Wind Mapping

 

Management of Meteorological Variables and Wind Mapping

INTRODUCTION

Once established the feasibility for the installation of wind farms in a determined country or region from the political, legal and economic point of view, the next step is to deeply study the geographic region about which the wind potential is intended to be analyzed.
In order to determine which region is the most adequate to start the studies, what must be known is the distribution of the atmospheric capes in the area, the direction or main directions of the winds and the local conditions of the area, as the obstacles (buildings, trees, etc.), the ground surface and the orography of the area. In the present conditions of the Argentine Republic, such study must be based on reports or previous analyses which could be available or which have to be done, on reports by the National Meteorological Service and satellite images, among others. It is important to say that nowadays there is not a unique wind map of the Argentine Republic, which would be the basic instrument for any kind of consult or preliminary study. The analysis of this generic information with which we count on at present only gives a generic data and it must be considered as such.
It is also advisable to count on the local assessment of a meteorologist, who can give a better interpretation as regards historic values of the region, wind tendencies, pressure, average temperature and humidity, as well as data concerning the meteorological phenomena which usually take place in the area (such as rain and/or snow storms, extreme temperatures, strong winds, etc.)
Once this preliminary study has been done a series of processes must be followed so as to know the area in detail and which, finally, allow to get a scientific and meteorologically sustained answer to the question How much wind can we expect in the determined area?
So as to answer the previous question, it must be taken into account that the wind conditions for an area are defined by the wind profiles of this area, the average wind speed and direction, the wind speed distribution and direction, and the wind daylight and seasonal patterns.

DEVELOPMENT

It must be considered that to determine the period in which the measurement in the area will be done, the lasting of that period depends on the kind of project that is intended to be carried out. If the intention is to develop a complete wind map of a region, it must be considered the measurement taking for at least 10 years (i.e., long-term); otherwise, if it is a preliminary analysis of the wind resource, it must be considered the realization of the measurement, at least, for a year in its initial phase (i.e., short and medium-term).
After this, the quantity and setting of the anemometers to be installed must be defined, taking into account that it is advisable the average surface monitored by each anemometer to be of 2500 km². Technically, it is recommended the use of head anemometers, calibrated every 6 months in a  certified wind tunnel. Once calibrated and installed the anemometers in the region, it is advisable that they work, for their optimal performance, during nearly 4 weeks before the measurement starts, without their data to be considered for the study which has to be done. These data, must only be taken into account so as to be able to study the correct equipment calibration and the acquisition process of data, as well as the correct functioning of the electronic equipments of meteorological measurement and data store.
Calibrating the anemometers every 6 months and doing a bimonthly follow-up of each of them in the field during the data collection period, will minimize the error introduction or the loss of data. The errors in this kind of wind study must be understood as a very complex factor and can lead to the complete failure of the whole wind project. Taking into account that the energy found in the wind is proportional to the cubic wind speed (E~V³) and that, according to Betz’s law, theoretically the 59.3% of the wind energy can be extracted, the measurement that are done on the “wind resource” must be very accurate and free of possible errors.
Considering that the wind which is further from the surface has more speed and less turbulence, due to the fact that it is not affected by obstacles from the land, it is advisable to carry out the measurement as high as the standards allow. In general, it is advisable to carry out in the same spot two measurements at different heights, at 10 and at 30 meters, though measurements have been done with meteorological masts placed even at 50 or 100 meters high.
Once the measurement network of meteorological data has been installed on the land, the analysis of the air turbulence must be carried out. This analysis must be done measurement the vertical movement of the air (through the use of ultrasound anemometers), as well as the air temperature.
It is important to take into account the density of the air in the region, due to the fact that the air density in warmer regions lowers, and in colder regions it increases. It must be considered that it is better for the production of wind energy regions in which the air density is as high as possible, i.e., regions with cold temperatures (example: Argentina Patagonic region).
Air Density in Normal Conditions of Pressure and Temperature: 1.225 Kg. /m³.

Normal Conditions of Pressure and Temperature: 1013 hP and 20˚C (NCPT)

Data Collection and Processing

Considering that the meteorological data obtained is the “raw material” of the project and having used anemometric stations properly calibrated, installed and verified, the processing phase must begin, having a clear idea of what information must be obtained when this phase ends.
Because of that, and before the data processing, it is important to define adequate policies and protocols which allow to manipulate the information and to process it with highest levels of security and efficiency:

• It is advisable to establish the frequency for the data collection in the electronic equipments of the anemometers in 1 Hz.
• The meteorological values obtained must be averaged every 10 minutes (some anemometers allow to do this average internally).
• The policies of getting, collecting and transmitting the data coming from the anemometers installed in the measurement field must be defined.

It will be essential to do height extrapolations, which allow to estimate the winds that finally will be used for the wind electric energy production[5]. To extrapolate the winds there are two different equations:

1. Hellmann’s equation:

This equation allows to extrapolate the winds at a second height (h₂) though it is usually used only as an approximation.

v = wind speed
h = height above the ground [m]
α = Hellmann’s Exponent (For example: in Germany α=0.16)

2. Logarithmic Profile equation:

This equation must be applied only on average values, not on individual values, and must be used with measurements that imply long-term periods.

v = wind speed [m/s]
h = height from the ground [m]
d = thickness of the moving cape [m]
Z₀ = ground surface

So as to be conservative, in the analysis and interpretation of any of these equations, an error margin of +/- 10% must be considered.

The data processing in the first phase implies to obtain :

1. Weibull’s Curve.

f = density of frequency
v = wind speed (center of class) [m/s]
A = scale parameter [m/s]
c = shape parameter (note: c is k)

There is a relationship between Weibull’s parameters and the mean wind speed:

Increasing parameter c of Weibull with the height (empiric) Weibull c₂ = c₁ + 0.008 (h₂ – h₁)

2. Study of wind drafts

3. Maximum and minimum wind speeds

4. Compass card of the region

EVALUATION OF THE WIND RESOURCE

What is a wind map:

It is a representation of the magnitude and the direction of the winds of a region in graphic form, using cartography with a scale and determined symbolism.

Kinds of data needed

The data which is needed to draw the wind map of a region are of varied source and, depending on the method applied to do the job, they will have to be of different kind, having each method their compulsory data entry well defined. In this way, at the moment of developing the wind map which data is available must be reveled and in what way they can be used to apply which method.
Nevertheless, and not considering the method to be applied, the data necessary for the mapping can be summarized in the following list:

• Anemometric measurement or surface measurement.
• Orographic data.
• Topographic data.
• Data of land use/natural coverage.
• Satellite images.

The data measured on surface is of vital importance, due to the fact that it can be used to obtain the wind map of a region as well as to validate the results obtained through other methods which do not use measurements as entrance. On the other side, the surface data is still the most accurate at the moment of doing the project.

PARTS OF A WIND MAP

Data

The most frequent data represented on the map for a determined height are the mean wind speed (measured in m/s), the mean wind direction (expressed in arrows or characteristic symbols of plotting in meteorology), the mean energy density (measured in W/m²), the frequencies distribution, the compasses cards, the Weibull (A y k) parameters, the studies on wind drafts and the studies of turbulence, among others. Besides, the results must present not only the average historic data, but also the seasonal regimens and the daylight and night cycles of the resource.
There is another data which is used as entrance for the wind map models, but which can become very useful to be used and represented as a summary of the outcome. These are the ground surface map, the land use and vegetal coverage map, and topographic maps.
All these data will be represented at different heights, being nowadays the most common 30 and 50 m; though there are also atlases which represent the information at 10, 25, 30, 50, 75, 80, 100, 125 and 200 m. Really, once the calculus have been done and knowing how the wind profile behaves for an area, the values can be easily extrapolated in height through methods as the ones mentioned above.
Wind classes definition: Whatever the data represented on the wind map is, the objective is always the same: to reveal the wind potential in an area.

One of the data which is usually represented is the quantity of energy than can be obtained from a region.
This is measured in W/m² and there is a table of equivalencies between the wind speed and power, which is used in the USA, called Wind Class [1].

Models

In order to build wind maps data and models are needed. The models will be all those processes (programs, algorithms, methods) which allow to draw the wind behavior and distribution in an area or given region.
Once determined which ones will be the models to use and collected all the necessary data to feed the model, both things combine to become a wind map. In some cases, the models can be combined between themselves to get a more accurate result.
Kinds of climatic scales and their models: The models can be of macro scale, known as synoptic scale (more than 2000 km); meso scale (2 to 2000 km) or micro scale (up to the 2 km).
The most commonly used for the wind resource evaluation are those of meso and micro scale, both of them can be used separated or in combination. In general, the most common experiences are those in which both models are used together.
Generally, the models used –independently of the scale- can be of numerical or statistic type. In the case of the numerical models, are based in a group of more or less complex equations which model the physics reality of the climatic phenomena.
On the other side there are statistic models, which are characterized for applying principles of statistics and probabilities to solve the problem of how winds behave. Some of these methods are based on principles of traditional statistics and others use modern techniques of artificial intelligence, for example.
General numerical models: the numerical models can be classified in three different categories according to the way in how they model the reality (accuracy with which their equations model the physics behavior of the winds).

• Solving the fundamental equation models.
• Simplified physics models.
• Statistics analysis models.

Solving the fundamental equation models

These are models which solve the general equation of the flux movement of Navier-Stokes [3]. They include the description of the topography, of effects of the surface ground, they allow to model complex thermal effects and use geographic information, through the GIS systems. These are called meso scale models.
They allow the atmospheric representation or simulation in greater detail, at the same time they allow the modeling of a wider area than the rest of the numerical methods. These consider all –or almost all- the important meteorological phenomena. On the other side, they do not depend on data measured on surface. Known examples: KAMM (Karlsruhe Atmospheric Mesoscale Model, from the homonym university in Germany), MM5 (Mesoscale Model version 5 of NCAR/Penn - National Center for Atmospheric Research/ Pennsylvania), ETA (generated model every 12 hours created by the NCEP - National Center for Environmental Prediction and used by the National Meteorological Service of the Argentine republic) and MatMeso, among others.
At the same time, this kind of model requires the use of other methods so as to achieve a greater resolution and surface measurement if it is wished to validate that the outcome of the method is correct in all the cases.

Classification of the  meso scale phenomena(Fujita, 1986)

• Alfa Mesoscale (a): they have a dimension of between 200 and 2000 km with phenomena which can last between 6 hours and 2 days, as small hurricanes and weak anticyclones.
• Beta Mesoscale (b), which counts with sizes of between 20 and 200 Km. lasting between 30 minutes and 6 hours; there can be fields of local winds, mountain winds, breezes from the continent and the sea, connective complexes of meso scale and big electric storms.
• Gamma Mesoscale (c) of an estimated size of between 2 and 20 km, lasting between 3 and 30 minutes, representing phenomena like most of the electric storms and big size tornados.

In order to achieve a collection of wind resource data using a meso scale method the following steps must be followed: first, wind data and measures must be collected in height. In general, measurements of radio sound are used, though the measurements on surface can be considered to calibrate the model and estimate errors. The model is executed to simulate the winds of 10 to 15 years and, depending on the power of the calculus available and the region to be modeled, the resulting grid ca be between 1 and 5 Km. It is also possible to obtain a greater accuracy if a micro scale model is executed or one which allows a greater resolution within each point of the grid, for example the WAsP or WindMAP. After the execution of the model, the map of the wind resource is traced. In this map the data mentioned above can be represented.

Simplified physics models

They use a more reduced group of equations and –due to this- they model a smaller quantity of climatic phenomena. They are used to trace wind maps in low or medium complexity surfaces, getting maps equally useful and accurate, but requiring a lot lower potential of calculus. The advantages of this kind of methods are that they function with anemometric seasonal data of surface, with no need of height data and, besides, they are ideal for low complexity surfaces.
As a counterpart, their disadvantages are that they do not model the reality completely, they can only represent some aspects of the wind behavior and other meteorological variables. Then, they are not capable of modeling complex meteorological phenomena, but very important ones, as the breeze from the sea or the continent, or the wind produced by thermal effects, like the mountain winds; they do not take into account the splitting of the air flux produced by the irregular surface. It depends on anemometric measurements on surface, what implies that if the measurements are not enough or they are done in a wrong way, the model will generate an incorrect result. The anemometric non reliable measurements can not be used without using correction techniques which can introduce new errors in the calculus.

Models based on GIS

These kind of models are based on completely different functioning principles. For their functioning they use wind measurements in height which are extrapolated to low altitude. Moreover, they are based on the GIS (Geographical Information System) technology for the collection of data and the drawing of the part of the region to be analyzed.
In 1995 the NREL - National Renewable Energy Laboratory started to develop a new method of wind mapping based on the GIS technology. The model is called WRAM (Wind Resource Assessment Model). It produces maps of great quality and was used to develop the wind maps of several union states (North Dakota, South Dakota and Vermont; part of Minnesota, Iowa and Nebraska); apart from several international atlases like Dominican Republic, Mongolia,  Philippines and regions of Chile, China, Indonesia y Mexico. This method needs of wind values previously calculated and, really, it is no other thing than a method of representation, more than of calculus.

Models from the point of view of the principles

From the point of view of the physic-mathematics principles, the numericalal models are classified in:

• Based on the Jackson-Hunt’s theory
• Based on the uniform mass model

In the first case, these models tend to satisfy the Navier-Stokes’ equations [3]. Their basic characteristic is the description of two fundamental principles: the mass conservation and the moment conservation. Due to this, this kind of model is very sophisticated and has a very good output: an error among the 8 and the 10%.
In the case of the models based on the mass uniform model, they only describe (different from the previous ones) the mass conservation. They are less sophisticated and has a similar output –under determined conditions- to the most complex models. Examples of this kind of model are the WindMAP and the WAsP.
It can be deduced from the description of both models that the mass conservation principle is the most important determinant of the wind variation, always referring to surfaces of low or moderated complexity.

Graphics of the combined models

The graphic shows the steps to develop a wind map using a meso scale method and one of micro scale together, as the Wind Atlas style.

EXAMPLES OF THE MOST KNOWN MODELS

• KAMM (Karlsruhe Atmospheric Mesoscale Model) [meso scale] [4]
• Wind Atlas Analysis and Application Program (WAsP) [Simplified] [4]
• MesoMAP [meso scale] [7]
• WRAM Method [GIS] [8]
• WindMAP [Simplified] [7]
• WindSCAPE [Mix]: Raptor [micro] + TAPM [meso]

WIND MAP EXAMPLE: EUROPE

WIND MAP EXAMPLE: EUROPE OFFSHORE

WIND MAP EXAMPLE: DENMARK

CONCLUSIONS

One must be conservative in the interpretation not only of the data obtained as a consequence of the measurements done in the field by the measurement equipments but also with the extrapolations which are done, in height as well as on the surface. It is convenient to estimate between a 10% and a 20% less in the obtained data, and with those values to do the calculus and estimations.
Finally, the report of the “wind potential” of the region must present a technical and meteorologically sustained detail of the following information:

• The preliminary analysis of the region. (In this case it is advisable to count with the wind map of the country)
• Equipment installation process.
• Wind Mapping.
• Results and final report.

It does not matter what kind of wind project will be started, the evaluation phase of the potential of a region is one of the most important ones. According to its result the feasibility or not of a future project will be determined; also which is the best place within a region to establish a new wind complex.
The wind maps (atlas, resource evaluations, or whatever name they are assigned) are fundamental instruments to start any work of planning the installation of a wind farm. But all of them depend, at the same time, on fundamental incomes which will allow their creation: the meteorological data, of whatever kind they are.

REFERENCES

1. [1]. Wind Energy Danish Assoc. WindPower.org. Wind class standard definitions “Wind Class”. 11/Feb/2004,  <http://www.windpower.org/es/stat/unitsw.htm>
2. [2]. Brower, M., B. Bailey, and J. Zack. The New US Wind Resource Atlas [cdrom]. In: European Wind Energy Conference & Exhibition 2003. [Madrid], European Wind Energy Association, 2003.
3. [3]. Cambridge University Press. Foundations of Fluid Mechanics. Navier-Stokes Equations [on line]. 14/Aug/2004. <http://www.navier-stokes.net/>
4. [4]. Frank, H. P., O. Rathmann, N. G. Mortensen, and L. Landberg. The Numericalal Wind Atlas: The KAMM/WasP Method. [Roskilde, Denmark]: Information Service Department, RisØ National Laboratory, June 2001.
5. [5]. Gasch, R., and J. Twele. Wind Power Plants. Fundamentals, Design, Construction and Operation. [Berlin, Germany]: Solarpraxis AG, 2002.
6. [6]. Manwell, J. F., J. G. McGowen, and A. L. Rogers. Wind Energy Explained: Theory, Design and Application. [West Sussex, England]: John Wiley & Sons Ltd, 2002.
7. [7]. Brazil Mining and Energy Ministry. Mapas do Potencial Eólico Anual [cdrom]. In: Atlas Do Potencial Eólico Brasileiro. [Brasilia, Federative Republic of Brazil], e-dea Technologies/ Christianne Steil, 2001.
8. [8]. Nielsen, J., S. Innis, and K. Pollock. Renewable Energy Atlas of the West. <http://www.EnergyAtlas.org>
9. [9]. RisØ National Laboratory. Wind Energy Department.    Wind Resource Atlas for Denmark. [Denmark]: 23/Jan/2004. <http://www.risoe.dk>

Eng. Luis Mariano Faiella
Eng. Alejandro J. Gesino
Research & Development Area
Argentine Wind Energy Association
www.argentinaeolica.org.ar

Croatian Center of Renewable Energy Sources (CCRES)