Worldwide close to the wind

Our experts are operating worldwide. Our selection of special projects provides an overview about the most relevant aspects. The filtering function enables an easy access to single projects.


(Deutsch) Eignung der Gondelanemometerdaten von Windkraftanlagen für die Ermittlung der energetischen Verfügbarkeit – Bachelorarbeit

Sorry, this entry is only available in German.

Development of a measurement strategy and supervision of wind measurements

Site: Medium size project in Tunisia, mountainous region

Scope: A local company was in charge of the installation and operation of met masts. The equipment had already been procured. aj was to

  • advise on number of masts
  • advise on measurement positions
  • check and document the masts independently
  • check the measured data continuously
  • assess the site conditions with respect to turbine specification
  • prepare wind resource and energy production assessments


  • Although the site itself was not very complex, the complex surroundings led to varying and hardly predictable wind flow conditions across the site. Namely, flow detachment was detected on one of the mast positions in an important wind direction.
  • The placement of the masts was first defined for an initial stage. Care had to be taken to avoid misleading results due to local features.
  • As some data of the first and second mast had come in, the remaining open questions regarding the wind flow and the most relevant contributions to the uncertainty could be addressed. One of the masts was relocated to cover these best possible.
  • The choice of the mast location not only addressed the needs of the wind resource assessment, but also of turbine specification.
  • It was found that some booms had not been mounted as predefined. The wind direction offsets were not correct. The mounting of the anemometers was such that the mast structure influenced the measured data significantly. Together with the installation team, changes of the boom orientations and lengths were agreed. An additional anemometer was mounted to help quantifying all impacts on the measurement and assessing the wind shear with confidence. Finally, sufficient quality and purposefulness of the measured data could be achieved.
  • It was found that cable ties had been used that were not UV resistant. They were replaced before any measurement cables could have broken.
  • A measurement documentation was prepared by aj that was able to satisfy the requirements of investors and turbine suppliers.
  • A wind resource assessment with good confidence and state of the art uncertainty could be prepared by aj based on the data.



Assessment of the uncertainty of methods for the extrapolation of wind data to the long term

Sites: Several sites in different European regions with relatively long measured wind data sets


Calculation of the average wind speed of the entire measurement period based on the data of individual 12-months periods


  • The standard error of the long term mean wind speed determined with the MCP method and the MDP method developed by aj based on time series of 12 months is about 0.1 m/s.
  • With linear regressions over monthly means, the uncertainty is higher. Strongly deviating results occurred in some years with this method.
  • Based on hourly means, the long term mean wind speeds determined with the MCP method and the MDP method are usually very similar.
  • The MCP method allows for calculating a long term time series, whilst the MDP method yields a more accurate long termn frequency distribution and wind rose.
  • Gaps in the reference data set can have an important impact on the result.
Wind park optimisation of 40 planned wind turbines

Site: Bosnia-Herzegovina


  • due to a pronounced main wind direction the planning of the wind turbines was modified to one single row instead of two with short distances between the turbines
  • better placing on a mountain ridge possible


Development of an optimum wind farm layout using four met masts

Site: Western Canada


  • The influence of the nearby sea is strongly overestimated by the model, this could be solved through a modified modelling of the roughness lengths.
  • Poor quality of the met mast data, inconsistent anemometer data
  • Due to the measurement heights being close together, no reliable vertical extrapolation is possible.
  • A first layout of the wind farm was developed. Due to a pronounced main wind direction, the wind turbines ware placed in several rows, with an adequate distance between these. However, the distance between neighbouring turbines could be strongly reduced.
Consultancy on wind resource over the time of development of a wind farm project

Site: 300 MW project in Spain


  • initial estimate of wind resource
  • advice on measurement positions
  • optimisation of wind farm lay-out
  • input to turbine specifications
  • regular check on measured data
  • independent documentation of wind measurements
  • continuous further development of measurement strategy
  • preparation of wind resource and energy production assessments
  • selection of locations for masts for the power curve verifications


  • Although the wind farm area has a relatively simple structure, the wind resource was found to vary across it in a way that is not predicted by wind flow models and that does not correspond to the impression of the site. The reason is in the large size of the area, in which the regional (mesoscale) variation of the wind field becomes relevant.
  • Therefore, the measurement strategy consisted of a tall mast in a central position and a Lidar instrument which was first verified against that mast, then moved consecutively to a number of positions in the area.
  • The measurement positions were defined step by step depending on the outcome of the data captured in the meanwhile.
  • The wind resource was determined by combining the outcome of mesoscale wind flow calculations with the results of the wind measurements.
  • Due to the strong directionality of the wind field, the wind farm was arranged in rows with relatively narrow spacing within the rows but large distances between rows. With this, the wake losses were kept very low for all turbines. At the same time, the cost of infrastructure was minimised.
  • Icing was found to be an issue for the site, despite of its Mediterranean character.
  • Wind resource and energy production assessments could finally be prepared with good confidence.
  • The final results were not far from the initial estimate.
  • The largest uncertainties were found with the power performance of the wind turbines and the extrapolation of the measured data to the long term. The best way to reduce the uncertainty of the entire project was to prolong the mast measurement by one year.


Re-assessment of wake losses

Site: small wind farm in Germany; the client assumes distinctly higher wake losses than calculated in the original assessment report

Procedure: profound analysis of SCADA data, adapted model calculations


  • the met station data used in the wind assessment report turned out to be less appropiate for the site, which is influenced by very local effects, thus the wind rose of the met station is less representative
  • the frequency distribution of the wind speed of the met station data is also not representative for the site
  • thrust coefficients of the power curve of the wind turbines are too low as has often been observed by aj for this specific turbines manufacturer
  • very low turbulence at the site, i.e. the calculated wake losses are too low

Final result: wake losses more than doubled compared to the original calculations

Check of the performance and in particular the yaw behaviour of wind turbines

Site: Small wind farm in the state of Bavaria, Germany


  • Wind measurement by Sodar
  • Analysis of SCADA data of the wind turbines


  • The power curves as determined on the turbines are sufficiently in agreement with the warranted levels. However, improvements are possible.
  • The yaw algorithm of one of the investigated turbines is far from optimal.
  • A check after a corresponding software change showed a significant improvement.
  • One of the turbines is on average not aligned optimally with the wind direction. A change would result in an increase of the energy capture by several percent.


Assessment of the wind shear

Site: 200 MW project in Morocco

Issue to be resolved:

The existing wind measurements with masts only cover a small part of the height range of the rotors of the planned wind turbines. The course of the wind speed above the masts is therefore unknown. This results in a significant uncertainty of the energy prediction and a difference between the results of different experts.

Solution: Wind measurement with Sodar over several months


  • The wind shear across the entire heigt range of the rotor is very homogenous.
  • The main cause for the difference in the results between the existing wind resource assessments could be resolved.
  • The total uncertainty of the energy prediction for the entire project was reduced by about 1 %.
  • The data availability of the Sodar instrument at the site was very good.
  • The results of the verification of the Sodar instrument performed in North Germany are valid for the Moroccan site.
  • The parabolic sound reflectors should be cleaned from sand about every two months.
  • Operation of the instrument by local personnel was possible without problems.


Performance check of wind turbines

Site: medium size wind farm in Belgium


  • combination of analyses of SCADA data and a nearby wind measurement
  • comparison to reanalyses data, met station data and aj production indices


  • inconsistencies related to the power curve could be detected in the past, including longer periods with reduced availabilities and a variety of curtailments
  • more recently, the operational behaviour of the turbines has improved; however, some inconsistencies still occur around rated wind speed which need to be analysed further with respect to an additional load on the turbines
  • realistic performance of the wind turbines in the wake of neighbouring turbines, i.e. no under-performance with respect to the power curve
Re-assessment of energy production

Site: medium size wind farm in Northern Germany; the real energy production is much lower than calculated in the original assessment report

Procedure: profound analysis of SCADA data


  • reduction of the long term mean wind resource, albeit its contribution to the overall result is small compared to the following aspects
  • thrust coefficients too low
  • power curves are not reached in the wake of neighbouring turbines
  • strong deflection of the flow due to the specific wind farm layout in terms of very low distances between wind turbines
Re-Assessment of the energy losses due to curtailment for the protection of bats

Site: Medium size wind farm in the state of Hessen, Germany


  • Analysis of SCADA data of the wind turbines
  • Comparison with model calculations
  • Parametric studies


  • The calculations with the model of aj determine the real energy losses very well.
  • Too high values of the outside air temperature measured by the turbines lead to unnecessary shutdowns. The energy losses are thus signficantly higher.
  • Some turbines show too low wind speeds. This also leads to unnecessarily increased shutdown periods and too high energy losses.
  • The wake impact within the wind farm must be taken into account in the model calculations.
  • The energy losses are much higher in years with low wind resource compared to high wind years.


Comparison of the wind data measured on a tall met mast with common long term reference data

Site: North Germany


  • The on average decreasing wind speed level observed in met station data and production indices during the investigated period is confirmed by the met mast.
  • The reanalysis data do not show the same pattern.
  • The long term course of wind speed is similar in all measurement heights.
  • The measured data correlate best with met station data in low heights and with reanalysisd data in high levels. However, the magnitude of the correlation does not provide a means for evaluating the agreement of the long term average course in the time series.
Comparison of a reanalysis data source with measured wind data

Sites: Many sites in different regions of Germany


After a strong, but systematic correction, the reanalysis data show very realistic wind roses, wind speed frequency distributions and diurnal courses of the wind speed in larger heights above the ground at all sites, except for those with complex structure on a small scale.

3 wind measurements in parallel: met mast, Sodar and Lidar

Site: Black Forest, Germany


  • the vertical wind profile of the met mast showed implausible values
  • data of a Sodar instrument from an external company were also considered implausibe, therefore a Sodar instrument of the manufacturer AQ Systems was used afterwards
  • strong consistency between the AQ Systems Sodar and the Lidar