water depth, Extreme values, numerical simulations, numerical study, return period
EVALUATION OF THE STAND-ALONE WIND AND WAVE measurement system
S FOR THE HORNS REV 150 MW OFFSHORE wind farm
IN DENMARK Sшren Neckelmann, M.Sc., Mech. Eng., and Jan Petersen, Ph. D, M.Sc., Civil. Eng. ELSAMPROJEKT A/S, Kraftvжrksvej 53, DK-7000 Fredericia. E-mail: [email protected]
and [email protected]
, Phone: +45 7923 3333, Fax: +45 75 56 44 77 DENMARK ABSTRACT The main purpose of the work discussed in this paper was to investigate the wind and sea climates for the first of the two offshore wind farms in the Elsam area, i.e. the 150 MW Horns Rev wind farm to be built and commissioned by the end of year 2002. The Horns Rev site is located at a reef approx.14 km off Jutland in the North Sea
in a very harsh environment. The water depth
s at the site vary between 6 and 12 m. In-depth knowledge of the environment at the offshore site was crucial to enable the evaluation of vital information such as wind farm ENERGY PRODUCTION
prognoses, assessment of structural loads and the prediction of global long-term seabed changes and local scouring around the foundations. Meteorological and marine measuring systems
facilitating a very high data recovery rate are described, and the Initial Results
of more than 6400 hours of measuring data recorded since May 1999 are evaluated. Practical problems encountered during the monitoring period at the offshore location are presented. KEYWORDS Wind Energy, Offshore, Meteorological and Marine Measurements, Assessment of Climatic Conditions. INTRODUCTION Since the start of the modern development of wind turbines for electricity production, offshore siting of wind farms has held the promise of vastly increasing the viable wind resources without the obvious disadvantages of large-scale onshore wind farms with regard to visual impact and noise. The energy plan of the Danish Government "Energy 21" calls for an installed wind power capacity of 5500 MW by 2030 4000 MW offshore and the rest onshore. As the density of wind turbines in the onshore areas has almost reached the limit and the 1500 MW onshore goal was reached in 1999, the utilisation of the offshore resources is the only way to fulfil the goals set up in "Energy 21". A feasibility study
finished in 1995 has identified sites for up to 8000 MW of offshore wind farms in the Danish coastal waters and concluded that the wind energy technology
and wind turbine size have matured to a point where large-scale offshore demonstration wind farms are technically and economically feasible. On this background, the utility groups Elsam/Eltra and Elkraft, serving the western and eastern part of Denmark respectively, were ordered by the Danish government in early 1998 to construct five offshore wind farms in the period from 2002 to 2008 - each with a rated power of 150 MW - and to operate the farms for a 20-year period. The main purpose of the present work has been to establish a monitoring program
to investigate the wind and sea climates for the first of the two offshore wind farms in the Elsam area, i.e. the 150 MW Horns Rev wind farm to be built and commissioned by the end of year 2002. Based on measurement data and numerical studies of the wind and sea climate the further aim has been to evaluate vital information such as prognoses for wind farm energy production, structural loads (wind, waves, currents and ice), predicted global long-term seabed changes and local scouring around the foundations. Furthermore, the accessibility of the location - limited by its exposure to large waves from westerly directions is to be deduced.
The wind farm at Horns Rev will be erected approx. 14 km west of the Jutland coast in the harsh environment of the North Sea. The water depths at the erection site for the wind farm vary between 6 and 12 m. The site is characterised by a large undisturbed over water fetch in the western sectors towards the North Sea. The geographical layout of the Horns Rev wind farm and the measurement systems are shown in Figure 1.
Figure 1. Geographical layout of the wind farm and the measurement systems at Horns Rev.
DESCRIPTION OF THE MEASUREMENT SYSTEMS Making offshore measurements is a challenging task, and considerable resources have to be invested to obtain a high data coverage for a prolonged period. Facing this fact well proven robust and above all simplistic systems for the Power supply
and measurement systems have been carefully selected and taylored for years of stand-alone operation.
A square lattice mast was erected on a mono pile close to the wind farm site at Horns Rev and mid May 1999 the meteorological measurement programme was initiated and later in June 1999 the marine measurement programme started.
Today the site at Horns Rev is represented with more than 6400 hours of measuring data comprising more than 3.5 million simultaneously sampled signals from three different systems.
The meteorological measurements include four-level wind speed
s as well as wind directions, temperatures, atmospheric pressure, solar radiation
and lightning detection. The four measurement levels are 62, 45, 30 and 15 m above DNN (Danish Normal Zero), see Figure 2.
A redundant and completely separate Data Logger
system comprising wind speed, wind direction and atmospheric pressure is placed in the top section of the mast at level 58 m ensuring maximum data coverage in failure situations as well as being used for validation of the main system measurements.
Figure 2. The meteorological mast at Horns Rev is a square cross section
lattice mast raised on a Ш1700 mm monopile. All heights refer to DNN.
To keep the power consumption at an absolute minimum
, two standard data loggers were selected to scan the sensors. The sampling rate is 1Hz, and before storing in a flash RAM unit the data are condensed to 10-min average values together with maximum, minimum and standard deviation
s for the averaging period. For the wind speed sensors a 5-s gust is calculated as a current mean of five consecutive 1 Hz values. The storage capacity of the flash RAM unit is equivalent to 9 months of data. The main components of the power supply have been designed as simply as possible comprising a matrix of 2 x 5 50 W solar cells charging 25 NI-Cd batteries with a capacity of 1070 Ah. The power system
is capable of delivering 16 W continuously with a safety factor of 2.5 all year around. The marine measurement system includes two wave buoys (wave riders) continuously measuring wave heights and periods, and an acoustic doppler current profiler measuring currents, water level
, temperature and salinity. The positions of these devices are shown on Figure 1. All recorded data from the meteorological system and the marine current/WL system are transferred onshore by a GSM communication system
and subjected to ongoing Quality Control
before being stored in an SQL database. The wave rider measurements are continuously transmitted to a land-based receiver (27 MHz band). Presentation of the recorded and post-processed data are transferred from the database to a WEB-based geographic Information System
(GIS-system) facilitating fast and simultaneous flow to internal and external project partners.
DATA COVERAGE AND PRACTICAL PROBLEMS The data coverage rate for the measurement systems is shown Table 1 below.
Table 1. Data coverage rate for the meteorological and the marine measurement systems. Data from Horns Rev May 1999 to January 2000.
Measuring System Met. Main System Met. Redundant System Wave Rider, Pos. South Wave Rider, Pos. North Current/WL
No. of Hours 6160 6436 5018 3984 4450
Data Coverage 95.6% 99.9% 97.5% 77.4% 88%
Despite the poor accessibility to the measuring site the practical problems encountered during the first nine months of operation of the meteorological systems have been few and mainly restricted to damage due to lightning and Mechanical damage
caused by an extreme hurricane over Denmark on December 3, 1999.
The top anemometer at level 62 m was struck by lightning in October 1999, and replacement of the sensor was necessary. Disassembly of the anemometer revealed severe damage to the rotor cup and to the bearing system caused by the induced overheating. The mast was equipped with lightning detectors, but unfortunately no lightning was detected during the measurement period. A technical investigation of the damaged anemometer revealed that the lightning induced current was beneath the detection range (~ 1000 A) of the sensors which explains why no lightning was recorded for October 1999.
During the storm on December 3, 1999, parts of the electrical system were pushed out of their sockets as large waves impinged on the walls of the measurement shed located at level 6.2. This caused a disconnection of the power supply to the primary measuring system and hence a total system shut-down. Fortunately, the redundant measuring system continued to record data ensuring full data coverage from the top section of the mast during the following 10 days of "power out".
The problems encountered during operation of the marine system have had a variety of different sources. The low coverage of the southern wave rider is not yet fully understood, but initial investigations point towards a fabrication error, making the device sensitive to low temperatures. Also problems have been experienced with radio amateurs using the transmitting frequency of the southern wave rider.
The primary reason for the 12% non-coverage of the current/WL measuring device can be contributed to a BIOS series fault in the PC used for data acquisition
. To transfer the recorded data on shore, GSM communication has proven a reliable technique at least for the Horns Rev site. The only periodic problems have been encountered mainly due to external failures in the GSM net-system. Accessibility to an offshore installation is always a critical issue, and entering the mast from a small boat has proven a difficult task in the North Sea. As a consequence several service operations have been obstructed by sea conditions with wave heights in excess of Hs=1.3 m which is the critical limit for accessing the mast.
PRESENTATION OF MEASURED WIND DATA To give an overview of the wind records compiled from May 1999 to January 2000 at Horns Rev the main character
istics of the data are evaluated in the following section. When evaluating the outcome of the analysis the following facts should be kept in mind:
1. The data record is not representative of an annual average since the measuring period comprise less than nine months of data. 2. In general, 1999 was an extraordinarily poor wind year. 3. The wind speed data has not been corrected due to over-speeding caused by boom distortion. The rotors of the cup anemometers have been raised more than 55 cm above the mounting booms and the expected over-speeding amounts to less than 1% according to . 4. To reduce the considerable effect of flow distortion (mentioned in ) from the lattice mast wind speed sensors placed at opposite sides of the mast are used when evaluating the statistical values for the lower levels of the mast (45, 30 and 15 m). For these levels wind speed data are selected from the upwind side of the mast. 5. It should be noted that the actual measuring levels are influenced by the changing sea levels due to the tide and sea swells at the Horns Rev site. In the analysis below there have been no attempts at neutralising the influence from the changing sea levels.
Statistical Values In Table 2 are given the main statistical values for the measuring period.
Table 2. Main statistical values for the measuring period May 1999 to January 2000. Data from Horns Rev May 1999 to January 2000.
Height Mean Wind Speed Max. Wind Speed Mean Wind Direction Mean Turbulence Intensity Weibull Scale Parameter Weibull Shape Parameter Wind Shear (to nearest)
[m/s] [m/s] [є] [-] [m/s] [-]
62 9.7 45.4 194.9 0.088 10.9 2.2 0.16
45 9.2 43.1 194.7 0.097 10.4 2.2 0.10
30 8.8 40.7 195.3 0.101 10.0 2.2 0.10
15 8.2 39.5 0.119 9.3 2.2 -
Wind Speed Distribution The wind speed distribution at 62 m shown in Figure 3 is evaluated as the relative number of observations by 1 m/s wind speed bins for all direction sectors at 62 m height whereas the directional distribution is found as the relative number of observations by direction sectors. As expected the wind mainly comes from the south western sector, and when looking at all sectors the distribution can be fitted well by a Weibull distribution. For the separate direction sectors the conformity to Weibull distributions are somewhat less convincing.
Probability [ - ] 1 5 9 13 17 21 25 29 33
0.10 0.05 0.00 Wind Speed [m/s]
62 m - All Sectors
NNW WNW W WSW SSW
N 0.1 5 0.1 0.05 0 1 S
NNE ENE E ESE SSE
Figure 3. Observed wind speed distribution for all sectors and the directional distribution at 62 m. Data from Horns Rev May 1999 to January 2000.
Mean Wind Speed by Direction Wind speeds averaged by direction at the four measurement levels are shown in Figure 4.
Wind Speed [m/s]
12 10 8 6 4 2 0 N NNE ENE ENE ESE SSE S SSW WSW WSW WNW NNW Direction Sector
62 m 45 m 30 m 15 m Figure 4. Observed wind speeds averaged by directions at four heights. Data from Horns Rev May 1999 to January 2000.
As was expected the highest wind speeds are detected in the western sectors dominated by the long fetches of the North Sea. The shorter offshore distance of 18 km to the western coast and the land masses of Jutland dominates the eastern winds leading to the relatively low wind speeds from these sectors.
Wind Speed Profile Wind speed profiles offshore are expected to conform to logarithmic profiles, but when looking at the observed average wind speeds at the four levels in Figure 5, the figures reveals that wind speeds at 62 m are somewhat higher than could be expected from a logarithmic prediction.
Height [m] (Logaritmic scale)
Wind Speed [m/s]
Figure 5. Semi-logarithmic plot of observed wind speeds by height. Wind speeds are averaged for all sectors. Data from Horns Rev May 1999 to January 2000.
 discusses different factors which will lead to distortion of the profile away from logarithmic predictions, and it is worth noting that  has found the same tendency for other offshore sites in Denmark. At the moment no further analyses have been initiated to clarify this subject in the present work Turbulence Intensity Figure 6 shows the relationship between wind speed and turbulence intensity at the four measurement levels.
Turbulenc e Intens ity [ - ]
0.40 0.30 0.20 0.10 0.00 0
Wind Speed [m/s ]
62 m 45 m Figure 6. Variation of 30 m turbulence intensity with 15 m wind speed. Turbulence intensity is the weighed average for all direction sectors at each level. Data from Horns Rev May 1999 to January 2000.
The turbulence intensity is averaged by 1 m/s wind speed bins for all direction sectors and calculated as the ratio between the standard deviation of the wind speed and the wind speed. The turbulence intensity decreases with increasing wind speed to about 12 m/s and then begins to increase slightly at higher wind speeds. The increase in turbulence intensity is more pronounced at level 15 m, probably due to the response to increasing wave height. It should be noted that the depicted scatter in turbulence intensity at 15 m for wind speeds above 25 m/s are derived from a very low number of observations in these wind speed bins. Figure 7 shows the variation of mean turbulence intensity with wind direction at the four measurement levels.
Turbulenc e Intens ity
0.16 0.14 0.12 0.10 0.08 0.06 0.04 0.02 0.00
Direc tion Sec tor
62 m Figure 7. Variation of 45 m turbulence intensity with direc30 m tion sector. The turbulence 15 m intensity is the weighed avera- ge for all wind speed bins at each level. Data from Horns Rev May 1999 to January 2000.
According to  the observed higher turbulence intensities in the eastern sectors are probably caused by the land masses of Jutland since the fetch is below 20 km. Further analysis is, however, necessary to uncover these relationships. The observed vertical profile of the turbulence intensity is shown in Figure 8 as the average turbulence intensity derived at the four measurement levels. The observed increase in turbulence level with declining height is as was expected, however, there is a
clear tendency towards higher turbulence intensities at 15 m than might be expected assuming a linear relationship as observed in . In agreement with  the profile indicate that the turbulence intensity continues to decrease with increasing height.
70 62 60
50 45 40
20 15 10
0.090 0.100 0.110 Turbulenc e Intens ity [ - ]
Figure 8. Variation of turbulence intensity with height. The turbulence intensity is the weighed average for all wind speed bins and direction sectors at each level. Data from Horns Rev May 1999 to January 2000.
PRESENTATION OF MEASURED MARINE DATA Table 3 gives an overview of the most important parameters measured stating the mean and maximum values
. The WL is referred to the DNN (Danish Normal Zero) system.
Table 3. Mean and max. values of the most interesting recorded marine parameters.
Wave Rider South (note that Hmax has only been measured since Nov. 1, 1999) Wave Rider North (note that Hmax has only been measured since Nov. 1, 1999) Current (middle of water column) Water Level (rel. to DNN)
Unit Hs Hmax T02 Tp Hs Hmax T02 Tp Abs. current North Component East Component
Mean 1.15 m 2.46m 3.97 s 6.24 s 1.29 m 2.55 m 4.32 s 6.74 s 0.35 m/s 0.49 m
Max. 4.23 m 6.54 m 7.78 s 16.7 s 4.48 m 7.40 m 8.17 s 16.7 s 0.91 m/s -0.85 m/s / 0.83 m/s -0.42 m/s / 0.75 m/s -1.31 m / 2.80 m
Waves Since waves are generated by the wind shear and wind pressure on the sea surface, a clear dependency on these two phenomena can be expected. Figure 9 illustrates this relationship between the wind speed (1 hour average) in level 62 and the significant wave height Hm0. Beside the wind speed it is evident from this figure that also the fetch plays an important role for the wave generation, waves being smaller from eastern directions where the fetch is limited (15-40 km). The pronounced scatter in the figure can mainly be contributed to the fact that the waves do not react instantaneously to fast changes in wind speed.
4.50 4.00 3.50 3.00 2.50 2.00 1.50 1.00 0.50 0.00 0.00
North North-East East South-East South South-West West North-West
Wind speed (m/s)
Figure 9. Measured relationships at Horns Rev (wave rider south) between wind speed (level 62) and significant wave heights for 8 directions. Data from Horns Rev May 1999 to January 2000.
For wind speeds up to about 20 m/s there is an almost linear relationship between wind speed and wave height. For higher winds the waves seems to reach an upper limit of about 4 m. This limit is governed by the water depth which, when sufficiently shallow, forces the waves to break. Hence the extreme wave conditions are not limited by the wind but rather by the water level. The relation between the significant wave height Hm0 and the maximum wave height Hmax is illustrated in Figure 10 by the ratio Hmax/Hm0. Despite the inherent scatter, a weak trend towards a falling Hmax/Hm0-ratio can be observed for higher waves. This can again be explained by breaking of the largest waves.
Figure 10. Ratio between Hmax and
Hm0 (wave rider south data). Data
from Horns Rev May 1999 to January 2000.
Figure 11 shows the relation between wave height (Hm0) and wave period (T02). The measured values can roughly be regarded as grouped in two clusters an upper cluster where waves are generated by the wind and a lower cluster (Hm0 0.5 m) where the waves originate from swells. The wind generated part of the waves shows that there is a well-defined relationship between wave height and wave period.
Hm0 (m) Exceedance Probability
Figure 11. Relationship between
Hm0 and T02 (wave rider south data).
Data from Horns Rev May 1999 to
9.0 January 2000.
Finally, the exceedence probabilities of the significant wave heights measured at pos. south and north are plotted in Figure 12. As expected the waves are higher on the northern side of the reef, since the southern position (the wind farm) is somewhat protected from NW-waves by the reef itself.
A key point in selecting the right operation and
maintenance (O&M) strategy for an offshore wind
farm is the accessibility of the wind turbines. The
most natural choice of transport for offshore
operations is by boat. However with a limiting wave
height of Hs=1.2-1.3 m for boat operations it can
be seen on Figure 12 that the accessibility is only
Pos. North Pos. South
Figure 12. Probability of exceeding for measured significant wave heights at positions south and north. Data from Horns Rev May 1999 to January 2000.
0.001 0.0001 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 Hm0 (m)
Currents The Horns Rev site is dominated by a strong N-S going tidal current with maximum velocities of about 1 m/s. In extreme storm conditions with winds from the west, the wind induced current in the eastward direction has been measured to nearly 0.8 m/s (the December 3rd hurricane). Figure 13 illustrates the measured current magnitudes and their flow directions. W
N E 0 0.2 0.4 0.6 0.8 1 m/s
Figure 13. Measured current magnitudes and flow directions at Horns Rev (1999-07-01 2000-01-31). Data from Horns Rev May 1999 to January 2000. S
Numerical Simulations Models Prior to commencing the marine measurement programme, a numerical study was performed. The aim of this study was primarily to establish estimates of the long term extremes of waves, currents and water level. However an important part of the study was also to estimate the accessibility of the turbines for different limiting sea conditions (values of Hs) and `wave window lengths'. The input for this study was a 15-year continuous meteorological data series covering the North Sea. These meteorological data were used as input to the MIKE OSW model (Danish Hydraulic Institute's offshore wave model) to simulate a 15-year consecutive wave train in the proximity of the wind farm. Since the water depth is rather limited in and around the wind farm, the OSW waves were used as input to a shallow-water wave model which could then simulate the waves inside the wind farm site. Table 4 lists the extreme value analysis results
for 10-, 50- and 100-year return periods.
Table 4. Extreme values from numerical study.
Parameter Hs Hmax Currents Water level
10-year return period 5.0 m 8.1 m 0.79 m/s -1.7 m / +3.0 m
50-year return period 5.3 m 8.1 m 0.88 m/s -2.4 m / +3.5 m
100-year return period 5.4 m 8.1 m 0.92 m/s -2.7 m / +3.7 m
Comparing Table 4 with the measured values (Table 3) over the 8-month measuring period, the only divergence to be observed is that the measured currents exceed the long term predictions. However, this is fully expected as the current meter is located close to the met. mast on a water depth somewhat smaller than the average water depth in the wind farm site, which was used in the numerical simulations.
The analysis of `wave windows' was made on a monthly basis. In Figure 14 the accessibility, assuming that Hs permanently has to be below 1.3 m in time periods of 12, 24 and 72 hours, is plotted for each month. The accessibility in the winter season can be as low as 20% whereas the summer offers an accessibility of 70-80 %.
Accessibility for Hs<1.3 m
0% 1 2 3 4 5 6 7 8 9 10 11 12 month
Figure 14. Accessibility at Horns Rev for Hs<1.3 m. Data from Horns Rev May 1999 to January 2000.
HIGH-LIGHTS FROM THE HURRICANE "ADAM" On December 3, 1999 Denmark was hit by a hurricane exhibiting extremely fierce gusts measurements from several places recorded speeds as high as 40-50 m/s. With wind speeds recorded of this magnitude the hurricane was the fiercest in Denmark for more than a 100 years. The Horns Rev site was hit by the hurricane during the afternoon with wind speeds (10-min. avg.) increasing from 15 m/s at noon to more than 45 m/s at 18.00 hours. The measured maximum 1 Hz value of 58.5 m/s and a maximum 5-s running mean value of 53.4 m/s indicate very high gust values.
Figure 15 shows the wind speeds and air pressure during the culmination of the hurricane.
Wind Speed [m/s] Air Pressure [hPa]
50 980 40 970 30 960 20
10:00 12:00 14:00 Time
10 min. Avg. 5 sec. Avg. Air Press.
Figure 15. Measured wind speed and air pressure during the passage of the hurricane. Data from Horns Rev, December 3, 1999.
The zenith of the hurricane lasted for about 2.5 hours with two peaks divided by a distinct drop in wind speed in between.
SUMMARY A 62 m meteorological mast and two marine systems have been established at Horns Rev approx.18 km off the Jutland North Sea coast. More than 6400 hours of simultaneous meteorological and marine data have been recorded since May 1999. Simple and reliable stand-alone techniques have demonstrated their suitability for power supply. Data logging and GSM transmission have been used enabling a very high data recovery rate. Data records from an 8-month measuring period have been evaluated and the first results have been presented. The practical problems we have encountered have been presented along with their solutions. Measurement data from an extremely powerful hurricane striking Denmark on December 3, 1999 have been presented.
REFERENCES  Pedersen, B.M., K.S. Hansen , S. Шye, M. Brinch, O. Fabian, 1992: Some experimental investigations of the influence of the mounting arrangements on the accuracy of cup-anemometer measurements. Journ. Of Wind. Eng. And Ind. Aerod., vol 39, pp. 373-383  J. Hшjstrup, 1999: Vertical Extrapolation of Offshore Wind Profiles. European Wind Conference, 1- 5 March 1999, Nice, France, pp. 1220-1223.  R.J. Barthelmie, M.S. Courtney, B. Lange, M. Nielsen, A.M. Sempreviva. J. Svenson. F. Olsen, T. Christensen, 1999: Offshore Wind Resources at Danish Measurement Sites. European Wind Conference, 1- 5 March 1999, Nice, France, pp. 1101-1104.
S Neckelmann, J Petersen