Search

The input torque of a wind turbine contains an abundance of information about the operating conditions of the turbine. At the same time it is also a critical input for the efficiency determination of the drivetrain. An appropriate method of torque measurement plays an important role in wind turbines research and helps to further reduce the COE (cost of energy) of the wind turbines. However, a number of challenges are currently restricting the breadth and depth of torque measurement applications in the testing and operation of wind turbines. The research described in this dissertation studies the major challenges posed by the different aspects of torque measurement, including the measuring principle, the calibration and the uncertainty. The efficiency determination of the drivetrain is also studied as an important application, which has the highest demand on the accuracy of torque measurement and its calibration among applications in the wind energy industry. The research proposes new approaches and improvements to cope with the above mentioned challenges. A new method of torque measurement based on the rotary encoders is proposed and realised during a test campaign. During the same test campaign, improvement of the traditional torque measurement based on strain gauges is also demonstrated, where the influence of non-torque loads is greatly reduced by having more measuring points for the torque measurement. A new method is also proposed to address the problem of insufficient accuracy in the drivetrain efficiency determination of a wind turbine on the test bench. With the proposed method, the dependency of the determined efficiency on the accuracy of torque and electrical power measurement can be effectively reduced. As a result, the efficiency can be determined with an uncertainty considerably lower than that of the torque measurement. The method takes advantage of a specially designed test sequence whereby the test bench and the wind turbine drivetrain take turns to run in motor mode and drive the other one which operates in generator mode. The same test sequence is also adopted to develop a method of torque calibration. The method establishes a relationship between the torque and the electrical power using measurements from the two tests where the turbine drivetrain operates in different modes. The calibration uncertainty introduced by the power loss in the drivetrain is thus reduced. Detailed uncertainty analysis for the efficiency determination and torque calibration is carried out in this research to confirm the benefit as well as quantify the effectiveness of the methods proposed. Future work and further applications of the methods proposed are presented at the end of the dissertation.

In der Antriebsstrangentwicklung wird vermehrt auf 1D-Simulations-Tools zurückgegriffen. Die übliche Modellierung des Fahrers in der Simulation als PIRegler zeigt in verschiedenen Punkten Schwächen. Im vorliegenden Beitrag wird eine alternative Modellierung des Fahrers vorgestellt und mit dem PI-Regler verglichen. Der wesentliche Unterschied besteht in der gleichzeitigen Berücksichtigung von Weg- und Geschwindigkeitsabweichungen gegenüber der reinen Geschwindigkeitsregelung. Es ergeben sich signifikante Vorteile durch den Einsatz des aufgezeigten Fahrermodells, die sich besonders bei Überlastanforderungen an das Fahrzeugmodell zeigen. Solche Belastungen können vermehrt bei der Simulation von Hybridantrieben auftreten, wenn diese mit Messdaten konventioneller Antriebsstränge beaufschlagt werden.