What Is The Power Curve Verification Of Wind Turbines?
A wind turbine power curve is a graph that represents the relationship between the power output of a wind turbine and the wind speed. It shows how much electrical power a wind turbine can produce at different wind speeds.
The power curve allows wind farm developers and operators to estimate the annual energy production at a site and calculate the capacity factor of a turbine. It is a crucial tool for assessing the performance of a wind turbine.
The accuracy of the power curve is very important. Even small errors can lead to large discrepancies in energy production estimates, especially for large wind farms. Therefore, power curves must be precisely measured and verified through power performance testing.
Power curves are used for a variety of applications in the wind energy industry including:1
- Estimating annual energy production for a wind project
- Validating power performance claims of turbine manufacturers
- Assessing warranty obligations related to minimum power output
- Optimizing wind farm operations and maintenance
- Understanding impacts of degradation over time
Power Curve Testing Standards
The primary international standard for wind turbine power performance measurement and power curve testing is IEC 61400-12 published by the International Electrotechnical Commission (IEC). The latest version, IEC 61400-12-1 Ed. 3.0 b:2022, provides the requirements for turbine power performance measurements and procedures for analyzing, reporting, and applying correction to measurements [IEC 61400-12-1].
Other key standards organizations involved in establishing power curve testing standards include MEASNET (Measures for Wind Energy Assessment), CENELEC (European Committee for Electrotechnical Standardization), and AWEA (American Wind Energy Association). MEASNET aims to create a forum for discussion on best practices for site assessment and power performance measurements. CENELEC publishes the EN 61400-12 series aligned with IEC standards. AWEA provides guidelines used in the U.S. wind industry.
These standards aim to provide uniform procedures and uncertainty levels for power curve measurements. They cover aspects like sensor calibration, data acquisition, analysis methodology, power performance evaluation, and reporting. Adherence to standards ensures consistency and accuracy.
Testing Methods
Two common methods for testing the power performance of wind turbines are met towers and remote sensing devices such as lidar or sodar. Met towers rely on anemometers and wind vanes mounted at different heights to measure wind speed and direction. The anemometers should be placed at hub height and ideally at other heights along the rotor diameter to capture the wind profile. The data collection duration must be at least 3 months to capture seasonal variations.
Analysis of met tower data involves correlating the wind speed measurements at each height to the power output of the turbine. This establishes the power curve over a range of wind speeds. Lidar uses laser pulses to remotely measure wind velocity and can provide measurements across the entire rotor disk area. This provides advantages over met towers for complex terrain sites. The analysis techniques for lidar data are similar, correlating the wind speed and direction measurements to turbine power output.
Uncertainty Analysis
Uncertainty is prevalent in power curve analysis and can introduce error in forecasting wind turbine power output. The main sources of uncertainty include wind speed measurement uncertainty, performance variability, and statistical uncertainty in the curve fitting methodology (Simmons, 2018). Wind speed sensors have inherent uncertainties, and the wind field itself can vary across the rotor plane and over time, introducing more variability. Turbine performance can also change due to blade soiling, component wear, and grid fluctuations.
There are statistical methods to quantify these uncertainties such as the IEC 61400-12-1 guidelines, which provide a framework for calculating the standard uncertainty of power curve measurements (Lee et al., 2020). Uncertainty analysis provides confidence intervals around the power curve, which represents the range of potential turbine performance.
Strategies to reduce uncertainty include using redundant wind speed sensors, applying shear corrections, monitoring performance regularly, and increasing the sample size used to derive the curve fit. Measurement best practices, improved models, and quantifying the various uncertainty components can help increase confidence in power curve forecasts.
Extrapolation Methods
Extending the power curve beyond the tested wind speed range requires extrapolating the measured power curve data. This is necessary to estimate the wind turbine’s power production at higher wind speeds that exceed the limits tested during power performance verification. There are several methods used for power curve extrapolation including:
Linear Extrapolation – A simple approach that extends the measured power curve by drawing a straight line from the last tested wind speed and power output. This assumes a continued linear relationship beyond the tested range. However, the power curve begins to flatten out at higher wind speeds as the turbine reaches its rated power limit. So linear extrapolation tends to overpredict power output at very high wind speeds (Wind Rose Excel, n.d.).
Exponential Extrapolation – Fits an exponential function to the measured data to predict the power curve’s behavior at higher wind speeds. The exponential curve levels off gracefully near the turbine’s rated power capacity, avoiding the overprediction of linear extrapolation. This method provides a more realistic model of the flattening power curve (Lee et al., 2020).
Gaussian Mixture Models – Uses statistical Gaussian mixture models trained on the measured power curve data to extrapolate the relationship at higher wind speeds. This data-driven approach can capture the subtle nonlinearities in the power curve better than simple linear or exponential models (Zhang et al., 2023).
Overall, extrapolation methods allow wind turbine designers and operators to estimate power potential beyond the tested wind speed range. But they rely on assumptions of how the power curve continues beyond the measurement data. More advanced statistical and machine learning techniques may provide improved extrapolation accuracy.
Power Performance Testing
Power performance testing (PPT) involves measuring the power curve of a wind turbine to verify that it meets the expected performance specifications under real-world operating conditions. PPT typically follows the standards outlined in IEC 61400-12-1 or a similar national or international standard.
There are two main approaches to PPT – full turbine testing and component testing. Full turbine testing involves testing the entire wind turbine as installed in the field. This provides the most accurate representation of how the turbine will perform under normal operation. Anemometers and other meteorological equipment are used to measure wind speed and other environmental conditions. These are correlated with the power output readings from the turbine to generate the power curve across a range of wind speeds.
Component testing focuses on testing individual components like blades, gearboxes, generators in a lab setting. While useful for quality control and research, component testing does not account for the complex interactions and performance of the full turbine assembly in real-world conditions. Most independent PPT and certification involves full turbine field testing according to IEC 61400-12-1 guidelines (UL, DNV).
Power Guarantees
Wind turbine manufacturers typically provide contractual guarantees for the performance of their turbines in the form of power curve warranties. These warranties assure the owner that the turbines will generate electricity according to their published power curves within a certain tolerance band. Performance shortfalls below the guaranteed production levels require the manufacturer to provide financial compensation to the owner.
Power curve warranties are usually structured with tiered tolerance levels, with tighter tolerances in the main operating range near the rated power and widening bands in the cut-in and cut-out wind speeds. For example, according to one analysis, warranties often guarantee production within 2% of the warranted power curve between 90-110% of rated wind speed, 5% tolerance between 50-90% and 25% tolerance between 25-50% and 110-125%.
Financial compensation is typically paid based on the net present value of lost energy production over the warranty period, which may span 5-10 years. Warranties usually stipulate an availability guarantee of 97-98% to qualify for compensation. Advanced monitoring systems are used to verify and validate performance claims.
Monitoring and Verification
Long-term monitoring of wind farm power curves is critical for detecting underperformance and verifying power production claims. Operational power curves derived from long-term data provide insights into day-to-day performance variability and deviations from expected output. Marvuglia et al. (2012) developed a monitoring approach using machine learning models trained on reference power curves to identify anomalies and underperformance issues. This enables performance verification and early fault detection through continuous monitoring of operational power curves.
Dai et al. (2022) also emphasized the importance of obtaining real power curves from operational data to reflect the actual performance characteristics of wind turbines. Since manufacturer power curves represent idealized performance, long-term monitoring reveals real-world factors impacting output. Regular comparison of operational power curves to baseline expectations facilitates performance verification and troubleshooting.
Advancements
Recent advancements in wind turbine technology are enabling more efficient and cost-effective wind energy production. One key area of innovation is in Lidar systems, which use laser beams to remotely measure wind speeds and directions across the full sweep of the rotor. Lidar provides enhanced wind mapping and inflow measurements, allowing turbines to optimize performance through improved pitch and yaw control.1
Machine learning techniques are also being applied to wind turbine controls optimization. By leveraging artificial intelligence algorithms, turbine controllers can continuously tune parameters in response to changing wind conditions to maximize power output. According to one study, machine learning-based control increased annual energy production by 2%.2
There have also been advancements in power curve modeling using machine learning. These data-driven models provide more accurate predictions of wind turbine performance compared to traditional methods. Improved power curve forecasts enable better assessment of site wind resources and estimated energy production.3
Conclusion
In summary, the power curve is a fundamental concept in understanding and evaluating wind turbine performance. Accurate power curve measurement and verification ensures turbines are operating as expected and meeting power production guarantees. The standards and best practices for power curve testing continue to advance, providing more accurate models that account for real-world conditions.
Looking forward, innovations in lidar, machine learning, and physics-based modeling show promise for even better power curve predictions. As wind turbines grow larger and more complex, improved testing and analytical techniques will be key to maximizing performance and energy production. However, uncertainty and variability due to factors like turbulence and atmospheric stability will continue to challenge accurate power curve measurement.
Overall, power curve verification remains an active area of research and development. With the growth of wind energy and rising performance expectations, power curve accuracy will only increase in importance for manufacturers, operators, investors, and other stakeholders.
