied. d. For Ireland Ireland and Slovenia, the tariff levels depend on the plant size. In Slovakia the FIT rates are set annually depending on the index of national core in�ation [9] ation [9].. In accordance with the FIT system, the price and disposal of the produced electrical energy in WPPs are guaranteed over a �xed period of time. For majority of European countries that period is within the range from 10 to 20 years [5] [5],, which is considerably shorter compared to lifetimes of wind turbines, usually designed to be from 20 to 30 years [10]. [10]. Upon Upon expirati expiration on of the period period of guaran guarante teed ed price, price, a WPPwill offer offer the produ producedelect cedelectric rical al energy energy to the open market, where the price of electrical energy is usually formed on 1 h basis, according to the principle of demand and supply. The prices of electrical energy on the open market mainly follow follow the diagram diagram of consump consumption tion and may vary consider considerably ably within one day compared to the corresponding average daily price [11,12].. [11,12]
290
Z. Ðurisic et al. / Renewable Energy 46 (2012) 289e 296
Fig. 1. Level and duration of FIT of onshore WPPs commissioned in 2009 in several European countries.
Production of electrical energy from conventional power plants is mainly planned and can be adjusted to the market demands whereby the best price can be obtained for a produced kWh. Production of electrical energy in a WPP can be, by applying well developed models for short and long term wind forecasting [13], envisaged to a great extent, but it cannot be planned since it is governed by the wind. In other words, production pro �le of a WPP follows time variations of the wind speed which cannot be planned. Nevertheless, a wind is not a stochastic phenomenon, but it is a consequence of the complex thermal and other physical phenomena in the atmospheric layer close to earth surface. Daynight variations as well as seasonal variations induce periodic variations of the physical phenomena which have an in �uence on the wind characteristics. These periodic climatic variations statistically create typical daily and seasonal pro �les of wind speed. For the purpose of assessing economy of production of a WPP, it is necessary that during the planning phase,in addition to the making an estimate of its average annual production, typical time diagram of the expected WPP production power on the daily and monthly levels are analyzed. The basic idea of this work is introducing a correlation index between typical time diagram of WPP production and typical diagram of the price of electrical energy at the open market. This index may be useful during evaluation of economic parameters of a WPP project which will one part of its lifetime operate under the open market conditions. 2. Gross income estimate of a WPP
Gross income during the designed lifetime is one of the key elements for assessing the economy of a WPP project. In order to assess the gross income, it is necessary to know the expected Annual Electricity Production (AEP) for an average year during plant s lifetime and the price of electrical energy per produced MWh. In this analysis it will be assumed that during the expected lifetime of N LT years, the WPP operates N FIT years with subventions according to the FIT model and N OM ¼ N LT N FIT years at the open market, i.e. without subventions. Total gross income during the designed lifetime of a WPP can be represented by the following relation: ’
W LT ¼ W FIT þ W OM ;
(1)
where W FIT is the gross income during the period of guaranteed subventions (N FIT ) and W OM is the gross income during the period without subventions (N OM ). 2.1. Gross income estimate of a WPP during FIT period
During application of the FIT system, the gross income can simply be estimated according to relation:
W FIT ¼ N FIT $P FIT $ AEP ;
(2)
where P FIT (Euro/MWh) is guaranteed � xed price according to the FIT model over the � xed period of N FIT years, AEP (MWh) is average net estimated annual WPP production. AEP is usually estimated on the basis of the measurements of wind potential by applying some of the dedicated software, such as WAsP [14]. 2.2. Gross income estimate of a WPP under open market conditions
The price of electrical energy at the open market is usually formed on 1 h basis, according to the principle of demand and supply. Hourly price pro�le at the electricity market mainly follows the hourly consumption pro�le. Therefore, for the purpose of estimating gross income of a WPP for an average year under open market conditions it is essential to estimate not only AEP but also to estimate time diagram of WPP production. Estimate of an idealized gross income W OMideal of a WPP over the period N OM can formally be expressed by the following mathematical relation:
W OMideal ¼ N OM $P OMa g $ AEP $C WPM ; v
(3)
where P OMavg is the expected average annual price of electrical energy at the electricity market over N OM years, C WPM ( Correlation Wind Power Market Index) is index of correlation between time diagram of variation of the price of electrical energy at the electricity market and diagram of electrical energy production of WPP. The gross income estimate given by relation (3) is idealized because it assumes that all energy produced by the WPP is sold at the electricity market at the maximum hourly price. There are two main reasons which cause that the real gross income is lower than the estimate given by relation (3). Firstly, under real conditions there is a day ahead WPP production forecasting error. This error in
Z. Ðurisic et al. / Renewable Energy 46 (2012) 289e 296
the production estimate lowers the price due to penalization factors [15]. Secondly, the auctioned price of the produced MWh is obtained on the basis of different market indices [16] and, as a rule, it is lower compared to the maximum realized market price for the corresponding hour, causing that the realized income is smaller than the idealized. Taking these real factors into account can be mathematically made equivalent to the average coef �cient of income reduction m < 1, thus the real gross income is:
W OM ¼
m$W OMideal
¼
m$N OM $P OMav g $ AEP $C WPM :
(4)
Value of the reduction coef �cient m is dependent on the qualities of the forecast of production and forecast of hourly prices at the market. The aim of this work is de �ning a model for estimation of a dimensionless index C WPM which is of key importance for assessing the economy of a WPP under conditions of the open market of electrical energy. For a �xed pro�le of prices at the electricity market, index C WPM is de �ned by the wind pro�le exclusively, thus it is indicator of the wind quality at a particular location. De�ning this coef �cient allows investors to obtain a better insight into the economy of a WPP project. This coef �cient may also be of signi�cance for the donor of subvention during the period of subvention of production (e.g. applying FIT model) since projects having higher C WPM provide a higher degree of return of the means of subvention. In addition, since the correlation between the price at the market and diagram of consumption is quite strong, correlation index C WPM indirectly represent the correlation index with consumption demands. For these reasons, it can be expected that a WPP having higher C WPM will act positively as regards the losses in the distribution network, unloading of the element of the distribution system, as well as organizing production units and providing power reserve in a power system. 3. Mathematical model for estimation of correlation index C WPM
Correlation index is a unique indicator for each WPP at certain electricity market. Its estimate can be carried out on the basis of the available data on wind speed at the target location for a period of at least one year and historical data on hourly prices of electrical energy at the electricity market where the produced electrical energy of a WPP is to be sold. A quantitative estimate of this index can be done on the basis of the following relations:
(5)
C WPM ¼ C WPMd $C WPMs ;
where: C WPMd is correlation index of the average daily diagram of WPP production with the average daily diagram of the price of electrical energy at the electricity market, C WPMs is seasonal correlation index which determines the correlation between the average monthly production of a WPP and average monthly prices of electrical energy at the electricity market. Index C WPMd can be calculated applying the following relation:
C WPMd ¼
24 j ¼ 1 E j $P OMj
P
24$E a g $P OMa g v
v
24
¼
1 e $ p ; 24 j ¼ 1 j OMj
X
291
(6)
where: E j is average hourly production power of a WPPfor j-th hour of an average day expressed in MW, E avg is average annual production power of a WPP, P OMj is price of the produced MWh at the electricity market for the corresponding j-th hour of the average day. Values e j ¼ E j/E avg and pOMj ¼ P OMj/P OMavg are the corresponding normalized values of E j and P OMj respectively. Index C WPMs can be calculated according to the following relation:
12 m ¼ 1 T m $E m $P OMm
P
C WPMs ¼
8760$E a g $P OMa g v
v
12
¼
X t $e $ p m
m
OMm;
(7)
m¼1
where: E m is average monthly production power of WPP for month m, P OMm is average monthly price of one MWh at the electricity market for month m, T m is number of hours in month m. Values em ¼ E m/E avg and pOMm ¼ P OMm/P O M avg are the corresponding normalized values of E m and P OMm respectively and t m ¼ T m/8760 is relative number of hours in month m with respect to one year. It is important to note that in Eqs. (6) and (7) the normalized values e and p, on both daily and monthly levels, are not dependent upon the total production of WPP and price of one MWh at the market, but only on the pro �le of the production diagram and pro�le of the consumption diagram. Also, it is important to note that C WPM can be lower, equal, or higher than one. The higher index C WPM the wind is of a better quality, i.e. it is better correlated with the market prices of electrical energy. A WPP production pro �le and the pro �le of prices at an electricity market are different on the year by year basis, thus their correlation is variable. In order to estimate a unique correlation index over the lifetime of a WPP, it is necessary to de �ne a representative pro�le of the WPPproduction and a representative pro�le of prices at the analyzed electricity market. 3.1. Estimate of the time diagram of a wind turbine production
In order to make an estimate of the time diagram of WPP production it is required to possess measurement data on wind speed at the target location for the period of at least one year. On the basis of these data and power curve of the selected turbine it is possible to make an estimate of the wind turbine production for each hour of the analyzed year. The power curve of wind turbine is usually given for a �xed air density (standard is 3 r0 ¼ 1.225 kg/m ). In order to determine production power of a wind turbine for a 10 min interval, it is necessary to know the actual air density and wind speed at the hub height of the selected wind turbine. In this analysis it has been assumed that the wind turbine was located at the place of the measurement mast. If measurements of the wind speed are performed at the heights which are lower than the hub height of the selected wind turbine, it is necessary to perform extrapolation of the wind speed according to some of the models for height pro �le of the wind speed [17,18]. For each 10 min interval i it is necessary to calculate the corresponding wind speed V i at the hub height. Then, calculation of the corresponding effective wind speed is done according to the following relation [19]:
V eff i ¼ V i
1 3
ri
r0
(8)
;
where ri is the actual air density at the hub height of the wind turbine. The air density can be calculated on the basis of the air temperature and air pressure measurements by using the following equation: ri
¼
pi kg=m3 ; RT i
(9)
where: pi is the airpressure (Pa) over the 10-min time interval i, T i is the air temperature over the 10 min time interval i in degrees of Kelvin, and R is the speci �c gas constant for air (287J/kgK). The electrical power of the wind turbine for each 10 min interval i is estimated according to the following relation:
292
Z. Ðurisic et al. / Renewable Energy 46 (2012) 289e 296
E i ¼ E powercur e V effi ; v
(10)
where E power curve(V ef �) is the standard power curve of the wind turbine for the � xed air density r 0 ¼ 1.225 kg/m3). It should be mentioned that the previous approach to estimation of the production power is acceptable for a pitch-controlled wind turbine. For a stall-controlled wind turbine the power output predicted by the given power curve for the measured wind speed is calculated � rst and then the power output is adjusted to the following equation [19]:
E i ¼ E power cur e ðV i Þ v
ri r0
:
(11)
The average annual production power ( E avg ) of a WPP is estimated on the basis of the following relation:
E a g ¼ v
1 N E : N i ¼ 1 i
X
(12)
where N is the total number of observed 10 min intervals. The capacity factor (CF) of a WPP at annual level is calculated according to the following relation:
CF ¼
E a g ; E n v
(13)
where E n is rated power of WPP. Estimates of the production and CF of a WPP according to relations (10e13) concern an idealized production since the losses and unavailability of wind turbine are not taken into account. If it is assumed that these losses could be modeled by a � xed coef �cient and that they are linearly distributed in time, then in relations (6) and (7) they would be compensated, thus for C WPM calculation they do not have to be taken into account. Owing to the variation of the production pro �le on a year by year basis, it is required to analyze measurement data taken over a long period of time. For this purpose, in addition to the short-term measurements, one can use the long term measurements of the regional reference meteorological station.
3.2. Time diagrams of the prices of electrical energy at electricity markets
The prices of electrical energy at electricity markets are usually de�ned at hourly level on the basis of demand and supply according to the day-ahead principle. Daily variations of hourly prices mainly follow variations of the daily consumption, as is shown in [16] by a comparative analysis of the average hourly prices at the auctions in Spain (OMEL) and Germany (EEX) and the corresponding average hourly loads in Spain and Germany. In addition to the structure of consumption, daily diagram is in �uenced by the structure of production. Over the past several years the ever increasing presence of intermittent renewable energy sources (such as WPP and PV) has a signi �cant in�uence on the pro�le of price of electrical energy at some auctions in the regions having installed signi�cant capacities of renewable energy sources. These effects have been analyzed in [20 e23]. Fig. 2 shows average normalized hourly prices of electrical energy at EEX for an average day for every year over the period 2007e2011. On the basis of the average daily diagrams, the corresponding normalized average day over the period of several years has been formed. During peak hours the electrical energy is on average more expensive by up to 30% compared to the average daily price for an average day, while during minimum load it is on average cheaper by up to 40%. It is clear that WPPs located in the regions where winds predominantly blow during midday hours will accomplish higher income per MWh than WPPs located in the regions where winds predominantly blow during early morning hours. Fig. 3 shows average normalized price of electrical energy at EEX per month for each year over the period 2007 e2011. The seasonal pro�le of prices at EEX is characterized by relatively small deviations of average monthly prices from the average annual value. In addition, a relatively large deviation of average monthly prices on the year by year basis is noticeable. 4. The example
This example presents the calculation of the correlation index for the project Bavanistansko Polje e Serbia of rated power 188 MW
Fig. 2. Average normalized hourly electricity price at EEX for each year over the period 2007 e2011.
Z. Ðurisic et al. / Renewable Energy 46 (2012) 289e 296
293
Fig. 3. Average normalized monthly electricity price at EEX over the period 2007e2011.
[24]. The analyses have been carried out assuming that WPP Bavanistansko Polje has been realized by using wind turbines Vestas V90, 2 MW, H ¼ 105 m. On the basis of two-years-long measurements (2009e2010) of wind speed at the location of WPP and the methodology described in Section 3.1, estimates of the daily pro�le of the WPP power production have been obtained. The estimated net AEP of WPP Bavanistansko Polje is 484.2 GWh. Fig. 4 shows the estimated normalized average hourly productions of WPP Bavanistansko Polje for years 2009 and 2010. It can be concluded that the average hourly production pro�les of WPP Bavanistansko Polje for years 2009 and 2010 are very similar. Thus, it can be assumed that the daily production pro �le of WPP Bavanistansko Polje is very stable on the year by year basis. Fig. 5 shows concurrently the normalized average hourly diagram of the estimated WPP production and normalized average
hourly prices of electrical energy at EEX for an average day over the two years period 2009e2010. Fig. 5 shows concurrently the normalized average monthly productions of WPP Bavanistansko Polje and normalized average monthly prices of electrical energy at EEX. By a comparative analysis of the diagram of Fig. 5 one can conclude that the correlation between the average daily variation of the price of electrical energy and average daily variation of the production of WPP Bavanistansko Polje is relatively weak. On the basis of relation (6) the calculated average daily correlation index is C WPMd ¼ 0.973. On the basis of relation (7) the calculated average seasonal correlation index is C WPMs ¼ 0.999. Finally, the correlation index between the production of WPP Bavanistansko Polje and prices at EEX electricity market is, according to relation (5), C WPP C WPMd ¼ 0.973. Based on the previous analysis and relation (4), one z
Fig. 4. The estimated normalized average hourly productions of WPP Bavani stansko Polje for years 2009 and 2010.
294
Z. Ðurisic et al. / Renewable Energy 46 (2012) 289e 296
Fig. 5. The estimated normalized average hourly productions of WPP Bavani stansko Polje and normalized hourly prices of electrical energy at EEX for an average day over 2009e2010.
may conclude that for making an estimate of the gross income of WPP Bavanistansko Polje, under condition of selling the produced energy at EEX market, relevant is the average annual price of MWh reduced by 2.7%. It is possible to perform the preceding analysis for any market where the energy produced by a WPP could be sold and then calculate the corresponding correlation indices. Under free market conditions, it is possible to sell energy from a WPP at the market which has the statistical correlation which matches best the production pro�le of the WPP. 5. Variation of the correlation index on the year by year basis
Correlation index C WPM is a statistical parameter which is subject to certain variations on the year by year basis due to variations of the production pro�leof a WPP and marketpro�le. On the basis of the data on average hourly prices of electrical energy at EEX
for each year, for the period 2007e2011 (Fig. 2) the corresponding statistical deviation has been calculated. The results are shown in Fig. 7. It can be concluded that the average daily pro�le of the price of electrical energy is quite stable on the year by year basis, within the analyzed period of �ve years, the standard deviation of average hourly values being within the range from 3 to 10%. Fig. 8 shows standard deviation of the average monthly prices at EEX on the year by year basis over the period 2007e2011. Compared to the daily pro �le, the seasonal pro�le is not stable on the year by year basis, the dispersion being more pronounced over the cooler period of the year. Having in mind statistically small dispersions of the average monthly prices fro the annual average and a high discretion on the year by year basis, for EEX market one can adopt the assumption that the seasonal correlation index is C WPMs ¼ 1. It should be mentioned here that this conclusion couldn t in advance be made generally applicable to all markets. ’
Fig. 6. The normalized average monthly productions of WPP Bavani stansko Polje and normalized average monthly prices of electrical energy at EEX.
Z. Ðurisic et al. / Renewable Energy 46 (2012) 289e 296
295
Fig. 7. Standard deviation of the average daily pro �le of prices at EEX on the year by year basis over the period 2007 e2011.
From the previous analysis of EEX market, it can be adopted that C WPM ¼ C WPMd. From Figs. 2and 4 one may conclude that daily pro�les of the EEX market prices and production of the analyzed WPP are very stable, thus the correlation index is very stable on the year by year basis. 6. The limiting values of the correlation index
For each electricity market it is possible to de �ne theoretical limiting values of the correlation index C WPM . The limiting values C WPMmin and C WPMmax are functions of the capacity factor CF of a WPP, which is de �ned by relation (13). Theoretically minimal correlation index would be obtained if all AEP of a WPP was generated during hours of minimal price at the considered electricity market. C WPMmax is de�ned in the analogous way. Calculation of the limiting values of the correlation index is
carried out on the basis of the histograms of the normalized prices at the considered auction. Fig. 6 shows range of possible values of the correlation index C WPM at EEX for different values of CF of a WPP. Fig. 9 is unique for auction EEX. Similar diagrams can be made for any electricity market. For certain WPP these diagrams allow estimation of the corresponding limiting values of the correlation index if the capacity factor of the WPP is known. On the basis of the methodology described in Section 3 it is possible to estimate correlation index C WPM . By a comparative analysis of the estimated C WPM and the corresponding C WPMmin and C WPMmax it is possible the quality of wind in view of the daily and seasonal pro �les. E.g. for WPP Bavanistansko Polje, analyzed in Section 4, the estimated values CF ¼ 29.4% and C WPM ¼ 0.973 have been obtained. For the calculated CF , onthebasisof Fig. 9 it can be seen that the correlation index at EEX can be within the range 0.7e1.25.
Fig. 8. Standard deviation of average monthly prices at EEX over the period 2007 e2011.
296
Z. Ðurisic et al. / Renewable Energy 46 (2012) 289e 296
Fig. 9. Range of values of the correlation index C WPM at EEX for different values of capacity factor ( CF ) of a WPP.
7. Conclusion
For the purpose of assessing the economy of the project of a WPP under open market conditions, it is necessary to analyze the correlation between the production diagram and diagram of the market prices, in addition to making an estimate of AEP . The paper de�nes correlation index between the production pro�le and prices pro�le at an electricity market. This index represents a measure of quality of the wind as regards the daily and seasonal variations. For the purpose of estimation this index ( C WPM ) one should have measurement data concerning average daily and seasonal pro�les of the wind at the WPP location, as well as the average daily and seasonal pro�le of prices at the considered auction. For each electricity market it is possible to de �ne theoretical limiting values of the correlation index C WPM for a WPP which are functions of the capacity factor CF of the WPP only. By a comparative analysis of the estimated C WPM of a WPPand the corresponding limiting values C WPMmin and C WPMmax the quality of wind as regards the daily and seasonal pro�les can be estimated. References [1] El-Kordy MN, Badr MA, Abed KA, Ibrahim SMA. Economical evaluation of electricitygeneration considering externalities. RenewableEnergy 2002;25:317e28. [2] Owen AD. Renewable energy: externality costs as market barriers. Energy Policy 2006;35:632e42. [3] Saidur R, Islam MR, Rahim NA, Solangi KH. A review on global wind energy policy. Renewable and Sustainable Energy Reviews 2010;14:1744e62. [4] Madlener R, Stagl S. Sustainability-guided promotion of renewable electricity generation. Ecological Economics 2005;53:147e67. [5] Klein A, Merkel E, P�uger B, Held A, Ragwitz M, Resch G, et al. Evaluation of different feed-in tariff design options e best practice paper for the international feed-in cooperation. The Report of a research project funded by the German Ministry for the Environment, Nature Conservation and Nuclear Safety (BMU). 3rd ed.; December 2010. [6] Couture T, Gagnon Y. An analysis of feed-in tariff remuneration models: implications for renewable energy investment.Energy Policy 2010;38:955e65. [7] Ciarreta A, Gutiérrez-Hitab C, Nasirovc S. Renewable energy sources in the Spanish electricity market: instruments and effects. Renewable and Sustainable Energy Reviews 2011;15:2510e9.
[8] Michalak P, Zimny J. Wind energy development in the world, Europe and Poland from 1995 to 2009; current status and future perspectives. Renewable and Sustainable Energy Reviews 2011;15:2330e41. [9] Ragwitz M, Stricker E, Resch G. Recent experiences with feed-in tariff systems in the EU e A research paper for the International feed-in cooperation. A report commissioned by the German Ministry for the Environment, Nature Conservation and Nuclear Safety (BMU); November 2010. [10] Crawford RH. Life cycle energy and greenhouse emissions analysis of wind turbines and the effect of size on energy yield. Renewable and Sustainable Energy Reviews 2009;13:2653e60. [11] Viehmann J. Risk premiums in the German day-ahead electricity market. Energy Policy 2011;39:386e94. [12] Schlueter S. A long-term/short-term model for daily electricity prices with dynamic volatility. Energy Economics 2010;32:1074e81. [13] Foley AM, Leahy PG, Marvuglia A, McKeogh EJ. Current methods and advances in forecasting of wind power generation. Renewable Energy 2012;37:1 e8. [14] Mortensen NG, Landberg L, Troen I, Petersen EL. Wind atlas analysis and application program (WAsP). Roskilde, Denmark: Riso National Laboratory; 1993. [15] Brunetto C, Tina G. Wind generation imbalances penalties in day-ahead energy markets: the Italian case. Electric Power Systems Research 2011;81: 1446e55. [16] Falbo P, Fattore M, Stefani S. A new index for electricity spot markets. Energy Policy 2010;38:2739e50. [17] Lackner MA, Rogers AL, Manwell JF, McGowan JG. A new method for improved hub height mean wind speed estimates using short-term hub height data. Renewable Energy 2010;35:2340e7. [18] Gualtieri G, Secci S. Wind shear coef �cients, roughness length and energy yield over coastal locations in southern Italy. Renewable Energy 2011;36: 1081e94. [19] Dahmouni AW, Salah MB, Askri F, Kerkeni C, Nasrallah SB. Assessment of wind energy potential and optimal electricity generation in Borj-Cedria, Tunisia. Renewable and Sustainable Energy Reviews 2011;15:815e20. [20] Chao H. Ef �cient pricing and investment in electricity markets with intermittent resources. Energy Policy 2011;39:3945e53. [21] Woo CK, Horowitz I, Moore J, Pacheco A. The impact of wind generation on the electricity spot-market price level and variance: the Texas experience. Energy Policy 2011;39:3939e44. [22] Cutler NJ, Boerema ND, MacGill IF, Outhred HR. High penetration wind generation impacts on spot prices in the Australian National Electricity Market. Energy Policy 2011;39:5939e49. [23] Cruz A, Munoz A, Zamora JL, Espínola R. The effect of wind generation and weekday on Spanish electricity spot price forecasting. Electric Power Systems Research 2011;81:1924e35. [24] Obradovic M, Ðurisi c Z, Zindovi c M. Development of a 188 MW wind farm Bavanistansko Polje in Serbia. Proc. of European wind energy Conference (EWEC 2009), Marseille, France, Marth; 2009. “
”
