Wind speed deviations in complex terrain

Wind farm sites within complex terrain are subject to local wind phenomena, which have a huge impact on a wind turbine’s annual energy production. To reduce investment risk, an extensive site evaluation is therefore mandatory. Stationary long-term measurements are supplemented by CFD simulations, which are a commonly used tool to analyse and understand the three-dimensional wind flows above complex terrain. Though being under heavy research, such simulations still show a 10 huge sensitivity for various input parameters like terrain, atmosphere and numerical setup. Within this paper, a different approach aims to measure instead of simulate wind speed deviations above complex terrain by using a flexible, airborne measurement system. An unmanned aerial vehicle is equipped with a standard ultrasonic anemometer. The uncertainty of the system is evaluated against stationary anemometer at different heights and shows very good agreement, especially in mean wind speed (<0.12 ms) and mean direction (< 2.4 °) estimation. A test measurement was conducted above a forested and hilly 15 site to analyse the spatial and temporal variability of the wind situation. A position dependent difference in wind speed increase up to 30 % compared to a stationary anemometer is detected.


Measurement System "WindLocator"
2.1 Design 65 The measurement system, which has been used for the measurement campaigns within this paper, has two main, independent components: a powerful carrier system and a sensor unit, which consists of a commercially available ultrasonic anemometer and a self-developed compensation and data acquisition unit.
The foldable, commercial carrier system is a battery powered octocopter with a flight time of 25 mins and a maximum takeoff-weight of 12.5 kg. Including the sensor unit, the complete system only weighs 8.5 kg and therefore has a considerable 70 performance reserve. Flights at turbulent air as well as during gust speeds of 25 ms -1 have successfully been tested. A realtime-kinematics (RTK) GPS is included to perform high accuracy positional navigation and speed estimation. The open source flight controller has been adapted for an easy setup of specific measurement strategies, which then are autonomously being followed. Although a completely unobserved operation is technically possible, European laws at this moment require an operator to be within sight. 75  The Gill WindMaster 3D ultrasonic anemometer is placed on top of the compensation unit centred above the rotor plane. 80 Mounting the sensor on top of the UAV has several advantages. First of all, the rotational symmetry of the system allows wind measurement independent from yaw angle and wind direction. Additionally, this setup results in a horizontally centred mass during hovering and therefore leads to relatively small moments to be compensated by the UAV. This improves flight performance and flight time. Aside of that, the downwash above the rotors is less turbulent then below.
The distance of the sensor's measurement volume to the rotor plane is 750 mm and is considered as a trade-off between 85 manoeuvrability and reasonable interaction between wind sensor and propeller induced flows. Except for the power supply, the self-developed compensation and data acquisition unit is completely independent from the 90 UAV. If requirements concerning the carrier system change, the compensation unit as a whole can be reapplied easily on a new aircraft. It weighs 420 gr and contains all necessary sensors as well as an additional RTK-GPS for an accurate position and speed estimation by means of sensor fusion. Based on analytical calculations and various synthetic experiments, a compensation algorithm was developed, that efficiently reduces measurement errors due to movements of the airborne system as well as it's rotors. Additional telemetry transmits measurement data such as wind speeds and directions live to a ground 95 station for in situ analysis. The anemometer data is additionally saved to an internal storage with a rate of 10 Hz. https://doi.org/10.5194/wes-2020-25 Preprint. Discussion started: 17 March 2020 c Author(s) 2020. CC BY 4.0 License.

Validation of the system
All following calculations and measurements have been evaluated based on data, that has been processed by the compensation unit. The system validation in general was conducted on several levels of detail, beginning with the Guide to the Expression of Uncertainty in Measurement (GUM) to evaluate the standard uncertainty of a single measurement datum. The GUM allows 100 the calculation of the standard uncertainty without the necessity of a true reference value. Error estimation is done by creating a mathematical model of the WindLocator, including relevant influences and their uncertainties and combining them into the system's standard uncertainty, which is +/-0.37 ms -1 in our case.
After several synthetic tests with a fixated UAV to evaluate rotor influences (Figure 2), the WindLocator's compensation unit was tested during an indoor flight at zero-wind conditions ( Figure 3). Because no stable GPS signal could be received during the indoor flight, the WindLocator was set to "Altitude Hold", which utilizes the internal barometer to maintain an altitude of around 4 m during our test and automatically stabilizes pitch and roll axis for minimum horizontal movements. Nevertheless, small sensor inaccuracies made pilot interventions necessary to remain at sufficient distance to walls. After compensation, the wind data is given out in a global north-east-down coordinate system and is therefore independent from the specific orientation of the UAV. 110 https://doi.org/10.5194/wes-2020-25 Preprint. Discussion started: 17 March 2020 c Author(s) 2020. CC BY 4.0 License. Figure 4 shows the data in all three measured directions at a resolution of 10 Hz. Peaks, e.g. in vN-direction at second 17 (-1.5 ms -1 ) and 55 (-1.31 ms -1 ) are a result of the UAV's horizontal translation due to operator intervention. As expected, mean wind speeds during the indoor flight are very close to zero. Standard deviations up to 0.23 ms -1 meet our expectations according to GUM, but clearly show the influence of manual operator control and of the sensor being rather close to the turbulent downwash induced by the rotors.

Figure 4 North-East-Down wind speed components of indoor flight at zero wind conditions
After showing that under zero-wind conditions, mean values are in good agreement with our expectations, a measurement 120 setup was created to compare the performance of the WindLocator with a stationary anemometer. In flat, agricultural terrain 2 km west of Aachen (Northrhine-Westfalia, Germany), a stationary anemometer of the same type as the UAV's anemometer was mounted at a height of 3 m above ground level. Data acquisition and storage for the stationary anemometer were realised at 10 Hz by a self-developed data acquisition system, which uses time stamps synced with an internet time server. The UAV time stamps are derived from GPS time signals. The UAV was set to hold position at a height of 3 m. A distance of 4 m to the 125 stationary anemometer rectangular to the main wind direction was chosen to avoid interactions of the two measurement systems ( Figure 5).

Figure 5 WindLocator hovering close to stationary anemometer
Four ten-minute measurements with ~6000 data points each have been conducted, with a short break to switch batteries after 130 the second measurement point. In opposite to the indoor tests, all three wind components are combined into a single wind speed v for every measurement datum to improve comparability to the stationary anemometer. However, the vertical component vD in general has a minor impact on the resulting wind speeds.
The following diagrams (   For all measurement points (see Table 4), a very good agreement of the ten-minute-mean wind speed between WindLocator 145 and reference has been achieved, especially when taking into account the turbulent wind situation during such a low-altitude flight. Turbulence intensities (TI) up to 44 % have been calculated for the stationary reference. Although there are absolute differences of +1 % (measurement 1) to +6 % (measurement 2/4), the WindLocator already gives a good estimation of the prevailing turbulence intensity. An analysis of wind directions during this experiment was not yet possible, because an accurate orientation of the stationary measurement system could not be guaranteed. This is taken into account for the next experiment at a 134 m met mast under more realistic conditions ( Figure 10).

Figure 10: 134 m met mast at Windtest Grevenbroich GmbH test site in Germany
The measurement system was tested on a sunny day close to a met mast on a small plateau. Four measurements of 8-10 minutes have been conducted and are compared to the velocity data of a cup anemometer at 134 m and the directional data of a wind vane at 130 m above ground level. The WindLocator was flown to a height of 134 m based on barometer and GPS data and was then moved closer towards the met mast using the onboard camera system. Because the flight was performed without 160 autopilot, distances to the met mast and exact height vary throughout the measurements (see Table 5). Additionally, that table contains wind speed data analogue to Table 4 as well as information concerning the accuracy of wind direction estimations.
For all following calculations, the WindLocator data was averaged to 1 Hz for better comparability to the met mast. The results during the met mast experiment show a slightly different picture compared to the ground level measurements. Still comparable to the former test is the reasonable performance of the turbulence intensity estimation of the UAV. Additionally, measurements 2 ( Figure 12) and 4 ( Figure 14) show a good correlation of the WindLocator and the corrosponding reference speed However, mean wind speed deviations for the first and third measurement are not only higher than before, but also vary a lot more compared to the other measurements of that day. Serious deviations mainly occur during the first half of the 170 measurements ( Figure 11; Figure 13) in a temporary manner, e.g. seconds 180 to 270 for measurement 3.    Despite its challenging but beneficial design with the anemometer mounted above the rotor plane, the WindLocator performed very well throughout the tests, especially concerning the calculation of averaged measurement quantities like speed and direction. When the system uses its GPS based hover mode without "disturbance" by a pilot, mean wind speed differences compared to a reference were below 0.12 ms -1 and wind direction differences smaller than 2.4 °. Because the 195 airborne measurement system and a reference cannot measure at the exact same place at the exact same time, remaining uncertainties always also might be a result of spatial deviations in the wind situation, which will be discussed in more detail in the following section.

Test site and measurement setup
The aim of the campaign is to investigate mean wind speed deviations above complex terrain by using the WindLocator. The 200 area for this measurement campaign is a small hill in the south of Northrhine-Westphalia in the german Eifel and was chosen for the following reasons: • With a yearly mean wind speed of 6.5-7 ms -1 at 100 m above ground level, the area has rather high wind speeds compared to the rest of the county. Main wind direction is southwest.
• The terrain is considered to be complex ( Figure 19). The slope around the hill at most parts is greater than 40 degrees. 205 Forests extend to the south and west of the hill. A small village is located to the northeast.
• The region in general was considered to be suitable for wind turbines and has a good accessibility.

Figure 19 terrain model of the test area and measurement points (red)
To reduce experimental complexity, wind speed deviations within a two-dimensional plane above the described complex 210 landscape are going to be investigated. The plane to be surveyed is roughly 500 m x 500 m and placed on the middle of the hill with equal distance to the edges in the south and west. All planned measurement points, which are displayed as red circles in Figure 19, are at the same height above sea level and around 100 m above the lift-off point. Additional information is summarized in Table 6. The measurement workflow begins with the activation of the WindLocator's autopilot. The system then automatically flies to the first, predefined measurement point and holds position for five minutes before heading for the next point. For this feasibility study, five minutes were considered to be a reasonable trade-off between limited battery time and statistical coverage for each point. After four measurement points, the UAV automatically returns to its lift-off position for a battery change. Afterwards, 220 the measurement process continues with the next point. During post-processing, measurement points are automatically detected within the data for further analysis.
Aside of the WindLocator, an additional stationary ultrasonic anemometer was in use during the campaign. It was placed at a height of 3 m on free grassland, nearly centred under the measurement plane and also acquires data at 10 Hz. Figure 20 gives an overview of the measured resulting wind speeds v from the moving WindLocator and the stationary reference on ground. After the data acquisition was started, the UAV heads for the first measurement point, where it is holding position for five minutes 100 m above ground, before heading for the next waypoint. A measured wind speed deviation between WindLocator and reference is expected because of the differences in height and horizontal position of both systems. During the battery swap after four measurement positions, obviously no WindLocator data is available. The stationary reference 230 instead measures non-stop. Measuring 16 points of five minutes, making it 80 minutes of usable measurement data, has taken around two hours in total.

Figure 20: Comparison of resulting wind speeds from WindLocator (compensated) and Reference at 10Hz
Over all UAV measurement points, an absolute variation of mean wind speeds between 2 and 6 ms -1 has been detected ( Figure  235 21). As it was already implied in the introduction, this obviously is not only a result of spatial deviations due to complex terrain, but also a consequence of wind variation over time. Otherwise, the stationary reference (assuming it to be an indicator for the overall wind situation) would not have shown any significant differences in wind speed over time.  This is clearly not the case, especially when looking at the normalised reference wind speed, calculated by dividing the mean wind speed value of each point by the maximum mean value of all measurement points ( Figure 22). The strong correlation (R=0.86) between relative mean speeds of WindLocator and reference data is an indicator, that the ground level stationary anemometer for this particular campaign is a suitable reference to also track temporal changes of the overall wind situation.
The remaining differences between WindLocator and reference tend to be local wind speed deviations, e.g. due to terrain. 245 Taking this into account, a very simple measurement strategy seems suitable as a first approach: the time offset between single UAV measurement points can be compensated by referencing them to the stationary anemometer data. The result is a measured wind speed increase compared to the reference and is summarized in Table 7. 250  This measurement strategy, nevertheless, depends on various parameters concerning the stationary reference, its positioning and expected spatial variations and therefore cannot be considered to be valid in general. Advanced measurement strategies and criteria are currently under development. Figure 24 combines the measured results on the one hand and the UAV's GPS data on the other hand to a spatial distribution. 260 The purple arrow indicates the mean wind direction. Each red arrow represents a measurement point, showing the measured horizontal wind direction and indicating with its length the wind speed increase compared to the stationary reference. For a better insight, the wind speed increase then is interpolated linearly between measurement points to create a contour plot. The background shows the digital terrain model data. The highest increase in wind speed compared to the stationary anemometer is located towards the ridge at the upwind side, which meets our expectations concerning of a speed-up effect at a steep hill. The following area of lower wind speeds might then be a result of flow separation. Downwind of the hill, towards the southeast ridge, an additional area of higher wind speeds is located. Towards the plateau in northeast direction, lowest wind speed increase has been measured. 270

Conclusion
Within this paper, a UAV-based measurement system called WindLocator, its validation and its experimental application above complex terrain were presented. The measurement system consists of a powerful octocopter, a commercial ultrasonic anemometer centred above the rotor plane and a self-developed compensation and data acquisition unit. The latter was the enabler to efficiently reduce wind measurement errors due to movements of the UAV and rotor influences. This has been 275 shown in two test scenarios at different wind and turbulence conditions. At both tests, very good agreement with reference data could be achieved. Mean wind speeds have been estimated with a maximum difference of 0.12 ms -1 , wind directions with a maximum difference of 2.4 ° during position-controlled hovering.
Though rotor influences are a challenge, turbulence intensity estimation was reasonably good. Nevertheless, the compensation unit is under continuous development to improve accuracy at all relevant flight situations. 280 The biggest advantage of an airborne measurement system is its flexibility, allowing measurements at any arbitrary point in a wind field above any kind of landscape. This could make the WindLocator to a potential alternative for CFD simulations in complex terrain, delivering an analogue result for a specific weather situation without long computation times or modelling uncertainties. To do so, temporal and spatial variations of wind speed have to separated.
A hilly and forested region in the germen Eifel is investigated concerning its local wind speed deviations by using the 285 WindLocator. Although more advanced measurement strategies are currently under development, for this specific case, a very simple strategy was sufficient due to a good time correlation between reference and UAV: while the WindLocator automatically was flying from point to point, a stationary reference at ground level was used to compensate the time offset between single measurement points. The result was a plane of four times four measurement points, including information of wind speed increase compared to the reference and three-dimensional wind directions. Spatial differences of approximately 290 +/-30% compared to a mean value have been found at plausible locations, underlining the necessity of intensive site evaluation in complex terrain.