WIND ENERGY. EXPLAINED. Theory, Design and Application. Second Edition. J. F. Manwell and J. G. McGowan. Department of Mechanical and Industrial. Wind energy's bestselling textbook- fully revised. This must-have second edition includes up-to-date data, diagrams, illustrations and thorough. View Table of Contents for Wind Energy Explained. Wind Energy Explained: Theory, Design and Application. Author(s). J.F. Manwell · J.G.
|Language:||English, Indonesian, German|
|ePub File Size:||21.58 MB|
|PDF File Size:||18.68 MB|
|Distribution:||Free* [*Register to download]|
Library of Congress Cataloging-in-Publication Data Manwell, J. F. Wind energy explained: theory, design, and application / James Manwell, Jon McGowan. Wind energy explained: Theory, Design, and application [Book Review]. Article ( PDF Available) in IEEE Power and Energy Magazine 1(6) 51 · December with 4, Reads by J.F. Manwell, J.G. McGowan, and. Download as PDF, TXT or read online from Scribd Wind energy explained: theory, design, and application / James Manwell, Jon McGowan, Anthony Rogers .
There is a pressure gradient force in the vertical direction. These include the classic references Putnam as well as books by Eldridge In addition to the pressure gradient and gravitational forces.
Spera and Burton et al. The circulation of the atmosphere that results from uneven heating is greatly influenced by the effects of the rotation of the earth at a speed of about kilometers per hour at the equator. The variation in incoming energy sets up convective cells in the lower layers of the atmosphere the troposphere. In a simple flow model. These affect prevailing near surface winds. As shown in Figure 2.
Design and Application There are a number of other sources of information on wind characteristics as related to wind energy. At the same time. It should be noted that this model is an oversimplification because it does not reflect the effect that land masses have on the wind distribution. These include pressure forces. These include the work of Justus Hiester and Pennell Johnson The Coriolis force per unit mass. The direction of the Coriolis force is perpendicular to the direction of motion of the air.
Coriolis force. The resultant of these two forces. Above the boundary layer. The resulting wind. This force decreases as the height above the ground increases and becomes negligible above the boundary layer defined as the near earth region of the atmosphere where viscous forces are important. This imposes a further force on the wind.
In reality. The gradient wind is also parallel to the isobars and is the result of the balance of the forces: Friction at the surface causes the wind to be diverted more toward the low-pressure region. Smaller scale atmospheric circulation can be divided into secondary and tertiary circulation see Rohatgi and Nelson.
Examples of tertiary circulation. During the day. Tertiary circulations are small-scale. These different surfaces can affect the flow of air due to variations in pressure fields. Secondary circulations include the following: The oceans act as a large sink for energy.
All these effects lead to differential pressures which affect the global winds and many of the persistent regional winds. Reproduced by permission of Alternative Energy Institute. An understanding of these wind patterns. These include sea breezes and mountain winds. The direction reverses at night. Figure 2. Secondary circulation occurs if the centers of high or low pressure are caused by heating or cooling of the lower atmosphere.
Wind Characteristics and Resources 27 land masses. These include the following: A review of each of these categories as well as comments on wind speed variation due to location and wind direction follows.
Meteorologists generally conclude that it takes 30 years of data to determine long-term values of weather or climate and that it takes at least five years to arrive at a reliable average annual wind speed at a given location. The ability to estimate the inter- annual variability at a given site is almost as important as estimating the long-term mean wind at a site. They can have a large effect on long-term wind turbine production.
Time scale Figure 2. Inter-annual Inter-annual variations in wind speed occur over time scales greater than one year. As will be discussed in later sections. Reproduced by permission of ASME 2. Design and Application Spring and summer maxima occur in the wind corridors of Oregon.
Wind Characteristics and Resources 29 Figure 2. A typical diurnal variation is an increase in wind speed during the day with the wind speeds lowest during the hours from midnight to sunrise. Researchers are still looking for reliable prediction models for long-term mean wind speed. Annual Significant variations in seasonal or monthly averaged wind speeds are common over most of the world.
Winter maxima occur over all US mountainous regions. The largest diurnal changes generally occur in spring and summer. It is interesting to note that this figure clearly shows that the typical behavior of monthly variation is not defined by a single year of data. Diurnal Time of Day In both tropical and temperate latitudes.
Spring maxima occur over the Great Plains. Daily variations in solar radiation are responsible for diurnal wind variations in temperate latitudes over relatively flat land areas. The complexities of the interactions of the meteorological and topographical factors that cause its variation make the task difficult. This variation can be explained by mixing or transfer of momentum from the upper air to the lower air.
It is generally accepted that variations in wind speed with periods from less than a second to ten minutes and that have a stochastic character are considered to represent turbulence.
Design and Application Figure 2. For wind energy applications. Turbulence and its effects will be discussed in later sections of this chapter. As illustrated in Figure 2. More details on these factors as related to turbine design are discussed in Chapters 6 and 7 of this text.
Reproduced by permission of Alternative Energy Institute vary with location and altitude above sea level. Turbulence can be thought of as random wind speed fluctuations imposed on the mean wind speed. Texas Rohatgi and Nelson. These fluctuations occur in all three directions: Ten-minute averages are typically determined using a sampling rate of about 1 second. Short-term variations usually mean variations over time intervals of ten minutes or less.
Short-term Short-term wind speed variations of interest include turbulence and gusts. Although gross features of the diurnal cycle can be established with a single year of data.. Wind turbine structural loads caused by gusts are affected by these four factors. Wyoming Hiester and Pennell. Wind Characteristics and Resources 31 10 January Wind speed.
The graph shows monthly and five-year mean wind speeds for two sites 21 km apart. Short-term direction variations are the result of the turbulent nature of the wind. Seasonal variations may be small.
Variations in Wind Direction Wind direction also varies over the same time scales over which wind speeds vary. These short-term variations in wind direction need to be Figure 2. Power from the wind is proportional to the area swept by the rotor or the rotor diameter squared for a conventional horizontal axis wind machine. Crosswinds due to changes in wind direction affect blade loads. Wind Characteristics and Resources 33 considered in wind turbine design and siting.
From the continuity equation of fluid mechanics. Yawing causes gyroscopic loads throughout the turbine structure and exercises any mechanism involved in the yawing motion. The wind power density is proportional to the density of the air.
For standard conditions sea- level. The actual power production potential of a wind turbine must take into account the fluid mechanics of the flow passing through a power-producing rotor. Horizontal axis wind turbines must rotate yaw with changes in wind direction. The wind power density is proportional to the cube of the wind velocity. The average wind power density. Design and Application Table 2. In practice. It is important to distinguish between the different types.
Table 2. If annual average wind speeds are known for certain regions. Using estimates for regional wind resources. Some sample qualitative magnitude evaluations of the wind resource are: The energy pattern factor is calculated from: More accurate estimates can be made if hourly averages.
The economic potential is the technical potential that can be realized economically. The World Energy Council One such estimate World Energy Council. On a global basis. Implementation potential takes into account constraints and incentives to assess the wind turbine capacity that can be implemented within a certain time frame. The majority of this land was in the West. Elliot et al. For worldwide wind resource assessments. These authors conclude that.
This is based on the meteorological potential. More recent work on the world technical and economic potential of onshore wind energy is summarized in the paper of Hoogwijk et al. In comparison. For offshore wind farms. Numerous wind resource estimates have been made for the potential of wind energy in the United States. In order to provide this fraction of the US electrical demand about billion kWh per year. In this study Gustavson based his resource estimate on the input of the solar energy reaching the earth and how much of this energy was transformed into useful wind energy.
As noted in Volume 1 of Wind Energy: The technical potential is calculated from the site potential. Wind Characteristics and Resources 35 of wind energy potential that can be estimated. They also feature expanded data collection methods and improved analysis techniques. The estimates published in the s are much more realistic than earlier ones. This is equivalent to the available wind resource. In addition to variations due to the atmospheric stability. This variation of wind speed with elevation is called the vertical profile of the wind speed or vertical wind shear.
There are at least two basic problems of interest with the determination of vertical wind profiles for wind energy applications: Rotor blade fatigue life is influenced by the cyclic loads resulting from rotation through a wind field that varies in the vertical direction.
It should be noted that these are separate and distinct problems. For example.. Air density. Design and Application 0. The density of dry air can be determined by applying the ideal gas law. On the other hand. In wind energy engineering the determination of vertical wind shear is an important design parameter since: Instantaneous variation in wind speeds as a function of height e.
These factors will be discussed in the next sections. Seasonal variation in average wind speeds as a function of height e. Air density as a function of moisture content can be found in numerous books on thermodynamics such as Balmer Atmospheric stability is usually classified as stable.
The stability of the atmospheric boundary layer is a determining factor for the wind speed gradients e. Air pressure decreases with elevation above sea level. The international standard atmosphere assumes that the sea-level temperature and pressure are The pressure in the international standard atmosphere up to an elevation of m is very closely approximated by: The negative sign results from the convention that height.
As will be shown in the following analysis. Of course. Moist air is slightly less dense than dry air. If the atmosphere is approximated as a dry no water vapor in the mixture ideal gas.
The first law of thermodynamics for an ideal gas closed system of unit mass undergoing a quasi-static change of state is given by: A summary of how the atmospheric temperature changes with elevation assuming an adiabatic expansion follows. Above zi the temperature profile reverses. Assume the standard rate of 0. For comparative purposes. Using conventional nomenclature. The air is heated near the ground.
The surface layer of air extending to zi is called the convective or mixing layer. The concept of atmospheric stability is illustrated by considering the upward displacement of a small element of air to an altitude with a lower ambient pressure.
The small element of air being lifted in this example will cool at the dry. The temperature profile before sunrise the solid line decreases with increasing height near the ground and reverses after sunrise dashed line. Different temperature gradients create different stability states in the atmosphere. When one does exist.
Each component is frequently conceived of as consisting of a short-term mean wind. The lateral component perpendicular to U is v z. A summary and examples of these properties follows. Turbulent wind consists of longitudinal. More details concerning them are given in Appendix C and in the texts of Rohatgi and Nelson and Bendat and Piersol The sample would be denser and would tend to return to its original level.
This atmospheric state is called stable. If the test element of air had the same temperature as the surrounding air at the start. The longitudinal component. This explains the need for the daily balloon soundings taken at major airports worldwide to determine the actual lapse rate.
Turbulent wind may have a relatively constant mean over time periods of an hour or more. These features are characterized by a number of statistical properties: To generalize. One should note that the standard international lapse rate seldom occurs in nature. It is defined by the ratio of the standard deviation of the wind speed to the mean wind speed.
This time period is usually taken to be ten minutes. Assuming that the sample interval is dt. In this calculation both the mean and standard deviation are calculated over a time period longer than that of the turbulent fluctuations. In equation form: The sample rate is normally at least once per second 1 Hz.
The short-term mean wind speed can then be expressed in sampled form as: For the sake of clarity. The turbulence intensity. The data.
The lateral and vertical components can be decomposed into a mean and a fluctuating component in a similar manner. The length of this time period is normally no more than an hour. In general. Note that the short-term mean wind speed. Design and Application superimposed fluctuating wind of zero mean. Wind Characteristics and Resources 41 14 13 Wind speed. The Gaussian probability density function that represents the 0.
The normal probability density function for continuous data in terms of the variables used here is given by: The probability density function that best describes this type of behavior for turbulence is the Gaussian. Experience has shown that the wind speed is more likely to be close to the mean value than far from it.
A measure of the average time over which wind speed fluctuations are correlated with each other is found by integrating the autocorre- lation from zero lag to the first zero crossing. Gusts are relatively coherent well correlated rises and falls in the wind.
The autocorrelation function can be used to determine the integral time scale of turbulence as described below. The single resulting value is known as the integral time scale of the turbulence.
As described in Appendix C. While typical values are less than ten seconds. The graph also includes the von Karman power spectral density function described above for comparison. Wind Characteristics and Resources 43 Multiplying the integral time scale by the mean wind velocity gives the integral length scale. Of most importance here. L is the integral length scale. This is referred to as the von Karman psd elsewhere in this text.
The mean wind velocity is A suitable model that is similar to the one developed by von Karman for turbulence in wind tunnels Freris. These sinusoidal variations will have a variety of frequencies.
The power spectral density of the sample wind data above is illustrated in Figure 2. Appendix C gives details of how to determine psds. Based on the autocorrelation function illustrated above. Other psds are also used in wind engineering applications see the discussion of standards in Chapter 7.
A number of power spectral density functions are used as models in wind energy engineering when representative turbulence power spectral densities are unavailable for a given site.
This type of analysis originated in electrical power applications. The integral length scale tends to be more constant over a range of wind speeds than is the integral time scale. Since the average value of any sinusoid is zero. The first is that the average power in the turbulence over a range of frequencies may be found by integrating the psd between the two frequencies.
The actual wind speed. There are two points of particular importance to note regarding power spectral densities psds. Power spectral densities are often used in dynamic analyses. In wind energy studies. The first approach. It is based on a combination of theoretical and Figure The summary that follows will present some of the current models that are used to predict the variation in wind speed with elevation above ground.
The integration is from the lower limit of z0 instead of 0 because natural surfaces are never uniform and smooth. If one assumes a smooth surface.
U the horizontal component of velocity. Wind Characteristics and Resources 45 empirical research. The second approach. This yields: In this region the pressure is independent of z and integration yields: Note that U is used here.
Supervisory and control systems for wind turbines
Near the surface of the earth the momentum equation reduces to: Combining Equations 2. A summary of each of these laws and their general application follows. Both approaches are subject to uncertainty caused by the variable. Near the surface the pressure gradient is small. When this is used. Its basic form is: The log law is often used to extrapolate wind speed from a reference height. Uref see Spera. Correlations Based on Both Surface Roughness z0 and Velocity Wind researchers at NASA proposed equations for a based on both surface roughness and the wind speed at the reference elevation.
His expression has the form: It has been found that a varies with such parameters as elevation. Some researchers have developed methods for calculating a from the parameters in the log law. A review of a few of the more popular empirical methods for determining representative power law exponents follows. Many researchers. Correlation Dependent on Surface Roughness The following form for this type of correlation was proposed by Counihan The wind velocity at 30 m is set out in Table 2.
Wind Characteristics and Resources 47 Table 2. Note that at 10 m. Many authors define non-flat terrain as complex terrain this is defined as an area where terrain effects are significant on the flow over the land area being considered.
Design and Application wind power developers to accurately know the wind speed characteristics at turbine hub height generally between 60 to m — and across the rotor. Some of the effects of terrain include velocity deficits. These were developed for flat and homogenous terrain.
In the previous section. This conclusion was based on the use of experimental data sets from: The influence of terrain features on the energy output from a turbine may be so great that the economics of the whole project may depend on the proper selection of the site.
Non-flat terrain has large-scale elevations or depressions such as hills.. Elevation differences between the wind turbine site and the surrounding terrain are not greater than about 60 m anywhere in an Flat terrain is terrain with small irregularities such as forest.
For these three types of terrain. Note that some of these rules include wind turbine geometry: To qualify as flat terrain. Hiester and Pennell.
This section presents a qualitative discussion of a few of the more important areas of interest on the subject of terrain effects.
Recent work on this subject e. Flow in such terrain is divided into two classifications: Non-flat or complex terrain. For man-made obstacles. This type of flow. Man-made obstacles are defined as buildings. Natural obstacles include rows of trees. Wind Characteristics and Resources 49 Figure 2. Flow conditions in mountainous terrain are complex because the elevations and depressions occur in a random fashion.
The elevation difference between the lower end of the rotor disc and the lowest elevation on the terrain is greater than three times the maximum elevation difference h within 4 km upstream see Figure 2.
An important point to be made here is that information on wind direction should be considered when defining the terrain classification. The distinction between the two is made with comparison to the planetary boundary layer.
This affects the local wind profile. For small-scale flows this classification is further divided into elevations and depressions. Note that the estimates in the figure apply at a level equal to one building height. Small-scale Features Researchers Hiester and Pennell. A summary of each follows. The results of an attempt to quantify data from man-made obstacles are shown in Figure 2.
When the prevailing wind is not perpendicular. Depressions Depressions are characterized by a terrain feature lower than the surroundings. In addition to diurnal flow variations in certain depressions. The ratio of length to height should be at least Ridges are elongated hills that are less than or equal to m above the surrounding terrain and have little or no flat area on the summit.
Examples of the results for ridges follow. This classification includes features such as valleys. Characterization studies of this type of flow in water and wind tunnels. Steeper slopes give rise to stronger wind flow. The slope of a ridge is also an important parameter. The change in speed of the wind is greatly increased if depressions can effectively channel the wind. Wind Characteristics and Resources 51 Figure 2.
The following types of large depressions have been studied under this terrain classification: Design and Application depressions.
Here the mountains can effectively channel and accelerate the flow. They include mountains. The flow over these features is the most complex. An example of a large depression with the prevailing winds in alignment is shown in Figure 2. Large-scale Features Large-scale features are ones for which the vertical dimension is significant in relation to the planetary boundary layer.
The large number of parameters affecting the wind characteristics in a valley. This occurs when moderate to strong prevailing winds are parallel to or in alignment within about 35 degrees with the valley or canyon. These include orientation of the wind in relation to the depression. There are a number of ways to summarize the data in a compact form so that one may evaluate the wind resource or wind power production potential of a particular site.
Later sections of this text will describe how such curves can be estimated from analytical models of the wind turbine system. This section will review the following topics: Normally these curves are based on test data. Two typical curves. These include both direct and statistical techniques. This data could include direction data as well as wind speed data. Wind measurements and instrumentation are discussed in a later section of this chapter.
Wind Characteristics and Resources 53 2. Resource Characterization. The next section summarizes the use of the three non-statistical methods. Design and Application As also discussed in Chapter 1. The following four approaches will be considered: These data can be used to calculate the following useful parameters: In the following sections. It is most convenient to use the same size bins.
The data must first be separated into the wind speed intervals or bins in which they occur. Suppose that the data are separated into NB bins of width wj. An example of velocity duration curves. Wind Characteristics and Resources 55 5 The energy from a wind machine. As defined in this text. This histogram was derived from one year of hourly data. Design and Application Sample Histogram Number of Occurrences 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 Wind speed. Avelocity duration curve can be converted to a power duration curve by cubing the ordinates.
Port Elizabeth. This type of figure gives an approximate idea about the nature of the wind regime at each site. The total area under the curve is a measure of the average wind speed. The following steps must be carried out to construct velocity and power duration curves from data: The steeper the curve. South Africa. The difference between the energy potential of different sites is visually apparent.
An example of a curve of this type is shown in Figure 2. Wind Characteristics and Resources 57 3. This mathematical function was previously mentioned as a means of characterizing turbulence see Section 2. MW Turbine power 1. These techniques have been discussed by a number of authors including Justus If time series measured data are available at the desired location and height.
Note that the losses in energy production with the use of an actual wind turbine at this site can be identified. For statistical analysis. The probability density function may be used to express the probability of a wind speed occurring between Ua and Ub: As discussed next. That is: Mean wind speed.
It should also be noted that the probability density function can be superimposed on a wind velocity histogram by scaling it to the area of the histogram. It can be shown that: Design and Application Also. The Weibull distribution is based on two parameters and. The Rayleigh distribution uses one parameter: As shown.
The Weibull probability density function and the cumulative distribution function are given by: Wind Characteristics and Resources 59 0.
Weibull Distribution Use of the Weibull probability density function requires knowledge of two parameters: The probability density function and the cumulative distribution function are given by: Both of these parameters are functions of U and sU. Methods to determine k and c from U and sU are presented below. For example: Wind Characteristics and Resources 61 ii Empirical Lysen.
Ke defined as the total amount of power available in the wind divided by the power calculated from cubing the average wind speed is given by: Log—Log Plot Rohatgi and Nelson.
Extreme winds are normally described in terms of recurrence or return period. Another important consideration is the anticipated extreme wind speed. It equals the Rayleigh distribution. Extreme wind speeds are of particular concern in the design process. The slope of the straight line gives k. Using the Weibull distribution and assuming that c and k are known. This is the highest wind speed expected over some relatively long period of time.
The average and standard deviation of the resulting set would be used to find the parameters b and m. In the case of the values used in Figure 2. Note an additional subscript. Determination of extreme wind speeds by actual measurement is difficult. The probability density function for the Gumbel distribution is shown in Equation 2.
The most common statistical model for estimating extreme wind speeds is the Gumbel distribution. Gumbel Distribution 0. It is possible to estimate extreme wind speeds. The cumulative distribution function. Design and Application interval. The highest 10 min wind speed with a recurrence period of 50 yrs. The capacity factor of a wind turbine at a given site is defined as the ratio of the energy actually produced by the turbine to the energy that could have been produced if the machine ran at its rated power.
Two examples using Rayleigh and Weibull distributions form a basis for the analysis that follows. Wind Characteristics and Resources 63 2. The rotor power coefficient is defined by: This simplifies the previous integral as follows: For a numerical example. Idealized wind turbine. Design and Application The analysis. The integral can now be evaluated over all wind speeds. The average wind machine power.
As will be discussed in the next chapter. The density is raised to the first power.
2 jf mcgowan jg and rogers al manwell wind energy
For this example: Wind speed probability is given by a Rayleigh distribution. For an ideal machine. Elliot Wind Characteristics and Resources 65 As summarized in the review of Landberg et al.
Equation 2. This review summarized the following methods: In evaluating available wind data. The updated wind resource values are depicted on gridded maps.
The wind power classes range from Class 1 for winds containing the least energy to Class 7 for winds containing the most energy. Studies e. The wind atlas integrated the pre wind measurements with topography and land form characteristics to determine US wind resource estimates.
Design and Application A detailed review of all of these methods is beyond the scope of this text. Note that the 30 and 50 m heights correspond to the range of hub heights of many wind turbines then operating or under development. The use of available wind resource data is an important part of any resource assessment activity. These original atlases provided a general description of the wind resource within a large area mesoscale. In this section. The atlases depicted the annual and seasonal wind resource on a state and regional level.
In the atlas. This table was constructed using the following assumptions: Almost as soon as these results were published. These are documents that contain data on the wind speeds and direction in a region.
This work resulted in the publication of a new wind energy resource atlas in Elliot et al. An intensive program was therefore initiated by the Pacific Northwest Laboratories PNL to better characterize the wind energy potential in the United States. An example of the type of data summary contained in this atlas is given in Figure 2. Wind Characteristics and Resources 67 Figure 2. It should be noted that the results of this categorization also indicate the certainty of the wind resource based on data reliability and their area distribution.
Mean wind speed was estimated assuming a Rayleigh distribution of wind speeds and standard sea-level air density. They do not account. Class 2 areas are marginal and class 1 areas are unsuitable for wind energy develop- ment. Areas designated as class 4 or greater are generally considered to be suitable for most wind turbine applications.
Class 3 areas are suitable for wind energy development if tall towers are used. Various versions of wind resource maps for individual states are readily. In the most recent work on this subject. The approach for developing these maps consisted of three steps: This report contains. Details of this process and validation of the maps are summarized by Elliot and Schwartz The work of Elliot et al. In most cases. Wind Characteristics and Resources 69 available on the Internet see http: The increased detail as compared to the wind atlas work is readily apparent.
Wind resource. Hills and ridges. This includes such terrain as urban districts. This section provides information for regional wind resource assessments. Major orographic barriers such as the Alps. Sea coast. This section. Design and Application 2.
This condition defines a sea location 10 km offshore roughness class 0. Open sea. Open plain. In order to characterize these resources. Sheltered terrain. Asia to the east. The distribution of land and sea with the Atlantic Ocean to the west. Each terrain type was also associated with a roughness class. This is described as flat land with a few windbreaks roughness class 1.
The models and the analysis. It should be noted that the methodology for characterizing the wind resource was different than that used by PNL for the US wind atlases. It also gives a methodology for the local estimation of the mean power produced by a specific wind turbine at a specific site. These are characterized by the summit of a single axisymmetric hill with a height of m and a base diameter of 4 km.
This section contains the documentation meteorological and statistical part of the atlas. The atlas is divided into the following three parts: Large temperature differences between the polar air in the north and subtropical air in the south. The five topographical conditions are: As shown in Table 2. This describes a location with a uniform wind direction and land surface with a few windbreaks roughness class 3.
There is no one publication. Determining the wind resource. The different terrains have been divided into four types. Wind Characteristics and Resources Table 2. For this figure national wind atlases have been published for the countries marked in dark gray.
Singh et al. In this case. Landberg et al. Sandia National Laboratory. The scope of this work as of is shown in Figure 2. In a summary of world wind energy resource assessment technology. As more countries measure the wind resource for determining energy or power potential.
These have included Mexico. It is expected that they will enable wind energy researchers to better predict locations suitable for wind power sites. Resource assessments in these countries have focused on the development of rural wind power applications. Generally an estimate of the error i. Most of the currently used forecasting models consist of all or most of the following Landberg et al. For physical models these include the layout of the wind farm and the power curves of the wind turbines.
A detailed review of these models is beyond the scope of this chapter. The overall aim of this project is to develop accurate and robust models that substantially outperform current state- of-the-art forecasting models for both onshore and offshore applications.. In this area. Forecasting models require a basic set of inputs.. Microscale and mesoscale models also require information about the terrain i.
Short-term forecasting models usually estimate the energy production of a wind farm from one hour up to 48 hours ahead. Sometimes measurements of the production of individual turbines or the wind farm are used. Wind Characteristics and Resources 73 Recent work in wind energy forecasting has been concentrated on the next few hours or one to two day timeframe..
In most wind energy applications such information is not available. There are three types of instrument systems used for wind measurements: For each wind energy application. This function and the specific wind turbine power curve WTPC are needed to calculate the amount of available energy and the likely electric power output produced in the specific conditions in a region and with the technology employed [ 2 ]. The analysis of uncertainty plays an important role in the wind power industry because it can elucidate the error and the degree of reliability in a study, e.
During wind turbine power characterization, an extra source of uncertainty is attributable to the statistical process involved in power curve fitting. Given the frequent use of this international standard [ 3 ], several studies have aimed to develop improved statistical techniques, such as those used in power performance tests for small wind turbines SWT and wind power estimation [ 4 — 6 ].
All of these studies recommended techniques for improving the reliability of resource assessments and SWT power performance tests. Statistical calculus is used widely in wind resource assessments; therefore, it is necessary to reduce the sources of uncertainty to obtain reliable assessments [ 7 ]. One source of uncertainty is the error associated with the wind speed measurement process, but it is not possible to analyze its effect on wind power production.
This information is lost after the arithmetic mean is calculated to construct the mean ensemble, which limits the dispersion concept to the standard deviation for each mean calculated over time. This mean ensemble and the consequent dispersion concept assume that the wind speed with time can be represented by a normal distribution, which is not necessarily true. Several studies have aimed to detect and reduce the sources of uncertainty. Probability distribution models for wind resource variability have been studied [ 8 ] and a state of the art method for resource assessment was proposed [ 9 ].
To obtain reliable assessments, several techniques are used to measure wind speed, which may complement each other, but no previous studies have considered the quality of the measure employed for power output estimation.
Previously, it was demonstrated that because a longer time is used to obtain the mean ensemble, then the parameters that define the PDF may also change [ 10 ]. The reliability of the power assessment depends on the accuracy of the PDF parameters because they represent the wind speed conditions. However, it is not always easy to obtain accurate estimates due to the limited availability of the data [ 11 ], as well as the wide variety of PDFs that can be fitted [ 12 , 13 ].
A previous study considered the influence of the measurement quality on resource assessments [ 14 ], but it was assumed that the wind speed mean ensemble follows a normal distribution. In previous research, a wide range of PDFs have been used to represent wind speed conditions around the world [ 12 , 13 ], some of which were represented as bimodal probabilistic models [ 15 ].
In terms of the sources of uncertainty related to the goodness-of-fit of the PDFs that represent the wind speed conditions, several studies have aimed to determine how well the wind speed data are represented in specific conditions and locations by different PDFs [ 16 — 19 ].
As well as, to obtain reliable resource assessments [ 10 , 12 ]. However, due to the variable characteristics of the wind speed sampling method, it is impossible to analyze the influence of the error attributable to the meteorological device on the power resource assessment. Achieving the goal of reliable resource assessments is not an easy task because the process involves several sources of uncertainty, which are related to physical variables and the statistical process used for power estimation.
In the present study, we focus on uncertainty analysis during the early stage of power output estimation. Our proposed method may complement the development of reliable resource assessments because it yields an interval for the amount of energy that might be produced. The output of this method is the error propagation, which considers the WTPC and wind speed site conditions represented by a PDF, as well as the uncertainty associated with the meteorological device used to measure the wind speeds, where the latter element has never been included in previous methods.The highest average wind speeds in the United States are generally found along seacoasts, on ridgelines, and on the Great Plains;  however, many areas have wind resources strong enough to make a small wind turbine project economically feasible.
This section presents a qualitative discussion of a few of the more important areas of interest on the subject of terrain effects. Figure 1. Free yaw systems meaning that they can self-align with the wind are often used on downwind wind machines. Modern Wind Energy and its Origins 13 Figure 1.