Investigation of estimation algorithms as an alternative to distance sensors in wind turbines

Abstract:
One of the latest trends for high power wind turbines > 10 MW is to reduce the weight of the nacelle associated with constructional issues. One way is to increase the diameter of the generator and the pole-pairs in order to achieve high torque through low speeds. However, due to this relatively long diameter, axial vibrations are to be expected. In the proof of concept of an active vibration damping system, distance sensors have so far been used to detect vibrations, but additional sensors always increase the costs and the probability of failure. This paper developed a concept for the detection of axial vibrations to replace the distance sensors. For this purpose, parameter identification methods Model Reference Adaptive Systems, Recursive Least Squares, and Extended Kalman Filter are developed to estimate the rotor flux linkage. The result from these methods is first compared with the simulation. Then the methods are verified in the laboratory on the demonstration generator.