Download Wind Resource Assessment and more Essays (university) Mechanical Engineering in PDF only on Docsity!
General rights Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.
- Users may download and print one copy of any publication from the public portal for the purpose of private study or research.
- You may not further distribute the material or use it for any profit-making activity or commercial gain
- You may freely distribute the URL identifying the publication in the public portal
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from orbit.dtu.dk on: Oct 27, 2018
Wind resource assessment using the WAsP software (DTU Wind Energy E-0135)
Mortensen, Niels Gylling
Publication date: 2016
Document Version Publisher's PDF, also known as Version of record
Link back to DTU Orbit
Citation (APA): Mortensen, N. G. (2016). Wind resource assessment using the WAsP software (DTU Wind Energy E-0135). Technical University of Denmark (DTU). DTU Wind Energy E, No. 0135
Department of
Wind Energy
E Report 2017
46200 Planning and Development of Wind Farms:
Wind resource assessment using the WAsP software
Niels G. Mortensen
DTU Wind Energy E-0135 (ed.2)
December 2017
Contents
- 1 Introduction
- 1.1 Observation-based wind resource assessment
- 1.2 Numerical wind atlas methodologies
- 1.3 Wind resource assessment procedure
- 1.4 Energy yield assessment procedure
- 2 Meteorological measurements
- 2.1 Design of a measurement programme
- 2.2 Quality assurance
- 3 Wind-climatological inputs
- 3.1 Wind data analysis
- 3.2 Observed wind climate
- 4 Topographical inputs
- 4.1 Elevation map
- 4.2 Land cover map
- 4.3 Sheltering obstacles
- 5 Wind farm inputs
- 5.1 Wind farm layout
- 5.2 Wind turbine generator
- 6 WAsP modelling
- 6.1 Modelling parameters
- 6.2 WAsP analysis
- 6.3 WAsP application
- 6.4 Validation of the modelling
- 6.5 Special considerations
- 7 Additional technical losses
- 8 Modelling error and uncertainty
- 8.1 Prediction biases
- 8.2 Sensitivity analysis
- 8.3 Uncertainty estimation
- 9 Wind conditions and site assessment
- 9.1 Extreme wind and turbulence intensity
- 9.2 IEC site assessment
- References
- Acknowledgements
- A WAsP best practice and checklist Appendices
- B Note on the use of SAGA GIS
- C Digitisation of the land cover (roughness) map
- D The Global Wind Atlas
46200 Planning and Development of Wind Farms
The general course objectives, learning objectives and contents for DTU 46200 are listed below for reference. The full course description is given in the DTU Course Catalogue. The present notes are related to the wind resource assessment and siting parts only.
General course objectives
The student is provided with an overview of the steps in planning and managing the development of a new wind farm. The student is introduced to wind resource assessment and siting, wind farm economics and support mechanisms for wind energy. An overview of the various environmental impacts from wind farms is offered.
Learning objectives
A student who has met the objectives of the course will be able to:
- Describe the methodologies of wind resource assessment and their advantages and limitations.
- Explain the steps in the selection of a site for measurement of the wind resource and good practice for measurement of the wind resource.
- Calculate the annual energy production using the WAsP software for simple wind farm cases in terrain within the operational envelope of the WAsP model.
- Identify and describe factors adding to the uncertainty of the wind resource and wind farm production estimates.
- Estimate the most important key financial numbers of a wind project and explain their relevance.
- Identify the main environmental impacts from a wind farm and suggest mitigation measures.
- List the three most common policy tools for support of wind energy projects.
- Explain the steps in the development of a wind farm layout considering annual energy production, wind turbine loads and environmental impact.
- Explain the main steps in developing the grid connection of a wind farm.
Contents
An introduction to market, policy and support mechanisms relevant to wind energy. Wind resources and wind conditions: anemometry; design and siting of meteorological stations; wind distributions; observed, generalised and predicted wind climates; observational and numerical wind atlases, elevation maps and land cover, roughness classes and roughness maps; sheltering obstacles; wind farm wake effects, micro-scale flow modelling (WAsP), wind resource mapping; wind farm layout; wind farm annual energy production.
The procedure for obtaining an environmental permit for a wind farm. The various types of environmental impacts from a wind farm. Introduction to wind farm economics. Introduction to grid connection. The students will work in groups of 3 or 4. The group work will be documented in a report and will be presented orally by all course participants.
The result of wind resource assessment is therefore an estimate of the mean wind climate at one or a number of sites, in the form of:
- Wind direction probability distribution (wind rose), which shows the frequency distribution of wind directions at the site, i.e. where the wind comes from,
- Sector-wise wind speed probability distribution functions, which show the frequency distributions of wind speeds at the site.
Wind resource assessment provides important inputs for the siting, sizing and detailed design of the wind farm and these inputs are exactly what the WAsP software provides.
When it comes to the siting of individual wind turbines, a site assessment (IEC 61400-1) is usually carried out. This will provide estimates for each wind turbine site of the 50- year extreme wind, shear of the vertical wind profile, flow and terrain inclination angles, free-stream turbulence, wind speed probability distribution and added wake turbulence. This additional information may be obtained by using the WAsP Engineering software.
1.1 Observation-based wind resource assessment
Conventionally, wind resource assessment and wind farm calculations are based on wind data measured at or nearby the wind farm site. The WAsP software (Mortensen et al .,
- is an implementation of the so-called wind atlas methodology (Troen and Petersen, 1989); this is shown schematically in Figure 3.
WAsP analysis: from wind data to generalised wind climate
- Time-series of wind speed and direction → observed wind climate (OWC)
- OWC + met. mast site description → generalised wind climate (wind atlas)
WAsP application: from generalised to predicted wind climate
- Generalised wind climate + site description → predicted wind climate (PWC)
- PWC + power curve → annual energy production (AEP) of wind turbine
Wind farm production: from predicted wind climate to gross AEP
- PWC + wind turbine (WTG) characteristics → ‘WAsP gross’ AEP of wind farm
- PWC + WTG characteristics + wind farm layout → wind farm wake losses
- ‘WAsP gross’ AEP – wake losses → ‘WAsP net’ production of wind farm
Post-processing: from ‘WAsP net’ AEP to net AEP ( P 50 and P x )
- ‘WAsP net’ AEP – technical losses → net annual energy production ( P 50 )
- Net AEP – uncertainty estimate → Net AEP Px
Figure 3. Wind atlas methodology of WAsP (Troen and Petersen, 1989). Meteorological models are used to calculate the generalised wind climatology from the measured data – the analysis part. In the reverse process – the application of wind atlas data – the wind climate at any specific site may be calculated from the generalised wind climatology.
Note, that the WAsP software estimates the ‘WAsP gross’ and ‘WAsP net’ AEP only (steps 1-7 in Figure 3); the post-processing steps (8-9) must be carried out separately. The wind farm assessment tool (WAT) contain simple tools to aid in these calculations.
As can be deduced from Figure 3, WAsP is then based on two fundamental assumptions : first, the generalised wind climate is assumed to be nearly the same at the predictor (met. station) and predicted sites (wind turbines) and, secondly, the past (historic wind data) is assumed to be representative of the future (the 20-y life time of the wind turbines). The reliability of any given WAsP prediction depends very much on the extent to which these two assumptions are fulfilled.
1.2 Numerical wind atlas methodologies
WAsP has become part of a much larger framework of wind atlas methodologies, which also encompasses mesoscale modelling and satellite imagery analysis. This framework has been developed over the last two decades at Risø and DTU (Frank et al. , 2001; Badger et al ., 2006; Hansen et al ., 2007) in order to be able to assess the wind resources of diverse geographical regions where abundant high-quality, long-term measurement data does not exist and where important flow features may be due to regional-scale topography. Figure 4 is a schematic presentation of this entire framework.
Figure 4. Overview of state-of-the-art wind atlas methodologies (Hansen et al., 2007).
Wind resource assessment based on mesoscale modelling, the numerical wind atlas, can provide reliable data for physical planning on national, regional or local scales, as well as data for wind farm siting, project development, wind farm layout design and micro-siting of wind turbines. However, bankable estimates of power productions from prospective wind farms require additional on-site wind measurements for one or more years.
The present course notes thus describe mainly the ‘grey’, ‘green’ and ‘yellow’ parts of the diagram above, i.e. what is referred to as the observational wind atlas methodology. Different inputs to the WAsP modelling are described in Sections 2 to 5 ; the modelling itself is described in Section 6 , and the modelling errors and uncertainties in Section 8. Section 7 lists the different types of additional losses in the wind farm and Section 9 contains a very brief cookbook approach to site assessment using WAsP Engineering.
In addition to the present course notes, the WAsP help system (Mortensen et al .,
- contains a Quick Start Tutorial section which illustrates the essentials of the WAsP software user interface.
1.4 Energy yield assessment procedure
We can focus on the energy yield assessment procedure in a similar way as above and identify the following steps (Figure 6):
1. Site wind climate = Site wind data ± [long-term extrapolation effects] Using a long-term extrapolation procedure , the site wind data are referenced and adjusted according to the long-term climatology of the area. 2. Reference AEP = Wind climate at hub height plus [power curve] The reference AEP is calculated using the predicted wind climate at hub height at the mast location and the site-specific wind turbine power curve. However, most of the time this step is surpassed and the gross AEP is calculated directly. 3. Gross AEP = Reference AEP ± [terrain effects] Using a flow model , the observed wind climate at the mast site is transformed to the predicted wind climates at the wind turbine sites of the wind farm. The ‘flow modelling’ part of Figure 6 includes both vertical and horizontal extrapolations. 4. Potential AEP = Gross AEP – [wake losses] Using a wake model , the wake losses at each turbine site are estimated and subtracted from the gross AEP. This corresponds to the WAsP ‘net AEP’. 5. Net AEP = Potential AEP – [technical losses] The additional technical (operational) losses in the wind farm are subsequently estimated and subtracted from the potential AEP to get the net AEP value ( P 50 ) at the point of common coupling (PCC). 6. P 90 AEP = P 50 AEP – 1.282×[uncertainty estimate] The aggregate uncertainty of the entire energy yield process is estimated and the net AEP is adjusted to obtain a net value corresponding to a certain probability of exceedance, e.g. the P 90 value as shown above.
By dividing the prediction process into these steps we have isolated the different model calculation results and it is therefore fairly straightforward to compare different methods and models (Mortensen et al ., 2012, 2015). Figure 6 illustrates the steps in the procedure.
Figure 6. Overview of steps in the wind farm energy yield assessment procedure.
These steps and their definitions are not universally agreed or even used; however, IEC and Measnet working groups are addressing these issues at the moment.
2 Meteorological measurements
WAsP predictions are mostly based on the observed wind climate at the met. station site; i.e. time-series data of measured wind speeds and directions over one or several years that have been binned into intervals of wind direction (the wind rose) and wind speed (the histograms). Therefore, the quality of the measurement data has direct implications for the quality of the WAsP predictions of wind climate and annual energy production. In short, the wind data must be accurate , representative and reliable.
2.1 Design of a measurement programme
It is beyond the scope of these course notes to describe best practice for wind measure- ments in detail, but the aspects discussed below are particularly important.
If possible, the measurement programme should be designed based on a preliminary WAsP analysis of the wind farm site. Such design ensures that the measurements will be representative of the site, i.e. that the mast site(s) represent the relevant ranges of elevation, land cover, exposure, ruggedness index, etc. found on the site. In short, we apply the WAsP similarity principle (Landberg et al ., 2003) as much as possible when siting the mast(s). This design analysis may conveniently be based on SRTM elevation data and land cover information from satellite imagery such as Google Earth.
It is equally important in the design stage to use an observed wind climate that resembles the wind climate that may be observed at the wind farm site; e.g. by using data from a nearby met. station or modelled data from the region. A representative wind rose is particularly valuable as this may be used to determine the design of the mast layout; e.g. the optimum boom direction is at an angle of 90° (lattice mast) or 45° (tubular mast) to the prevailing wind direction. The height of the top (reference) anemometer should be similar to that of the wind turbine hub height; preferably > 2/3 h hub.
Anemometers should be individually calibrated according to international or at least traceable standards. Several levels of anemometry should be installed in order to obtain a high data recovery rate (above 90-95%) and for analyses of the vertical wind profiles. Air temperature (preferably at hub height) and barometric pressure should be measured in order to be able to calculate air density, which is used to select the appropriate wind turbine power curve data set.
It is extremely valuable – and sometimes required for bankable estimates – to install two or more masts at the wind farm site; cross-prediction between such masts will provide assessments of the accuracy and uncertainty of the flow modelling over the site. Two or more masts are also required in complex and steep terrain, where ruggedness index (RIX) and ∆RIX analyses – as well as WAsP CFD calculations – are necessary.
2.2 Quality assurance
For projects where the measurement campaign has already been initiated or carried out, it is important to try to assess the quality of the collected wind data, as well as to ensure the quality of any and all site data used for the analysis. A site inspection trip is required and should be part of any (commercial) WAsP study – whether it is a second opinion, due diligence or feasibility study.
A number of WAsP Site/Station Inspection checklists and forms exist for planning the site visit and for recording the necessary information. The positions of the met. mast(s) and turbine sites are particularly important. Bring a handheld GPS (Global Positioning
Figure 8. Time traces of wind direction (upper, 0-360°) and wind speed (lower, 0- m/s) from Sprogø for the year 1979. The graphic in the lower left of the Climate Analyst window shows concurrent data in a polar representation (data courtesy of Sund & Bælt).
The Climate Analyst checks the time stamps and observation intervals upon input of each data file, and also checks for missing records in the data series. However, the main quality assurance (QA) is done by visual inspection of the time series and polar plot, as well as the resulting observed wind climate. Things to look out for are e.g.:
- Are there any spikes or sudden drops in the data series?
- Are there periods of constant data values in the data series?
- Are there periods of missing data in the data series?
- Are there any unusual patterns in the data series?
- Are there any unusual patterns in the polar scatter plot?
- Do the wind speed time traces follow each other for different anemometers?
- Do the wind direction time traces follow each other for different vanes?
- Do the measured and Weibull-derived values of U and P compare well?
- Does the calm class (0-1 ms−^1 ) in the histogram look realistic?
Finally, the observed wind climate is calculated and exported to an OWC file. The OWC file can subsequently be inserted into the WAsP hierarchy, as a child of a meteorological station member.
3.2 Observed wind climate
The observed wind climate (OWC) should represent as closely as possible the long-term wind climate at anemometer height at the position of the meteorological mast. Therefore, an integer number of full years must be used when calculating the OWC, in order to avoid any seasonal bias. For the same reason, the data recovery rate must be quite high (> 90-95%) and any missing observations should preferably be distributed randomly over the entire period.
Wind data series from prospective wind farm sites rarely cover more than one or a few full years, so they must be evaluated within the context of the long-term wind climate, in order to avoid any long-term or climatological bias. Comparisons to near-by, long-term meteorological stations or to long-term modelled data for the area can be made using simple (or complicated) measure-correlate-predict (MCP) techniques.
WAsP uses Weibull distributions to represent the sector-wise wind speed distributions and the so-called emergent distribution for the total (omni-directional) distribution. The difference between the fitted (and emergent) and the observed wind speed distributions should therefore be small: less than about 1% for mean power density (which is used for the Weibull fitting) and less than a few per cent for mean wind speed.
4 Topographical inputs
The topographical inputs to WAsP are given in a vector map , which can contain height contour lines, roughness change lines and lines with no attributes (say the border of the wind farm site). In addition, nearby sheltering obstacles may be specified in a separate obstacle group , which can be shown on the map too.
Map coordinates and elevations must be specified in meters and given in a Cartesian map coordinate system. The map projection and datum should be specified in the Map Editor so this information is embedded in the map file. All metric coordinates used in the WAsP workspace should of course refer to the same map coordinate system. Obstacle distances and dimensions must likewise be given in meters.
The Map Editor can do the transformation from one map coordinate system to another; the Geo-projection utility program in the Tools menu can further transform single points, lists of points and lists of points given in an ASCII data file.
4.1 Elevation map
The elevation map contains the height contours of the terrain. These may be digitised directly from a scanned paper map – as described in the Map Editor Help facility – or may be obtained from a database of previously digitised height contours, established by e.g. the Survey and Cadastre of a country or region. Alternatively, they can also be generated from gridded or random spot height data using contouring software.
The elevation map should extend at least several (2-3) times the horizontal scale of significant orographic features from any site – meteorological mast, reference site, wind turbine site or resource grid point. This is typically 5-10 km. A widely cited rule for the minimum extent of the WAsP map is max(100× h , 10 km), where h is the height of the calculation point above ground level; this is usually sufficient for the elevation map.
The accuracy and detail of the elevation map are most critical close to the site(s), therefore it is recommended to add all spot heights within the wind farm site and close to the meteorological mast(s); one can also interpolate or digitise extra height contours if necessary. The contour interval should be small (≤ 10 m) close to calculation sites, whereas the contour interval can be larger further away from these sites (≥ 10 m).
Non-rectangular maps (circular, elliptic, irregular) are allowed and sometimes preferred, e.g. in order to reduce the number of points in the map, while at the same time retaining model calculation accuracy. There is no limitation to the size of the map, but the calculation time is proportional to the computer memory used for the map data.
The final elevation map should be checked for outliers and errors by checking the range of elevations in the map. An elevation map generated from a gridded data set could also be compared to a scanned paper map of the same area. If comparing the relief to Google Earth (GE), it should be borne in mind that the GE representation of the 3D terrain is usually much smoother than the WAsP representation.
Figure 10. Land cover map for a site in Northern Portugal. The thin white lines show a land cover classification derived from the EU Corine 25-ha vector data set. A transformation table is needed for translating the land cover codes to roughness lengths.
Roughness lengths must be specified in meters and the roughness length of water surfaces must be set to 0.0 m in WAsP! This is because WAsP also uses this value as a flag value: ‘0 m’ indicates a water surface, whereas a small roughness length value means a smooth land surface (snow, sand, bare soil or the like).
The final roughness map should be checked systematically for errors, since these may give rise to erratic results in the WAsP calculations. Check the range of roughness length values in the main window of the Map Editor, and check for dead ends and cross points in the map display window (View > Nodes, Dead ends and Cross points). Some map editing may be needed to eliminate any dead ends and cross points.
When there are no more dead ends and cross points in the map, the consistency of the roughness values must be checked (View > Line Face Roughness errors) – there must be no line face roughness (LFR) errors! Finally, the roughness classification and values should be verified against a scanned paper map or by viewing the roughness change lines in Google Earth. The map may also be verified during a visit to the site.
All maps and images of the terrain are snapshots of the state of the terrain surface. The land cover and roughness length information used should of course correspond to the modelling scenario: use a historic (or present-day) map for modelling the meteorological mast(s) and use a present-day (or future) map for modelling the wind farm sites.
Roughness maps from gridded data
High-resolution gridded (raster) land cover data exist for many parts of the world. The current version of WAsP cannot employ such data directly, so it is necessary to construct the roughness change lines (vector map) from the raster data, e.g. using the Map Editor. There is currently no standard procedure for making vector roughness maps from raster land cover data; however, some techniques have been demonstrated using GIS systems or the WAsP Terrain Workshop. In addition, work is in progress to make the WAsP models use gridded elevation and land cover data directly.
4.3 Sheltering obstacles
Terrain features such as houses, walls, shelter belts, or a group of trees, that are quite close to the WAsP calculation site may be treated as sheltering obstacles and modelled using the shelter model of WAsP, see Figure 3. The following simple rule-of-thumb may be used to determine which model to use:
- If the point of interest (anemometer, turbine hub or other calculation point) is closer to an obstacle than about 50 obstacle heights ( H ) and its height lower than about 3 obstacle heights – then treat the feature as a sheltering obstacle and use the shelter model.
- If the point of interest is further away than 50 H and/or higher than 3 H , then treat the feature as a roughness element, i.e. adding to the roughness of the terrain.
5 Wind farm inputs
The wind farm inputs to WAsP consist of the layout of the wind farm (turbine site coordinates) and the characteristics of the wind turbine generator(s): hub height, rotor diameter and the site-specific power and thrust curves.
5.1 Wind farm layout
WAsP does not contain any advanced layout design tools, so the layout must be done free-hand or calculated in e.g. MS Excel. Turbine site coordinates may then be copied and pasted into WAsP. Free-hand layouts may be established quickly in the vector map by pressing the Ctrl-key and then dragging a turbine site (cloning) to a new turbine position. Distance circles around the turbine positions can be shown in the Spatial view as an aid in keeping a certain distance between the turbine positions.
769000 770000 771000 772000 773000 774000 Red Belt Easting [m]
718000
719000
720000
721000
Red Belt Northing [m]
2600
2700
2800
2900
3000
3100
3200
Row 1
Row 2
Row 3
Row 4
Figure 11. Sample wind farm layout in Zafarana, Egypt (Mortensen et al., 2005). The layout is mainly determined by the available land, the wind resource and aesthetics. Row 1-3 follow simple arcs, in which the positions were calculated using MS Excel; row 4 follow a terrain feature and the positions were originally drafted by hand.
If measurements of atmospheric pressure have not been carried out, the Air Density Calculator of WAsP can be used to estimate the air density from site elevation and the annual average air temperature at the site. A comparison of measured and WAsP-derived mean air densities for 10 sites in South Africa (Mortensen et al ., 2014) and eight sites in NE China (Mortensen et al ., 2010) is shown in Figure 13.
Figure 13. Measured and estimated mean air densities for 10 stations in South Africa (Mortensen et al., 2014) and eight stations in NE China (Mortensen et al., 2010).
If measurements of air temperature are not carried out, an annual average air temperature at the site must be estimated and input to the air density calculator.
The appropriate (nearest) performance table in the Wind turbine generator member of WAsP is then selected using the calculated or estimated site air density.
Wind atlas structure
The generalised wind climate (wind atlas data set) is specified for five standard heights above ground level and five land cover (roughness) classes. These standard conditions should span the characteristics of all calculation sites in the project; WAsP is then able to interpolate between these conditions. However, the standard settings may also be adapted to the project in question.
The default heights in the WAsP wind atlas are 10, 25, 50, 100 and 200 m a.g.l. If the wind turbine hub heights or anemometer heights are somewhere between these values, one or more of these heights may be adapted to the project characteristics; e.g. 10, 20, 40, 62 and 100 m, see Figure 14. A maximum of five heights can be specified.
The default roughness classes in the wind atlas correspond to roughness lengths ( z 0 ) of 0, 0.03, 0.10, 0.40 and 1.5 m. If large parts of the terrain has a roughness length somewhere between or outside of these values, e.g. like the low values of many desert surfaces, one or more of these values may be adapted to the project.
Figure 14. Sample wind atlas data set where the heights are adapted to site conditions. Data courtesy of the Wind Atlas for South Africa (WASA) project.
Atmospheric stability
WAsP contains a stability model which employs separate mean and RMS heat flux values for over-land and over-water conditions; where over-water conditions are defined as areas of the vector map where the roughness length z 0 = 0 m. The default wind profile over land corresponds to slightly stable conditions and the wind profile over water to near-neutral stable conditions; the exact heat flux values are given in the WAsP project configuration.
The default heat flux values were originally determined for the European Wind Atlas (Troen and Petersen, 1989) but have also worked well for many other parts of the world. Note, that the heat fluxes are not identical to what can be measured using e.g. a sonic anemometer, but have the same qualitative meaning.
The mean (and rarely RMS) heat flux value may be adjusted to site conditions – in order to tweak the wind profile slightly – but this should only be done following careful analysis and improvement of the elevation and land cover map. Furthermore, mast flow distortion might be evaluated and taken into account in the analysis. Since the WAsP heat flux values cannot be determined objectively (in the present version), they must be based on careful wind profile analysis, see Figure 18.
6.2 WAsP analysis
WAsP analysis is the transformation of the wind climate observed at a meteorological station to the generalised (also sometimes called regional ) wind climate. A typical WAsP analysis as it appears the WAsP hierarchy is shown in Figure 15.
The generalised wind climate may be dynamic , as indicated in Figure 15 by the small mast in the icon; or static, if the wind atlas data set is simply a previously calculated data file. The dynamic generalised wind climate may contain a map, which is then specific to the met. station site. The met. station can contain an obstacle group, which is then specific to the met. mast. The map and obstacle group should of course correspond to the conditions in the time period when the wind measurements were taken. Dynamic generalised wind climates are preferred over static because they reflect changes directly.
