Notes Topic 12 - 1 INTRODUCTION 2 WIND RESOURCES 3 WIND...

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: 1 INTRODUCTION 2 WIND RESOURCES 3 WIND TURBINE COMPONENTS AND CONCEPTS 4 WIND TURBINE AERODYNAMICS 5 WIND TURBINE BLADE DESIGN AND BLADE MANUFACTURE 6 WIND TURBINE MECHANICAL DESIGN 7 GENERATORS 8 GRID CONNECTION AND POWER CONDITIONING 9 OPERATION CONTROL OF WIND TURBINES 10 CONTROL FOR SAFETY 11 INSTALLING SMALL TURBINES 1 Introduction 1 2 Wind Resources 1 3 Wind Turbine Components and Concepts 1 4 Wind Turbine Aerodynamics 1 5 Wind Turbine Blade Design and Blade Manufacture 1 6 Wind Turbine Mechanical Design 1 7 Generators 1 8 Grid Connection and Power Conditioning 1 9 Operation Control of Wind Turbines 1 10 Control for Safety 1 11 Installing Small Turbines 1 12 WIND FARM PLANNING AND DESIGN 3 12.1 Choosing and securing a wind farm site ............................................................................................... 3 12.1.1 What is a wind farm........................................................................................................................ 3 12.1.2 Appropriate and Best Wind Farm Practice for Wind Farm Design................................................ 7 12.1.3 What to look for in choosing a wind farm site ............................................................................... 8 12.1.3.1 Wind Resource ......................................................................................................................... 8 12.1.3.2 Site and adjacent infrastructure.............................................................................................. 21 12.1.3.3 Obtaining land tenure............................................................................................................. 21 12.1.3.4 Environmental issues ............................................................................................................. 23 12.2 Designing the wind farm layout .......................................................................................................... 24 12.2.1 Turning Wind Resource Information into Real World Energy Production Estimates ................. 24 12.2.1.1 The practical energy losses from a wind farm and the Certified Power Curve ..................... 25 12.2.1.2 Estimating the long term wind resource ................................................................................ 29 12.2.1.3 The difference in wind resource experienced at each turbine location.................................. 32 12.2.1.4 Wind turbine technical availability and wind farm capacity factor ....................................... 34 12.2.1.5 Wake effects and turbine spacing .......................................................................................... 35 12.2.1.6 Electrical line losses............................................................................................................... 37 12.2.2 Site Wind Characteristics Affecting the Choice of Wind Turbine ............................................... 37 Figure 12-1 Part of the Blayney 10MW wind farm in NSW consisting of fifteen 600kW Vestas wind turbines......................................................................................................................................................... 3 Figure 12-2 A 1980s vintage wind farm in California, United States, showing HAWT and Darrius wind turbines......................................................................................................................................................... 3 Figure 12-3 Offshore wind farm in Denmark..................................................................................................... 4 Figure 12-4 A wind cluster of three NEG-Micon 750kW turbines near Randers, Denmark. ............................ 4 12-1 Figure 12-5 A “wind wall” in the United States................................................................................................. 5 Figure 12-6 A 750kW NEG Micon wind turbine owned by a farmer in Denmark. Such individual or community ownership is becoming more common in Europe. ................................................................... 5 Figure 12-7 A wind farm of over 100 turbines near Lem, Denmark.................................................................. 6 Figure 12-8 Standard summer and winter isobaric charts for the south west of Western Australia. ................. 9 Figure 12-9 Wind Roses for selected Australian Bureau of Meteorology sites across Australia..................... 11 Figure 12-10 Predicted average wind speed atlas at 30m for Western Australia. ............................................ 12 Figure 12-11 Turbulence Intensity Iu(z) measured at a wind farm site in Western Australia, at different heights and binned by wind speed. ............................................................................................................ 14 Figure 12-12 Output of the Esperance Ten Mile Lagoon wind farm from directions of higher and lower turbulence................................................................................................................................................... 15 Figure 12-13 View of partially completed Albany Wind Farm, showing adjacent turbulence increasing cliff line.............................................................................................................................................................. 15 Figure 12-14 Typical system load curve for a large Australian interconnected system, shown for summer and winter. ........................................................................................................................................................ 18 Figure 12-15 Good example of diurnal pattern of wind farm output. Here shown with actual system load for a 48 hour period starting 8:00am on 2nd February, 2002, for the Albany Wind Farm. (Source: Data courtesy of Western Power)....................................................................................................................... 20 Figure 12-16 Variations in long term average wind speed at Esperance, WA, at 10m height......................... 24 Figure 12-17 Sources of energy losses from a constant speed wind turbine. Based on research undertaken for the German “Growian” research turbine.................................................................................................... 26 Figure 12-18 Power curve for a Westwind 20kW turbine measured at Exmouth............................................ 27 Figure 12-19 Certified Power Curves for various commercially available turbines. Information from public available documents and developers must obtain up to date curves for accurate power predictions. ....... 28 Figure 12-20 Influence of using different shear exponents to raise 10m wind data to hub height in predicting the energy yield of the experimental 3MW Growian wind turbine. .......................................................... 30 Figure 12-21 Definition of flat country in the surroundings of a wind turbine. ............................................... 32 Figure 12-22 Example of energy yield output from a wind resource extrapolation/modelling computational model, also showing wake estimates using eddy/viscosity model............................................................. 33 Figure 12-23 Aerodynamic array efficiency for a 16 turbine square wind farm array and various turbine separation efficiencies................................................................................................................................ 36 Figure 12-24 Figure showing wind direction correlation. ................................................................................ 41 Figure 12-25 Figure showing 315 to 330 degree sector correlation for wind speeds....................................... 42 Table 12-1 Values of surface roughness zo, power law exponent n and turbulence intensity Iu(10m) for various ground terrains............................................................................................................................... 13 Table 12-2 Rules of thumb for turbine spacing in a wind farm....................................................................... 37 Table 12-3 Wind speed parameters for wind turbine classes according to the IEC wind turbine standard IEC 61400-1 Wind turbine generator systems (taken from [Reference: 11])..................................................... 38 Table 12-4 Long-term prediction of target wind resource using method of ratios........................................... 40 Table 12-5 Comparison of methods illustrated by Examples 12-1 and 12-2. .................................................. 43 Exercise 12-1 Example adjustments to measured wind speed data to reflect the long term average, using the simple “method of ratios” technique.......................................................................................................... 39 Exercise 12-2 Simplified example of measure-predict-correlate method for long term wind resource estimation. .................................................................................................................................................. 40 12-2 12 WIND FARM PLANNING AND DESIGN Figure 12-3 Offshore wind farm in Denmark. 12.1 Choosing and securing a wind farm site 12.1.1 What is a wind farm In its simplest form a wind farm is a collection of devices that extract kinetic energy in the wind that can then be used to perform some kind of useful work. Where and how the term came into common use is unknown though it may have been a product of the use of more than one windmill or turbine on open farmland in Europe. Certainly it has been in widespread use for some time. In recent practice the term has been used to describe a collection of wind turbines generating electricity though their numbers, types, their location and spread can vary substantially, Figure 12-1, Figure 12-2 and Figure 12-3. The wind energy industries have grown in relative isolation and it is not surprising that regional variations in nomenclature have occurred. In some parts of Europe for example a wind farm can be termed a “wind park”, while some countries tend to use the term “wind farm”. Figure 12-1 Part of the Blayney 10MW wind farm in NSW consisting of fifteen 600kW Vestas wind turbines. (Source: Photo courtesy of Western Power) Other terminology related to wind farm variants can be found. The term “wind cluster” is sometimes used to describe small wind farms of less than five turbines, Figure 12-4, while the term “wind wall” has been used for many closely spaced turbines in areas in the United States, Figure 12-5. Other such terminology no doubt exists. Figure 12-4 A wind cluster of three NEG-Micon 750kW turbines near Randers, Denmark. (Source: Photo courtesy of Paul Ebert, Western Power) Figure 12-2 A 1980s vintage wind farm in California, United States, showing HAWT and Darrius wind turbines. (Source: Photo courtesy of Western Power) (Source: Photo courtesy of Western Power) 12-3 12-4 Figure 12-5 A “wind wall” in the United States. Wind farm designs can vary from single turbines through to hundreds of machines installed on open farmland, ridge tops, breakwaters, deserts, in forests, on industrial sites and even in the ocean. Presently installations of between 5 and 30MW with up to 50 turbines are most common though individual projects over 100MW or including more than 100 turbines have and still are been undertaken throughout the world, Figure 12-7. In Australia projects as large as 150MW have been pursued for example. As turbines continue to get larger in capacity there is also a general trend to larger installed MW wind farms, but using a smaller number of machines. (Source: Photo courtesy of Western Power) In Australia, the term wind farm is used as a broad description for the installation of any number of wind turbines in a discrete regional area as a single project, usually owned and operated by a single developer. While nearly always these projects have more than one turbine, a project consisting of a single machine is often also called a wind farm in Australia. Usually the terminology is reserved for developments with turbines of medium to large size though groups of small turbines less than 50kW each exist in Australia and it is becoming more common to refer to these as “mini” wind farms. Figure 12-7 A wind farm of over 100 turbines near Lem, Denmark. Wind farms in Australia usually are given regional names that relate to their locality and relative novelty – such as the Windy Hill Wind Farm in Queensland, the Toora Wind Farm in Victoria, The Woolnorth wind farm in Tasmania or the Crookwell Wind Farm in New South Wales. In locations in Europe where wind turbine density is higher, only the larger installations tend to have such general names. The size of a wind farm is dependent on many things. On large interconnected grid systems it is not unusual to have tens of thousands of MWs of installed fossil fuel generation capacity. With even large MW sized wind turbines thousands of machines can be installed on such systems giving rise to, potentially, enormous wind energy developments. There are also many thousands of smaller, independent grid systems operational throughout the world and on these the amount of wind energy that can be installed is limited only by the amount of energy that can be technically consumed by those systems. Given certain incentives it is also common for individuals or groups (cooperatives) to invest in wind energy projects, giving rise to thousands of individual machines or small wind clusters dotted around the landscape – this is particularly the case in parts of mainland Europe (see Figure 12-6). Figure 12-6 A 750kW NEG Micon wind turbine owned by a farmer in Denmark. Such individual or community ownership is becoming more common in Europe. (Source: Photo courtesy of Western Power) In Australia, wind farms and renewable energy projects are heavily influenced by project economics due to the low cost of fossil fuel and electricity, although Federal legislation has resulted in a penalty for electricity retailers who do not supply a certain percentage of their sales from renewable electricity sources. This has stimulated a market for renewable energy supply and while wind energy is a competitive option the economics are such that wind farm projects are mostly targeted in areas of high wind speeds on land close to the coast, or at higher elevations. This is very different in other parts of the world where higher costs of electricity and legislation can make even marginal wind sites commercially attractive. Typically a good wind farm site in Australia can be characterised as having mean annual wind speed above 7 m/s at turbine hub height. In New Zealand there are wind farms operating with means above 10 m/s, which is an exceptional wind resource, though wind turbines have to be specially adapted to run under such constant high loads. While wind speed is not the only criteria to judge the financials of a wind project, certainly it is a primary issue and this leads to marginal sites being defined as having wind speeds of about 7 m/s or less. In most places in the world wind farm developments begin at the windiest and easiest sites first. With time and experience this shifts to more marginal sites although land availability, public pressure and changing political climates can provide increased pressure for such change. It is likely that a similar scenario will occur in Australia but this will depend heavily on legislation. In certain countries the political climate and electricity structure is such that wind farms can be commercial virtually anywhere, and this has led to high turbine density and extraordinary growth in the industry. (Source: Photo courtesy of Western Power) 12-5 12-6 In purely technical terms there is very little difference between wind farms. characterised by the following technical components; • • • • Generally they can be Some of this criticism is warranted. In the 1980s, enormous wind farms were built in the United States with little regard to aesthetics, the environment, or the communities that had to live with them (see Figure 12-2). In other parts of the world wind farms have been noisy, have significantly affected local populations of birds and have negatively impacted on the general public. Such developments are inappropriate in that they lead to ongoing opposition to the technology. Turbines and foundations, Maintenance access roads or access points, Electrical interconnection including transformers, switchgear and protection equipment, and, Communications and monitoring equipment. However, the placement of these components as a wind farm design relies on a range of technical, social, economic and environmental issues. While technical considerations drive longevity, reliability, energy production and ultimately the economics of a wind farm development, it is foolish of a developer or designer to believe they are the only design considerations. Wind farms also have to fit into a social environment that is increasingly complex and to do this requires what is known as an “Appropriate Wind Farm Design”. 12.1.2 Appropriate and Best Wind Farm Practice for Wind Farm Design Public opinion fuels political agendas and for wind energy to be successful it requires the support of both. In a country with few wind farm developments there is always a novelty value associated with the first few and people are generally happy to see renewable energy projects going ahead because of the environmental benefits. This easy ride for the developer typically does not last. [Reference:1 ] Large renewable energy installations have a uniquely difficult time gaining statutory and local public approval. In the United States for example this can be harder than for fossil fuel plant [Reference:2 ]. Renewable plant requires a high quality renewable energy resource plus the infrastructure to deliver the electricity to the customer. There is little point for example building a medium sized wind farm if there is no electrical transmission close by, as such a project simply cannot afford to build such a line. Requiring both resource and infrastructure together narrows down the number of potential sites dramatically and raises the probability of objections to the development with little room for the developer to compromise. Often renewable energy projects are targeted for areas that are free of development, such as open spaces that are not necessarily zoned as industrial. This is particularly the case for wind farms as usually where the good winds are there is little existing development. Often these prospective wind farm sites are “wild” areas that can be of great natural beauty. Having the appropriate infrastructure close by also usually means this area is close to a population centre, and not remote. Local people are likely to have a great deal to say about such developments and if stiff opposition arises, this can lead in turn to political opposition and difficulties in obtaining approvals. Opposition to wind farms can be very strong. In England, for example, few wind farm developments gain approval despite some of the best winds in Europe and the general perception that wind energy is sustainable and clean [Reference:1 ]. Other areas of the world have seen similarly strong opposition, with the main opposition arguments based on; • • • • • • It is now common practice to mould wind farms to suit the particular needs of a proposed site and this is what is referred to as an “Appropriate Wind Farm Design”. While it broadly means examining all the issues that arise to acknowledge and investigate opponents fears, some of which are locally unique, it more generally means undertaking a range of work involving; • • • • • Visual amenity studies, Noise studies, Flora and fauna studies, Land use studies, and Consultation with communities (called “Stakeholder Management”). Representative bodies for wind energy throughout the world have recognised that such work should be considered normal practice if sustained wind farm growth is to occur. The Australian Wind Energy Association (AusWEA) has published recommended practices [Reference: 3] that seek to ensure that this occurs. These and other issues will be examined in more detail in subsequent topics. 12.1.3 What to look for in choosing a wind farm site Land is a critical component in a wind farm development and the issues that arise in its selection can ultimately prove the most problematic. Wind farms have been built on all manner of sites covering a broad spectrum of land uses and many have issues unique to themselves. In general though the following issues will arise. 12.1.3.1 Wind Resource Obviously a good wind resource is required and one which gives a suitable energy yield and economic return to the project’s investors. In theory a wind farm developer seeks a site with strong consistent winds, which is free of turbulence from obstructions or topography in the prevailing wind directions, which has a manageable maximum wind speed and which blows the strongest when the energy is worth the most. Such locations exist though in practice a compromise has to occur. Finding such a site also requires a process that can take between six months to several years, depending on the wind farm size and acceptable risk. This section expands on the material covered in Topic 2. Winds are produced by large scale or “macro” weather events that are essentially repeatable. Variations on the strength and position of these occur regularly though in general they provide a reliable indication of where the winds will be and from what direction. The affect on the visual amenity of an area, Unacceptable turbine noise, The effect on flora and fauna (especially birds), Shadow “flicker” from sunlight passing through or reflecting off the blades, Disruption to television and radio reception, and Failure of wind turbines to deliver reliable power & make a difference to the environment. In Australia, macro weather patterns are dominated by high and low-pressure systems that move in a west to east pattern across the country. Latitudinal differences occur in the relative positions of these systems depending on the time of year and hence there can be seasonal differences in wind speeds and directions at 12-7 12-8 different locations. An extreme example of this is the tropical cyclones that occur from time to time in the northern parts of Australia. An example of consistent weather patterns is the subtropical ridge axis, which is a belt of semi-permanent high-pressure systems that encircle the southern hemisphere between latitudes of 25 and 40 degrees south, and its effect on the South-West of Australia. In Figure 12-8 a standard summer and winter pattern across this area is shown. In summer the weather is dominated by high pressure systems around 40 degrees south which move west with the wind speeds and direction changing during the passage as shown in the figure, though generally the flow is easterly. In winter the highs sit around 30 degrees south, enabling strong cold fronts to penetrate into the area causing a NW-SW flow. Figure 12-8 Standard summer and winter isobaric charts for the south west of Western Australia. Winter Isobaric Charts (Source: Taken from Reference: 4) Regional areas are known to be windy historically and often can be identified through long-term wind resource information, of which the Bureau of Meteorology (BOM) is a good source in Australia. Figure 129 for example shows wind roses for various BOM sites. It seems intuitive that areas of higher latitudes close to the coastline, or at higher elevation, will have a better wind resource and this is generally the case as shown in Figure 12-9. However, wind strengths and consistencies are influenced by many factors such as the land/sea interface and topography, and often areas that at first would be considered poor in resource can prove to be very good. How these areas are identified relies on experience and the use of various assessment tools, some of which will be described in this course. Summer Isobaric Charts 12-9 12-10 Figure 12-9 Wind Roses for selected Australian Bureau of Meteorology sites across Australia. Figure 12-10 Predicted average wind speed atlas at 30m for Western Australia (Source: http://www.bom.gov.au/) Caution needs to be used when using long-term wind data that has been acquired for meteorological purposes. This is because the height of the monitoring mast used, the type of equipment and the measurement location have not been chosen or placed specifically for wind energy purposes and data can be somewhat misleading. As an example, the wind atlas of Western Australia produced in the late 1980s [Reference: 5] and shown in Figure 12-10 is based on, in large part, long-term Bureau of Meteorology data. Figure 12-10 appears to indicate most of the state as being marginal for wind energy with the small Southwest area the best. All of the wind farms built in Western Australia up to 2000 were, however, outside of that Southwest area and have proved commercially viable. (Source: Taken from [Reference: Error! Bookmark not defined.]) Atmospheric Boundary Layer Recall from Topic 2 that the wind speed profile can be expressed by various functions. Using a logarithmic function, the wind shear profile can be written as: Equation 12-1 Wind atlases and meteorological information will, nevertheless, point to locations likely to have a good wind resource. As will be shown, such data is also an invaluable asset in predicting the longer term output from wind farms and hence the project’s long term economic merit. Once macro scale regions are identified as having a potentially good wind resource there are a number of ways of locating the best sites on a finer or micro level. To do this it is necessary first to understand the effects of surface roughness and topography on wind speeds and turbulence, and hence energy yields. This leads to assessment tools that are useful in the identification of prospective parcels of land for development. where subscripts denote the different heights. Equation 12-1 is very useful and it is tempting to use it with abandon in wind energy developments. However, the equation is specifically accurate only for what is known as steady state and neutrally stratified 12-11 12-12 atmospheric conditions that in reality rarely, if ever, exist for any length of time at a wind farm site. If wind speed information is available from monitoring equipment at the proposed site the equation can nevertheless assist the designer in predicting wind speeds at different heights. Alternatively, using a power exponent function, the wind speed profile can be expressed as: Equation 12-2 have some understanding of the predicted wind speeds and directions from the closest known wind speed measurement source. There are also numerous numerical modelling tools available for predictions of flow to analyse speed up and these are discussed more shortly. The topography and the physical attributes that exist on it are known as the orography of the site, and the wind energy industry uses this term widely. Understanding and using the sites orography is a key component in wind farm design. Table 12-1 shows typical turbulence intensity Iu values at 10m height for various terrains. Usually Iu(z) falls with increasing height and Figure 12-11 shows this from measurements at a wind farm site in Western Australia. Turbulence also usually drops with increasing wind speed and this can also be seen in the figure. Figure 12-11 Turbulence Intensity Iu(z) measured at a wind farm site in Western Australia, at different heights and binned by wind speed. where n is known as the power law exponent. As Equation 12-2 is in structure independent of surface roughness it is used extensively by wind farm designers who can measure n using two anemometers on their monitoring masts, and then use this to scale wind speeds to different heights. When choosing a wind farm site it is important that surface roughness be considered as siting of turbines close to rougher elements will lead to decreased energy yield, and a poorer wind energy project due to increased shear in that flow. Table 12-1 recalls Table 2-1 of Topic 2 and shows values of surface roughness along with power law exponent and turbulence intensity for different classes of terrain. Table 12-1 Values of surface roughness zo, power law exponent n and turbulence intensity Iu(10m) for various ground terrains. Terrain classification Cities with buildings 10 – 50m high Villages, suburbs with low buildings, trees Forests Surface roughness zo (m) 1 – 10 Power law exponent n 0.35 – 0.5 0.4 – 2 0.27 – 0.4 Turbulence intensity Iu(10m) 0.42 – ∞ 0.3 – 0.6 0.7 – 6 0.29 – 0.5 0.4 – 1.9 Farmland with many trees and buildings 0.1 – 0.35 0.22 – 0.27 0.21 – 0.8 Farmland with few trees and buildings 0.02 – 0.1 0.14 – 0.2 0.15 – 0.21 Crops up to 1.2m high 0.04 – 0.2 0.15 – 0.22 0.17 – 0.25 Grass and shrubs 0.25 – 1.0m high 0.04 – 0.1 0.15 – 0.2 0.17 – 0.25 Grass 0.02 – 0.1m high 0.003 – 0.01 0.1 – 0.13 0.12 – 0.14 Bare soil (not ploughed) 0.001 – 0.01 0.1 – 0.13 0.1 – 0.14 Sand and desert (no dunes) 0.0003 – 0.02 0.1 – 0.15 0.08 – 0.15 Snow 0.00001 – 0.001 0.09 – 0.15 0.08 – 0.15 Water (increases with increasing wave size) 0.00001 – 0.001 0.09 – 0.15 0.07 – 0.1 0.0001 0.1 0.08 Smooth ice (Source: Data courtesy of Western Power) Turbulence also has an interesting affect on turbine power output, making it generally become erratic as the turbulence intensity increases. This manifests itself as large variations in wind turbine power output with variations of more than 50% in a few seconds common with high turbulence (see Figure 12-12). Such behaviour depends in part on the frequency of turbulent motion of the fluid particles and the inertia of the turbine’s rotor and generator, which affects its response time. As will be seen in later topics, such changes in turbine output can have a detrimental affect on electrical grids with the affect decreasing with larger and more spread out wind farms. Topographic speed-up and turbulence Topographic speed-up can be used to the wind farm designer’s advantage. Often a prospective wind farm area will have land available at higher elevation and by examining topographic maps, with experience, it is relatively easy to identify ridges and hills that may experience wind speed-up. To do this it is important to 12-13 12-14 It is often the case that orography that may cause turbulence may also result in significant wind speed-up. Many wind farms have been positioned in areas where turbulence is high or potentially high but this has been chosen deliberately as those areas can also have better winds that increase the commercial return. An example is the Albany Wind Farm in Western Australia shown in Figure 12-13, where the 80m high cliff line evident can produce significant turbulence. However, in such circumstances it is possible to have a successful wind farm provided sensitive siting and turbine selection is undertaken - in the Albany wind farm case, turbines were positioned back from the cliffs and on high towers outside of the turbulent zone. Figure 12-12 Output of the Esperance Ten Mile Lagoon wind farm from directions of higher and lower turbulence. Severe Wind Events It has already been discussed that extreme variations to the normal weather patterns across Australia occur and some knowledge of this is important for the wind farm designer and developer during the siting selection process. Wind turbines operate in a formidable environment and can be destroyed if subjected to winds or turbulence to which they were never designed. While the public opinion of the wind energy industry can be harmed by such failures, they can also result in serious injury and can be economically disastrous. The project’s designer has to ensure such events are understood and that the turbines are sited and selected accordingly. The methods of doing so are discussed in following topics but in the initial site selection process it is normal practice to investigate severe weather events, usually through the Bureau of Meteorology or its data archives, or through historical records in the region of interest. (Source: 10-minute averaged data courtesy of Western Power) When selecting a wind farm site, the orography that can give rise to increased turbulence need to be investigated and their possible effects quantified. Typical examples include cliff-lines, buildings and steep slopes above about 15 degrees – anything that causes the wind flow to change direction rapidly that causes strong flow shear, vorticity and turbulence. Macro and micro flow modelling As the size of the wind energy industry has grown so has the research effort to maximise the efficiency and economic return from wind farms. Much of this effort has gone into predicting wind farm energy yield and to do so numerous computational modelling tools have been developed. Many of these are available commercially and can be found scattered throughout the wind energy literature and brochures. They provide an important tool for the wind farm designer and developer. Wind flow over an area of land can be complex as is the theory behind the physical processes involved. Most designers do not have the expertise or time to really come to terms with the intricacies involved though a general understanding is important as unusual or misleading results can occur simply through an incorrect keystroke or input data. Realising this, the wind farm designer learns with experience what to look for from modelling results and what the limitations and uses of such software are. Figure 12-13 View of partially completed Albany Wind Farm, showing adjacent turbulence increasing cliff line. Wind farm modelling software has been available for many years though in the early days of the industry were found to be accurate only in specific types of terrains. Most models today are more accurate, the result of years of theoretical research coupled with experimentation and fieldwork to verify results. With increasing accuracy has also come increased computational complexity, although computing speed has increased faster and modern computers have little trouble evaluating even the most complex of wind farm sites. All models use mathematical and/or empirical relationships to predict the flow over the area of interest and thereby allow wind turbines to be positioned to best effect. Typically these models rely on digital information of topography as well as knowledge of surface roughness, obstacles such as buildings and wind speed. This information is combined and often a numerical iteration is performed over a grid dividing up the area with the result being predictions of wind speeds and hence energy yields. Methods for adjusting for height, topographic speed-up, surface roughness and turbulence as discussed earlier are incorporated in the algorithms. (Source: Photo courtesy of Western Power) 12-15 12-16 Some models attempt to predict the wind speed information based on macro scale meteorological events and it is these that provide a useful tool for the analysis of very large areas in the search for good wind farm sites. These models can be used over areas hundreds of kilometres across and usually assume a simpler laminar wind flow regime and neglect the effects of turbulence, which on such a broad scale can be of second order importance. Developers routinely use macro models at an early stage in a project to identify the best areas to install wind-measuring equipment, though as enormous amounts of data are used the computational time can be lengthy and costly. Modelling software that allows positioning of turbines on a micro scale, or what is termed “micro-siting”, can be quite different to macro scale models and will be discussed in proceeding topics. Site surveys Once a prospective wind farm site has been identified than a site survey is necessary. Before doing so it is best to obtain cadastral (property boundaries, road infrastructure etc.) and topographic information that covers the area and typically this is obtained from local land authorities in electronic or paper form for a fee. Large wind farm developments often use a Geographic Information System (GIS) which allows an electronic site map to be produced, with contours, land forms, roadways and the like on different electronic “layers” which can be turned on and off at will. This GIS map can be built up as more becomes known about the site and used for micro-siting of turbines. Often local knowledge can be useful when identifying areas within a region that are specifically windy – local areas are called “Windy Hill” and “Sea Breeze Point” for a reason. Local experience can sometimes be based on a single wind speed event, such as a particularly bad storm and caution needs to be taken with general comments such as “it’s always windy here”. Wind prospecting allows the general merits of a region to be confirmed and the site for more complex wind monitoring to be selected. It can also confirm the results of any macro-scale wind modelling or be used to calibrate those models. Wind monitoring is always carried out for larger scale wind farm developments where a great deal of capital is involved and small increases in energy yield can give rise to large changes in project economics. There is nothing more accurate than actually measuring the wind resource at the intended hub height, though turbines can have 100m high towers and placing anemometry equipment at such heights safely requires the use of significant tower infrastructure. Usually fixed triangular lattice type towers are used which can be climbed using safety equipment with harnesses. Wind monitoring is typically carried out between 6 months and 2 years, or simply a long enough period to provide the developer with information that can lead to both accurate wind turbine micro-siting and longterm energy yield predictions. This data must be of high quality, reliable and accurate, and a quality assurance system needs to be in place to ensure that this occurs. Obtaining financing for wind farm projects and the ultimate economic success will depend on just how accurate and auditable this monitoring is and a great deal of effort is needed in this regard. Time value of energy and capacity credit – does the wind blow at the right times? Electricity consumption changes throughout the day on a diurnal pattern. On large interconnected grids this consumption profile depends on the day of the week, the weather and the time of year. Typically Australian electrical systems have a consumption profile similar to that shown in Figure 12-14 Figure 12-14 Typical system load curve for a large Australian interconnected system, shown for summer and winter. While many things are looked at during an initial site survey, it is important that an area free of obstructions is found – this is known as having a good or clear “fetch”. Depending on the size of the wind farm planned, it may be possible to remove such obstacles to improve the fetch or to avoid them altogether. Measurements – Wind prospecting and monitoring In the proceeding sections it is clear that there are inaccuracies and areas of caution involved with choosing a site in terms of wind resource using broad scale techniques. While it is likely that a reasonably accurate prediction of wind speeds can be made there is no better alternative than to actually measure. Usually there are two types of wind measurements, broadly defined as; • • Wind prospecting: Measurements of wind speed, direction and sometimes turbulence aimed at identifying the wind resource in a particular region over a scale of many tens of kilometres, and Wind monitoring: Measurements of wind speed, direction, turbulence and air temperature specifically at multiple heights including expected hub height, aimed at obtaining data for the micrositing of wind turbines into a wind farm over a scale of less than ten kilometres. Wind prospecting usually involves cheaper tilt down towers at heights up to about 50m over a period of between 6 months to one year, and often with instruments at two heights to obtain the wind shear. Sometimes this type of assessment is all that is required if the wind farm is small scale or in a region where other wind monitoring work or wind farms have been installed, as then a good picture of the long term wind conditions is already known. As more sophisticated wind prospecting involves more expense, the tower height and measurement complexity has to mirror the size of wind farm development and project risk. 12-17 (Source: Data courtesy of Western Power) The typical profile has changed enormously in the last twenty years in Australia and this mirrors changes in societal patterns of living and working. Air conditioning, for example, now has much higher penetration into 12-18 the average household market and on a hot summers day can cause a dramatic rise in consumption, which skews the diurnal pattern, helping to give rise to profiles that have significant peaks in summer. Figure 12-15 Good example of diurnal pattern of wind farm output. Here shown with actual system load for a 48 hour period starting 8:00am on 2nd February, 2002, for the Albany Wind Farm. In a perfect electrical system there would be a flat consumption profile. This means that the generation plant would deliver electricity at a constant amount with little variation, which suits large scale thermal generation as it takes considerable time to bring stages of these on line. Peaky profiles make it necessary to install generation plant that can change and respond quickly, or to install energy storage such as pumped hydro, which can be stored during the “troughs” and released at the “peaks”. Electrical systems have evolved to compensate for more “peaky” consumption profiles with a generation portfolio based on a range of technologies with differing orders of merit. Typically such systems include a large fossil fuel based generation source that provides base load power, or the majority of the non-fluctuating component of the electrical demand on a time scale of many hours. Such base load plant includes gas, coal or oil fired power stations and nuclear facilities up to sizes as large as 5000 MW. The next orders of merit are generally given to generation sources that can provide quicker responses, to eventually get to what is known as peaking plant that is brought on line for those periods of highest load, which can fluctuate very quickly. Typically such peaking plant can be gas turbines, diesel generators or hydro facilities. Changes in electrical consumption that require smaller and faster acting plant also give rise to higher costs. This is due to their lower economies of scale, higher operational costs and the fact that for much of the time the plant can be unutilised. This means that during the day the cost of electricity varies as various merit order plant is brought on and off line and consumption rises and falls. As a result of the time varying cost of electricity, what generation plant is on at what times can become complex commercially. As an example, Australia has a National Electricity Market (NEM) which operates in Queensland, New South Wales, Victoria and South Australia in which generators in those states bid a forward price per MWh for the right to generate into that connected system. A centralised dispatch office monitors consumption and the bids, and dispatches generating plant accordingly. The National Electricity Market Management Company (NEMCO) oversees this process. Large, fossil fuel based thermal plant can generate at low cost and hence provides the majority of base load in NEM. At times of peak demand the bidding value can rise very high, which makes it profitable for quicker acting generation plant to come on line at that time. Also, electrical systems have to be controlled to keep voltage and frequency within acceptable limits and special ancillary services are available and controlled by NEMCO to be used for this purpose. (Source: Data courtesy of Western Power) In some places in the world renewable plant is given special status because of its environmental benefits and does not have to fit exactly such a market or consumption pattern – that is, the wind farm’s output is given the same value regardless of the time of day. This is not yet the case in Australia so a wind farm which produces electricity at peak times of the day may have an advantage commercially over another which produces more electricity, but at the wrong times of the day. In looking at prospective wind farm sites, such diurnal patterns in wind speed are therefore important to study. An example of a good diurnal output pattern required under such circumstances is shown in Figure 12-15. Another related and important aspect of wind farms is their “capacity credit”, or their ability to replace conventional generation. Capacity credit involves the reliability of a generation plant– or how much capacity can be relied on from the plant at a given time and for how long. A large thermal power station once brought on line can produce its rated output for weeks and therefore has high capacity credit, while a wind farm is at the mercy of the fluctuating wind resource and may not be as reliable. As an example, a wind farm output can change from maximum capacity to very little in the space of half an hour with the passage of a weather front. Higher capacity credit plants traditionally have been more valuable to an electrical system operator because of this reliability, while wind farms have been given no capacity credit. As experience has grown, however, it appears that geographically dispersed wind farms on a given system can produce reliable and predictable quantities of power [Reference: 6]. Wind farm developers may need to evaluate this in terms of the wind resource, its predicability and reliability for their own projects, as this may influence negotiations with system operators and, therefore, the economics of the project. A discussion of the complexities and problems of fitting wind farms into the Australian electricity market from a systems and operation perspective is given in Arnott [Reference: 7] and Outhred [Reference: 8]. 12-19 12-20 12.1.3.2 Site and adjacent infrastructure A site with an incredible resource may be impossible physically to develop, while there may be no customers to sell the electricity to or means of transporting it to where there are. Wind farms are very reliant on infrastructure available close to and at the site. The issues involved with infrastructure broadly fall into the following categories; • • • • Electrical transmission/distribution connection, Proximity of services such as roads, concrete supply, cranes and heavy transport, The bearing capacity of the soil for large, tall structures, and The practicality of moving in large pieces of turbine equipment and erecting them. The most critical but often overlooked issue is electrical connection that the developer must approach early, especially for larger wind farms. Not all electrical lines are equal and their ability to take the amount of electricity generated by the wind farm will depend on the line itself, the amount of load it carries and where it fits into the local electrical system. In Australia, electrical lines are categorised as being Transmission or Distribution largely on the voltages they carry. As a rule, electrical infrastructure that transports large quantities of electricity long distances is termed Transmission (66, 132, 330 and 500kV), while those feeding directly to customers at street level is known as Distribution (240V, 415V, 3.3kV, 6kV, 11kV, 22kV and 33kV). Transmission systems can transmit more electrical power due to their voltage and conductor size although whether a wind farm can connect to them will depend primarily on the existing load on that line, its length and the voltage and frequency stability on that line. Generally distribution systems are more difficult to connect to. These issues will be discussed further in later topics. The building of long transmission lines is not normally possible for wind farms unless they are very large, as the cost is prohibitive. In Australia, wind farms have been connected to both transmission and distribution infrastructure and electrical connections have been nearly always less than 20km. The ability to deliver and physically move equipment onto the site will depend on the road infrastructure and heavy haulage equipment available. Large wind turbines have components as heavy as 50 tonnes and loads up to 50m in length which require wide, gently sloped and stable roads for transport. The movement of craneage for such large turbines is also reliant on roads, where surface cross-fall must be small otherwise tall jibs can cause the crane to become unstable during movements between turbine sites. When assessing proposed sites the ability to source and bring such equipment on site needs to be looked at. Wind farms also need materials such as concrete, reinforcing steel and water – the later for dust suppression, cleaning and wetting. Getting these into some sites can be problematic and expensive. The ability to build foundations will require some initial civil assessment that usually includes local knowledge as to what types of soils and bearing strengths are known. Areas involving materials such as water, rock and moving loose sand will require more specialised foundations and hence greater cost. 12.1.3.3 Obtaining land tenure 12-21 Wind farms require land and no single issue creates more difficulties for wind farm developers and designers. Obtaining tenure over suitable land for the twenty or so years that the wind farm will produce for involves issues that can be legally difficult, protracted and publicly sensitive. To work through those requires the appropriate wind farm design approach discussed earlier. When assessing prospective sites it is important to assess the land’s other uses, both present and future. Land is zoned by the local (or State) authority according to its purpose, ownership and strategic plan for the area and this needs assessing. There is little point for example building a wind farm if a housing estate is planned close by which will experience unacceptable noise from the wind farm, or turbines interfere with approaches to landing strips for aircraft. Similarly, other planned developments close by could significantly affect the wind resource at the site with examples being tree farming or tall structures such as industrial buildings. There may also be conflicting and recognised recreational uses – for example, hang-gliding, para-gliding or kite flying may be officially sanctioned in the area. An area used for recreation will invoke a strong response from the public. Wind farms in Australia have been built on a variety of land types. This includes Crown Land, which is essentially owned and controlled by the Government, and private or leasehold land. The Windy Hill wind farm near Ravenshoe, Qld, is built on private land and a legal agreement exists under which the landowner allows the turbines to operate on the property and for this receives a monetary sum. Such agreements on farmland are common for wind farms in Australia. A wind farm usually has only a small impact on the yield of farmland and there appears to be no evidence that wind turbines have a long-term negative influence on stock grazing around the machines. In broad acre farming involving large harvesting machinery, if spaced incorrectly wind turbines can make it more difficult to harvest as the machinery cannot negotiate the corners around turbines and therefore some of the land becomes unproductive. To prevent this it is sometimes better to follow straight placement lines or fences for turbine positions on such land. On Crown Land, usually negotiations are with the relevant State land body and sometimes with the local authority, which has management or vesting over the land. The Exmouth wind farm in Western Australia was built on Crown Land. Access to Crown Land for wind farms also involves a lease arrangement involving a monetary sum usually set at acceptable commercial rates, although because of wider political scrutiny such access can usually take a great deal of time to procure. ‘Viewsheds’ in Australia are often classified according to their importance historically or nationally, or how beautiful or important culturally they are. Many windy locations in this country are classified as very important as they are wild places free of development or because they are nationally recognised. The Great Ocean Road in Victoria for example includes the Twelve Apostles in a very windy area although, as symbols of Australia recognised internationally, it would be unacceptable to develop a wind farm adjacent to these. Just what is and isn’t acceptable of course requires public debate. A wind farm has a typical lifetime of 20 years and its owner will need some kind of tenure on the site. This means negotiating with the landowner or recognised land authority for a legal agreement, such as a lease or wind farm agreement that will involve a negotiating process. Such tenure also has to include access rights and noise buffers and in Australia can involve discussion with Native Title claimants whom may have an interest in the land in question. The wind farm designer has to be aware of the issues that arise from such commercial discussions as they can lead to changes to wind turbine placements and alignments, and hence energy yields and turbine selection. 12-22 12.2 Designing the wind farm layout 12.1.3.4 Environmental issues Australia has a unique environment that, despite its seemingly tough nature, is very sensitive to disturbance. The native birds, animals and plants that make up this are precious national resources that have to be both protected and encouraged. The building, operation and decommissioning of any industrial plant in Australia will effect the local environment and a wind farm is no exception. Despite being a renewable energy resource with environmental benefits, to put 100m high structures into a site means that some part of that environment will be disturbed. The wind farm designer has an obligation to see that any environmental disturbance is minimised and acceptable to the approving authorities and to the general public. When developing a wind farm site the obligation is usually on the wind farm developer to prove whether the disturbance to the area’s environment is acceptable. For the wind farm designer this means being aware of the environmental issues that exist and altering the wind farm design to accommodate these. In Australia such work is commonplace and has been very successful and the results are that the overall environmental impact of wind farm developments in this country has been small. In other parts of the world this has not always been the case. In other countries wind farms have been implicated in regard to significant effects on birds, soil erosion, land degradation and land contamination through liquid hydrocarbons. 12.2.1 Turning Wind Resource Information into Real World Energy Production Estimates The financial success of a wind farm project depends heavily on its energy yield as this gives rise to the project cash flow for servicing debt and creating profit. Typically wind farm projects in Australia have very tight budgets and slim returns meaning that energy yield estimates need to be very accurate. This is particularly the case for predictions of energy yield over many years of operation as normally financial analyses are calculated over the lifetime of the plant. For the designer this creates challenges, though by recognising the issues involved and proceeding step-wise through the analyses it is possible to obtain an accurate yield estimate within an acceptable risk margin. The history of wind farm development has seen many early projects fail financially in part because of the inability of their designers to recognise all the factors that influence energy yield. For example, a wind farm design can be based on too little wind resource information that fails to capture long-term trends resulting in unrealistic long-term energy estimates. Figure 12-16 offers a simple example, using long term Bureau of Meteorology information from Esperance in Western Australia, that shows just how variable the wind resource can be over a long time scale. Also shown in the figure is the output from the Ten Mile Lagoon wind farm. Note that wind monitoring in the high periods of Figure 12-16 could lead the designer to assume that the wind resource is better than it really is long term. Figure 12-16 Variations in long term average wind speed at Esperance, WA, at 10m height. When assessing a prospective site it is usual practice to undertake a review of existing environmental conditions to ascertain what issues exist and the likely cost of accommodating these. Depending on the size of development and the location, this can involve discussions with local authorities, experts or long-term residents in relation to general environmental issues including; • • • • Environmental status – Protected, locally/nationally/internationally recognised, local planning strategy or use status, Flora – endangered or rare species, issues of disease, introduction of weeds and effects on crops, Fauna – avian and ground based fauna, endangered or rare species, migratory patterns, local or adjacent protected areas that could be influenced, rehabilitation, quarantine status for stock protection and effects of wind farm on animal grazing, and, Land – wind and water erosion, soil born diseases and issues of access restrictions, seasonal variations in access, ground water contamination. Such issues can appear overwhelming and can lead to a site being identified as inappropriate to develop. However, in Australia wind farms have been successfully built in World Heritage Areas, on operational farmland and in areas including sensitive vegetation. Such success depends heavily on the ability of the designer to work within the sites environmental constraints and the tools necessary to do this are discussed in subsequent topics. (Source: Data from the Bureau of Meteorology) Another error made in yield predictions is to neglect the practical issues surrounding the operation of wind turbines, which only become evident in the field. These include the production time lost to breakdown, very high wind events and the electrical losses in wind farm cabling. As wind turbines extract energy from the airflow the wake of one turbine can affect those down-stream, also causing a loss in energy yield that can be very significant. 12-23 12-24 It is generally recognised that there will always be some uncertainty when it comes to yield estimates for a wind farm. Historically, such uncertainty has been declining as experience is gained with the way wind turbines operate and how to better design them, what maintenance and operational issues accrue with time and how to better predict long term wind resources changes over the years. Figure 12-17 Sources of energy losses from a constant speed wind turbine. Based on research undertaken for the German “Growian” research turbine. 12.2.1.1 The practical energy losses from a wind farm and the Certified Power Curve The energy yield of a wind farm is very dependent on the wind resource but the reality of designing a low cost, practical machine, as well as land tenure, environmental and other siting restrictions, means that other significant issues also come into play. Theoretically a wind turbine should be able to extract energy from the wind flow with an efficiency defined by the Betz limit (as covered in Topic 3). Aerodynamic, mechanical and electrical design inefficiencies which are the result of practical engineering issues lower this substantially as do output limitations in terms of rated power and cut-out wind speed designed to protect the turbine under high wind speed events. Such design issues are neatly summarised in Figure 12-17 that shows the calculated losses for various parameters for the 3MW Growian research turbine from Germany [Reference: 9], based on an assumed wind speed frequency distribution and operational machine restrictions. In that figure the actual energy yield for this constant speed turbine was less than 40% of that in the original wind flow. In practice, wind turbine designers can make adjustments to those parameters in Figure 12-17 to improve energy yield and even to tailor the turbine to suit a particular site. However, often such changes come at a significant cost penalty and do not necessarily fit easily into wind turbine production lines where even small changes to the standard production model can be time consuming and uneconomic. Because of this, most turbine manufacturers have for a certain turbine model a number of versions available that are tailored for particular wind and turbulence regimes. This will be discussed more in proceeding sections. To come to terms with design losses and to specify more precisely what the energy production is from their machines, it is common practice for manufacturers to obtain a Certified Power Curve. Such a curve is usually produced by a certification agency of which there are a number throughout the world to a recognised testing standard (for example [Reference: 10 ] and see Appendix 4 in [Reference: 3]) typically based on 10minute average wind speed and power output data. Such testing can be time consuming and hence costly as certification requires the full range of wind speeds to be reached for a suitable period and in the past has been limited to larger turbines. However, testing procedures specifically for smaller turbines are now available and it is becoming more common for these to also have certified performance figures. Figure 12-18 shows the average power curve from the a 20kW Westwind turbine at the Exmouth Advanced Mini Wind Farm, to the standard in [Reference: 10 ] while Figure 12-19 shows Certified Power Curves for a number of commercially available wind turbines. To come to terms with the variability in output obvious in Figure 1218, it is normal practice to use the method of bins with the wind speed bin simply averaged at the end of the test period. Power curves are typically produced for set conditions such as temperature and air density and the wind farm designer has to make adjustments for conditions that differ at the proposed wind farm site. Power curves are normally based on wind speed measured at the turbine hub height. 12-25 (Source: Taken From [Reference: Error! Bookmark not defined.]) 12-26 Figure 12-18 Power curve for a Westwind 20kW turbine measured at Exmouth. Figure 12-19 Certified Power Curves for various commercially available turbines. Information from public available documents and developers must obtain up to date curves for accurate power predictions. (Source: Data with kind permission from Westwind wind turbines and ACRE) (Source: P. Ebert, Western Power) Recall from Topic 3 that once a power curve P is known it is a simple exercise to obtain a measure of a turbine’s energy yield, E, by using the known wind speed frequency distribution, Φ, as; Equation 12-3 n E = ∑ Pi Φi i =1 where the subscript i represents the ith wind speed bin, providing the bin width of both P and Φ are equivalent, the range covers the operational wind speeds of the turbine and the wind speed was measured at turbine hub height. Equation 12-3 provides the means to estimate the theoretical energy yield for a single turbine provided the wind speed measurements have been taken at or close to the final turbine placement position. The equation cannot, however, accurately predict the long term energy yield from that turbine or a wind farm because it takes no account of: • • • • • 12-27 Changes in the long term wind resource at the site, The differences in wind resource experienced at each turbine location, The availability of the turbines. Wake effects from one turbine to the next, and, The electrical line losses up to the point of metering. 12-28 These points significantly affect the energy yield and result in a rewriting of Equation 12-3 for an m turbine wind farm to give; Figure 12-20 Influence of using different shear exponents to raise 10m wind data to hub height in predicting the energy yield of the experimental 3MW Growian wind turbine. Equation 12-4 m n E = ∑ ∑ Pi , j Φi , j − ( wake loss ) j ∗ availability − (electrical losses ) j =1 i =1 An explanation of these new terms in Equation 12-4 will be covered within the remainder of Section 12.2.1. 12.2.1.2 Estimating the long term wind resource Wind prospecting and monitoring is usually performed over a limited time interval because it is too costly to continue it longer and project financiers cannot sustain longer time periods without some kind of financial return. This means that corrections are usually performed to the measured data to incorporate long term changes in the wind resource and to predict what the longer-term energy yield will be. The effect of wind monitoring period on long term energy yield accuracy is difficult to predict but in Figure 12-16 it is obvious that measurements would need to be over decades to get an accurate reflection of how the wind resource varies longer term. Fortunately, in many locations throughout Australia the Bureau of Meteorology has had recording instruments in place for many decades and this is invaluable as reference data. Usually the long term resource is estimated by correlating the measured wind monitoring data from the target wind farm site, WT(u,θ,z,t), with concurrently measured but longer term reference data, WR(u,θ,z,t). Here the wind resource information W is subscripted as T for target and R for reference and is shown as a function of wind speed, direction, height and time – the four parameters which, essentially, define it. What is sought in this analysis is a wind speed and direction histogram, which represents the yearly average across the site over the period the wind farm will be operating. With this wind speed information Equation 12-3 can be used to give the long-term energy yield. Often wind farm designers neglect the effect that wind direction can have on energy yield. As will be seen, the wake effects between wind turbines can be significant although site land restrictions may mean that there is little alternative than to place machines close together. Knowledge of the way the wind resource varies in terms of direction can, therefore, influence the energy yield significantly as placement can be made to align with the directions of highest energy yield and lower wake losses. It is often at this stage that the designer has to adjust WT to align with the hub height of the turbine, if measurements were not made at that height. This can be accomplished using Equation 12-1 or Equation 12-2 but care needs to be exercised as discussed with those equations – Figure 12-20 for example shows the large differences in annual energy yield predictions for the experimental 3MW Growian turbine [Reference: XXXX] when using Equation 12-2. Many designers break WT into wind direction bins (called sectors) and calculate the average shear in each, using this to adjust to the height required and this has been found to yield more accurate results. 12-29 (Source: Taken from [Reference: XXXX]) With WT adjusted to be level with the turbine’s hub then z no longer becomes an issue in the analyses and the two data sets can be signified as WT(u,θ,t) and WR(u,θ,t). However, care has to be taken with the reference mast height, as lower masts are particularly susceptible to the influences of ground surface roughness and elements, such as buildings, affecting those recordings. There is little that can be done about this but sometimes if a certain sector is obviously badly affected it can be ignored in the analyses. The approach from here usually depends on the time interval over which the reference data was measured. Bureau of Meteorology data in Australia was for many decades measured at 3 hourly intervals only and hence a direct correlation with 10-minute target data is not possible. Under such circumstances some designers chose a simple correction to the target data known as the “method of ratios” which in its purest form can be written as; Equation 12-5 where the term in square brackets is known as the “ratio” and the averages shown are over a certain seasonal period T. This equation looks complex but simply assumes that the ratio of short- to long-term changes in wind resource at the reference site is the same as the target site for that period. In practice, wind direction is not normally corrected using the method of ratios and the equation can be written purely for wind speeds as; 12-30 Equation 12-6 Using Equation 12-6 the reference wind speed data is averaged over the seasonal period T, often this is monthly, and then compared to the long-term reference average for that same period to provide the required ratio. The target data for that same seasonal period is adjusted using this ratio with either the total seasonal average or the discrete 10-minute data for that same period T being adjusted. The advantage of this more complex method is that, given many years of suitable reference data, the wind farm designer can obtain for the same period an estimate of the winds at the target wind farm site rather than just for the period of wind monitoring measurements. This then means that both wind direction and speed are available for that longer period which significantly improves the accuracy of the predictions, and that a shorter wind-monitoring period can be used. The disadvantage to the measure-correlate-predict method is the computational effort required and that not always a simple correlation exists either in direction or for wind speeds in certain sectors. To overcome this latter problem requires experience with applying the method. 12.2.1.3 The difference in wind resource experienced at each turbine location As wind direction is not normally corrected the designer may need to assume that from year to year it changes little and therefore affects yield results little. In some situations with very predominant winds this is an accurate assumption and similarly for a single wind turbine installation. The designer will also have to assume and that the wind speed ratios are independent of wind direction and this, unfortunately, is seldom the case. Typically a comparison with the reference wind rose or direction frequency histogram will be needed to see if there are substantial differences between reference and target wind directions or a longer windmonitoring period established to measure and quantify this. For a wind farm consisting of more than one turbine only one of the machines can sit exactly at the windmonitoring site. Siting restrictions may also mean that it is impossible to place a turbine at that location, or the wind monitoring mast may be required to be left in position for future testing of turbine performance. The further away from this point the more chance there is that the wind resource will differ and by how much depends on things such as changes in surface roughness elements, height, obstructions to the flow and distance from topographic speed-up elements. A far more accurate estimate of long term wind resource is the “measure-correlate-predict” method which is more widely used throughout the wind energy industry and which includes wind direction in the analyses [References: 11 & Reference: 12]. For this method it is best to obtain reference data on a similar time base to that measured at the proposed wind farm site, and in Australia this usually means obtaining 10-minute average Automatic Weather Station data from the Bureau of Meteorology. Figure 12-21 Definition of flat country in the surroundings of a wind turbine. The measure-correlate-predict method is based on wind direction sectors with the long-term reference data adjusted to reflect what is would have been if measured at the target site. The method achieves this by assuming that the relationships between the measured target data and the reference data, found by comparing these for the period the reference data was measured, are representative of a longer term relationship. It then uses these relationships to alter or “shift” the longer term reference data to the target site. This is best described in steps. Step 1: Measure - Wind speeds and directions are measured at the target site for a suitable period. For the reference site all data obtained to date, including this period, are also obtained. Step 2: Correlate A - Wind directions are correlated to obtain the relationship between the reference and target sites correlation coefficient can be obtained if the two locations are close, but sometimes a more sophisticated curve can be used if the correlation is not so obvious. Correlate B - For the same period as Step 2, corresponding reference and target site wind speed measurements are binned by wind direction (typically 15 degree bins) and a similar correlation coefficient obtained for wind speeds between both. This means that the relationship between the two measured wind speeds is known so that the wind speed at one can be estimated with knowledge of the other. Step 3: Predict – All of the long-term reference data is adjusted using, firstly, the sector wide wind speed correlation coefficients on wind speed and then the wind direction coefficients. This essentially “alters” the reference data to reflect that which it would be if it were measured at the target location. (Source: Taken from [Reference: XXXX]) On a flat landscape with little change in surface roughness measurements at one point will be representative of others on the site for a distance of tens of kilometres. This is particularly the case at inland sites in 12-31 12-32 relatively flat country as no particular topographic feature, such as a land/sea interface or steep topography, is responsible for the wind resource being there. According to [Reference: XXXX], flat country can be defined according to Figure 12-21 where: • • • hc is the maximum vertical difference across the site and the ratio of this to the distance between the highest and lowest ground point is less than 0.032 within 4km upwind and 0.8km downwind of the turbine, differences in elevation across the site do not exceed 60m within a radius of 11.5km, and, the height of the rotor relative to the lowest ground point within a distance of 4km upwind is at least three times greater than hc. Most wind farm sites in Australia rely on particular landscape features or land/sea positioning and, hence, wind resource will usually vary across the site and this means that energy yield will depend on turbine positioning. There is no simple way to adjust the wind resource across a more complex site. In Topic 12.1 micro modelling techniques were introduced and it is these that provide the wind farm designer with the best means to deal with such changes in wind resource. There are several proprietary modelling products available on the market to do this. Typically these micro models rely on detailed topographical information and user input of surface roughness elements and wind monitoring information. Figure 12-22 Example of energy yield output from a wind resource extrapolation/modelling computational model, also showing wake estimates using eddy/viscosity model. Most models use theoretical fluid mechanic principles that have been verified by empirical results from field work to calculate the wind flow over the terrain in question in fine detail and, as discussed previously, have become accurate enough that even quite complex terrain can be dealt with. Usually the models calculate the wind speed and direction histograms at user-defined points, such as a grid that covers the area. These are then in turn used with the input Certified Power Curve for the turbine in question resulting in application of Equation 12-3 and an energy yield estimate. Figure 12-22 gives an output from such a model, showing estimated wind energy yield based on input topographic, turbine and wind resource information. 12.2.1.4 Wind turbine technical availability and wind farm capacity factor Wind turbines like all machines have periods where they are not able to operate and this has to be estimated by the wind farm designer as this represents a loss in energy production. For wind turbines this “down-time” can be due to a number of reasons, such as; • • • • • regular or planned maintenance unplanned maintenance (unexpected turbine faults) grid problems lack of wind, and winds too high In Topic 3, the concepts of availability and capacity factor for a single wind turbine were introduced. The following section seeks to clearly define those terms within the context of an operational wind farm. In the electricity generation industry the term availability, A, has been introduced to characterise a power plant’s ability of generating energy or of executing certain functions and is defined in [Reference: Error! as: Bookmark not defined. Equation 12-7 where T is a time period and subscripts V and N corresponds to plant available and nominal reference time respectively. Nominal reference time is simply the total elapsed time during the availability scrutiny and is often a calendar year. Typical availabilities of fossil fuel based generating plant are between 80 and 90%, meaning that for 10-20% of the time they are unavailable to generate. For wind turbines the definition of TV has taken some time to evolve and there is still some debate between manufacturers and interested groups about how it should be defined. This is because, unlike fossil fuel plants, a wind turbine’s energy resource, the wind, is out of the control of the operator – this means there can be periods under which the turbine could operate were the wind blowing. The question then arises as to whether a wind turbine’s availability figure should include low wind speed periods or just be periods when the wind is above cut-in – as the former depends heavily on the wind regime at the wind farm site it may be misleading in that a high availability may not necessarily mean that wind turbine has higher energy yield. (Source: WindFarmer software from Windops Ltd) Regardless of the period over which TV is ultimately calculated, modern wind turbines have very good availability with the majority of machines now exceeding 95%. 12-33 12-34 A parameter related to availability but more often used in the wind energy industry to characterise a wind farm’s energy yield is the capacity factor, C, which is defined as; lost due to wake effects. Figure 12-23 shows array efficiency for various array spacings for a 16-turbine square wind farm layout [Reference: XXXX]. Wakes have several practical ramifications for the wind farm designer. For a single wind turbine unless the surrounding topography’s surface roughness and/or topography changes markedly around the turbine then wind direction information has little influence on energy yield. With more than one turbine the influence of wakes will vary with wind direction and, therefore, such direction must be taken into account in energy yield estimates and this is of the most important reasons for measuring wind direction during wind monitoring programs. Equation 12-8 where is the wind farms average power output and PR the rated wind farm power output. As an example, the Esperance Ten Mile Lagoon wind farm has a C of 32%, meaning that, on average, the wind farm is producing 32% of its rated power. Another way of calculating C is to simply use the annual energy yield E, as in Topic 3; Wakes also lead to increased fatigue damage on downstream turbines and a lowering in turbine availability and, consequently, energy yield. In the early days of wind farm design, some multiple turbine wind farms were built on a square grid pattern with turbines as close as 3D. It was found that the turbines in the central part of the grid in these farms experienced substantially increased downtime due to this problem. Figure 12-23 Aerodynamic array efficiency for a 16 turbine square wind farm array and various turbine separation efficiencies. Equation 12-9 While C is stated often in wind energy circles, it really has very little technical use as its value depends heavily on the rated wind speed of the turbine being used. Hence, if two wind turbines are being evaluated for a site and both have different rated wind speeds, then the capacity factor can be skewed in favour of one or the other. A far better approach in this respect is to use the energy yield E and evaluate the turbines using an economic analyses to give a cost per unit of electricity. 12.2.1.5 Wake effects and turbine spacing When an element of wind passes through a wind turbine, the blades extract energy from it and as a result its momentum is lowered and the amount of turbulence in it increases. As that element continues downstream in the turbine’s wake, were it not to encounter any further flow disturbance with time it would redevelop under the influence of an infinite undisturbed flow-field surrounding it. Such redevelopment can take up to 20 rotor diameters (20D), or more [Reference: XXXX], and depends greatly on the ambient turbulence intensity, Iu, with higher turbulence leading generally to a quicker wake recovery. If another wind turbine lies within that wake, depending on how close it is to the upstream machine it will experience both a decrease in momentum flux in the flow and an increase in turbulence. Both of these lead to a lowering in the energy extracted by that wind turbine while the increased turbulence can lead to greater fatigue loads in both frequency and magnitude. If the wind farm consists of many turbines and, therefore, the flowfield within the wind farm is influenced by multiple wind turbine wakes, then the effects can be felt on many of the machines and compounds depending on the number of machines upwind. (Source: Taken from [Reference: XXXX]) The loss of energy yield due to wake effects is quantified by what is known as the Array Efficiency, which is a term describing the amount of energy extracted by the real wind farm divided by that which would have been extracted were there to be no wake losses, and is expressed as a percentage. Typical wind farm designs have array efficiencies between 90 and 95%, meaning that around 10% of the energy in the flow is typically The wake behind a turbine is very complex immediately behind the rotor and this area is known as the near wake region, which extends about 1D downwind of the turbine. As the flow develops further downstream the turbulence within it and that ambient from the surrounding air acts on the wake to erode and mix it to the point where it eventually merges with that of the original flow. This profile development area between the near wake and this point is the area to which downstream turbines could possibly be placed. Numerous analytical and empirical wake models [Reference: XXXX] have been developed for this regions because of this and many of the computer micro-modelling wind farm software packages include these to estimate the effects of wakes on energy yield. Figure 12-22 includes the turbine outputs which show wake losses in percent calculated using what is known as a Eddy-Viscosity wake model (discussed in [Reference: XXXX]). 12-35 12-36 Practical experience with what wakes mean for wind farms has resulted in a number of rules of thumb for wind turbines placement that are given in Table 12-2. This Table shows separation distances in the direction of, and normal to, the prevailing wind direction. For sites with a variable wind direction the turbines would have to be between 7D and 10D apart while for a strongly directional site they could be as close as 3D. Many wind turbines have failed in storm conditions and nowadays suppliers offer models with different blades lengths, tower types and strengths to conform to standards for different site conditions. The International Electrotechnical Commission (IEC) publishes standards which specify four different classes of turbine which are designed to suite different wind conditions, and these are shown in Table 12-3 taken from [Reference: 11]. To meet the standard the turbine must be designed to withstand these conditions. Table 12-2 Rules of thumb for turbine spacing in a wind farm. Turbine separation In strongly predominant wind direction 7 – 10 rotor diameters In direction normal to predominant designed for a given wind resource, particularly in relation to high winds and an ability to survive extreme wind speed events. The wind farm will also require insurance and, again, the risk profile of the development will be better and therefore insurers more likely to lower premiums for a project if the machine’s characteristics match those of the resource in which it will operate. 3 – 5 rotor diameters Table 12-3 Wind speed parameters for wind turbine classes according to the IEC wind turbine standard IEC 61400-1 Wind turbine generator systems (taken from [Reference: Reference: 11]). Parameters Obviously it is not always possible to separate turbines by a large amount to minimise wake effects as this can lead to increased electrical cabling length, more land disturbance and area required and, overall, greater project cost. The figures in Table 12-2 have been found, generally, to be tradeoffs between wake losses and extra separation costs and under such distances an array efficiency of around 90-95% would be expected. 12.2.1.6 Electrical line losses Often overlooked by wind farm developers, the electrical losses that occur from electrical cabling and wind turbine transformers can be considerable and lead to a substantial lowering of wind farm output. The effect on the wind farm’s economics of such losses depend almost entirely on where the electrical output of the wind farm is measured and metered – if, for example, the output is measured at the wind turbine’s low voltage terminals then such losses will be negligible. If, however, the power meters are measured at the connection point to the grid, then such losses can be significant. Electrical line losses vary but a value of around 3-5% for cabling and transformer losses is fairly typical. Electrical line losses will be discussed further in Topic 13. 12.2.2 Site Wind Characteristics Affecting the Choice of Wind Turbine The job of selecting the right turbine for a wind farm development is largely dictated by the final electricity generation costs. This is because a number of wind turbine standards (see for example Appendix 4 of [Reference: XXXX]) have been developed over the years and these provide the wind farm designer comfort in that the machines would be expected to operate according to that standard; that is, efficiently and safely. There are, however, a number of issues that should be examined beforehand in relation to the wind resource at the site that may influence the decision to purchase a certain turbine type, being; Reference wind speed, Uref (m/s) Annual average wind speed, Uave (m/s) 50 year return gust speed, 1.4Uref (m/s) 1 year return gust speed, 1.05Uref (m/s) Class I Class II Class III Class IV 50 42.5 37.5 30 10 8.5 7.5 6 70 59.5 52.5 42 52.5 44.6 39.4 31.5 In Table 12-3, Uref refers to the 10-minute wind speed at hub height with a return period of 50 years, which is defined simply as five times the annual mean wind speed. The IEC standard also uses turbulence intensity, Iu, which it specifies as two levels A (higher) and B (lower). Note that extreme wind states are typically given in the wind industry as a 50 year return value. Other standards such as Gemanischer Lloyd from Germany, and Dutch and Danish standards, also exist to specify site conditions a turbine must be capable of meeting and turbine manufacturers choose these or the more widely targeted IEC standards, or both, to apply in their designs. Following a testing procedure for each turbine type, which is undertaken by testing authorities, certificates of compliance are issued and are typically made available by manufacturers to intending purchasers. Note that Australia does not and may never have its own technical standards for wind turbines. It does, however, have Australian Standards for the design of structures for wind loading (AS1170 part 2) and it is sometimes a requirement of approval bodies to have such standards applied to wind farm projects by a structural engineer. These standards divide the continent into wind class sectors, depending on the likelihood of events such as cyclones, and various factors are used to take account of the terrain around the intended wind farm. These issues may make a certain turbine type or size more appropriate and this will in turn affect the wind farm design itself. Wind farm financiers will also take comfort in a wind turbine which is specifically In bad storms with high, gusty winds, a wind turbine usually protects itself by some form of action. With pitch regulated turbines this could mean stopping the rotor and facing into the wind with all blades at 90 degrees of pitch, or for a stall controlled turbine stopping and yawing to 90 degrees out of the wind. However, caution must be exhibited if such mechanisms are disabled through loss of the grid supply that is a common occurrence in Australia during storms. Because of this it is normal practice in Australia to specify that protection must be capable of being applied without grid supply, and turbine manufacturers can do this through energy storage mechanisms built into the turbines such as batteries or hydraulic accumulators that in turn activate the required mechanisms. 12-37 12-38 • • • Annual average wind speed at the site Turbulence intensity at the site Maximum wind speeds expected A wind farm designer cannot be expected to know the intricacies of all the wind turbine standards so it is normal practice to specify to the turbine supplier what is known about the site wind resource, so that an appropriate wind turbine can be put forward. Typically a wind turbine of a lower Class will be more expensive as it is designed for much harsher conditions and, therefore, is built more strongly. Such turbines usually have shorter blades so that energy yield will be lower. Note that wind turbines installed in Australia are usually IEC Class I or II (see Table 12-3), due to the relatively high wind speeds experienced here compared to mainland Europe. A final note is required in regard to the calculation of extreme wind speeds based on site data that often the developer calculates to specify to the turbine supplier. As has been seen previously, a Weibull probability distribution is usually calculated based on measurements of measured site to supply a frequency distribution for energy yield calculations. In such distributions there is usually a paucity of information at higher wind speeds and hence the distribution is inaccurate at these wind speeds. Several means of estimating the extreme wind speed in terms of various return periods are available. A popular method is that attributed to Gumbel outlined in [Reference: 10], which uses a separate probability distribution for measured high wind speed events to predict the maximum wind speeds expected on a specified return period. Variations on this occur with one being to correlate measurements to the closest reference site, as in the Measure-Correlate-Predict method, so that a much longer period of measurements is established and the accuracy of predictions improved. ----//---- Exercise 12-1 Example adjustments to measured wind speed data to reflect the long term average, using the simple “method of ratios” technique. Problem: A wind farm is proposed at a site, the “target”, over which wind speed and direction information has been monitored for 12 months starting 1 January 2003 at a height of 30 and 50m at 10 minute intervals. A Bureau of Meteorology (BOM) 10m mast is located 20km from the target and has been sampling wind speed at 3 hourly intervals for 42 years and can be used as the “reference”. Predict the long-term monthly wind speed averages for the target site by using the simple “method of ratios”. A 50m-hub height is required. Comment on how now an estimate of the long-term wind speed and direction frequency histogram can be produced for the target site including the corrections. What further comments can be made about wind direction, which has not been corrected in any way? Solution: As the target data has wind speed information at the hub height of the turbine no shear calculations or corrections are necessary. The BOM recordings are three hourly only and hence a direct correlation with the measured ten minute averages is not possible. Table 12-4 Long-term prediction of target wind resource using method of ratios Reference site (BOM site) Target site (wind farm site) Year 2003 monthly average wind speed at 10m in m/s u R (month ) January February March April May June July August Septembe r October Novembe r Decembe r Long term monthly average wind speed at 10m in m/s u R (month )longterm Ratio of 2003 to long term monthly average s Year 2003 monthly average wind speed at 50m in m/s uT (month ) Corrected monthly average at 50m in m/s using Equation 12-6 uT (month )longterm 5.5 5.7 5.2 4.9 4.7 5.5 6.0 6.2 5.9 5.1 5.3 4.9 5.0 5.1 5.2 5.9 6.4 6.3 0.93 0.93 0.94 1.02 1.09 0.95 0.98 1.03 1.07 7.3 7.3 6.5 6.3 6.2 7.5 8.0 8.3 8.1 6.8 6.8 6.1 6.4 6.8 7.1 7.8 8.5 8.7 5.6 6.0 5.8 5.7 1.04 0.95 7.5 7.7 7.8 7.3 5.4 5.2 0.96 6.8 6.5 7.3 7.2 Target averages Using the ratios in the table it is possible to adjust the target 50m and 30m discrete ten-minute average data for each equivalent month. This gives one year (52560 points) of corrected ten minute average data points for each height which can be used to make a frequency histogram and an energy yield estimate using Equation 12-3. As wind direction is not corrected, its frequency histogram remains unaltered and this creates some uncertainty in that the wind farm design may be very dependent on direction and hence if seasonal changes do exist this may lead to an inappropriate wind farm design. To check this a comparison between the target and reference wind direction data should be made or a longer wind-monitoring program established. The BOM reference data has been broken down into monthly averages for each of the 42 years and the longterm averages for each month calculated. This and the reference monthly averages for 2003 are shown in Table 12-4 below with the ratio of each calculated in accordance with Equation 12-6. The wind speed averages for each month as measured at 50m at the target site are also shown, as is the result of applying Equation 12-6 to give the estimated long-term target monthly averages. Exercise 12-2 Simplified example of measure-predict-correlate method for long term wind resource estimation. 12-39 12-40 Problem: The consulting engineer representing the banking institution financing the wind farm proposed in Exercise 12-1 is not satisfied with the long term wind resource estimations using the simple method of ratios. The engineer has requested the wind farm designer to provide calculations that include wind direction and generally provide a more accurate result. Discussions with the Bureau of Meteorology have shown that an Automatic Weather Station measuring hourly average wind speed and wind direction at a height of 10m was installed at the reference location in January 1991 until the present, and this data is available for purchase. If there was only a few years of such high resolution reference data available, how would that change the analyses? Solution: Step 1. Measure - We now have a period of wind resource measurement at the target site matched by a reference site, as well as 11 previous years from that reference site. The BOM site only measured hourly averages but this is adequate for the measure-correlate-predict method of long term estimation – if only three hourly BOM spot readings had been taken then this would not be the case as, being spot readings, they take no account of fluctuations in the wind resource between readings. The period of target site measurement is 1/1/03 to 31/12/03 which gives 52560 ten minute average wind speed and direction readings. This is averaged over each hour to give 8760 hourly averages, matching exactly the time and date stamps on the BOM data. Step 2. Correlate A – For each hourly average for the period 1/1/03 to 31/12/03 we graph reference 50m data against BOM data, matching point to point. Only the 50m data is used as this is the hub height of the turbine of interest. This gives the graph in the figure below. (Source: P. Ebert, Western Power) A least squares fit forced through zero is put through the data, the slope of which becomes the correlation coefficient for wind directions. The correlation between reference and target site is quite strong as indicated by the small amount of scatter in the graph, meaning that wind directions measured at both sites are very similar with time. Correlate B – Target data for the period 1/1/03 – 31/12/03 is binned by wind direction into twenty-four 15° sectors as 0 to 15°, 15 to 30°, 30 to 45° ……. 315 to 330°, 330-345°, 345-360° and the time and date stamp on each individual data point in those sectors noted. For the data in each sector, wind speeds are graphed against the corresponding date and time stamped BOM data with the final result for the 315 to 330° sector is shown in the figure below. Figure 12-25 Figure showing 315 to 330 degree sector correlation for wind speeds Figure 12-24 Figure showing wind direction correlation. (Source: P. Ebert, Western Power) A least squares fit forced through zero is calculated for the data for each sector graph, the slope of which becomes the correlation coefficient for all wind speeds in that sector. The final result is 24 straight-line slopes, which become the correlation coefficients for wind speeds. Step 3. Predict – Now using all the 12 years of BOM data available, firstly the wind direction is adjusted using the correlation coefficient of the Correlate A step above – this essentially “shifts” this data to be as it would be if measured at the target site. Following this, all wind speeds are binned by 15° sectors and the 24 12-41 12-42 correlation coefficients calculated in Correlate B above are used to adjust the wind speeds accordingly to, again, “shift” the reference site wind speeds to the target site. 5. Dear, S. & Lyons, T. J. (1991). “Western Australian Wind Atlas: Atlas Program”, Minerals and Energy Research Institute of Western Australia, Supplement to Report no. 64, project number E162. Providing time and date stamps are maintained with each BOM data point the final result is 12 years of hourly averages “shifted” to be as if they were measured at the target site. This data can then be used to produce wind speed and direction histograms and with Equation 12-3 a more accurate long-term energy yield estimate produced for the financing institution. As a comparison with the method of Exercise 12-1, Table 12-5 gives the resulting monthly averages which show a significant difference to those in the previous example, which has a corresponding influence on the projected long term energy yield and hence project financial viability. This indicates most probably that ground topography is a significant influence on the wind direction and wind speeds at the target site. 6. “Modelling Utility-Scale Wind Power Plants. Part 2: Capacity Credit”, Milligan, M. R., Wind Energy, Volume 3, Number 4, October-December 2000. January February March April May June July August September October November December Target average 8. Outhred, H. 2002 “Power System Operation and Network Issues for Wind Farms”, Proceedings of the Australian Wind Energy Association Conference, 23-26 July, Glenelg, South Australia. 9. Hau, E. (2000) “Wind Turbines: Fundamentals, Technologies, Applications, Economics”, Springer Publishers. Table 12-5 Comparison of methods illustrated by Examples 12-1 and 12-2. Month 7. Arnott, I. (2002). “Intermittent Generation in the National Electricity Market”, proceedings of the Australian Wind Energy Association Conference, Adelaide, 23-26 July, 2002. Corrected monthly averages at 50m height in m/s “Measure-correlate-predict” method 6.8 6.7 6.8 6.9 6.1 6.4 6.4 6.5 6.8 6.7 7.1 7.0 7.8 7.8 8.5 8.3 8.7 8.5 7.8 7.7 7.3 7.3 6.5 6.7 7.22 7.21 10. IEA Expert Group Study: Recommended practices for Wind Turbine Testing and Evaluation 1. Power Performance Testing, 3 Edition 1992. “Method of ratios” method 11. Burton, T., Sharpe, D., Jenkins, N. and Bossanyi, E. (2001) “Wind Energy Handbook”, Wiley. 12. Manwell, J.F., McGowan, J.G. & Rogers, A.L. 2002 “Wind Energy Explained – Theory, Design & Application”, Wiley & Sons Providing there are more years of reference data available than target data, it is always an advantage to correlate and correct wind resource information in this way. Long-term prediction accuracy of course depends on the number of reference data years available and as a check the wind farm designer should obtain lower resolution but longer-term reference data to see how other years have varied in the past. References: 1. Ebert, P. R. (1999) “Stakeholder Management: Ignore it and your wind farm project may never happen”, proceedings of the Australian Wind Energy Conference, June 28-30, Newcastle, NSW. 2. “Siting Struggles: The Unique Challenge of Permitting Renewable Energy Plants”, Kahn, R. D., The Electricity Journal, March 2000 3. “Best Practice Guidelines for the Implementation of Wind Energy Projects in Australia”, the Australian Wind Energy Association, March 2002. 4. Wind,Waves, Weather – Perth Waters. Boating Weather Series by the Australian Bureau of Meteorology, July 1993. 12-43 12-44 ...
View Full Document

This note was uploaded on 06/09/2011 for the course PV 5053 taught by Professor Aasd during the Three '11 term at University of New South Wales.

Ask a homework question - tutors are online