PO017

Verifying production losses due to yaw misalignment: iSpin yaw

Nick Janssen, Eduardo Gil Marín
ROMO Wind, Aarhus, Denmark

Abstract

Up to today, it has been difficult to validate the effect of a turbine's yaw misalignment on production losses. Nevertheless, yaw misalignment is widely agreed to be one of the major factors for underperformance of a wind turbine. Typically the effect on a yaw misalignment is verified by measuring a power curve on a turbine with a yaw misalignment, then adjusting the turbine, and afterwards measuring a second (improved) power curve. The major drawback of this method, is that both power curves are measured in different time periods, and therefore different atmospheric conditions. This makes is difficult to isolate the effect of yaw misalignment on a power curve.

ROMO Wind performed measurements in a unique project, where a turbine was subjected to an artificially induced yaw misalignment forcing the turbine to be misaligned to the wind direction from -10 degrees to +10 degrees on a 6 hourly cycle for a 2 month period. This means that the test turbine has continuously measured data at different yaw misalignments, yet similar atmospheric conditions. It was concluded that a turbine's performance scales with a factor cos^2 of the yaw misalignment. Furthermore, it was concluded that performance can accurately be monitored using ROMO Wind's iSpin technology.

Method

Traditionally, a turbine affected by yaw misalignment will operate with this yaw misalignment for a number of months. After the yaw misalignment is detected, the turbine is typically adjusted to a zero degree yaw misalignment. If we think outside of the box, then a yaw misalignment time series could also alternate between -x and +x with a known period, by applying a continuously changing offset into the controller. High-wind speed bins in the power curve will then be filled up for all yaw misalignment bins after only a single storm. This methodology has been applied to a multi-MW wind turbine.

Results

The AEP from power curves in different yaw misalignment bins was compared to the AEP from the zero-yaw misalignment power curve. This gives six different relative AEPs (from a total of 7 power curves) for six different yaw misalignment bins. The percentage-wise decrease in AEP (i.e. the lost production) was computed and an excellent fit was observed compared to the cosine-squared model.

Conclusions

It was concluded that the power output of a turbine is affected by yaw misalignment. A higher yaw misalignment means a lower power production. The power and yaw misalignment relate to each other according to the cosine-squared model. In this study, the model was validated with real data and it was concluded that the model is highly accurate. This implicates that following the cosine-squared model is an easy way to determine the annual losses in power due to a yaw misalignment. It is essential that this model is implemented on an industry-wide scale to determine annual production losses due to yaw misalignment, as well as to provide the ability to act accordingly to reduce these losses.

Objectives

There exists a large amount of doubt concerning the effect of yaw misalignment on wind turbine performance. Most wind turbine owners are unsure about the magnitude of the effect of yaw misalignment on power output, and therefore hesitate in their decisions. For this reason it is nearly impossible to make a proper risk analysis to base a yaw misalignment correction decision on. From previous publications on ROMO Winds database of measurements exceeding 300 wind turbines, over 50% of the measured turbines had a yaw misalignment exceeding 4 degrees. This is a real problem, which delegates will lean how to address, and thereby reduce annual production losses and increase revenue.