Romo Wind takes wind data accuracy and transparency to the next level
Get more and detailed insights on Yaw Misalignment distribution in several wind turbine types.
Determining the true performance of each wind farm turbine is extremely difficult if not impossible. Nowadays performance measurements in wind farms are limited to single turbines with free wind sectors, defined flat areas and using dedicated met masts. Although lidars have the potential to replace met masts they are limited to the same constraints. Hence the only wind measurement source available on each turbine is the nacelle anemometry, which provides neither meaningful nor trustworthy input for a performance measurement. Here the spinner anemometer technology iSpin can change the game and provide more transparency and insights in turbine performance.
Due to its unique position and measurement principle in front of the rotor it is possible to overcome the limitations of conventional nacelle anemometry in two ways: First of all it is possible to measure the main aspects which define the wind input. Secondly the iSpin measurements show high robustness for local flow conditions, different from those of e.g. a reference measurement on a prototype site. Based on the iSpin technology an approach is presented using these capabilities for performance monitoring and performance comparisons of wind farm turbines. Contrary to conventional nacelle anemometry iSpin allows the usage of 360° inflow data even in wake of other turbines or behind obstacles. This enables a holistic overview of the performance of individual turbines and the wind farm itself.
During their operational life wind turbines are subjected to many factors influencing the loads and finally the performance characteristics the systems will show (see Figure 1). Wind turbines are of course affected by the site specific environmental conditions and – in case of a wind farm – by conditions introduced by other turbines in the wind farm. Furthermore the set-up of the turbine itself, characterised e.g. by the blades, the drive train and the turbine control logic is defining the response of the turbine to external and internal conditions and events. Other issues which will appear during turbine life are components degradation, components failure and malfunctioning turbine control, sensors or actuators. All this can lead to a behavior and finally a performance characteristic outside manufacturer’s specification.
The difficult thing is to get an objective grip on the performance of wind turbines. Questions like: “Does a turbine changes its performance characteristic over time?” or “How do the turbines perform relative to each other?” are difficult to answer. For customers as well as for manufacturers it is important to be able to identify a performance outside the “normal” range and to be able to differentiate between the drivers. In other words: Is there an under performance and what is causing the under performance? Are the environmental conditions different than expected or forecasted during the site evaluation process and is this causing less annual energy production? Or is the turbine under performing, due to e.g. degradation, issues with components and problems with the turbine control or misleading measurements.
Figure 1: Performance drivers of a wind turbine
Wind turbines are energy producing devices, i.e. it is important for the customer and the manufacturer to know how efficiently a turbine converts energy from the given wind conditions. This power performance characteristic is commonly expressed as electrical power (output) versus horizontal wind speed (input) measured under free inflow conditions at a distance of 2 to 4 rotor diameters in front of the turbine. Here is where the big dilemma in the wind industry lies so far.
On the one hand it should be monitored that every turbine’s performance characteristic is within the specification, but on the other hand it is impossible to measure the wind quantities at all turbines and at all sites, using met-masts or other accepted forward looking measurement systems. This is a known fact and the current way to handle it is to use nacelle anemometry wind speed measurements for performance monitoring of each individual turbine. Additionally – in certain, limited cases – the input-output relationship, the power curve, is determined by using met-masts, ground based lidars, nacelle lidars or spinner anemometer. Those power curve verification measurements are performed at prototype sites or at dedicated turbines in a wind farm to verify that the turbine power curve is fulfilling contractual obligations. The results from the verification measurements are considered representative not only for the individual turbine, but for all turbines of the evaluated wind farm.
Usually during prototype measurement scaling factors, offset and nacelle transfer function (“NTF”) for nacelle anemometry are developed for the specific turbine type and anemometry in relation to the reference measurement system, which is in most cases an IEC compliant wind met mast. As a guideline for deriving these values the methods and criteria described in the IEC standards 61400-12-1 and 12-2 are normally used. This also means that only free inflow sectors are considered to avoid any unpredictable influence to the met-mast respectively another forward looking measurement system and the test turbine from wake effects of other turbines or any other obstacles in the area.
Unfortunately, the existing nacelle based wind speed measurements – although considered as not really meaningful and trustworthy – are quite often the only source of information for the input quantity “wind” in relation to the output quantity “power”. Due to their position behind the rotor the conventional nacelle anemometers are unfortunately very sensitive to changes in the flow conditions compared to those of the free inflow sector at the prototype site. Flow conditions differing from those at the prototype site, caused by environmental or wake effects, will lead to different inflow passing the rotor and nacelle structure and finally reaching the nacelle anemometer. Furthermore those sensors are not able to measure the turbulence intensity (the variation of the wind speed) and the inflow angles, which are affecting the power characteristics as well as the fatigue loads.
Figure 2 shows how the formerly linear relationship between nacelle anemometer measurements (y-axis) and reference system measurements (x-axis) changes when the flow conditions change, e.g. after a yaw misalignment correction or application of vortex generators.
Figure 2: Change of nacelle anemometer measurement characteristic after yaw misalignment correction: Scatter plot (left), deviation (right)
First of all it can be seen that in general the nacelle anemometer – despite air density correction and filtering for free inflow sector data – overestimates wind speeds compared to the reference measurement system. Instead of a ratio of 1 the ratio between nacelle anemometer and the reference system readings is rather 1.05. This already indicates that the scale, offset and nacelle anemometry NTF, once established for this turbine type, is not applicable for this specific site and provides therefore higher wind speeds. Additionally and more important, the correction of the yaw misalignment causes the nacelle anemometer to measure even higher wind speeds in the “after” period compared to the “before” period, where the turbine was operated misaligned. Obviously the nacelle anemometer’s NTF is not able to cope with the changes in the flow conditions after the correction of yaw misalignment.
What does this mean to the power curve evaluation based on nacelle anemometer readings in this specific example? At least two things: As a first consequence the performance of the wind turbine based on the nacelle anemometer is underestimated compared to the reference measurement. Due to higher wind speed values shown by the nacelle anemometer the power curve is shifted to the right. As a second consequence – after the yaw misalignment correction – the nacelle wind speed to reference wind speed relationship even increases in the partial load range, i.e. the air flow reaching the nacelle anemometer seems to be less disturbed by the rotor and the nacelle structure and is therefore faster. This causes the nacelle anemometer based power curve, i.e. the SCADA power curve to get even worse after the correction and therefore the effect of yaw misalignment correction appears to be minor or negative although in reality it is effective and significantly positive.
By using the iSpin technology1, with its unique position and measurement principle in front of the rotor it is possible to overcome the limitations of conventional nacelle anemometry:
First of all, there are the advanced measurement capabilities of the spinner anemometer iSpin itself. Table 1 shows the aspects of the wind field affecting the turbine performance and which of those can be measured with conventional nacelle anemometry and which with the iSpin system. Except for wind shear and veer – for which investigations are ongoing – all relevant factors describing the wind input to a turbine for a performance evaluation can be measured with the iSpin system.
Table 1: Measurement capability of nacelle anemometer and spinner anemometer iSpin
|Quantity||Conventional nacelle anemometry||iSpin|
|Wind speed||NTF sensitive to different inflow condition||NTF shows robustness even in wake|
|Turbulence intensity||No possibility to measure||Key capability of iSpin|
|Flow inclination||No possibility to measure||Key capability of iSpin|
|Yaw misalignment||Indirect measurement, very sensitive to measurement position||Key capability of iSpin|
|Shear||No possibility to measure||Evaluation ongoing|
|Veer||No possibility to measure||Evaluation ongoing|
The second important point is that the robustness against flow conditions differing e.g. from those of the prototype site is significantly increased. Tests at wind farms in different terrains indicate that even when considering 360° inflow – this means also including wake situation caused by other turbines – the scatter and the characteristic of the power curve remain nearly the same as for the free inflow sector. Figure 3 shows as an example a result of the Nørrekær Enge2 test experiment. Here the scatter and power curve characteristics of turbine number 4 are almost similar for free inflow and 360° inflow conditions.
Figure 3: iSpin power curves at turbine no. 5 of the wind farm Nørrekær Enge: Free inflow (left) and 360 degree inflow (right)
Another example is shown in Figure 4 (left diagram). In this study the relation between iSpin and
reference wind speed measurements was analysed before and after the application of two measures
(vortex generator installation and yaw misalignment correction) to improve the turbine performance.
Although the vortex generators were installed near to the blade root the resulting slopes of both periods
(“before VG” and “after VG”) remained almost the same. The same robustness of the iSpin transfer
function, i.e. measurement characteristic can be seen for the yaw misalignment correction which was
performed at this turbine some weeks after the vortex generator installation.
Figure 4: iSpin and nacelle anemometry measurement characteristic before and after vortex generator installation and after yaw misalignment correction
Contrary to the nacelle anemometry measurements (see figure 4 right diagram) it becomes obvious that the iSpin based wind speed measurements are not affected by the vortex generator installation or the yaw misalignment correction. This means that the measurement characteristic of the iSpin system remains stable and can be used for performance change detection application. All in all this leads to the conclusion that – with the quantities able to measure and once being calibrated at a site using a reference measurement system like a met-mast – iSpin allows precise wind measurements which can be used for evaluation of the individual power curves of all wind turbines within the wind farm itself and for other wind farms in comparable sites.
What does it mean to use iSpin measurement data to generate power curves and get an idea about the performance of each individual turbine or a complete wind farm? Figure 5 is showing the power curves using 360° inflow and being measured with the iSpin system and the nacelle anemometry, i.e. provided by the turbine SCADA systems for the Nørrekær Enge wind farm located in Denmark. In addition to this power curves the IEC 61400-12-1 compliant power curve – measured with a met mast in front of turbine number 4 – is shown as well. For this evaluation 9 of the wind farm turbines have been used (T2 to T6 and T10 to T13), de-rated turbines were excluded.
Figure 5: Comparison of power curves based on iSpin and nacelle anemometer measurements
From the graph at least four observations can be made: First of all none of the SCADA power curves match the IEC power curve. This means that both, the calibration factor and the nacelle anemometry NTF having been established once for the turbine type are no longer applicable and definitely not usable for 360° inflow. Secondly, a very large variation in power curves between the different turbines can be seen. In contrast the iSpin based 360° power curves do match the IEC power curve very well – the average difference between the 360° iSpin power curves and the IEC power curve was 0.1% – and all power curves show similar characteristics, with a variation in AEP within 2.1% for the given annual wind speed at site.
What does this mean in terms of interpretation of power curves? As a consequence an operator only looking on the nacelle anemometer power curves might conclude that these turbines are underperforming and might start to invest in optimization measures, where none are necessary. The iSpin power curves show a very consistent picture of power curves, which are well in the nominal performance range indicated by the IEC compliant power curve.
Looking into Figure 5 it is obvious that by using the SCADA based power curves all turbines seem to perform very different. With the knowledge about the low robustness of the nacelle anemometer calibration factor and NTF it is very difficult – not to say almost impossible – to identify which turbine is the best performing and which performs worst. Translating the spread of power curves into AEP variation this does mean a band around an average power performance (which is not including the IEC power curve!!) of +/-8% (see red band in figure 6). This is very different for the iSpin based power curves. Here the AEP band around the average power curve is +/-2% (see green band in figure 6), significantly improving the ability to identify turbines operating differently from others.
Figure 6: AEP variation bands found at NKE test site for iSpin based PCs and SCADA based PCs
As an example for identifying turbines operating outside a nominal range turbine number 13 can be used (see orange line in Figure 7). For this turbine – for which a yaw misalignment of 6.8° was detected using the iSpin system – the power curve and consequently the AEP differed from the average power curve and AEP band based on the turbines not affected by any yaw misalignment. All power curves of those unaffected turbines were within the AEP band.
Figure 7: Identification of a turbine outside normal operation
The Nørrekær Enge field experiment demonstrated for a flat terrain site that it is possible to apply the iSpin wind speed measurements to generate a reference turbine performance characteristic including a band for the wind farm and to compare the other turbines in the wind farm to this reference. To transfer this approach to other wind farms, iSpin systems should be installed on all turbines in the wind farm, but only at one of them an IEC compliant reference measurement to generate the iSpin free wind speed calibration factor and NTF has to be performed. Preferably after commissioning an accredited 3rd party consultant should perform power curve verifications according IEC 61400-12-1 and 12-2 on one wind turbine in the wind farm, using an IEC compliant met-mast set-up and calibrated iSpin equipment. After some plausibility checks the free flow calibration factor and NTF derived at this reference turbine can then be applied to the other turbines to measure the wind and capture the power curves as well. Figure 8 shows a possible process flow to generate a site and turbine specific reference performance characteristic including tolerance band, i.e. perform the iSpin Guardian approach.
Figure 8: Possible process flow for iSpin Guardian approach
In order to achieve good results and low uncertainty in the generation of the AEP band, the requirements and data treatment listed in Table 2 should be fulfilled respectively performed when applying the iSpin Guardian approach. Although not essential, the best and most meaningful results will be achieved if the iSpin technology is factory fitted by the turbine manufacturer and directly connected to the turbine control system before delivery to the wind farm. This would allow the highest accuracy and repeatability of the sensor installation and the best timewise correlation with the turbine power signal.
Table 2: Requirements and data treatment for iSpin Guardian approach
The iSpin Guardian approach is a scalable approach and therefore not limited only to the performance evaluation of a specific wind farm. After gathering reference measurements, i.e. free wind speed calibration factors and NTFs for a specific turbine type at several different sites, it should be possible to generate a turbine type specific performance characteristic which can be used for fleet wise evaluation of the turbine performance.
One interesting aspect of creating more transparency in performance is how it might influence warranties on performance and reduces risks in business case calculations. Table 3 shows what is currently offered concerning warranties and what is considered possible when using the iSpin technology. Costs, effort and suitable installation opportunities for wind met masts do limit power performance verifications of all wind turbines in a wind farm following accepted standards and practices. Here the iSpin technology enables that the very limited turbine performance spot check, only applied for a short period, can be replaced by turbine lifetime performance monitoring of all turbines in a wind farm
Table 3: Performance transparency and effect on warranties
|Warranties||Offered today||Technically possible|
As described in this article the knowledge of environmental conditions being provided by the iSpin system can help to differentiate whether a deviating turbine performance and finally the energy production at the specific site is mainly due to underperformance of a turbine or whether this is caused by the site and wind farm arrangement itself. But this is not the only application for which the iSpin measurement data can be used. Wind turbines are designed following certain type classes and turbulence intensity distributions. This means that a set of pre-defined and standardised environmental conditions is used to create the design load envelope – including fatigue and ultimate loads – which the components have to withstand. When it comes to real sites the environmental conditions and further influencing factors caused e.g. by the wind farm turbines themselves will be different from those standardised conditions. Therefore those conditions are checked against the standardised environmental conditions, the load envelope or the design itself during the site assessment procedures of the manufacturers. The information about the site conditions and turbine interaction effects being used in the site assessment are based on historical data respectively models with uncertainties included. But what happens if the wind regime and the terrain properties change (especially in complex terrain) or obstacles occur or disappear and how does all this affect each individual turbine over time? As explained before conventional nacelle anemometry is not suitable to capture those conditions and their changes over time, because the measurement characteristic gets influenced by the changes and relevant quantities cannot be captured by them. Here iSpin measurement data including wind quantities as listed in Table 1 and including air pressure and air temperature measurements can help in many ways. iSpin measurements at every wind farm turbine or even at dedicated turbines can be used to identify differences to originally used site conditions and document under which wind conditions the turbines have been operated over time. This enables, for example, the check or adaptation of operating strategies for the wind farm and individual turbines, e.g. optimisation of park efficiency. Another important aspect is the usage of historical iSpin measurement data from each individual wind farm turbine to support the decision about life time extension and inspection intervals. As a third aspect the availability of iSpin measurement data can and will help to make more informed decisions about planned wind farm extensions.
As the spinner anemometer NTF is, in this evaluation, a significant source of uncertainty, the “iSpin Guardian” approach could be improved even more by using the calibrated free spinner wind speed instead of free horizontal wind speed. The turbine type specific power curve would be generate using the free spinner wind speed, i.e. the free wind speed without taking the deformation due to rotor induction effects into account. Although the resulting spinner wind speed power curve cannot be directly compared to the warranted power curve based on met mast free wind speeds, the performance characteristic measured with this approach is unique for a specific turbine type and therefore comparable and traceable. This would be achieved firstly by proving during the reference measurement, that the turbine is matching the warranted power curve according to IEC 61400-12-1 and 12-1, secondly by mapping the spinner wind speed power curve of the reference turbine to the reference power curve(s) according to IEC and thirdly by comparing spinner wind speed power curve of the wind farm turbines with the spinner wind speed power curve of the reference, i.e. the “iSpin Guardian”. This concept would considerably reduce measurement uncertainties and therefore allow monitoring and managing the performance of wind farms in a far more granular way than is possible today. Even more importantly it has the potential to allow financiers and insurers to reduce risk premiums in their business calculations and therefore generate improved returns for the wind turbine industry
If you have further questions, please do not hesitate to contact us at email@example.com.